1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
DRAFT - DO NOT CITE OR QUOTE
&EPA
EPA/635/R-10/003C
www.epa.gov/iris
TOXICOLOGICAL REVIEW
OF
DICHLOROMETHANE
(METHYLENE CHLORIDE)
(CAS No. 75-09-2)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
December 2009
NOTICE
This document is an Interagency Science Consultation draft. This information is distributed
solely for the purpose of pre-dissemination peer review under applicable information quality
guidelines. It has not been formally disseminated by EPA. It does not represent and should not
be construed to represent any Agency determination or policy. It is being circulated for review
of its technical accuracy and science policy implications.
U.S. Environmental Protection Agency
Washington, DC

-------
32
33	DISCLAIMER
34
35	This document is a preliminary draft for review purposes only. This information is
36	distributed solely for the purpose of pre-dissemination peer review under applicable information
37	quality guidelines. It has not been formally disseminated by EPA. It does not represent and
38	should not be construed to represent any Agency determination or policy. Mention of trade
39	names or commercial products does not constitute endorsement or recommendation for use.
40
41

-------
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
CONTENTS—TOXICOLOGICAL REVIEW OF DICHLOROMETHANE
(CAS No. 75-09-2)
CONTENTS—TOXICOLOGICAL REVIEW OF DICHLOROMETHANE	ii
LIST OF TABLES	vii
LIST OF FIGURES	xiii
LIST OF ACRONYMS	xvii
FOREWORD	xix
AUTHORS, CONTRIBUTORS, AND REVIEWERS	xx
1.	INTRODUCTION	1
2.	CHEMICAL AND PHYSICAL INFORMATION	3
3.	TOXICOKINETICS	5
3.1.	ABSORPTION	5
3.1.1.	Oral — Gastrointestinal Tract Absorption	5
3.1.2.	Inhalation—Respiratory Tract Absorption	5
3.2.	DISTRIBUTION	7
3.3.	METABOLISM	9
3.3.1.	The CYP2E1 Pathway	11
3.3.2.	The GST Pathway	14
3.4.	ELIMINATION	19
3.5.	PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS	21
3.5.1.	Probabilistic Mouse PBTK Dichloromethane Model (Marino et al., 2006)	26
3.5.2.	Probabilistic Human PBTK Dichloromethane Model (David et al., 2006)	 31
3.5.3.	Evaluati on of Rat PB TK Di chl oromethane Model s	38
3.5.4.	Comparison of Mouse, Rat and Human PBTK Models	39
HAZARD IDENTIFICATION	43
4.1. STUDIES IN HUMANS	43
4.1.1.	Introduction—Case Reports, Epidemiologic, and Clinical Studies	43
4.1.2.	Noncancer Studies	43
4.1.2.1.	Case Reports of Acute, High-dose Exposures	43
4.1.2.2.	Controlled Experiments Examining Acute Effects	44
4.1.2.3.	Observational Studies Focusing on Clinical Chemistries, Clinical
Examinations, and Symptoms	45
4.1.2.4.	Observational Studies Using Workplace Medical Program Data	51
4.1.2.5.	Studies of Ischemic Heart Disease Mortality Risk	55
4.1.2.6.	Studies of Suicide Risk	56
4.1.2.7 Studies of Infectious Disease Risk	57
4.1.2.8.	Studies of Reproductive Outcomes	57
4.1.2.9.	Summary of Noncancer Studies	60
4.1.3.	Cancer Studies	62
4.1.3.1.	Identification and Selection of Studies for Evaluation of Cancer
Risk	62
4.1.3.2.	Description of the Selected Studies	63
4.1.3.3.	Cellulose Triacetate Film Base Production Cohorts	63
4.1.3.4.	Cellulose Triacetate Fiber Production Cohorts	70
4.1.3.5.	Solvent-Exposed Workers—Hill Air Force Base, Utah	75
4.1.3.6.	Case-Control Studies of Specific Cancers and Dichloromethane	76
4.1.3.7.	Summary of Cancer Studies by Type of Cancer	83
ii	DRAFT - DO NOT CITE OR QUOTE

-------
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
4.2.	SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIO AS SAYS IN
ANIMALS—ORAL AND INHALATION	92
4.2.1.	Oral Exposure: Overview of Noncancer and Cancer Effects	92
4.2.1.1.	Toxicity Studies of Subchronic Oral Exposures: Hepatic Effects	93
4.2.1.2.	Toxicity Studies of Chronic Oral Exposures: Hepatic Effects and
Carcinogenicity	96
4.2.2.	Inhalation Exposure: Overview of Noncancer and Cancer Effects	102
4.2.2.1.	Toxicity Studies of Subchronic Inhalation Exposures: General,
Renal, and Hepatic Effects	103
4.2.2.2.	Toxicity Studies from Chronic Inhalation Exposures	107
4.3.	REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND
INHALATION	124
4.3.1.	Reproductive Toxicity Studies	127
4.3.1.1.	Oral (Gavage) Studies	127
4.3.1.2.	Inhalation Studies	128
4.3.2.	Developmental Toxicity Studies	129
4.3.2.1.	Oral (Gavage) Studies and Culture Studies	129
4.3.2.2.	Inhalation Studies	130
4.4.	OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES	132
4.4.1.	Short-term (2-Week) Studies of General and Hepatic Effects in Animals	132
4.4.2.	Immunotoxicity Studies in Animals	133
4.4.3.	Neurotoxicology Studies in Animals	135
4.4.3.1.	Neurotoxicology Studies—Oral Exposures	140
4.4.3.2.	Neurotoxicology Studies—Inhalational Exposure	141
4.5.	MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE
MODE OF ACTION	148
4.5.1.	Genotoxicity Studies	148
4.5.1.1.	In Vitro Genotoxicity Assays	148
Bacterial, Yeast, and Fungi mutagenicity assays	148
4.5.1.2.	In Vivo Genotoxicity Assays	158
4.5.2.	Mechanistic Studies of Liver Effects	167
4.5.3.	Mechanistic Studies of Lung Effects	170
4.5.4.	Mechanistic Studies of Neurological Effects	175
4.6.	SYNTHESIS OF MAJOR NONCANCER EFFECTS	177
4.6.1.	Oral Exposures	177
4.6.1.1.	Summary of Human Data	Ill
4.6.1.2.	Summary of Animal Data	178
4.6.2.	Inhalation Exposures	182
4.6.2.1.	Summary of Human Data	182
4.6.2.2.	Summary of Animal Studies	183
4.6.3.	Mode of Action Information	190
4.6.3.1.	Mode of Action for Nonneoplastic Liver Effects	190
4.6.3.2.	Mode of A ction for Nonneoplastic Lung Effects	191
4.6.3.3.	Mode of Action for Neurological Effects	191
4.6.3.4.	Mode of Action for Reproductive and Developmental Effects	192
4.6.3.5.	Mode of Action for Immunotoxicity	193
4.7.	EVALUATION OF CARCINOGENICITY	194
4.7.1.	Summary of Overall Weight of Evidence	194
4.7.2.	Synthesis of Human, Animal, and Other Supporting Evidence	195
in
DRAFT - DO NOT CITE OR QUOTE

-------
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
4.7.3. Mode of Action Information	204
4.7.3.1.	Hypothesized Mode of Action	204
4.7.3.2.	General Conclusions About the Mode of Action for Tumors in
Rodents and Possible Relevance to Humans	211
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES	213
4.8.1.	Possible Childhood Susceptibility	213
4.8.2.	Possible Gender Differences	215
4.8.3.	Other Susceptible Populations	215
5. DOSE-RESPONSE ASSESSMENTS	217
5.1.	ORAL REFERENCE DOSE (RID)	217
5.1.1.	Choice of Principal Study and Critical Effect—with Rationale and
Justification	217
5.1.2.	Derivation Process for Noncancer Reference Values	220
5.1.3.	Evaluation of Dose Metrics for Use in Noncancer Reference Value
Derivations	223
5.1.4.	Methods of Analysis—Including Models (PBTK, BMD, etc.)	224
5.1.5.	Rfl) Derivation—Including Application of Uncertainty Factors (UFs)	229
5.1.6.	Previous RfD Assessment	230
5.1.7.	RfD Comparison Information	230
5.2.	INHALATION REFERENCE CONCENTRATION (RfC)	234
5.2.1.	Choice of Principal Study and Critical Effect—with Rationale and
Justification	234
5.2.2.	Derivation Process for Reference Concentration Values	238
5.2.3.	Methods of Analysis—Including Models (PBTK, BMD, etc.)	238
5.2.4.	RfC Derivation—Including Application of Uncertainty Factors (UFs)	243
5.2.5.	Previous RfC Assessment	245
5.2.6.	RfC Comparison Information	245
5.3.	UNCERTAINTIES IN THE ORAL REFERENCE DOSE AND INHALATION
REFERENCE CONCENTRATION	250
5.4.	CANCER ASSESSMENT	262
5.4.1.	Cancer Oral Slope Factor	263
5.4.1.1.	Choice of Study/Data—with Rationale and Justification	263
5.4.1.2.	Derivation Process for Cancer Oral Slope Factor	263
5.4.1.3.	Dose-Response Data	266
5.4.1.4.	Dose Conversion and Extrapolation Methods: Cancer Oral Slope
Factor	266
5.4.1.5.	Oral Cancer Slope Factor	273
5.4.1.6.	Alternative Derivation Based on Route-to-Route Extrapolation	273
5.4.1.8.	Previous IRIS Assessment: Cancer Oral Slope Factor	276
5.4.1.9.	Comparison of Cancer Oral Slope Factors Using Different
Methodologies	276
5.4.2.	Cancer Inhalation Unit Risk	278
5.4.2.1.	Choice of Study/Data—with Rationale and Justification	278
5.4.2.2.	Derivation Process for Cancer Inhalation Unit Risk	279
5.4.2.3.	Dose-Response Data	279
5.4.2.4.	Dose Conversion and Extrapolation Methods: Cancer Inhalation
Unit Risk	280
5.4.2.5.	Cancer Inhalation Unit Risk	287
5.4.2.8. Previous IRIS Assessment: Cancer Inhalation Unit Risk	291
iv	DRAFT - DO NOT CITE OR QUOTE

-------
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
5.4.2.8. Comparison of Cancer Inhalation Unit Risk Using Different
Methodologies	292
5.4.3.	Differences Between Current Assessment and Previous IRIS PBTK-based
Assessment	294
5.4.4.	Application of ADAFs	296
5.4.4.1.	Application of ADAFs in Oral Exposure Scenarios	297
5.4.4.2.	Application of ADAFs in Inhalation Exposure Scenarios	297
5.4.5.	Uncertainties in Cancer Risk Values	298
6.	MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND
DOSE RESPONSE	310
6.1.	HUMAN HAZARD POTENTIAL	310
6.2.	DOSE-RESPONSE	313
6.2.1.	Oral RfD	313
6.2.2.	Inhalation RfC	314
6.2.3.	Uncertainties in Reference Dose and Reference Concentration Values	315
6.2.4.	Oral Cancer Slope Factor	316
6.2.5.	Cancer Inhalation Unit Risk	320
6.2.6.	Uncertainties in Cancer Toxicity Values	323
7.	REFERENCES	325
APPENDIX A: EXTERNAL REVIEW PANEL COMMENTS	A-l
APPENDIX B: HUMAN PBTK DICHLOROMETHANE MODEL	B-1
B-l. Human Model Description	B-l
B-2. Revisions to Parameter Distributions of David et al. (2006)	B-3
B-3. CYP2E1 and GST-T1	B-4
B-4. Analysis of Human Physiological Distributions for PBPK Modeling	B-8
B-4.1. Age	B-8
B-4.2. Gender	B-9
B-4.3. Body Weight	B-10
B-4.4. Alveolar Ventilation	B-12
B-4.5. Cardiac Output	B-l3
B-4.6. Fat Fraction	B-14
B-4.7. Liver Fraction	B-14
B-4.8. Tissue volume normalization	B-16
B-5. Summary of Revised Human PBTK Model	B-16
APPENDIX C: RAT DICHLOROMETHANE PBTK MODELS 	C-l
C-l. PBTK Models Selected for Comparison Against Dichloromethane Kinetic Data
Sets	C-l
C-2. Model Performance	C-4
C-2.1. Evaluation of Model Structure for Description of Carboxyhemoglobin Levels	C-4
C-2.2. Evaluation of Model Structure for Prediction of Uptake, Blood and Liver
Concentrations, and Expiration of Dichloromethane	C-5
C-2.3. Evaluation of Model Structure on Relative Flux of CYP and GST Metabolism of
Dichloromethane	C-9
C-2.4. Evaluation of Model Predictions of Oral Absorption of Dichloromethane	C-ll
C-3. Model Option Summary	C-l5
v
DRAFT - DO NOT CITE OR QUOTE

-------
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
APPENDIX D: SUMMARY OF BENCHMARK DOSE (BMD) MODELING OF
NONCANCER ENDPOINTS	D-l
D-l. Oral Reference Dose: BMD Modeling of Nonneoplastic Liver Lesion Incidence
Data For Rats Exposed to Dichloromethane In Drinking Water For 2 Years
(Serota et al., 1986a)	D-l
D-2. Inhalation Reference Concentration: BMD Modeling of Nonneoplastic Liver
Lesion Incidence Data For Rats Exposed to Dichloromethane by Inhalation For 2
Years (Nitschke et al., 1988a)	D-5
APPENDIX E: SUMMARY OF BENCHMARK DOSE (BMD) MODELING OF CANCER
ENDPOINTS	E-l
E-l. Oral Cancer Slope Factors: BMD Modeling of Liver Tumor Incidence Data for
Mice Exposed to Dichloromethane in Drinking Water for 2 Years (Serota et al.,
1986b; Hazelton Laboratories, 1983)	E-l
E-l.l. Modeling results for the internal liver metabolism metric	E-3
E-l.2. Modeling results for the whole body metabolism metric	E-5
E-2. Cancer Inhalation Unit Risk: BMD Modeling of Liver and Lung Tumor
Incidence Data for Male Mice Exposed to Dichloromethane via Inhalation for 2
Years (Mennear et al., 1988; NTP, 1986)	E-7
E-2.1. Modeling results for the internal liver metabolism metric, liver tumors	E-9
E-2.2. Modeling results for the internal lung metabolism metric, lung tumors	E-l 1
E-2.3. Modeling results for the whole body metabolism metric, liver tumors	E-13
E-2.4. Modeling results for the whole body metabolism metric, lung tumors	E-15
APPENDIX F. COMPARATIVE CANCER INHALATION UNIT RISK BASED ON FEMALE
MICE DATA	F-l
APPENDIX G: COMPARATIVE CANCER INHALATION UNIT RISK BASED ON
BENIGN MAMMARY GLAND TUMORS IN RATS	G-l
APPENDIX H: SOURCE CODE AND COMMAND FILES FOR DICHLOROMETHANE
PBTK MODELS	II-l
vi
DRAFT - DO NOT CITE OR QUOTE

-------
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
LIST OF TABLES
Table 2-1. Physical properties and chemical identity of dichloromethane	3
Table 3-1. Distribution of radioactivity in tissues 48 hours after inhalation exposure of mature
male Sprague-Dawley rats (n = 3) for 6 hours	7
Table 3-2. Brain and perirenal fat dichloromethane and blood CO concentrations in male Wistar
rats exposed by inhalation to dichloromethane at constant exposure concentrations
compared with intermittently high exposure concentrations	8
Table 3-3. Mean prevalences of the GST-T1 null (-/-) genotype in human ethnic groups	15
Table 3-4. GST-T1 enzyme activities toward dichloromethane in human, rat, mouse, and
hamster tissues (liver, kidney, and erythrocytes)	17
Table 3-5. Values for parameter distributions in a B6C3Fi mouse probabilistic PBTK model for
dichloromethane compared with associated values for point parameters in earlier
deterministic B6C3Fi mouse PBTK models for dichloromethane	29
Table 3-6. Internal daily doses for B6C3Fi mice exposed to dichloromethane for 2 years
(6 hours/day, 5 days/week) calculated with different PBTK models	30
Table 3-7. Results of calibrating metabolic parameters in a human probabilistic PBTK model for
dichloromethane with individual kinetic data for 42 exposed volunteers and MCMC
analysis	32
Table 3-8. Parameter distributions used in human Monte Carlo analysis for dichloromethane by
David et al. (2006)	 34
Table 3-9. Parameter distributions for the human PBTK model for dichloromethane used by
EPA	36
Table 3-10. Parameter values for the rat PBTK model for dichloromethane used by EPA	39
Table 3-11. Parameters in the mouse, rat, and human PBTK model for dichloromethane used by
the EPA	41
Table 4-1. Percentage of male General Electric plastic polymer workers reporting neurologic
symptoms or displaying abnormal values in measures of neurological function,
hepatic function, and cardiac function	54
Table 4-2. Ischemic heart disease mortality risk in four cohorts of dichloromethane-exposed
workers	56
Table 4-3. Suicide risk in two cohorts of dichloromethane-exposed workers	57
Table 4-4. Mortality risk in Eastman Kodak cellulose triacetate film base production workers,
Rochester, New York	66
Table 4-5. Mortality risk by cumulative exposure in Eastman Kodak cellulose triacetate film
base production workers, Rochester, New York	67
Table 4-6. Mortality risk in Imperial Chemical Industries cellulose triacetate film base
production workers, Brantham, United Kingdom: 1,473 men employed 1946-1988,
followed through 1994	 70
Table 4-7. Mortality risk in Hoechst Celanese Corporation cellulose triacetate fiber production
workers, Rock Hill, South Carolina: 1,271 men and women employed 1954-1977,
followed through 1990	 72
Table 4-8. Cancer mortality risk in Hoechst Celanese Corporation cellulose triacetate fiber
production workers, Cumberland, Maryland: 2,909 men and women employed
1970-1981, followed through 1989	 74
Table 4-9. Summary of cohort studies of cancer risk and dichloromethane exposure	85
Table 4-10. Summary of case-control studies of cancer risk and dichloromethane exposure	87
vii	DRAFT - DO NOT CITE OR QUOTE

-------
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
Table 4-11. Incidences of histopathologic changes in livers of male and female F344 rats
exposed to dichloromethane in drinking water for 90 days	94
Table 4-12. Incidences of histopathologic changes in livers of male and female B6C3Fi mice
exposed to dichloromethane in drinking water for 90 days	96
Table 4-13. Studies of chronic oral dichloromethane exposures (up to 2 years)	97
Table 4-14. Incidences of nonneoplastic liver changes and liver tumors in male and female F344
rats exposed to dichloromethane in drinking water for 2 years	98
Table 4-15. Incidences for focal hyperplasia and tumors in the liver of male B6C3Fi mice
exposed to dichloromethane in drinking water for 2 years	100
Table 4-16. Studies of chronic inhalation dichloromethane exposures	108
Table 4-17. Incidences of nonneoplastic histologic changes in male and female F344/N rats
exposed to dichloromethane by inhalation (6 hours/day, 5 days/week) for 2 years 110
Table 4-18. Incidences of selected neoplastic lesions in male and female F344/N rats exposed to
dichloromethane by inhalation (6 hours/day, 5 days/week) for 2 years	112
Table 4-19. Incidences of nonneoplastic histologic changes in B6C3Fi mice exposed to
dichloromethane by inhalation (6 hours/day, 5 days/week) for 2 years	114
Table 4-20. Incidences of neoplastic lesions in male and female B6C3Fi mice exposed to
dichloromethane by inhalation (6 hours/day, 5 days/week) for 2 years	115
Table 4-21. Incidences of selected neoplastic histologic changes in male and female Sprague-
Dawley rats exposed to dichloromethane by inhalation (6 hours/day, 5 days/week)
for 2 years	119
Table 4-22. Incidences of selected nonneoplastic histologic changes in male and female
Sprague-Dawley rats exposed to dichloromethane by inhalation (6 hours/day,
5 days/week) for 2 years	121
Table 4-23. Incidences of selected neoplastic histologic changes in male and female Sprague-
Dawley rats exposed to dichloromethane by inhalation (6 hours/day, 5 days/week)
for 2 years	123
Table 4-24. Summary of studies of reproductive and developmental effects of dichloromethane
exposure in animals	125
Table 4-25. Reproductive outcomes in F344 rats exposed to dichloromethane by inhalation for
14 weeks prior to mating and from GDs 0-21	128
Table 4-26. Studies of neurobehavioral changes from dichloromethane, by route of exposure and
type of effect	137
Table 4-27. Studies of neurophysiological changes as measured by evoked potentials resulting
from dichloromethane, by route of exposure	138
Table 4-28. Studies of neurochemical changes from dichloromethane, by route of exposure. 139
Table 4-29. Results from in vitro genotoxicity assays of dichloromethane with bacteria, yeast, or
fungi	149
Table 4-30. Results from in vitro genotoxicity assays of dichloromethane with mammalian
systems, by type of test	155
Table 4-31. Results from in vivo genotoxicity assays of dichloromethane in insects	158
Table 4-32. Results from in vivo genotoxicity assays of dichloromethane in mice	160
Table 4-33. Results from in vivo genotoxicity assays of dichloromethane in rats and hamsters
	164
Table 4-34. Comparison of in vivo dichloromethane genotoxicity assays targeted to lung or liver
cells, by species	165
Table 4-35. NOAELs and LOAELs in selected animal studies involving oral exposure to
dichloromethane for short-term, subchronic, or chronic durations	180
viii
DRAFT - DO NOT CITE OR QUOTE

-------
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
Table 4-36. NOAELs and LOAELs in animal studies involving inhalation exposure to
dichloromethane for subchronic or chronic durations, hepatic, pulmonary, and
neurologic effects	185
Table 4-37. NOAELs and LOAELs in selected animal studies involving inhalation exposure to
dichloromethane, reproductive and developmental effects	188
Table 4-38. Incidence of liver tumors in male B6C3Fi mice exposed to dichloromethane in a 2-
year oral exposure (drinking water) studya	197
Table 4-39. Incidences of liver tumors in male and female F344 rats exposed to dichloromethane
in drinking water for 2 years	198
Table 4-40. Incidences of selected neoplastic lesions in B6C3Fi mice exposed to
dichloromethane by inhalation (6 hours/day, 5 days/week) for 2 years	200
Table 4-41. Incidences of selected neoplastic lesions in F344/N rats exposed to dichloromethane
by inhalation (6 hours/day, 5 days/week) for 2 years	201
Table 4-42. Incidences of mammary gland tumors in two studies of male and female Sprague-
Dawley rats exposed to dichloromethane by inhalation (6 hours/day, 5 days/week)
for 2 years	202
Table 4-43. Comparison of internal dose metrics in inhalation and oral exposure scenarios, in
male mice and rats	204
Table 5-1. Incidence data for nonneoplastic liver lesions and internal liver doses, based on
various metrics, in male and female F344 rats exposed to dichloromethane in
drinking water for 2 years (Serota et al., 1986a)	225
Table 5-2. BMD modeling results for incidence of noncancer liver lesions in male and female
F344 rats exposed to dichloromethane in drinking water for 2 years, based liver-
specific CYP metabolism dose metric (mg dichloromethane metabolism via CYP
pathway per liter liver tissue per day)	227
Table 5-3. RfD for dichloromethane based on PBTK model-derived probability distributions of
human drinking water exposures extrapolated from nonneoplastic liver lesion
incidence data for male rats exposed via drinking water for 2 years, based on liver-
specific CYP metabolism dose metric (mg dichloromethane metabolized via CYP
pathway per liter liver tissue per day)	228
Table 5-4. Potential points of departure with applied UFs and resulting RfDs	232
Table 5-5. Incidence data for nonneoplastic liver lesions (hepatic vacuolation) and internal liver
doses, based on various metrics, in female Sprague-Dawley rats exposed to
dichloromethane via inhalation for 2 years (Nitschke et al., 1988a)	239
Table 5-6. BMD modeling results for incidence of noncancer liver lesions in female Sprague-
Dawley rats exposed to dichloromethane by inhalation for 2 years, based on liver
specific CYP metabolism metric (mg dichloromethane metabolized via CYP
pathway per liter liver tissue per day)	241
Table 5-7. Inhalation RfC for dichloromethane based on PBTK model-derived probability
distributions of human inhalation exposure extrapolated from nonneoplastic liver
lesion data for female rats exposed via inhalation for 2 years, based on liver-specific
CYP metabolism dose metric (mg dichloromethane metabolized via CYP pathway
per liter liver tissue per day)	242
Table 5-8. Potential points of departure with applied UFs and resulting RfCs	248
Table 5-9. Statistical characteristics of human equivalent applied doses in specific populations
of the GST-T1 _/" group	260
Table 5-10. Statistical characteristics of HECs in specific populations of the GST-T1group
	262
IX
DRAFT - DO NOT CITE OR QUOTE

-------
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
Table 5-11. Incidence data for liver tumors and internal liver doses, based on GST metabolism
dose metrics, in male B6C3Fi mice exposed to dichloromethane in drinking water
for 2 years	267
Table 5-12. BMD modeling results and tumor risk factors for internal dose metric associated
with 10% extra risk for liver tumors in male B6C3Fi mice exposed to
dichloromethane in drinking water for 2 years, based on liver-specific GST
metabolism and whole body GST metabolism dose metrics	269
Table 5-13. Cancer OSFs for dichloromethane based on PBTK model-derived internal liver
doses in B2C3F1 mice exposed via drinking water for 2 years, based on liver-
specific GST meabolism and whole body metabolism dose metrics, by population
genotype 	272
Table 5-14. Alternative route-to-route cancer OSFs for dichloromethane extrapolated from male
B6C3Fi mouse inhalation liver tumor incidence data using a tissue-specific GST
metabolism dose metric, by population genotype	274
Table 5-15. Cancer OSF based on a human BMDLio using administered dose for liver tumors in
male B6C3Fi mice exposed to dichloromethane in drinking water for 2 years	275
Table 5-16. Comparison of OSFs derived using various assumptions and metrics, based on
tumors in male mice	277
Table 5-17. Incidence data for liver and lung tumors and internal doses, based on GST
metabolism dose metrics, in male and female B6C3Fi mice exposed to
dichloromethane via inhalation for 2 years	280
Table 5-18. BMD modeling results and tumor risk factors associated with 10% extra risk for
liver and lung tumors in male and female B6C3Fi mice exposed by inhalation to
dichloromethane for 2 years, based on liver-specific GST metabolism and whole
body GST metabolism dose metrics	283
Table 5-19. IURs for dichloromethane based on PBTK model-derived internal liver and lung
doses in B6C3Fi male mice exposed via inhalation for 2 years, based on liver-
specific GST metabolism and whole body metabolism dose metrics, by population
genotype	286
Table 5-20. Upper bound estimates of combined human IURs for liver and lung tumors
resulting from lifetime exposure to 1 (J,g/m3 dichloromethane, based on liver-specific
GST metabolism and whole body metabolism dose metrics, by population genotype
	289
Table 5-21. Inhalation units risks based on human BMDLio values using administered
concentration for liver and lung tumors in B6C3Fi mice exposed by inhalation to
dichloromethane for 2 years	291
Table 5-22. Comparison of IURs derived by using various assumptions and metrics	293
Table 5-23. Comparison of key B6C3Fi mouse parameters differing between prior and current
PBTK model application	295
Table 5-24. Application of ADAFs to dichloromethane cancer risk following a lifetime (70-
year) oral exposure	297
Table 5-25. Application of ADAFs to dichloromethane cancer risk following a lifetime (70-
year) inhalation exposure	298
Table 5-26. Summary of uncertainty in the derivation of cancer risk values for dichloromethane
	299
Table 5-27. Statistical characteristics of human internal doses for 1 mg/kg-day oral exposures in
specific populations	308
Table 5-28. Statistical characteristics of human internal doses for 1 mg/m3 inhalation exposures
in specific subpopulations	309
x	DRAFT - DO NOT CITE OR QUOTE

-------
464
465
466
467
468
469
470
471
472
473
474
475
476
All
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
Table 6-1. Comparison of OSFs derived by using various assumptions and metrics, based on
liver tumors in male mice	319
Table 6-2. Comparison of IURs derived by using various assumptions and metrics	322
Table B-l. Parameter distributions used in human Monte Carlo analysis for
dichloromethane by David et al. (2006)	B-2
Table B-2. Parameters for BW distributions as functions of age and gender	B-10
Table B-3. Parameter distributions for the human PBTK model used by the EPA	B-18
Table C-l. Parameter values used in rat PBTK models	C-3
Table C-2. Effect of PBTK model configuration on predicted dichloromethane metabolite
production in the liver of rats from inhalation of 200 or 1,000 ppm dichloromethane
for 4 hours (Andersen et al., 1991) or 2,000 or 4,000 ppm dichloromethane for
6 hours/day, 5 days/week, for 2 years (NTP, 1986)	C-10
Table C-3. Observations and predictions of total expired dichloromethane resulting from
gavage doses in rats	C-l5
Table D-l. Incidence data for nonneoplastic liver lesions and internal liver doses, based
on various metrics, in male and female F344 rats exposed to dichloromethane in
drinking water for 2 years (Serota et al., 1986a)	D-l
Table D-2. BMD modeling results for incidence of noncancer liver lesions in male and female
F344 rats exposed to dichloromethane in drinking water for 2 years, based liver-specific
CYP metabolism dose metric (mg dichloromethane metabolism via CYP pathway per liter
liver tissue per day)	D-2
Table D-3. Incidence data for nonneoplastic liver lesions (hepatic vacuolation) and internal liver
doses, based on various metrics, in female Sprague-Dawley rats exposed to
dichloromethane via inhalation for 2 years (Nitschke et al., 1988a)	D-5
Table D-4. BMD modeling results for incidence of noncancer liver lesions in female
Sprague-Dawley rats exposed to dichloromethane by inhalation for 2 years, based
on liver specific CYP metabolism metric (mg dichloromethane metabolized via
CYP pathway per liter liver tissue per day)	D-6
Table E-l. Incidence data for liver tumors and internal liver doses, based on GST
metabolism dose metrics, in male B6C3Fi mice exposed to dichloromethane in
drinking water for 2 years	E-l
Table E-2. BMD modeling results and tumor risk factors for internal dose metric
associated with 10% extra risk for liver tumors in male B6C3Fi mice exposed to
dichloromethane in drinking water for 2 years, based on liver-specific GST
metabolism and whole body GST metabolism dose metrics	E-2
XI
DRAFT - DO NOT CITE OR QUOTE

-------
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
Table E-3. Incidence data for liver and lung tumors and internal doses, based on GST
metabolism dose metrics, in male B6C3F1 mice exposed to dichloromethane via
inhalation for 2 years	E-7
Table E-4. BMD modeling results and tumor risk factors associated with 10% extra risk
for liver and lung tumors in male B6C3Fi mice exposed by inhalation to
dichloromethane for 2 years, based on liver-specific GST metabolism and whole
body GST metabolism dose metrics	E-8
Table F-l. Incidence data for liver and lung tumors and internal doses, based on GST
metabolism dose metrics, in female B6C3Fi mice exposed to dichloromethane via
inhalation for 2 years	F-l
Table F-2. BMD modeling results and tumor risk factors associated with 10% extra risk for liver
and lung tumors in female B6C3Fi mice exposed by inhalation to dichloromethane for 2
years, based on liver-specific GST metabolism and whole body GST metabolism dose
metrics	F-3
Table F-3. IURs for dichloromethane based on PBTK model-derived internal liver and lung
doses in B6C3Fi female mice exposed via inhalation for 2 years, based on liver-specific
GST metabolism and whole body metabolism dose metrics, by population genotype	F-4
Table F-4. Upper bound estimates of combined human IURs for liver and lung tumors resulting from
lifetime exposure to 1 |ig/m3 dichloromethane, based on liver-specific GST metabolism and whole
body metabolism dose metrics, by population genotype, using female mouse data for derivation of
risk factors	F-5
Table G-l. Incidence data for mammary gland tumors and internal doses, based on different dose
metrics, in male and female F344 rats exposed to dichloromethane via inhalation for
2 years	G-l
Table G-2. BMD modeling results associated with 10% extra risk for mammary gland tumors in
F344 rats exposed by inhalation to dichloromethane for 2 years, based on AUC for
dichloromethane in blood	G-3
Table G-3. IURs for dichloromethane, based on benign mammary tumors and PBTK model-
derived internal doses in F344N rats exposed via inhalation for 2 years, based on AUC for
dichloromethane in blood dose metric	G-4
Xll
DRAFT - DO NOT CITE OR QUOTE

-------
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
LIST OF FIGURES
Figure 3-1. Proposed pathways for dichloromethane metabolism	10
Figure 3-2. Schematics of PBTK models (1986-2006) used in the development of estimates for
dichloromethane internal dosimetry	22
Figure 3-3. Schematic of mouse PBTK model used by Marino et al. (2006)	27
Figure 3-4. Schematic of human PBTK, used by David et al. (2006)	 31
Figure 3-5. Schematic of rat PBTK model used in current assessment	38
Figure 5-1. Exposure response array for oral exposure to dichloromethane	219
Figure 5-2. Process for deriving noncancer oral RfDs and inhalation RfCs using rodent and
human PBTK models	221
Figure 5-3. PBTK model-derived internal doses (mg dichloromethane metabolized via the CYP
pathway per liter liver per day) in rats and humans and their associated external
exposures (mg/kg-day), used for the derivation of RfDs	226
Figure 5-4. Comparison of RfDs derived from selected point of departures for endpoints
presented in Table 5-4	233
Figure 5-5. Exposure response array for chronic (animal) or occupational (human) inhalation
exposure to dichloromethane (log Y axis)	235
Figure 5-6. Exposure response array for subacute to subchronic inhalation exposure to
dichloromethane (log Y axis)	236
Figure 5-7. PBTK model-derived internal doses (mg dichloromethane metabolized via the CYP
pathway/L liver/day) in rats and humans versus external exposures (ppm)	240
Figure 5-8. Comparison of RfCs derived from selected point of departures for endpoints
presented in Table 5-8	249
Figure 5-9. Comparison dichloromethane oxidation rate data with alternate kinetic models... 253
Figure 5-10. Sensitivity coefficients for long-term mass CYP- and GST-mediated metabolites
per liver volume from a daily drinking water concentration of 10 mg/L in rats	257
Figure 5-11. Sensitivity coefficients for long-term mass CYP- and GST-mediated metabolites
per liver volume from a long-term average daily inhalation concentration of 500 ppm
in rats. (KA is not included since it has no impact on inhalation dosimetry.)	257
Figure 5-12. Frequency density of human equivalent applied doses in specific populations in
comparison to a general population (0.5- to 80-year-old males and females) estimate
for an internal dose of 15.1 mg dichloromethane metabolized by CYP per liter liver
per day; all groups were restricted to the GST-T1 population)	260
Figure 5-13. Frequency density of HECs in specific populations in comparison to a general
population (0.5- to 80-year-old males and females) estimate for an internal dose of
128.1 mg dichloromethane metabolized by CYP per liter liver per day ; all groups
restricted to the GST-T1_/" population)	262
Figure 5-14. Process for deriving cancer OSFs and IURs by using rodent and human PBTK
models	265
Figure 5-15. PBTK model-derived internal doses (mg dichloromethane metabolized via the GST
pathway per liter liver per day) in mice and humans and their associated external
exposures (mg/kg-day) used for the derivation of cancer OSFs based on liver tumors
in mice	268
Figure 5-16. PBTK model-derived internal doses (mg dichloromethane metabolized via the GST
pathways per liter tissue per day) for liver (A) and lung (B) in mice and humans, and
their associated external exposures (ppm), used for the derivation of cancer
inhalation unit risks	282
Xlll
DRAFT - DO NOT CITE OR QUOTE

-------
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
Figure 5-20. Histograms for a liver-specific dose of GST metabolism (mg GST metabolites per
liter liver per day) for the general population (0.5- to 80-year-old males and females)
and specific age/gender groups, within the population of GST-T1+/+ genotypes, given
a daily oral dose-rate of 1 mg/kg-day dichloromethane	308
Figure 5-21. Histograms for liver-specific dose of GST metabolism (mg GST metabolites per
liter liver per day) for the general population (0.5- to 80-year-old males and females)
and specific age/gender groups, within the population of GST-T1+/+ genotypes, given
a continuous inhalation exposure to 1 mg/m3 dichloromethane	309
Figure B-l. Schematic of the David et al. (2006) PBTK model for dichloromethane in the
human	B-l
Figure B-2. Total CYP2E1 activity (Vmax) normalized to the average total activity in 14-18
year-old individuals (Vmax(14-18)) plotted against normalized body weight (BW) for
individuals ranging from six months to 18 years of age. Data source: Johnsrud et al.
(2003)	B-6
Figure B-3. Body-weight scaled CYP2E1 activity (Vmaxc) normalized to the average scaled
activity in 14-18 year-old individuals (VmaxC (14-18)) plotted against age individuals
ranging from six months to 18 years of age. Data source: Johnsrud et al. (2003)	B-7
Figure B-4. U.S. age distribution, 6 months-80 years (values from U.S. Census Bureau)	B-9
Figure B-5. U.S. age-specific gender distribution (values from U.S. Census Bureau)	B-9
Figure B-6. Function fits to age-dependent data for BW mean and standard deviations
for males and females in the United States (values from Portier et al., 2007)	B-l 1
Figure B-7. Example body weight (BW) histogram from Monte Carlo simulation for 0.5-
to 80-year-old males and females in the United States (simulated n = 10,000)	B-12
Figure B-8. Mean value respiration rates for males and females as a function of age (values from
Clewell et al., 2004)	B-l3
Figure B-9. Geometric standard deviations (GSDs) for respiration rates for males and
females as a function of age (values from Arcus-Arth and Blaisdell, 2007)	B-l3
Figure B-10. Fraction body fat (VFC) over various age ranges in males and females (data
from Clewell et al., 2004)	B-15
Figure B-l 1. Fraction liver (VLC) as a function of age (data from Clewell et al., 2004)	B-16
Figure C-l. Schematic of the Andersen et al. (1991) PBTK model for dichloromethane
in the rat	C-2
Figure C-2. Observations of exhaled dichloromethane (DCM) and carbon monoxide (CO) after a
bolus oral dose of 200 mg/kg in rats (data of Angelo et al., 1986b)	C-5
xiv
DRAFT - DO NOT CITE OR QUOTE

-------
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
Figure C-3. Observations and predictions (models A-D) of Gargas et al. (1986) data for
respiratory uptake by 3 rats of 100, 500, 1000, or 3000 ppm dichloromethane in a 9-
L closed chamber	C-6
Figure C-4. Observations (points) and predictions (curves, Models A-D) of Angelo et al.
(1986b) data for dichloromethane in blood following 10 and 50 mg/kg iv injection
in rats. Model predictions with doses at 56% of the nominal doses, i.e., 5.6 and 28
mg/kg, are shown for comparison as dashed lines for model D	C-7
Figure C-5. Observations and predictions (Models A-D) of Andersen et al. (1987) data for
dichloromethane in rat blood from inhalation of 200 and 1000 ppm dichloromethane for 4
hours	C-8
Figure C-6. Observations and predictions (Models A-D) of Angelo et al. (1986b) data for
percent of dichloromethane dose expired as dichloromethane following 10 and
50 mg/kg iv injection in rats. Model predictions with doses at 56% of the nominal
doses, i.e., 5.6 and 28 mg/kg, are shown for comparison as dashed lines for model
D	C-9
Figure C-7. Observations and Model D predictions of Angelo et al. (1986b) data for (A) percent
dose expired as dichloromethane, (B) blood dichloromethane, (C) percent expired as
carbon monoxide, and (D) liver dichloromethane in rats given a single dichloromethane
gavage doses of 50 and 200 mg/kg, using a numerically fitted GI absorption rate (KA) of
1.8	C-12
Figure C-8. Model predictions with of blood carboxy-hemoglobin (COHb, percent of
total Hb) from a single gavage dose of 526 mg/kg DCM in rats, compared to the
data of Pankow et al. (1991). Model simulations performed with Model D (heavy
read line, KA = 1.8/h) or with an alternate value of KA = 0.47/h, fit to these data	C-13
Figure D-l. Predicted (logistic model) and observed incidence of noncancer liver lesions in male
F344 rats exposed to dichloromethane in drinking water for 2 years
(Serota et al., 1986a)	D-3
Figure D-2. Predicted (log-probit model) and observed incidence of noncancer liver lesions in
female Sprague-Dawley rats inhaling dichloromethane for 2 years
(Nitschke 1988a)	D-7
Figure E-l. Predicted and observed incidence of animals with hepatocellular carcinoma or
adenoma in male B6C3Fi mice exposed to dichloromethane in drinking water
for 2 years, using liver-specific metabolism dose metric (Serota et al., 1986b; Hazelton
Laboratories, 1983)	E-3
Figure E-2. Predicted and observed incidence of animals with hepatocellular carcinoma or
adenoma in male B6C3Fi mice exposed to dichloromethane in drinking water
for 2 years, using whole-body metabolism dose metric (Serota et al., 1986b; Hazelton
Laboratories, 1983)	E-5
xv
DRAFT - DO NOT CITE OR QUOTE

-------
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
Figure E-3. Predicted and observed incidence of animals with hepatocellular carcinoma or
adenoma in male B6C3Fi mice exposed by inhalation to dichloromethane for 2 years,
using liver-specific metabolism dose metric (Mennear et al., 1988; NTP, 1986)	E-9
Figure E-4. Predicted and observed incidence of animals with carcinoma or adenoma in
the lung of male B6C3Fi mice exposed by inhalation to dichloromethane for 2
years, using liver-specific metabolism dose metric (Mennear et al., 1988; NTP,
1986)	E-ll
Figure E-5. Predicted and observed incidence of animals with hepatocellular carcinoma or
adenoma in male B6C3Fi mice exposed by inhalation to dichloromethane for 2 years,
using whole-body metabolism dose metric (Mennear et al., 1988; NTP, 1986)	E-13
Figure E-6. Predicted and observed incidence of animals with carcinoma or adenoma in
the lung of male B6C3Fi mice exposed by inhalation to dichloromethane for 2
years, using whole-body metabolism dose metric (Mennear et al., 1988; NTP,
1986)	I >15
Figure G-l. PBTK model-derived internal doses (daily average AUC for dichloromethane in
blood) in rats and humans, and their associated external exposures (ppm) used for the
derivation of cancer IURs, based on mammary tumors in rats	G-2
xvi
DRAFT - DO NOT CITE OR QUOTE

-------
719

LIST OF ACRONYMS
720


721


722
ACGIH
American Conference of Governmental Industrial Hygienists
723
ADAF
age-dependent adjustment factor
724
AEGL
acute exposure guideline level
725
AIC
Akaike's Information Criterion
726
ALT
alanine aminotransferase
727
AP
alkaline phosphatase
728
AST
aspartate aminotransferase
729
ATSDR
Agency for Toxic Substances and Disease Registry
730
AUC
area under the curve of a concentration versus time plot
731
BAER
brainstem-auditory evoked response
732
BMD
benchmark dose
733
BMDL10
95% lower bound on the BMD
734
BMR
benchmark response level
735
BW
body weight
736
CAEP
cortical-auditory-evoked potential
737
CASRN
Chemical Abstracts Service Registry Number
738
CHO
Chinese hamster ovary
739
CI
confidence interval
740
CMR
Chemical Marketing Reporter
741
CNS
central nervous system
742
COHb
carboxyhemoglobin
743
CV
coefficient of variation
744
CYP
cytochrome P450
745
DNA
deoxyribonucleic acid
746
EPA
U.S. Environmental Protection Agency
747
FEP
flash-evoked potential
748
FOB
functional observational battery
749
GD
gestation day
750
GSH
reduced glutathione
751
GST
glutathione S-transferase
752
HEC
human equivalent concentration
753
HPRT
hypoxanthine-guanine phosphoribosyl transferase
754
IARC
International Agency for Research on Cancer
755
ICD-9
International Classification of Diseases 9th ed.
756
IgM
immunoglobulin M
757
IRIS
Integrated Risk Information System
758
IUR
inhalation unit risk
759
LOAEL
lowest-observed-adverse-effect level
760
LOH
loss of heterozygosity
761
MCHC
mean corpuscular hemoglobin concentration
762
MCMC
Markov Chain Monte Carlo
763
Mg
milligrams
764
mRNA
messenger ribonucleic acid
765
NADPH
nicotinamide adenine dinucleotide phosphate
766
NIOSH
National Institute of Occupational Safety and Health
767
NLM
National Library of Medicine
xvii
DRAFT - DO NOT CITE OR QUOTE

-------
768
NOAEL
no-observed-adverse-effect level
769
NRC
National Research Council
770
NTP
National Toxicology Program
771
OR
odds ratio
772
OSF
oral slope factor
773
OSHA
Occupational Safety and Health Administration
774
PBTK
physiologically based toxicokinetic
775
PND
postnatal day
776
QCC
cardiac output
777
RfC
reference concentration
778
RfD
reference dose
779
SD
standard deviation
780
SEM
standard error of the mean
781
SEP
somatosensory-evoked potential
782
SMR
standardized mortality ratio
783
SRC
Syracuse Research Corporation
784
SSB
single-strand break
785
TWA
time-weighted average
786
UF
uncertainty factor
787
VPR
ventilation:perfusion ratio
xviii
DRAFT - DO NOT CITE OR QUOTE

-------
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to exposure to dichloromethane.
It is not intended to be a comprehensive treatise on the chemical or toxicological nature of
di chl oromethane.
The intent of Section 6, Major Conclusions in the Characterization of Hazard and Dose
Response, is to present the major conclusions reached in the derivation of the reference dose,
reference concentration and cancer assessment, where applicable, and to characterize the overall
confidence in the quantitative and qualitative aspects of hazard and dose response by addressing
the quality of data and related uncertainties. The discussion is intended to convey the limitations
of the assessment and to aid and guide the risk assessor in the ensuing steps of the risk
assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov (email address).
xix
DRAFT - DO NOT CITE OR QUOTE

-------
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGERS
Glinda S. Cooper, Ph.D.
Ambuja S. Bale, Ph.D., DABT
Office of Research and Development, IRIS Program
U.S. Environmental Protection Agency
Washington, DC
AUTHORS
Glinda S. Cooper, Ph.D.
Ambuja S. Bale, Ph.D., DABT
Andrew Rooney, Ph.D.
Paul Schlosser, Ph.D.
Allan Marcus, Ph.D.
Gene (Ching-Hung) Hsu, Ph.D., DABT
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC
John C. Lipscomb, Ph.D., DABT
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Cincinnati, OH
Peter McClure, Ph.D., DABT
Michael Lumpkin, Ph.D.
Fernando Llados, Ph.D.
Mark Osier, Ph.D., DABT
Daniel Plewak, B.S.
Syracuse Research Corporation
Syracuse, NY
Elizabeth Dupree Ellis, Ph.D.
Oak Ridge Institute for Science and Education
Center for Epidemiologic Research
Oak Ridge, TN
REVIEWERS
This document has been peer reviewed by EPA scientists.
xx
DRAFT - DO NOT CITE OR QUOTE

-------
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
INTERNAL EPA REVIEWERS
Ghazi Dannan, Ph.D
Karen Hogan, M.S.
Jennifer Jinot, Ph.D
Paul White, Ph.D
Samantha Jones, Ph.D.
Jamie Strong, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
David Herr, Ph.D
National Health and Environmental Effect Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
xxi
DRAFT - DO NOT CITE OR QUOTE

-------
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
1. INTRODUCTION
This document presents background information and justification for the Integrated Risk
Information System (IRIS) Summary of the hazard and dose-response assessment of
dichloromethane. IRIS Summaries may include oral reference dose (RfD) and inhalation
reference concentration (RfC) values for chronic and other exposure durations, and a
carcinogenicity assessment.
The RfD and RfC, if derived, provide quantitative information for use in risk assessments
for health effects known or assumed to be produced through a nonlinear (presumed threshold)
mode of action. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with
uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human
population (including sensitive subgroups) that is likely to be without an appreciable risk of
deleterious effects during a lifetime. The inhalation RfC (expressed in units of mg/m3) is
analogous to the oral RfD, but provides a continuous inhalation exposure estimate. The
inhalation RfC considers toxic effects for both the respiratory system (portal of entry) and for
effects peripheral to the respiratory system (extrarespiratory or systemic effects). Reference
values are generally derived for chronic exposures (up to a lifetime), but may also be derived for
acute (<24 hours), short-term (>24 hours up to 30 days), and subchronic (>30 days up to 10% of
lifetime) exposure durations, all of which are derived based on an assumption of continuous
exposure throughout the duration specified. Unless specified otherwise, the RfD and RfC are
derived for chronic exposure duration.
The carcinogenicity assessment provides information on the carcinogenic hazard
potential of the substance in question and quantitative estimates of risk from oral and inhalation
exposure may be derived. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
effects may be expressed. Quantitative risk estimates may be derived from the application of a
low-dose extrapolation procedure. If derived, the oral slope factor is a plausible upper bound on
the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk is a
plausible upper bound on the estimate of risk per (j,g/m3 air breathed.
Development of these hazard identification and dose-response assessments for
dichloromethane has followed the general guidelines for risk assessment as set forth by the
National Research Council (1983). EPA Guidelines and Risk Assessment Forum Technical
Panel Reports that may have been used in the development of this assessment include the
following: Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA, 1986a),
Guidelines for Mutagenicity Risk Assessment (U.S. EPA, 1986b), Recommendations for and
Documentation of Biological Values for Use in Risk Assessment (U.S. EPA, 1988a), Guidelines
for Developmental Toxicity Risk Assessment (U.S. EPA, 1991), Interim Policy for Particle Size
2/13/2008
1
DRAFT - DO NOT CITE OR QUOTE

-------
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
and Limit Concentration Issues in Inhalation Toxicity Studies (U.S. EPA, 1994a), Methods for
Derivation of Inhalation Reference Concentrations and Application of Inhalation Dosimetry
(U.S. EPA, 1994b), Use of the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA,
1995), Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for
Neurotoxicity Risk Assessment (U.S. EPA, 1998a), Science Policy Council Handbook. Risk
Characterization (U.S. EPA, 2000a), Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000b), Supplementary Guidance for Conducting Health Risk Assessment of Chemical
Mixtures (U.S. EPA, 2000c), A Review of the Reference Dose and Reference Concentration
Processes (U.S. EPA, 2002), Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens
(U.S. EPA, 2005b), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006a), and A
Framework for Assessing Health Risk of Environmental Exposures to Children (U. S. EPA,
2006b).
The literature search strategy employed for this compound was based on the Chemical
Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
scientific information submitted by the public to the IRIS Submission Desk was also considered
in the development of this document. The relevant literature was reviewed through April 2009.
2
DRAFT - DO NOT CITE OR QUOTE

-------
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
2. CHEMICAL AND PHYSICAL INFORMATION
Dichloromethane is a colorless liquid with a penetrating ether-like odor (Lewis,
1997).1 Selected chemical and physical properties of dichloromethane are listed in Table 2-1.
Table 2-1. Physical properties and chemical identity of dichloromethane

Physical property/chemical identity
Reference
CAS number
75-09-2
Lide (2000)
Synonyms
methylene chloride, methylene dichloride,
O'Neiletal. (2001)

methyl bichloride

Molecular weight
84.93
O'Neiletal. (2001)
Chemical formula
CH2C12
O'Neiletal. (2001)
Boiling point
40°C
Lide (2000)
Melting point
—95.1°C
Lide (2000)
Vapor pressure
1.15 x 102 mm Hg at 25°C
Boublik et al. (1984)
Density
1.3266 g/mL at 20°C
Lide (2000)
Vapor density
2.93 (air= 1.02)
Holbrook (2003)
Water solubility
1.30 x 104 mg/L at 25°C
Horvath (1982)
Other solubility
Miscible in ethanol, ether, and
International Agency for

dimethylformamide; soluble in carbon
Research on Cancer (IARC)

tetrachloride
(1999)
Partition coefficient
log Kow = 1.25
Hanschetal. (1995)
Flash point
Not flammable
U.S. Coast Guard (1999)
Auto ignition temperature
640°C
Holbrook (2003)
Latent heat of vaporization
3.30 x 105 J/kg
U.S. Coast Guard (1999)
Heat of fusion
16.89 cal/g
U.S. Coast Guard (1999)
Critical temperature
245.0°C
Holbrook (2003)
Critical pressure
6.171 x 106 Pa
Holbrook (2003)
Viscosity
0.430 cP at 20°C
Lewis (1997)
Henry's constant
3.25 x 10~3 atm m3/mol at 25°C
Leighton and Calo (1981)
OH reaction rate constant
1.42 x 1() 13 cm3/molecule sec at 25°C
Atkinson (1989)
Chemical structure
H

I
ci—c-ci
I
H
Dichloromethane is produced by two methods of manufacturing (International Agency
for Research on Cancer [IARC], 1999). The older method involves the direct reaction of
methane with chlorine either at high temperatures or at lower temperatures under catalytic or
photolytic conditions (Holbrook, 2003). The more common method used today involves an
initial reaction of hydrochloric acid with methanol to yield methyl chloride. Excess methyl
chloride is then reacted in the gas phase thermally with chlorine to produce dichloromethane
(Holbrook, 2003). This process can also be carried out catalytically or photolytically.
3
DRAFT - DO NOT CITE OR QUOTE

-------
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
Dichloromethane became an important industrial chemical in the U.S. during
World War II (Hardie, 1964). Dichloromethane has been used in paint strippers and removers,
as a propellant in aerosols, in the manufacture of drugs, pharmaceuticals, film coatings,
electronics, and polyurethane foam, and as a metal-cleaning solvent. Dichloromethane can also
be used in the decaffeination process of coffee and tea (ATSDR, 2000). The U.S. production
was 3.8 million pounds in 1941 and 8.3 million pounds in 1944 (Searles and McPhail, 1949).
Dichloromethane production rose sharply in the decades following the war due to the increased
demand for this substance for use mainly in paint strippers (Hardie, 1964; Searles and McPhail,
1949). U.S. production in 1947, 1955, 1960, and 1962 was approximately 19, 74, 113, and
144 million pounds, respectively (Hardie, 1964; Searles and McPhail, 1949). As other solvent
uses and its use in aerosol propellants became important, demand for this substance increased
further (Anthony, 1979). Dichloromethane production continued to rise dramatically through the
1970s; production capacities were 520 million pounds in 1973 and 830 million pounds in 1979
(Chemical Marketing Reporter [CMR], 1979, 1973).
After 1980, production of dichloromethane began to decline. Production capacities fell
from 722 million pounds in 1982 to 465 million pounds in 1997 (CMR, 1997, 1982). The total
U.S. production capacity for dichloromethane in 2000 was 535 million pounds (CMR, 2000).
The demand for dichloromethane decreased from 600 million pounds in 1979 to 200 million
pounds in 1999 (CMR, 2000, 1979). The decline in production of and demand for
dichloromethane over the past 2 decades has been attributed to increased regulation, the use of
alternative chemicals in aerosol spray cans, and concern over dichloromethane carcinogenicity
(Holbrook, 2003; ATSDR, 2000).
Dichloromethane in the environment will partition mainly to air (National Library of
Medicine [NLM], 2003). In air, dichloromethane exists as a vapor. Some of the
dichloromethane released to soil or water is expected to volatilize to air. In soil,
dichloromethane is expected to be highly mobile and may migrate to groundwater. The potential
for dichloromethane to bioconcentrate in aquatic or marine organisms is low. Dichloromethane
may biodegrade in soil or water under both aerobic and anaerobic conditions.
1 To avoid confusion, "dichloromethane" is used throughout this summary even if a specific paper used the term
"methylene chloride."
4	DRAFT - DO NOT CITE OR QUOTE

-------
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
3. TOXICOKINETICS
3.1. ABSORPTION
3.1.1.	Oral — Gastrointestinal Tract Absorption
There are currently no data available on absorption of dichloromethane following oral
intake in humans. However, after oral administration in animals, dichloromethane is rapidly and
nearly completely absorbed in the gastrointestinal tract (Angelo et al., 1986a, b; McKenna and
Zempel, 1981). Angelo et al. (1986b) reported that, following administration of single
radiolabeled oral doses (10, 50, or 200 mg/kg) to mature male F344 rats, 97% of the label was
detected in the exhaled air within 24 hours, indicating nearly complete absorption. At several
time points within 40 minutes of dose administration, less than 2% of the dose was found in the
lower part of the gastrointestinal tract, indicating that the majority of dichloromethane absorption
occurs in the upper gastrointestinal tract (Angelo et al., 1986b). Similar results were reported in
mature male B6C3Fi mice exposed to up to 50 mg/kg (Angelo et al., 1986a). In mature male
Sprague-Dawley rats administered a single dose (1 or 50 mg/kg) of radiolabeled
dichloromethane, less than 1% of the label was found in feces collected for 48 hours after dose
administration (McKenna and Zempel, 1981). Absorption of dichloromethane generally follows
first-order kinetics (Angelo et al., 1986a), and no evidence for a dichloromethane-specific carrier
has been presented. The vehicle appears to affect the rate, but not the extent, of gastrointestinal
absorption, with an aqueous vehicle resulting in a more rapid absorption of dichloromethane than
an oil-based vehicle (Angelo et al., 1986a).
3.1.2.	Inhalation—Respiratory Tract Absorption
Several studies in humans have demonstrated the absorption of dichloromethane
following inhalation exposure. In a study by Astrand et al. (1975), 14 male volunteers (ages 19-
29) were exposed to about 870 mg/m3 (250 ppm) or 1,740 mg/m3 (500 ppm) for 30 minutes
while resting or exercising on a bicycle ergometer. There was a pause of about 20 minutes
without exposure between rest and exercise periods. Uptake of dichloromethane was estimated
at about 55% while resting and about 40, 30, and 35% at respective workloads of 50, 100, and
150 watts. Blood levels of dichloromethane correlated directly with exposure concentrations,
and did not appear to increase when a workload was applied (Astrand et al., 1975). Similar
reports of rapid uptake and a direct correlation between dichloromethane exposure level and
blood levels in humans have been presented by other groups (DiVincenzo and Kaplan, 1981;
DiVincenzo et al., 1971).
With extended (1-2 hours or greater) exposure, uptake tends to reach a steady-state level,
at which point blood dichloromethane levels remain more or less constant (DiVincenzo and
Kaplan, 1981; DiVincenzo et al., 1972; Riley et al., 1966). DiVincenzo et al. (1972) reported
5
DRAFT - DO NOT CITE OR QUOTE

-------
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
that in humans exposed to 100 or 200 ppm of dichloromethane for 2 hours (without physical
exercise), dichloromethane was rapidly absorbed, reaching an approximate steady state, as
assessed by levels of unchanged dichloromethane in the expired air, within the first 15-
30 minutes of exposure. A later study by the same group (DiVincenzo and Kaplan, 1981)
similarly reported a rapid absorption of dichloromethane in volunteers exposed to 50-200 ppm
for 7.5 hours on each of 5 consecutive days. A steady-state level, as assessed by levels of
unchanged dichloromethane in the expired air, was reached quickly (1-2 hours), with exhaled
dichloromethane levels increasing with increasing exposure level. A similar pattern was seen
with blood dichloromethane levels. Estimated pulmonary uptake was 69-75% and did not vary
appreciably with exposure concentration. In another experiment in which one of the
investigators was seated during exposure to 100 ppm dichloromethane for 2 hours,
concentrations of dichloromethane in expired air reached an apparent plateau of about 70 ppm
within the first hour of exposure (Riley et al., 1966).
Body fat may influence absorption of dichloromethane, as evidenced by data from an
experiment involving 12 men ages 21-35, divided into two groups (n = 6 per group) based on
percent body fat (Engstrom and Bjurstrom, 1977). The mean percent body fat in the leaner
group was 7.8% (standard error of the mean [SEM] 1.9), range 2.3-13.6%), compared with
25.1% (SEM 2.8), range 18.3-36.2%), in the more overweight group. Total uptake of
dichloromethane during a light exercise period (50 watts2) for 1 hour with an exposure level of
750 ppm was positively correlated with percent body fat (r = 0.81), and the estimated amount of
dichloromethane in fat storage was also correlated with percent body fat (r = 0.84).
A pattern of absorption similar to that seen in humans has been seen in animals. Initially,
dichloromethane is readily absorbed following inhalation exposure, as evidenced by rapid
appearance of dichloromethane in blood, tissues, and expired air (Withey and Karpinski, 1985;
Stott and McKenna, 1984; Anders and Sunram, 1982; Carlsson and Hultengren, 1975; Roth et
al., 1975). For example, absorption of inhaled 500 ppm dichloromethane in anesthetized, mature
male F344 rats reached an apparent plateau within 10-20 minutes and was relatively constant for
up to 2 hours (Stott and McKenna, 1984). In these experiments, absorption was calculated from
measurements of exposure (nose only) and effluent concentrations and ventilation flow rate in
intact animals; double tracheostomized rats were used to measure absorption in the isolated
upper respiratory tract and the lower respiratory tract. At a ventilation rate of 53 mL/minute,
absorption expressed as mean percentage of dichloromethane available for absorption was 44%>
(standard deviation [SD] 10) in intact rats, 13.2%> (SD 3.6) in the upper respiratory tract, and
37% (SD 4.1) in the lower respiratory tract.
2 A watt is the International System Unit of power and is equal to one joule of energy per second. It is a measure of
the rate of energy use or production (i.e., the exercise effort that was exerted by the individuals in the study).
DRAFT - DO NOT CITE OR QUOTE
6

-------
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
3.2. DISTRIBUTION
Results from studies of animals show that, following absorption, dichloromethane is
rapidly distributed throughout the body and has been detected in all tissues that have been
evaluated. Twenty minutes after a single intravenous dose of 10 mg [14C]-dichloromethane/kg to
mature male B6C3Fi mice (Angelo et al., 1986a), total label was greatest in the liver
(6.72 |ig-equivalents/g tissue), with lower levels reported in the lung (1.82 |ig-equivalents/g
tissue), kidney (1.84 [j,g-equivalents/g tissue), and the remainder of the carcass
(1.90 |ig-equivalents/g tissue). By 4 hours post administration, levels in the liver had fallen to
3.08 [j,g-equivalents/g tissue, lung levels were 0.64 |ig-equivalents/g tissue, and carcass levels
were 0.23 |ig-equivalents/g tissue. The levels in the kidney rose sharply in the first hour
postexposure but then fell and remained steady at -1.60 |ag-equivalents/g tissue for the
remaining 3 hours of the study (Angelo et al., 1986a). McKenna et al. (1982) exposed groups of
mature male Sprague-Dawley rats to 50, 500, or 1,500 ppm [14C]-labeled dichloromethane for
6 hours and examined tissues at 48 hours for presence of radiolabel; results are shown in
Table 3-1. The greatest concentration of label was found in the liver, followed by the kidney and
lung.
Table 3-1. Distribution of radioactivity in tissues 48 hours after inhalation
exposure of mature male Sprague-Dawley rats (n = 3) for 6 hours

Mean ± SD, jig-equivalent dichloromethane/g tissue, by exposure level
Tissue
50 ppm
500 ppm
1,500 ppm
Liver
8.4 ± 1.5
35.6 ±7.5
44.2 ±3.5
Kidney
3.3 ±0.1
16.2 ±2.4
30.5 ±0.2
Lung
1.9 ±0.2
11.0 ± 1.3
16.5 ± 1.6
Brain
0.8 ±0.3
4.2 ± 1.3
6.7 ±0.2
Epidydimal fat
0.5 ±0.2
6.5 ±0.5
4.1 ±0.9
Skeletal muscle
1.1 ±0.1
4.4 ± 1.9
7.7 ±0.7
Testes
1.1 ±0.2
5.5 ± 1.3
8.1 ±0.5
Whole blood
1.1 ±0.2
8.1 ± 1.9
8.9 ± 1.7
Remaining carcass
1.3 ±0.2
5.9 ±0.9
8.6 ± 1.4
Source: McKenna et al. (1982).
As noted in the preceding section, adipose tissue may affect the uptake of
dichloromethane, and there is also evidence of a relation between adiposity and dichloromethane
storage. In the study by Engstrom and Bjurstrom (1977) involving 12 men ages 21-35 exposed
to 750 ppm dichloromethane during a 1 hour light exercise (50 watts) period, dichloromethane
was measured in body fat biopsy specimens at 1, 2, 3, and 4 hours postexposure. All specimens
were taken from the buttocks. The concentration of dichloromethane (per gram tissue) was
7
DRAFT - DO NOT CITE OR QUOTE

-------
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
negatively correlated with percent body fat, but the total estimated amount of dichloromethane in
fat tissue 4 hours postexposure was higher in subjects with a higher amount of fat (r = 0.84).
Carlsson and Hultengren (1975) exposed groups of 10 mature male Sprague-Dawley rats
to [14C]-dichloromethane for 1 hour at a mean concentration of 1,935 mg/m3 (557 ppm) and SD
of 90 mg/m3 (26 ppm). The initial levels were highest in the white adipose tissue (approximately
80 |ig dichloromethane per gram tissue) compared with approximately 35, 20, and 5 |ig-
equivalent dichloromethane/g tissue in the liver, kidney and adrenal glands, and brain,
respectively. These initial levels in the adipose quickly fell to less than 10 |ig-equivalent
dichloromethane/g tissue; more moderate declines were seen in the other tissues.
With acute 6-hour exposure scenarios, peak exposure concentrations may have a greater
influence on dichloromethane levels in the brain and perirenal fat than time-weighted average
(TWA) concentrations during the exposure period (Savolainen et al., 1981). In rats exposed over
a 6-hour period for 5 days/week to a TWA of 1,000 ppm dichloromethane consisting of two
1-hour peak concentrations (2,800 ppm) interspersed with exposure to 100 ppm, levels of
dichloromethane in the brain and perirenal fat were significantly higher than corresponding
levels in rats exposed to constant levels of 1,000 ppm. This difference was not seen with blood
carbon monoxide (CO) levels (Table 3-2). With constant exposure concentrations of 500 or
1,000 ppm, perirenal fat levels of dichloromethane approximately doubled following 2 weeks of
exposure compared with 1 week of exposure, indicating that some storage of dichloromethane in
fat tissue can occur with repeated exposure scenarios (Table 3-2). In contrast, brain levels of
dichloromethane in rats exposed for 1 week were higher than brain levels in rats exposed for 2
weeks. One possible explanation of these observations is that there is an induction of enzymes
involved in dichloromethane metabolism in liver and other tissues with repeated exposure and
dichloromethane in fat is poorly metabolized.
Table 3-2. Brain and perirenal fat dichloromethane and blood CO
concentrations in male Wistar rats exposed by inhalation to
dichloromethane at constant exposure concentrations compared with
intermittently high exposure concentrations



Exposure weeks

Exposure level3
1
2
1
2
1 2
(TWA, ppm)
Brain (nmol/g)
Perirenal fat (nmol/g)
Blood CO (nmol/g)
Control
0
0
0
0
40 ± 15 30 ±10
500, constant
30 ± 7
9 ± 3
436± 47
918 ± 215
675 ± 195 781 ±62
1,000, constant
33 ± 2
14 ± 3
1,316 ±209
2,171 ±219
876 ± 80 825 ± 56
1,000, with two 1-hour
111± 18
50 ± 15
2,295 ± 147
2,431 ± 146
728 ± 84 873 ± 90
peaks of 2,800 ppm





"Groups of 5 rats were exposed to 0, 50, or 1,000 ppm 6 hours/day or 100 ppm interspersed with two 1-hour
peaks of 2,800 ppm for 5 days/week for 1 or 2 weeks. Tissue concentration values are mean ± SD.
Source: Savolainen et al. (1981).
8
DRAFT - DO NOT CITE OR QUOTE

-------
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
Placental transfer
Dichloromethane is capable of crossing the placental barrier and entering the fetal
circulation. Anders and Sunram (1982) reported that when pregnant Sprague-Dawley rats (n =
3) were exposed to 500 ppm dichloromethane for 1 hour on gestational day (GD) 21, mean
maternal blood levels were 176 nmol/mL (SEM 50), while fetal levels were 115 nmol/mL (SEM
40); interestingly, the levels of CO, a metabolite of dichloromethane, were similar in both the
maternal blood (167 nmol/mL, SEM 12) and fetal blood (160 nmol/mL, SEM 31). Withey and
Karpinski (1985) also reported higher maternal compared with fetal dichloromethane levels
based on a study of five pregnant Sprague-Dawley rats exposed to 107-2,961 ppm of
dichloromethane. Maternal blood levels of dichloromethane were 2-2.5-fold higher than those
found in the fetal circulation.
Blood-brain barrier transfer
Dichloromethane is thought to readily transfer across the blood-brain barrier, as
evidenced by the detection of radioactivity in brain tissue 48 hours after exposures of rats to
radiolabeled dichloromethane at concentrations of 50, 500, or 1,500 ppm for 6 hours (McKenna
et al., 1982) (see Table 3-1), and the historical demonstrations that dichloromethane has transient
sedative and anesthetic properties in humans (for review of these reports, see Mattsson et al.
[1990] and Winneke [1974]). Dichloromethane is no longer used as an anesthetic gas because
the margin between anesthetic and lethal doses is narrow (Winneke, 1974).
3.3. METABOLISM
Metabolism of dichloromethane involves two primary pathways, outlined in Figure 3-1
(Agency for Toxic Substances and Disease Registry [ATSDR], 2000; Guengerich, 1997; Hashmi
et al., 1994; Gargas et al., 1986). Dichloromethane is metabolized to CO in a cytochrome
P450 (CYP)-dependent oxidative pathway that is predominant at low exposure levels. The CYP-
related pathway results in the addition of oxygen, followed by spontaneous rearrangement to
formyl chloride, and then to CO; each spontaneous rearrangement releases H+ and Cl~ ions. At
higher exposure levels, the CYP pathway becomes saturated and a second pathway begins to
predominate. Glutathione S-transferase (GST)-catalyzed addition of glutathione (GSH) is the
initial step in this pathway. The replacement of one of the chlorine atoms with the S-glutathione
group results in formation of S-(chloromethyl)glutathione and the release of H+ and CI ions.
Hydration of S-(chloromethyl)glutathione results in an S-glutathionyl methanol molecule, which
can spontaneously form formaldehyde or rearrange to form an S-glutathione formaldehyde
molecule, and then further rearrange to formate. Both formaldehyde and formate can then be
further metabolized to C02.
9
DRAFT - DO NOT CITE OR QUOTE

-------
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
CI
H-C-H
i
GS
S-(chloromethyl)
glutathione
GSTT
Dichloromethane
91
H-C-H
OH
H-C-H
I
GS
S-glutathionyl methanol
o
II
/Cs
hTc-h
Formaldehyde
I
I
CO,
G-S
H
O
.1
CI
JCYP2K1
Formic acid
co9
OH H
O
, JL,
cr h
Formyl Chloride
1
(minor pathway) q
G - S VH
CO,
CO
Carbon Monoxide
I
COHb
Carboxyhemoglobin
Figure 3-1. Proposed pathways for dichloromethane metabolism.
Adapted from: ATSDR (2000); Guengerich (1997); Hashmi et al. (1994); Gargas et al. (1986).
As described in the following discussion of the two pathways, a metabolic balance
appears to exist between them, with the CYP pathway tending to be relatively more active at
lower doses and the GST pathway metabolizing the majority of a dichloromethane dose at higher
exposure levels once the CYP pathway has become saturated. Exposure to other agents may
shift this balance. For example, pretreatment with compounds that deplete GSH (e.g., buthionine
sulfoximine, diethylmaleate, phorone) resulted in an increase in blood carboxyhemoglobin
(COHb) levels, following a single injection of dichloromethane, relative to animals that did not
receive GSH depletion, indicating a shift to the CYP pathway (Oh et al., 2002). Similarly, co-
exposure to agents that compete for CYP2E1 results in a shift toward the GST pathway and away
from CO production (Lehnebach et al., 1995; Pankow and Jagielki, 1993; Pankow et al.,
1991a, b; Glatzel et al., 1987; Roth et al., 1975).
10
DRAFT - DO NOT CITE OR QUOTE

-------
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
3.3.1. The CYP2E1 Pathway
There is considerable evidence of the importance of the CYP2E1 metabolic pathway in
studies in animals (Oh et al., 2002; Wirkner et al., 1997; Kim and Kim, 1996; Lehnebach et al.,
1995; Pankow et al., 1991a, b; Pankow and Hoffmann, 1989; Pankow, 1988; Glatzel et al., 1987;
Angelo et al., 1986a, b; Landry et al., 1983; Anders and Sunram, 1982; McKenna et al., 1982;
McKenna and Zempel, 1981; Rodkey and Collison, 1977; Carlsson and Hultengren, 1975; Roth
et al., 1975; Fodor et al., 1973) and humans (Takeshita et al., 2000; DiVincenzo and Kaplan,
1981; Astrand et al., 1975). These studies demonstrate that exposure to dichloromethane,
regardless of exposure route, results in the formation of CO, as assessed by direct measurements
of elevated levels of CO in expired air and increased levels of COHb in the blood.
The first step in the CYP2E1 pathway is the formation of formyl chloride (Figure 1).
Watanabe and Guengerich (2006) conducted a series of studies to investigate the downstream
metabolites of formyl chloride, and reported only marginal (3% maximum at pH 9) formation of
^-formyl GSH from formyl chloride in the presence of GSH. Therefore, most (>97%) of the
formyl chloride is metabolized further to carbon monoxide. Furthermore, CO formation from
formyl chloride was independent of GSH presence in the assay.
Results from numerous studies in rats in which CYP2E1 metabolism was blocked or
induced indicate that the generation of CO occurs as a result of metabolism of dichloromethane
by the CYP2E1 pathway (Figure 3-1). Co-exposure of rats to a high dose of ethanol
(174 mmol/kg), which is metabolized by CYP2E1, and dichloromethane (1.6, 6.2, 15.6 mmol/kg)
resulted in no increase in blood COHb, indicating that the metabolic pathway for CO formation
had been either blocked or saturated (Glatzel et al., 1987). Similar results have been seen with
coadministration of other known CYP substrates, including diethyldithiocarbamate (Lehnebach
et al., 1995), methanol (Pankow and Jagielki, 1993), benzene, toluene, and three xylene isomers
(Pankow et al., 1991b). Pretreatment of animals with CYP inducers (e.g., benzene, toluene,
xylenes, methanol, isoniazid), particularly those that induce CYP2E1, resulted in an increased
level of CO formation, as assessed by COHb formation or measurement in expired air, following
single exposures to dichloromethane (Kim and Kim, 1996; Pankow and Jagielki, 1993; Pankow
et al., 1991b; Pankow and Hoffmann, 1989; Pankow, 1988). Pretreatment with disulfuram, a
CYP2E1 blocker, resulted in a complete lack of formation of COHb following dichloromethane
exposure, indicating that CYP2E1 is the isozyme responsible for metabolism of dichloromethane
(Kim and Kim, 1996).
Evidence in hamster and rat studies suggests that the CYP2E1 pathway becomes
saturated at high dichloromethane exposure levels; comparable data from studies in mice were
not found. In hamsters, mean COHb percentages were elevated to a similar degree (about 28-
30%, compared with <1% in controls) in three groups exposed by inhalation to 500, 1,500, or
3,500 ppm dichloromethane for 6 hours (Burek et al., 1984). After 21 months of exposure by
this protocol, mean COHb percentages in the three exposure groups remained similarly elevated,
11	DRAFT - DO NOT CITE OR QUOTE

-------
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
indicative of saturation of the CYP2E1 pathway in hamsters at exposure levels >500 ppm and a
lack of accumulation of dichloromethane and CYP2E1 metabolites with chronic exposure.
McKenna et al. (1982) found that blood COHb levels in rats increased when inhalation exposure
concentration was increased from 50 to 500 ppm but that similar levels of COHb were reported
following exposure to 1,500 ppm as following exposure to 500 ppm; the peak blood COHb
percentages were approximately 10%. In rats exposed to 0, 50, 200, or 500 ppm for 6 hours/day,
5 days/week for 2 years, mean COHb percentages were 2.2, 6.5, 12.5, and 13.7%, respectively,
suggesting that saturation of the CYP2E1 pathway is approached at 200 ppm (Nitschke et al.,
1988a). In male F344 rats exposed for 4 hours to dichloromethane concentrations of about 150,
300, 600, 1,000, and 2,000 ppm, mean COHb percentages (estimated from a figure) were about
4% at 150 ppm and about 8% at each of the four higher exposure concentrations (Gargas et al.,
1986). McKenna and Zempel (1981) reported that increasing the oral dose of labeled
dichloromethane from 1 mg/kg to 50 mg/kg in rats resulted in a lower fraction of the total dose
being metabolized to CO. Single injections of 3 and 6 mmol/kg of dichloromethane in rats
resulted in nearly identical levels of blood COHb (Oh et al., 2002).
In human subjects exposed to dichloromethane in the workplace, saturation of CYP
metabolism appears to be approached in the 400-500 ppm range (Ott et al., 1983e). Blood
samples were drawn during working hours from 136 fiber production workers who were exposed
to dichloromethane, acetone, and methanol. TWA exposure concentrations for the workers were
determined by personal monitoring techniques, and percent COHb levels in the blood samples
were determined. Estimated TWA concentrations in the exposed workers showed a bimodal
distribution. The lower mode of exposure concentrations showed the highest frequency in the
150-200 ppm range, while the higher mode showed the highest frequency in the range of 450-
500 ppm. Plots of percent COHb against TWA exposure concentrations showed that saturation
begins to be apparent in the 400-500 ppm range of exposure concentrations.
The liver is the tissue most enriched in CYP2E1 catalytic activity, but CYP2E1 protein
and messenger ribonucleic acid (mRNA) have been detected in other human tissues, including
the lung, brain, kidney, pancreas, bladder, small intestine, and blood lymphocytes (Nishimura et
al., 2003). As such, the liver is expected to be the main site of CYP metabolism of
dichloromethane, but other tissues are also expected to metabolize dichloromethane via this
pathway. Of particular relevance given the neurologic effects seen with dichloromethane are the
distribution and inducibility of CYP2E1 in different areas of the brain (Miksys and Tyndale,
2004). Individuals with decreased CYP2E1 activity may experience decreased generation of CO
and an increased level of GST-related metabolites following exposure to dichloromethane. As a
result, these individuals may be more susceptible to the chronic effects of dichloromethane from
GST-related metabolites than individuals with higher levels of CYP2E1 activity. Conversely,
individuals with higher CYP2E1 activity may experience relatively increased generation of CO
12
DRAFT - DO NOT CITE OR QUOTE

-------
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
at a given dichloromethane exposure level and, therefore, may be more susceptible to the acute
toxicity of dichloromethane (from CO).
Results from studies examining human interindividual variation in CYP2E1 activities
(e.g., catalytic activities, protein levels, or mRNA levels) indicate that individuals may vary in
their ability to metabolize dichloromethane through the CYP2E1 pathway. In a study of liver
samples from 30 Japanese and 30 Caucasian individuals, two- to threefold variation was found in
the levels of CYP2E1 protein, whereas catalytic activity toward substrates associated with
CYP2E1 (e.g., 7-ethoxycoumarin) displayed a wider range of values, approximately 25-fold; no
clear gender-specific or ethnic differences were found in hepatic levels of CYP2E1 protein or
enzymatic activities associated with CYP2E1 (Shimada et al., 1994). In a study of
interindividual variation in 70 healthy human subjects (40 men and 30 women) given an oral
dose of chlorzoxazone, a therapeutic agent whose metabolism and blood clearance has been
related to CYP2E1 levels, a three- to fourfold range in plasma half-life and clearance values was
observed, with no clear or dramatic age- or gender-specific differences (Kim et al., 1995). A six-
to sevenfold range in chlorzoxazone hydroxylation activity was reported for a group of
69 healthy, smoking and nonsmoking male and female volunteers with mixed ethnic
backgrounds; the range was markedly increased when a group of 72 alcoholic inpatients was
included (Lucas et al., 1999). In studies of human liver microsomes, four- to sixfold ranges in
CYP2E1-dependent oxidation of trichloroethylene have been reported (Lipscomb et al., 2003,
1997). CYP2E1 protein levels in 50 specimens of human lymphocytes from healthy individuals
showed an approximate fivefold range (Bernauer et al., 2000), and a 3.7-fold range in liver
CYP2E1 mRNA levels was reported for a group of 24 patients with chronic hepatitis (Haufroid
et al., 2003). More recently, a threefold range was reported for maximal rates of hepatic
CYP2E1-catalyzed metabolism of dichloromethane, which were estimated with a modified
physiologically based toxicokinetic (PBTK) model originally developed by Andersen et al.
(1987) and kinetic data (e.g., dichloromethane breath and blood concentrations) for 13 volunteers
(10 males and 3 females) exposed to one or more concentrations of dichloromethane by
inhalation for 7.5 hours (Sweeney et al., 2004). In summary, most studies indicate a three- to
sevenfold variability in CYP2E1 activity, as assessed by various types of measurements, among
"healthy" volunteers. However, various clinical factors (i.e., obesity, alcoholism, use of specific
medications) or co-exposures (i.e., to various solvents) (Lucas et al., 1999) may result in greater
variation, and thus the potential for saturation at lower exposures, within the general population.
Several genetic polymorphisms for the human CYP2E1 gene have been described, but
clear and consistent correlations with interindividual variation in CYP2E1 protein levels or
associated enzyme activities have not been identified (Ingelman-Sundberg, 2004; Lucas et al.,
2001; Kim et al., 1995; Shimada et al., 1994). The most frequently studied CYP2E1
polymorphisms, Rsal/PstI, are located in the 5'-flanking region of the gene, and mutations are
thought to lead to increased CYP2E1 protein expression via transcription (Lucas et al., 2001).
13	DRAFT - DO NOT CITE OR QUOTE

-------
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
Available data indicate that the frequency of this polymorphism, as well as other CYP2E1
polymorphisms, varies among ethnic groups. For example, Stephens et al. (1994) examined
blood samples from 126 African-Americans, 449 European Americans, and 120 Taiwanese
subjects and found frequencies for a rare Rsal allele (C2) of 0.01 in African-Americans, 0.04 in
European Americans, and 0.28 in Taiwanese subjects. In a study of 102 Mexicans, the reported
mutation frequency at the Rsal C2 allele was 0.30 (Mendoza-Cantu et al., 2004).
3.3.2. The GST Pathway
The other major pathway for dichloromethane metabolism involves the conjugation of
dichloromethane to GSH, catalyzed by GST. This results in the formation of a GSH conjugate
that is eventually metabolized to C02 (Figure 3-1). The conjugation of dichloromethane to GSH
results in formation of two reactive intermediates that have been proposed to be involved in
dichloromethane toxicity, S-(chloromethyl)glutathione and formaldehyde. In studies with rat,
mouse, and human liver cytosol preparations in the presence of GSH, examination of metabolites
with 13C-NMR indicated that S-(chloromethyl)glutathione was an intermediate in the pathway to
formaldehyde (Hashmi et al., 1994). Formaldehyde formation from dichloromethane has been
noted in human (Bruhn et al., 1998; Hallier et al., 1994; Hashmi et al., 1994), rat, and mouse
(Casanova et al., 1997; Hashmi et al., 1994) cells in vitro. Formation of free hydrogen ion is also
hypothesized, although no direct evidence supporting this has been presented.
The GST pathway has approximately a 10-fold lower affinity for dichloromethane than
the CYP pathway (Reitz et al., 1989; Andersen et al., 1987). At lower exposure concentrations,
the CYP pathway is expected to predominate, but, as exposure concentrations increase, the GST
pathway is expected to gain in relative importance as a dispositional pathway for absorbed
dichloromethane. Based on in vitro studies with liver preparations, the estimated Michaelis-
Menten kinetic constants (Kms) in GST assays with dichloromethane were about 137 mM in a
B6C3Fi mouse preparation and about 44 mM in two human preparations (Reitz et al., 1989). In
contrast, estimated Kms in CYP assays were about 1.8, 1.4, and 2.0 mM in B6C3Fi mouse, F344
rat, and Syrian golden hamster preparations, respectively. In four human liver preparations,
estimated CYP Kms were about 2.6, 2.0, 0.9, and 2.8 mM (Reitz et al., 1989).
Early investigations indicated that in humans GSTs of the a-, [j,-, and 7i-cl asses were not
responsible for the metabolism of dichloromethane (Bogaards et al., 1993). Tissue samples that
metabolized substrates specific to those GST classes did not conjugate dichloromethane to GSH.
Later investigations identified the recently-characterized GST theta class (Meyer et al., 1991),
specifically GST-thetal-1 (GST-T1), as the GST isoenzyme responsible for the metabolism of
dichloromethane (Mainwaring et al., 1996; Blocki et al., 1994). In the absence of the GST-T1
gene, no deoxyribonucleic acid (DNA)-protein cross-links were formed by human liver cells
exposed to dichloromethane (Casanova et al., 1997), and formaldehyde production was not
detected in human erythrocytes (Hallier et al., 1994). In a mouse model with a disrupted GST-
14	DRAFT - DO NOT CITE OR QUOTE

-------
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
T1 gene, GST activity with dichloromethane in liver and kidney cytosol samples was
substantially lower compared with wild-type GST mice (Fujimoto et al., 2007).
A polymorphism of the GST-T1 gene has been demonstrated in humans. People with
two functional copies of the gene (+/+) readily conjugate GSH to dichloromethane. Individuals
having only one working copy of the gene (+/-) display relatively decreased conjugation ability.
Individuals with no functional copy of the gene (-/-) do not express active GST-T1 protein and
do not metabolize dichloromethane via a GST-related pathway (Thier et al., 1998). Results from
studies of GST-T1 genotypes in human blood samples indicate that average prevalences of the
GST-T1 null (-/-) genotype are higher in Asian ethnic groups (47-64%) than in other groups,
including Caucasians (19-20%), African-Americans (22%), and mixed groups (19%) (Raimondi
et al., 2006; Garte et al., 2001; Nelson et al., 1995) (see Table 3-3). Although information on the
age distribution of study subjects was not generally reported in these analyses, there is little
reason to expect effect modification by age since this is not a gene linked to early mortality.
Based on data collected by Nelson et al. (1995) and U.S. 2000 census data (and assuming Hardy-
Weinberg equilibrium), Haber et al. (2002) calculated U.S. average distributions of GST-T1
genotypes as follows: 32% +/+; 48% +/-; and 20% -/-.
Table 3-3. Mean prevalences of the GST-T1 null (-/-) genotype in human
ethnic groups
Ethnic group

Reference

Nelson et al. (1995)a
Garte et al. (200l)b
Raimondi et al. (2006)°
Chinese
64.4% (n = 45)
Not reported
Not reported
Korean
60.2% (n = 103)
Not reported
Not reported
Caucasian
20.4% (n = 442)
19.7% (n= 5,577)
19.0% (n = 6,875)
Asian
Not reported
47.0% (n = 575)
53.6% (n= 1,727)
African-American
21.8% (n= 119)
Not reported
Not reported
Mexican American
9.7% (n = 73)
Not reported
Not reported
Other
Not reported
Not reported
19.4 %(n= 1,485)
aNelson et al. (1995) examined prevalence of the null GST-T1 genotype from analysis of blood samples from
subjects of various ethnicities as noted above.
bGarte et al. (2001) collected GST-T1 genotype data in Caucasian (29 studies; 5,577 subjects) and Asian (3 studies,
575 subjects) ethnic groups; subjects were controls in case-control studies of cancer and various polymorphisms in
genes for bioactivating enzymes.
°Raimondi et al. (2001) collected GST-T1 genotype data from 35 case-control studies of cancer and GST-T1
genotype; data in this table are for control subjects. The "other" group in this study is defined as Latino, African-
American, and mixed ethnicities.
Results from a study of the distribution of activity levels for in vitro conjugation of
dichloromethane with GSH in 22 human liver samples are roughly reflective of these estimates
of the distribution of this polymorphism (Bogaards et al., 1993). No activity was found in 3/22
of the liver samples. Eleven of the samples showed low activity levels (0.21-0.41 nmol
15
DRAFT - DO NOT CITE OR QUOTE

-------
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
product/minute/mg protein), and eight samples showed high activity levels ranging from 0.82 to
1.23 nmol/minute/mg protein. In another study of seven human subjects, lysates of erythrocytes
showed high activities for producing formaldehyde from dichloromethane (presumably via GST-
Tl) in three subjects (15.4, 17.7, and 17.8 nmol product/minute/mg hemoglobin) and lower
activity in the other four subjects (4.3, 6.0, 7.2, and 7.6 nmol product/minute/mg hemoglobin)
(Hallier et al., 1994).
Comparisons of mice, rats, humans, and hamsters for the ability to metabolize
dichloromethane via the GST pathway in liver and lung tissues indicate that mice appear to be
the most active at metabolizing dichloromethane (Sherratt et al., 2002; Thier et al., 1998;
Casanova et al., 1997, 1996; Hashmi et al., 1994; Reitz et al., 1989). Reitz et al. (1989) reported
mean (± SD) GST enzymatic activity levels with dichloromethane as substrate (in units of nmol
product formed/minute/mg protein) in liver cytosol preparations to be: 25.9 ± 4.2 units in
B6C3Fi mice (n = 15 determinations per preparation); 7.05 ±1.7 units in F344 rats (n = 6); and
1.27 ± 0.21 units (n = 6) in Syrian golden hamsters. Mean GST activity levels in liver
preparations from four human subjects (accident victims screened for human immunodeficiency
virus and hepatitis B and C and obtained through a transplant center) were 2.62 ± 0.44 units (n =
10), -0.01 ± 0.04 units (n = 6), 2.71 ± 0.45 units (n = 6), and 3.03 ± 0.44 units (n = 6) (Reitz et
al., 1989). The finding that one of the four individuals was unable to conjugate dichloromethane
with GST was reflective of the estimated frequency of the GST-T1 null genotype in the U.S.
population (see Table 3-3). Mean GST activity levels in lung cytosol preparations showed a
similar rank order among species: 7.3 ± 1.4 units in mice (n = 4), 1.0 ± 0.1 units in rats (n = 4),
0.0 ± 0.2 units in hamsters (n = 4), and 0.37 ± 0.25 units in a pooled lung preparation from the
same four human subjects (n = 2). Reitz et al. (1989) noted that relative abilities of these animal
species to metabolize dichloromethane via the GST pathway correlated with their cancer
sensitivities in long-term inhalation bioassays: (1) B6C3Fi mice showed statistically significant
increased incidence of liver and lung tumors in a 2-year cancer bioassay (National Toxicology
Program [NTP], 1986); (2) rats showed much less evidence of increased incidence of liver
tumors, and no increased risk of lung tumors at equivalent exposure concentrations but showed
increased incidence of nonmalignant mammary tumors (NTP, 1986; Burek et al., 1984); and
(3) Syrian golden hamsters did not show tumorigenic responses at any site (Burek et al., 1984).
Thier et al. (1998) conducted a study evaluating the activity of GST-T1 after treatment of
dichloromethane in the cytosol of liver and kidney homogenates from hamsters (pooled male and
females), rats (pooled male and female), male mice, and female mice and for humans classified
as nonconjugators, low conjugators, or high conjugators of GST to dichloromethane. Little
information is provided about the human samples other than that 13 kidney cancer patients were
the source of the kidney samples; normal tissue identified by pathological exam was used. Blood
samples from 10 of these patients were collected and enzyme activities measured in erythrocytes
from 9 of these samples were reported. Results of conjugation of dichloromethane to GSH from
16
DRAFT - DO NOT CITE OR QUOTE

-------
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
these studies are presented in Table 3-4. As can be seen from the table, activity levels (expressed
as nmol/minute per mg of cytosolic protein) of humans varied considerably, with nonconjugators
(presumed to be GST-T1 ) having no detectable activity, low conjugators (presumed to be
GST-T1 ) having moderate activity, and high conjugators (presumed to be GST-T1+/+) having
approximately twice the activity seen in low conjugators. In the liver, the activity of rat GST
conjugation was over twofold that seen in human high conjugators, while levels in mice were
>11-fold (males) or 18-fold (females) greater than those of human high conjugators. In the
kidney, the activity of high-conjugator humans was approximately 1.8-fold that of rats and was
comparable to the activity of both male and female mice. The data in Table 3-4 show the
following order for GST-T1 activities with dichloromethane as substrate: in liver preparations,
mouse » rat > human high conjugators > human low conjugators > hamster > human
nonconjugators and, in kidney preparations, female mouse ~ male mouse ~ human high
conjugators > rat ~ human low > hamster > human nonconjugators. In addition, the data indicate
that activity levels in liver, kidney, and erythrocytes of human subjects are in correspondence
with the nonconjugator, low conjugator, and high conjugator designations.
Table 3-4. GST-T1 enzyme activities toward dichloromethane in human,
rat, mouse, and hamster tissues (liver, kidney, and erythrocytes)

Activity (nmol/min per mg protein)3
Activity (nmol/min per mL)a

Liver
Kidney
Erythrocytes
Human, nonconjugators
Not detectable (2)
Not detectable (1)
Not detectable (1)
Human, low conjugators
0.62 ±0.30 (11)
1.38 ±0.52 (8)
9.67 ± 2.49 (5)
Human, high conjugators
1.60 ±0.48 (12)
3.05 ±0.72 (4)
18.28 ±0.46 (3)
Rat
3.71 ±0.28 (8)
1.71 ±0.28 (8)
Not measured
Mouse, male
18.2 ±2.22 (5)
3.19 ±0.46 (5)
Not measured
Mouse, female
29.7 ±6.31 (5)
3.88 ±0.90 (5)
Not measured
Hamster
0.27 ± 0.20 (6)
0.25 ±0.21 (6)
Not measured
aMean ± SD with number of samples noted in parentheses.
Source: Adapted from Thier et al. (1998).
Sherratt et al. (2002) reported that, on a per mg basis, native recombinant mouse GST-T1
(purified after expression in Escherichia coli) was approximately twofold more active toward
dichloromethane than native recombinant human enzyme, as well as being approximately
fivefold more efficient (as assessed by the ratio of kcat/Km).
The distribution of GST-T1 in human tissues has been examined with antibodies raised
against recombinant human GST-T1 (Sherratt et al., 2002, 1997). Immunoblotting of sodium
dodecyl sulfate polyacrylamide gel electrophoresis gels loaded with tissue extracts from a
73 year-old man who had died with brochopneumonia and atherosclerosis indicated the
17
DRAFT - DO NOT CITE OR QUOTE

-------
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
following order of expression of GST-T1: liver - kidney > prostate - small intestine >
cerebrum - pancreas - skeletal muscle > lung - spleen - heart - testis (Sherratt et al., 1997).
It was estimated that the levels of cross-reacting materials in the cerebrum, pancreas, or skeletal
muscle extracts were about 10% of those in the liver, whereas levels in the lung, spleen, heart,
and testis were less than 5% of the levels in the liver. Comparison of the amounts of cross-
reacting material in soluble liver extracts from a B6C3Fi mouse and five human subjects (i.e.,
normal liver tissue samples from biopsies of secondary liver tumors) found that levels of GST-
T1 protein were higher in the mouse extracts than in any of the human liver extracts (Sherratt et
al., 2002). Densitometer analysis indicated that the GST-T1 level in the mouse liver extract was
about fivefold higher than those in human liver extracts displaying the highest level. Cross-
reacting material was not detectable in liver extracts from one of the five human subjects,
indicating that this individual may have been GST-T1 null (Sherratt et al., 2002).
Results from in situ hybridization with oligonucleotide anti-sense probes for GST-T1
mRNA levels and immunohistochemical studies with antibodies to GST-T1 have indicated that
there may be subtle differences between mice and humans in the intracellular localization of
GST-T1 in the liver. Mainwaring et al. (1996) reported that staining for GST-T1 mRNA was
higher in liver slices from B6C3Fi mice than in liver slices from F344 rats and that staining in
human liver samples was very low. Although the number of mouse and rat liver samples
examined in this study was not indicated in the available report, it was reported that slices from
five human liver samples were examined. No information was provided regarding the clinical
history of the sources of the human samples. In mouse liver, staining for GST-T1 mRNA was
enhanced in the limiting plate hepatocytes, in nuclei, in bile-duct epithelial cells, and in lesser
amounts in the centrilobular cells in general. In rat liver, a similar pattern was observed, except
no enhanced staining was observed in the limiting plate hepatocytes or in nuclei. Staining for
GST-T1 mRNA in the human liver samples showed an even distribution throughout the liver
lobule, and no mention of a specific nuclear localization was made (Mainwaring et al., 1996).
Quondamatteo et al. (1998), using antibodies to GST-T1, subsequently reported a similar
localization of GST-T1 protein in nuclei of cells in mouse liver slices. In another study using
antibodies raised against recombinant human GST-T1 or a peptide derived from the deduced
mouse GST-T1 primary sequence, Sherratt et al. (2002) reported that nuclear staining was
observed in all cells in mouse liver slices (from five individual B6C3Fi mice) showing the
presence of mouse GST-T1; staining in the cytoplasm was only detected in cells with very high
levels of GST-T1. In liver slices obtained from two human subjects (males, ages 60 and
61 years, with a secondary liver tumor and what was described as a "cavernous hemangioma"
without malignancy, respectively), the most intense nuclear staining was associated with bile
duct epithelial cells, but there was heterogeneity of staining within hepatocytes; some cells
showed nuclear staining, but others only exhibited cytoplasmic staining (Sherratt et al., 2002).
18
DRAFT - DO NOT CITE OR QUOTE

-------
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
In summary, the relative amount of dichloromethane metabolized via the GST pathway
increases with increasing exposure concentrations. As the high affinity CYP pathway becomes
saturated (either from high exposure levels of genetic or other factors that decrease CYP2E1
activity), the GST pathway increases in relative importance as a dispositional pathway for
dichloromethane. Two reactive metabolites (S-(chloromethyl)glutathione and formaldehyde)
resulting from this pathway have been identified. GST-T1 is the GST isozyme that catalyzes
conjugation of dichloromethane with GST. Interindividual variation in the ability to metabolize
dichloromethane via GST-T1 is associated with genetic polymorphisms in humans. Estimated
U.S. population prevalence of nonconjugators (-/- at the GST-T1 locus) is about 20%, but higher
prevalences (47-64%) have been reported for Asians (Raimondi et al., 2006; Haber et al., 2002;
Garte et al., 2001; Nelson et al., 1995). The prevalences for low (+/- at the GST-T1 locus) and
high (+/+) conjugators have been estimated at 48 and 32%, respectively (Haber et al., 2002).
The liver and kidney are the most enriched tissues in GST-T1, but evidence is available for the
presence of GST-T1 in other tissues at lower levels, including the brain and lung. In humans,
GST-T1 expression in the brain is lower than that seen in the liver or kidney but higher than in
the lung. Comparisons of mice, rats, humans, and hamsters for the ability to metabolize
dichloromethane via the GST pathway in liver (based on measurement of tissue-specific enzyme
activity) indicate the following rank order: mice > rats > or ~ humans > hamsters. This relative
ranking corresponds to the rank order of the strength of the association between inhalation
exposure to dichloromethane and liver tumors in long-term cancer bioassays with mice, rats, and
hamsters. In mouse liver tissue, GST-T1 appears to be localized in the nuclei of hepatocytes and
bile-duct epithelium, but rat liver does not show preferential nuclear localization of GST-T1. In
human liver tissue, some hepatocytes show nuclear localization of GST-T1 and others show
localization in cytoplasm, as well as in bile duct epithelial cells. The apparent species
differences in intracellular localization of GST-T1 may play a role in species differences in
susceptibility to dichloromethane carcinogenicity if nuclear production of
S-(chloromethyl)glutathione is more likely to lead to DNA alkylation than cytoplasmic
production.
3.4. ELIMINATION
Dichloromethane is eliminated mainly through exhalation either of the parent compound
or as the two primary metabolites CO2 and CO (Angelo et al., 1986a, b; McKenna et al., 1982;
DiVincenzo and Kaplan, 1981; DiVincenzo et al., 1972, 1971). In human studies,
dichloromethane is rapidly eliminated from the body following the cessation of exposure, with
much of the parent compound completely removed from the bloodstream and expired air by
5 hours postexposure in experiments using exposure levels of 90, 100, or 210 ppm (DiVincenzo
et al., 1972, 1971; Riley et al., 1966). Studies in rats have similarly demonstrated that
elimination from the blood is rapid, with elimination half-times in F344 rats on the order of 4-
19	DRAFT - DO NOT CITE OR QUOTE

-------
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
6 minutes following intravenous doses in the range of 10-50 mg/kg (Angelo et al., 1986a). In a
study using Sprague-Dawley rats, Carlsson and Hultengren (1975) demonstrated variability in
elimination rates between different types of tissues, with the most rapid elimination seen in the
adipose and brain tissue, while elimination from liver, kidneys, and adrenals proceeded more
slowly.
In a study using human volunteers, DiVincenzo and Kaplan (1981) reported a dose-
related increase in CO in the expired breath after inhalation exposure to 50-200 ppm of
dichloromethane, with a net elimination as CO on the order of 25-35% of the absorbed dose.
Similar results have been reported in animal studies. Following gavage administration of 50 or
200 mg/kg-day doses of [14C]-labeled dichloromethane in water to groups of six mature male
F344 rats for up to 14 days, >90% of the label was recovered in the expired air within 24 hours
of dose administration (Angelo et al., 1986b). Following administration of the first of 14 daily
50 mg/kg-day doses, radioactivity in parent compound, CO2, and CO in the 24-hour expired
breath accounted for 66, 17, and 16% of the administered radioactivity, respectively; similar
patterns were reported for 24-hour periods following administration of the seventh and
fourteenth 50 mg/kg-day dose. Following administration of the first 200 mg/kg-day dose,
radioactivity in parent compound, CO2, and CO in the 24-hour expired breath accounted for 77,
9, and 6%, respectively, of the administered radioactivity (Angelo et al., 1986b). In mature, male
Sprague-Dawley rats given a smaller dose (1 mg/kg) of [14C]-labeled dichloromethane,
radioactivity in parent compound, CO2, and CO in 48-hour expired breath accounted for 12, 35,
and 31%, respectively; these data indicate that, at lower dose levels, a greater percentage of the
administered dose was metabolized by the CYP pathway and eliminated in the expired breath,
compared with higher dose levels (McKenna and Zempel, 1981). Similar patterns of
radioactivity distribution in parent compound, CO2, and CO in expired breath were found in
mature male B6C3Fi mice following gavage administration of 50 mg/kg-day (in water), or
500 or 1,000 mg/kg-day (in corn oil), [14C]-labeled dichloromethane (Angelo et al., 1986a). For
example, radioactivity in parent compound, CO2, and CO in 24-hour expired breath accounted
for 61, 18, and 11% of the administered radioactivity, following administration of a single
50 mg/kg dose to a group of six mice (Angelo et al., 1986a). Exhalation rates were similarly
high following inhalation exposure of mature male Sprague-Dawley rats (>90%) (McKenna et
al., 1982) or following intravenous administration of dichloromethane to mature male F344 rats
(Angelo et al., 1986b).
Elimination of dichloromethane in the urine of exposed humans is generally small, with
total urinary dichloromethane levels on the order of 20-25 jag or 65-100 |ig in 24 hours
following a 2-hour inhalation exposure to 100 or 200 ppm, respectively (DiVincenzo et al.,
1972). However, a direct correlation between urinary dichloromethane and dichloromethane
exposure levels was found in volunteers, despite the comparatively small urinary elimination
(Sakai et al., 2002). Following administration of a labeled dose in animals, regardless of
20
DRAFT - DO NOT CITE OR QUOTE

-------
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
exposure route, generally <5-8% of the label is found in the urine and <2% in the feces
(McKenna et al., 1982; McKenna and Zempel, 1981; DiVincenzo et al., 1972, 1971).
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS
Several PBTK models for dichloromethane in animals and humans have been developed
from 1986 to 2006. These models are mathematical representations of the body and its
absorption, distribution, metabolism, and elimination of dichloromethane and select metabolites,
based on the structure of the Ramsey and Andersen (1984) model for styrene. The models'
equations are designed to mimic actual biological behavior of dichloromethane, incorporating in
vitro and in vivo data to define physiological and metabolic equation parameters. As such, the
models can simulate animal or human dichloromethane exposures and predict a variety of
dichloromethane and metabolite internal dosimeters (i.e., instantaneous blood and tissue
concentration, area under the curve [AUC] of concentration versus time plots, rate of metabolite
formation), allowing for the extrapolation of toxicity data across species, route of exposure, and
high to low exposure levels. The development of dichloromethane PBTK models has resulted in
either increased biological detail and functionality or refinement of model parameters with newly
available data. The former type of development provides more options for toxicity data
extrapolation, while the latter serves to increase confidence in model predictions and decrease
uncertainty in risk assessments for which the models were, or will be, applied. This section of
the document describes each of the models reported in the scientific literature and/or used by the
regulatory community (i.e., Occupational Safety and Health Administration [OSHA], EPA) and
their contribution to the advancement of predictive dosimetry and data extrapolation for
dichloromethane. In some instances, model development was accomplished by the addition of
new biological compartments (e.g., tissue systems). Diagrams of the compartmental structure of
the models are shown in Figure 3-2. Significant statistical advances in parameter estimation also
have been incorporated in model development. For this reason, some animal and human PBTK
models may be described as deterministic (Sweeney et al., 2004; Casanova et al., 1996; Reitz et
al., 1988a, b; U.S. EPA, 1988b, 1987a, b; Andersen et al., 1987; Gargas et al., 1986) in which
point estimates for each model parameter are used, resulting in point estimates for dosimetry.
Others may be described as probabilistic (Jonsson and Johanson, 2001; El-Masri et al., 1999;
OSHA, 1997), in which probability distributions for each parameter were defined, resulting in
probability distributions for dosimetry. The latter approach, particularly utilizing a Bayesian
hierarchical statistical model structure (described below) (David et al., 2006; Marino et al., 2006)
to estimate parameter values, allows for the introduction of intra- and interspecies variability into
model predictions and quantitative assessment of model uncertainty. Both deterministic (U.S.
EPA, 1988b, 1987a, b) and probabilistic (OSHA, 1997) applications have been used to develop
regulatory values. As discussed below, subsequent applications of the developed models for
cancer risk assessment have resulted in significantly different estimates of human cancer risk.
21	DRAFT - DO NOT CITE OR QUOTE

-------
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
3-2A
J__L
3-2 B
Gas
Exchange
Richly
Perfused
Slowly
Perfused
CO Sub
Model
1 t
Alveolar
Air
i t
CYP 	1
3-2C
Endogenous
Production

11 GsTnr
Gas

Lung
Exchange

Richly
Perfused
Slowly
Perfused

I t -X
Gas

Lung
Exchange

Richly
Perfused
Slowly
Perfused

CO Sub
Model
1 t
Alveolar
Air
1 t
i t GST^ir
Gas

Lung
Exchange

3-2 E
Richly
Perfused
Slowly
Perfused
Endogenous
Production
GST
11 ,s,nr
Gas
Exchange
Lung
Richly
Perfused
Slowly
Perfused
DC

3-2F
11 GsTnc
Gas

Lung
Exchange

Formaldehyde
I
_ DN Aprotein crosdinks
(mg) I t
Bone
Marrow
Richly
Perfused
Slowly
Perfused

Gas

Lung
Exchange

Richly
Perfused
Perineral
Fat
Subcutaneous
Fat
Working
Muscle
Resting
Muscle

Figure 3-2. Schematics of PBTK models (1986-2006) used in the
development of estimates for dichloromethane internal dosimetry.
Key references: Model A—Gargas et al. (1986); B—Andersen et al. (1987); C—Andersen et al.
(1991); D—Casanova et al. (1996); E—Sweeney et al. (2004); F—OSHA (1997); G—Jonsson
and Johanson (2001). Models C-G all build on the structure in model B. Models E and G have
been applied in humans; all others have been applied in humans and rodents (mice and/or rats).
CYP = CYP pathway metabolites; GST = GST pathway metabolites.
22
DRAFT - DO NOT CITE OR QUOTE

-------
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
The deterministic rat model of Gargas et al. (1986), based on previous work by Ramsey
and Andersen (1984) examining inhalation pharmacokinetics of styrene in rats, was the first
PBTK model for dichloromethane. It was comprised of four compartments (fat, liver, richly
perfused tissues, and slowly perfused tissues [Figure 3-2A]) and described flows and partitioning
of parent material and metabolites through the compartments with differential equations.
Metabolism, which was restricted to the liver compartment, was described as two competing
pathways: the GST pathway, described with a linear first-order kinetic model, and the CYP
pathway, described with a saturable Michaelis-Menten kinetic model. Rate constants for the
CYP and GST pathways in rats were determined by optimization of the model with in vivo gas
uptake data. COHb production was modeled both endogenously and from CYP-mediated
metabolism of dichloromethane. This model demonstrated the dose-dependent flux through the
competing CYP and GST metabolic pathways and the effect of CYP inhibition on COHb
generation.
Andersen et al. (1987) extended the rat model of Gargas et al. (1986) to include a lung
compartment, including CYP and GST metabolism pathways within the lung, in rats, mice,
hamsters, and humans (Figure 3-2B). Physiological flow rates were allometrically scaled among
species by % power of body weight (BW). Rate constants for the CYP and GST pathways in
rodents were determined by optimization of the model with in vivo gas uptake data. CYP rate
constants for humans were derived from data on dichloromethane uptake in human subjects
(number of subjects not reported). Human GST rate constants were derived by allometric
scaling of the animal GST rate constants. Model predictions compared favorably with kinetic
data for human subjects exposed by inhalation to dichloromethane (Andersen et al., 1987).
Using the mouse cancer bioassay data from NTP (1986), Andersen et al. (1987) compared the
linear body surface area-derived or the PBTK model-derived human liver and lung dose
surrogates associated with tumor development (mg dichloromethane metabolized via GST
pathway/volume tissue/day). They reported that PBTK model-extrapolated human liver and lung
internal doses were for inhalation exposure 167- and 144-fold lower and for drinking water
exposure 45- and 213-fold lower, respectively, than body surface area scaled internal doses. The
study authors suggested that the lower model-predicted human internal dose surrogates were due
to the need to saturate the CYP pathway before appreciable tumorigenic metabolite levels could
be attained, which is not captured by extrapolation based on body surface area.
U.S. EPA (1988b, 1987a, b) slightly modified the Andersen et al. (1987) model for mice
by using different alveolar ventilation and cardiac flow rates and used the mouse and human
models to derive human cancer risks from animal tumor incidence data. The flow rate
parameters in the Andersen et al. (1987) model were based on a human breathing rate of
12.5 m3/day (reflecting a resting rate), compared with the EPA value of 20 m3/day (reflecting
average daily activity level) and a mouse breathing rate of 0.084 rnVday (based on allometric
scaling of bioassay-specific BWs), compared with the rate commonly used by EPA,
23	DRAFT - DO NOT CITE OR QUOTE

-------
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
0.043 m3/day (U.S. EPA, 1987a). The internal dose metric used in the applications of the model
to cancer risk assessment was reflective of the amount of dichloromethane metabolized by the
GST pathway. In addition to using the mouse and human PBTK models to account for species
differences in dosimetry, a body surface area correction factor of 12.7 was applied to low-dose
slopes of estimated dose-response relationships for liver and lung tumors in mice to account for
presumed higher human responsiveness, relative to mice, to dichloromethane-induced cancer
(U.S. EPA, 1987a). The factor of 12.7 is the cube root of the ratio of human to mouse reference
BWs; this BW scaling factor was applied to adjust for interspecies toxicodynamic variability
(i.e., presumed differences in the lifetime impact in mice and humans of a given daily amount of
dichloromethane metabolically activated per liter of tissue) (Rhomberg, 1995). A human cancer
inhalation unit risk (IUR) of 4.7 x 10 7 per (|ig/m3), based on this analysis, was placed on IRIS in
September 1990.
The Andersen et al. (1987) models were also modified by addition of submodel structures
for estimation of new dosimeters of interest. Andersen et al. (1991) added the capability to
specifically describe the kinetics of dichloromethane, CO, and COHb in rats and humans with
the addition of the Coburn-Forster-Kane equation to describe CO and COHb kinetics
(Figure 3-2C). However, equations were not added for metabolism of dichloromethane to CO in
the lung. Casanova et al. (1996) extended the Andersen et al. (1987) mouse model to include a
submodel that predicted the formation of formaldehyde and DNA-protein cross-links in the liver
(Figure 3-2D).
Further refinements of the Andersen et al. (1987) models allowed for incorporation of
new data. New in vitro measurements of metabolic rate constants in human and animal tissues
were incorporated into the Andersen et al. (1987) models by Reitz and coworkers (Reitz, 1991;
Reitz et al. 1988a, b). Sweeney et al. (2004) modified the Andersen et al. (1987) human PBTK
model, adding extrahepatic CYP metabolism in richly perfused tissues (Figure 3-2E) to obtain a
better fit of the model to kinetics data for humans. Data for 13 volunteers (10 men and
3 women) who were exposed to one or more concentrations of dichloromethane for 7.5 hours
included dichloromethane concentrations in breath and blood, COHb concentrations in blood,
and CO concentrations in exhaled breath. Individual CYP Vmaxc (maximal velocity) values were
obtained by optimizing model predictions to match time-course data simultaneously for
dichloromethane concentrations in blood and exhaled breath for each individual. Resultant
individual values of CYP Vmaxc ranged from 7.4 to 23.6 mg/hour/kg0'7, indicating an
approximate threefold range in maximal CYP metabolic activity.
The significance of metabolic variability for the kinetics of dichloromethane in animals
and humans was explored by several investigators using PBTK models. Dankovic and Bailer
(1994) used the updated human model presented by Reitz et al. (1988a, b) to explore the
consequences of interindividual variability in vitro kinetic constants for the CYP and GST
pathways (based on data for four human subjects) and reported that predicted GST-metabolized
24	DRAFT - DO NOT CITE OR QUOTE

-------
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
doses to the lung and liver could range from about zero to up to fivefold greater than those
predicted with the values of these rate constants used in the Reitz et al. (1988a, b) model.
El-Masri et al. (1999) replaced parameter estimates in the mouse and human PBTK models
presented by Casanova et al. (1996) with probability distributions, including published
information on the distribution of GST-T1 polymorphism in human populations, and used Monte
Carlo simulations to estimate distributions of cancer potency of dichloromethane in mice,
distributions of the amount of DNA-protein cross-links formed in the liver of humans, and
distributions of human cancer risks at given exposure levels of dichloromethane. The analysis
showed that, at exposure levels of 1, 10, 100, and 1,000 ppm dichloromethane, average and
median cancer risk estimates were 23-30% higher when GST-T1 polymorphism was not
included in the model.
Given the demonstrated influence of population variability in dichloromethane
metabolism on PBTK model-derived cancer risk estimates (El-Masri et al., 1999; Dankovic and
Bailer, 1994), PBTK model development has included a more formal statistical treatment of data
for physiological and metabolic variability. Bayesian statistical approaches have been applied to
develop probabilistic PBTK models for dichloromethane. Probabilistic models account for
variability between individuals in model parameters by replacing point estimates for the model
parameters with probability distributions. Calibration or fitting of probabilistic PBTK models to
experimental toxicokinetic data is facilitated by a Bayesian technique called Markov Chain
Monte Carlo (MCMC) simulation, which quantitatively addresses both variability and
uncertainty in PBTK modeling (Jonsson and Johanson, 2003).
OSHA (1997) used MCMC simulation to fit probabilistic versions of the Reitz et al.
(1988a, b) and Andersen et al. (1991, 1987) mouse and human models, which included
probability distributions for all model parameters. GST- and CYP-mediated metabolism
occurred in the liver and lung compartments (see Figure 3-2F). The model parameters were
modified to focus on occupational exposure scenarios; that is, a parameter distribution for work
intensity (using data from Astrand [1989]) was added, which adjusted physiological flow rates as
a function of work intensity as measured in watts. In addition, updated measurements of
blood:air and tissue:air partition coefficients (Clewell et al., 1993) were used to describe
distributions for these parameters. The Clewell et al. (1993) blood:air partition coefficient of 23
is higher than the value of 8.29 reported by Andersen et al. (1987) and used by EPA (U.S. EPA,
1988b, 1987a, b). The newer Clewell et al. (1993) value for mice is the preferred value, since it
is much closer to the values for rats (19.4) and hamsters (22.5) rather than humans (9.7), as
reported by Andersen et al. (1987). Distributions of metabolic, physiological, and partitioning
parameters in the mouse and human models were updated by using Bayesian methods with data
for mice and humans in published studies of mouse and human physiology and dichloromethane
kinetic behavior.
25
DRAFT - DO NOT CITE OR QUOTE

-------
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
Jonsson et al. (2001) used additional human kinetics data to expand the PBTK model of
Reitz et al. (1988a, b) and added new model compartments (Figure 3-2G). These investigators
used MCMC simulation to develop a probabilistic model from the Reitz et al. (1988a, b) human
model by using published in vitro measurements of liver Vmax for the CYP pathway (Reitz et al.,
1989) and kinetic data for five human subjects exposed by inhalation to dichloromethane
(Astrand et al., 1975). A working muscle compartment was added to the basic Andersen et al.
(1987) and Reitz et al. (1988a, b) structure (see Figure 3-2G). Jonsson and Johanson (2001)
refined and extended this probabilistic model by including an additional fat compartment (to
provide a better description of the experimental data for the time course of dichloromethane in
subcutaneous fat), incorporating (with MCMC simulation) kinetic data for dichloromethane in an
additional 21 human subjects and including three GST-T1 genotypes/phenotypes
(nonconjugators -/-, low conjugators +/-, high conjugators +/+). Monte Carlo simulations were
then used with the refined probabilistic model to predict human liver cancer risk estimates at
several dichloromethane exposure levels using an algorithm similar to the one used by El-Masri
et al. (1999), using DNA-protein cross-links as the internal dose metric. The mean, 50th, 90th,
and 95th percentile human cancer risk values from Jonsson et al. (2001) and El-Masri et al.
(1999) were very similar, within onefold of one another for simulated exposure levels up to
100 ppm.
The most statistically rigorous and data-intensive PBTK model development was
performed by Marino et al. (2006) for mice and David et al. (2006) for humans. Development of
these models used multiple mouse and human data sets in a Bayesian hierarchical statistical
structure to quantitatively capture population variability and reduce uncertainty in model
dosimetry and the resulting risk values. EPA used these models in the derivation of reference
values and cancer risk estimates in the current assessment, and these models are described in
more detail below.
3.5.1. Probabilistic Mouse PBTK Dichloromethane Model (Marino et al., 2006)
Marino et al. (2006) used MCMC analysis to develop a probabilistic PBTK model for
dichloromethane in mice, using the Andersen et al. (1987) model structure as a starting point
(Figure 3-3). Metabolic kinetic parameters (Vmaxc, Km, kfC, Al, and A2) (Table 3-5) were
calibrated with this Bayesian methodology by using several experimental data sets. Distribution
parameters (i.e., means and coefficients of variation [CVs]) for other physiological parameters
(i.e., BW, fractional flow rates, and fractional tissue volumes) and partition coefficients were
taken from the general literature as noted by Clewell et al. (1993). Marino et al. (2006) noted
that using distributions for these latter parameters from the general literature (based on a large
number of animals) was better than updating them based on the relatively smaller number of
animals in the available dichloromethane kinetic studies. Clewell et al. (1993) determined
blood:air and tissue:air partition coefficients (means and CVs) with tissues from groups of male
26	DRAFT - DO NOT CITE OR QUOTE

-------
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
and female B6C3Fi mice. These partition coefficients were derived by using a vial equilibration
method similar to that used by prior investigators (Andersen et al., 1987; Gargas et al., 1986).
Tissue:air partition coefficients were approximately two to three times lower than previously
utilized values with the exception of the liver coefficient, which was similar to previous values
(Table 3-5). The blood:air partition coefficient (23) from Clewell et al. (1993) is higher than the
previously reported value of 8.3 (Gargas et al., 1986). The higher value is more in line with
values measured in rats (19.4) and hamsters (22.5) and, thus, is more reasonable than the older
value of 8.3. Table 3-5 shows mean and CVs for physiological parameters and partition
coefficients in the Marino et al. (2006) mouse model as well as values used in earlier
deterministic PBTK mouse models for dichloromethane.
i t GSTn
CYP
Gas
exchange

Lung

Fat
Richly
perfused
Slowly
perfused
Liver
GST

CYP
CO sub
model
1 t
Alveolar
air
i t
Blood
T
Endogenous
production
Figure 3-3. Schematic of mouse PBTK model used by Marino et al. (2006).
The Bayesian calibration of the cardiac output constant (QCC), ventilation:perfusion ratio
(VPR), and metabolic parameters was divided into three sequential steps: using kinetic data from
closed chamber studies with mice treated with an inhibitor of CYP2E1 (trans-1,2-
dichloroethylene) in order to minimize the oxidative pathway and enable a more precise estimate
of parameters for the GST pathway, followed by kinetic data for mice given intravenous
injections of dichloromethane to estimate metabolism parameters in the absence of pulmonary
absorption processes and, finally, kinetic data for naive mice exposed to dichloromethane in
closed chambers (Marino et al., 2006). The initial prior distributions were based on mean values
27	DRAFT - DO NOT CITE OR QUOTE

-------
1783	used by Andersen et al. (1987) for the metabolic parameters and by OSHA (1997) for the
1784	parameters for VPR, ratio of lung Vmax to liver Vmax (Al), and ratio of lung GST 1st order kinetic
1785	constant (lung KF) to liver KF (A2). Posterior distributions from the first Bayesian analysis were
1786	used as prior distributions for the second step, and posterior distributions from the second step
1787	were used as prior distributions for the final updating. Final results from the Bayesian
1788	calibration of the mouse probabilistic model are shown in Table 3-5.
28
DRAFT - DO NOT CITE OR QUOTE

-------
Table 3-5. Values for parameter distributions in a B6C3Fi mouse probabilistic PBTK model for
dichloromethane compared with associated values for point parameters in earlier deterministic B6C3Fi mouse
PBTK models for dichloromethane
Marino et al. (2006)a	U.S. EPA
Final posterior Final posterior (1988b, Andersen et al.
Parameter	Prior mean Prior CV	mean	CV	1987a, b)	(1987)
Fractional flow rates (fraction of QCC)h





QFC Fat
0.05
0.60

0.05
0.05
QLC Liver
0.24
0.96

0.24
0.24
QRC Rapidly perfused tissues
0.52
0.50

0.52
0.52
QSC Slowly perfused tissues
0.19
0.40

0.19
0.19
Fractional tissue volumes (fraction ofBW)h





VFC Fat
0.04
0.30
These parameters were taken from an
0.04
0.04
VLC Liver
0.04
0.06
extensive literature database derived from
0.04
0.04
VLuC Lung
0.0115
0.27
a large number of animals; therefore,
0.0119
0.0119
VRC Rapidly perfused tissues
0.05
0.30
further Bayesian updating does not
0.05
0.05
VSC Slowly perfused tissues
0.78
0.30
inform on the true mean and variance for
0.78
0.78
Partition coefficients0


these values.


PB Blood:air
23
0.15

8.29
8.29
PF Fat:blood
5.1
0.30

14.5
14.5
PL Livenblood
1.6
0.20

1.71
1.71
PLu Lung:blood
0.46
0.27

1.71
1.71
PR Rapidly per£used:blood
0.52
0.20

1.71
1.71
PS Slowly perfused:blood
0.44
0.20

0.96
0.96
Flow rates





QCC Cardiac output (L/hr/kg074)
28.0
0.58
24.2 0.19
14.3d
28.0e
VPR ventilation:perfusion ratio
1.52
0.75
1.45 0.20
1.0
1.0
Metabolism parameters





Vmaxc Maximum CYP metabolic rate (mg/hr/kg0 7)
11.1
2
9.27 0.21
11.1
11.1
Km CYP affinity (mg/L)
0.396
2
0.574 0.42
0.396
0.396
kfC First-order GST metabolic rate constant (kg0 3/hr)
1.46
2
1.41 0.28
1.46
1.46
Al Ratio of lung Vmaxc to liver Vmaxc
0.462
0.55
0.207 0.36
0.416
0.416
A2 Ratio of lung kfC to liver kfC
0.322
0.55
0.196 0.37
0.137
0.137
1789
1790	aMCMC analysis was used to update prior distributions (means and CVs) for flow rate and metabolic parameters in a sequential process with three sets of kinetic
1791	data from mouse studies, as explained further in the text. Final values for posterior distributions are given in this table.
1792	bSource: Andersen et al. (1987, 1991).
1793	°Source: Clewell et al. (1993).
1794	dBased on a mouse breathing rate of 0.043 m3/day.
1795	"Based on a mouse breathing rate of 0.084 m3/day.
29

-------
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
Marino et al. (2006) used the Bayesian-calibrated mouse model to calculate internal dose
metrics associated with exposure conditions in the NTP (1986) B6C3Fi mouse cancer inhalation
bioassay. The internal dose metric selected was milligrams (mg) dichloromethane metabolized
by the GST pathway per liter tissue per day. This is the same dose metric used in earlier
applications of PBTK models to derive human cancer IUR estimates based on cancer responses
in mice (OSHA, 1997; Andersen et al., 1987; U.S. EPA, 1987a, b). Its use is consistent with
evidence that dichloromethane metabolism via GST-T1 results in the formation of a reactive
metabolite that damages DNA and results in the formation of tumors (see section 4.7). The
model was used to calculate values for this internal dose metric in the lung and liver of mice in
the NTP (1986) study, using the mean values of the final distributions for the parameters in the
model. Resultant values were three- to four-fold higher than values calculated with the Andersen
et al. (1987) and U.S. EPA (1987a, b) versions of the model (Table 3-6). Marino et al. (2006)
noted that the difference could be primarily attributed to the changes in the partition coefficients
based on Clewell et al. (1993) as well as to the Bayesian updating of the metabolic parameters
(see Table 3-5).
Table 3-6. Internal daily doses for B6C3Fi mice exposed to
dichloromethane for 2 years (6 hours/day, 5 days/week) calculated with
different PBTK models
NTP (1986) 	PBTK model
Target organ
exposure level"
Marino et al. (2006)
U.S. EPA (1987a, b)
Andersen et al. (1987)
Liverb
Control
0
0
0

2,000 ppm
2,359.99
727.8
851

4,000 ppm
4,869.85
1,670
1,811
Lung1,
Control
0
0
0

2,000 ppm
474.991
111.4
123

4,000 ppm
973.343
243.7
256
a2,000 ppm = 6,947 mg/m3; 4,000 ppm = 13,894 mg/m3.
internal dose expressed as mg dichloromethane metabolized by the GST pathway per liter tissue per day.
Marino et al. (2006) noted that inclusion of extrahepatic CYP metabolism in the slowly
perfused tissue compartment in the mouse model had little impact on the formation of GST
metabolites in the liver and lung, especially at exposure levels used in the mouse NTP (1986)
bioassay. To support this contention, the Andersen et al. (1987) model was modified to include
10% of the liver rate of oxidative metabolism in the slowly perfused tissue compartment (as
suggested by Sweeney et al. [2004]), and the modified model was used to calculate the formation
of GST metabolites. If extrahepatic metabolism was included in the slowly perfused tissue
compartment, there was a 5-6% reduction in the formation of GST metabolites in the lung and
liver at an exposure level of 50 ppm. At 2,000 or 4,000 ppm, however, there was only a 0.77 or
0.37%) reduction, respectively. Marino et al. (2006) did not discuss the impact of including
30

-------
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
extrahepatic metabolism in the rapidly perfused tissue compartment; the same group of
investigators developed a human PBTK model that included CYP metabolism in the richly
perfused compartment (David et al., 2006).
3.5.2. Probabilistic Human PBTK Dichloromethane Model (David et al., 2006)
The basic model structure used by David et al. (2006) was that of Andersen et al. (1987)
with the addition of the CO submodel of Andersen et al. (1991), refinements from the Marino et
al. (2006) mouse model, and an inclusion of CYP metabolism in richly perfused tissue (Figure 3-
4). David et al. (2006) used Bayesian analysis to develop and calibrate metabolic parameters in a
human probabilistic PBTK model for dichloromethane, using kinetic data from several studies of
volunteers exposed to dichloromethane (n = 13 from DiVincenzo and Kaplan [1981]; n = 12
from Engstrom and Bjurstrom [1977]; n = 14 from Astrand et al. [1975]; n = 3 from Stewart et
al. [1972a], and group means for metabolism parameters from Andersen et al. [1991]). Exhaled
dichloromethane and CO and blood levels of dichloromethane and COHb were available in the
studies by Andersen et al. (1991) and DiVincenzo and Kaplan (1981). The other three studies
included two or three of these measures. The only available data for levels of dichloromethane
in fat came from the study of Engstrom and Bjurstrom (1977) (described in section 3.2 within
adipose tissue).
i t nr
Gas

Lung
Exchange

Richly
Perfused
Slowly
Perfused

CO Sub
Model
1 t
_&A/eolar
Air
1 t
Endogenous
Production
Figure 3-4. Schematic of human PBTK, used by David et al. (2006).
Values (means and SDs or CVs) for the model parameter distributions were selected from
multiple sources considered to provide the most current scientific evidence for each parameter
(David et al., 2006). Mean values for cardiac output (QCC), VPR, and all fractional tissue
volumes and blood flow rates were based on mean values used by EPA (U.S. EPA, 2000d) in a
PBTK model for vinyl chloride, as were values for CVs for all physiological parameters, except
CVs for VPR and fractional lung volume, which were set to those used by OSHA (1997). Means
31

-------
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
for the CO submodel parameters were set equal to those in Andersen et al. (1991), except for
those for the endogenous rate of CO production (REnCOC) and the background amount of CO
(ABCOC), which were based on data collected by DiVincenzo and Kaplan (1981). Means for
partition coefficients, the ratio of lung Vmax to liver Vmax, and the ratio of lung KF to liver KF
(A2) were those used by Andersen et al. (1987), whereas prior means for Vmaxc and Km were
those used by Andersen et al. (1991). The prior mean for the metabolic parameter for CYP
metabolism in the rapidly perfused tissue was set at 0.03, slightly lower than the value suggested
by Sweeney et al. (2004). Prior CVs for the metabolic parameters were set at 200%.
MCMC analysis was used to calibrate metabolic parameters in the human model in a
two-step approach: (1) posterior distributions were estimated separately by using data from each
of the five studies with kinetic data for humans exposed to dichloromethane (with durations
ranging from 1 to 8 hours and concentrations ranging from 50 to 1,000 ppm); and (2) posterior
distributions were estimated with combined data from the 42 individual subjects from the four
studies with individual subject data (DiVincenzo and Kaplan, 1981; Engstrom and Bjurstrom,
1977; Astrand et al., 1975; Stewart et al., 1972a). Results from the Bayesian calibration with the
combined kinetic data for individual subjects are shown in Table 3-7. This analysis resulted in a
narrowing of the distribution for the CYP2E1 metabolism parameters Vmax and Km, from a fairly
broad prior distribution with a CV of 200% for both parameters to 13.1 and 33.6%, respectively,
for Vmax and Km.
Table 3-7. Results of calibrating metabolic parameters in a human
probabilistic PBTK model for dichloromethane with individual kinetic data
for 42 exposed volunteers and MCMC analysis
Parameter
Prior distributions
Posterior distributions
Mean
(arithmetic)
CV
Mean
(arithmetic)
CV
Vmaxc —maximal CYP metabolic rate (mg/hr/kg0 7)
6.25
2
9.42
0.131
Km—CYP affinity (mg/L)
0.75
2
0.433
0.336
kfC—first-order GST metabolic rate (kg0 3/hr)
2
2
0.852
0.711
Al—ratio of lung Vmaxc to liver Vmaxc
0.00143
2
0.000993
0.399
A2—ratio of lung kfC to liver kfC
0.0473
2
0.0102
0.728
FracR—fraction of Vmaxc in rapidly perfused tissues
0.03
2
0.0193
0.786
Source: David et al. (2006).
The parameter statistics shown in Table 3-7 (values reported by David et al., 2006) are
summary statistics of the converged parameter chains obtained in that analysis, leaving out any
evaluation of correlation or covariance among the updated parameters. As such, these statistics
implicitly include both the inter-individual variability that would have been elucidated by the
32

-------
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
Bayesian analysis (variation between mean values for each individual for which data were
available) and uncertainty in those values.
David et al. (2006) further refined the human probabilistic model to reflect
polymorphisms in the GST pathway: homozygous positive (+/+) GST-T1 individuals,
heterozygous (+/-) GST-T1, and homozygous negative (-/-) GST-T1 individuals with no GST
activity. Distributions of GST activities for these genotypes in a group of 208 healthy male and
female subjects from Sweden were scaled to obtain distributions of kfCfor each genotype that,
when weighted by estimated frequencies of the genotypes in the U.S. population, would result in
an overall population mean equal to the kfC mean for the posterior distribution shown in
Table 3-7 (0.852 kg°'3/hour). The resultant mean kfC values were 0.676 kg°'3/hour (SD 0.123)
for heterozygous individuals and 1.31 kg0 3/hour (SD 0.167) for homozygous positive
individuals. The final parameter distributions used by David et al. (2006) are summarized in
Table 3-8.
As described in Appendix B, EPA undertook an evaluation of the David et al. (2006)
model and parameterization, focusing on the adequacy of the characterization of parameter
distributions in the full human population. EPA's conclusion is that the reported distributions for
physiological parameters in particular, but also key metabolic parameters, only represented a
narrow set of adults (with the exception of BW). The EPA therefore chose to use supplemental
data sources to define these distributions in a way that should fully characterize the variability in
the human population for individuals between six months and eighty years of age. The EPA
incorporated additional data concerning the variability in CYP2E1 activity among humans, based
on Lipscomb et al. (2003). The Lipscomb et al. (2003) study was based on in vitro analysis of
liver samples from 75 human tissue donors (activity towards trichloroethylene and measurements
of protein content) to estimate a distribution of activity in the population. These data support a
wider distribution in CYP2E1 activity than had been used in the David et al. (2006) model, with
approximately a sixfold range between the upper and lower bounds in Lipscomb et al. (2003) and
a twofold range in David et al. (2006). Thus the EPA replaced the David et al. distribution
parameters by using the same GM = 9.34, but GSD = 1.73. Further, since even the data available
to Lipscomb et al. (2003) were limited, and the log-normal distribution is naturally bounded to
be greater than zero, the EPA chose to use a non-truncated distribution for this parameter. (Since
the distribution form for CYP2E1 was set by David et al. (2006) to be log-normal and the U.S.
EPA chose to retain that form, even without specific bounds the distribution only includes values
greater than zero.) Finally, the scaling of CYP2E1 for individuals under the age of 18 was
adjusted based on the data of Johsrud et al. (2003); the EPA's analysis of these data indicate
CYP2E1 activity in children is better predicted when assumed to scale with body weight (BW)
raised to the 0.88 power, as compared to the more general power of 0.74, used by David et al.
CYP2E1 activity for individuals over the age of 18 is still assumed to scale as BW°74
33

-------
Table 3-8. Parameter distributions used in human Monte Carlo analysis for
dichloromethane by David et al. (2006)
Distribution

Parameter
Mean
(arithmetic)
SD
Source
BW
Body weight (kg)
70.0
21.0
Humans3
QCC
Cardiac output (L/hr/kg1174)
16.5
1.49
Humans3
VPR
Ventilation:perfusion ratio
1.45
0.203
Humans3
QFC
Fat
0.05
0.0150
Humans3
QLC
Liver
0.26
0.0910
Humans3
QRC
Rapidly perfused tissues
0.50
0.10
Humans3
QSC
Slow perfused tissues
0.19
0.0285
Humans3
Tissue volumes (fraction BW)



VFC
Fat
0.19
0.0570
Humans3
VLC
Liver
0.026
0.00130
Humans3
VLuC
Lung
0.0115
0.00161
Humans3
VRC
Rapidly perfused tissues
0.064
0.00640
Humans3
VSC
Slowly perfused tissues (muscle)
0.63
0.189
Humans3
Partition coefficients



PB
Blood:air
9.7
0.970
Humansb
PF
Fat:blood
12.4
3.72
Ratsb
PL
Livenblood
1.46
0.292
Ratsb
PLu
Lung:arterial blood
1.46
0.292
Ratsb
PR
Rapidly perfused tissue :blood
1.46
0.292
Ratsb
PS
Slowly perfused tissue (muscle :blood)
0.82
0.164
Ratsb
Metabolism parameters



Vmaxc
Maximum metabolism rate (mg/hr/kg )
9.42
1.23
Calibration'
Km
Affinity (mg/L)
0.433
0.146
Calibration'
Al
Ratio of lung VMax to liver Vmax
0.000993
0.000396
Calibration'
A2
Ratio of lung KF to liver KF
0.0102
0.00739
Calibration'
FracR
Fractional CYP2E1 capacity in rapidly perfused tissue
0.0193
0.0152
Calibration'
First order metabolism rate (/hr/kg°3)




Homozygous (-/-)
0
0
Calibration'
kfC
Heterozygous (+/-)
0.676
0.123
Calibration'

Homozygous (+/+)
1.31
0.167
Calibration'
1:1 US EPA, 2000d. Human PBTK model used for vinyl chloride.
bAndersen et al. (1987). Blood:air partition measured using human samples; other partition coefficients based on
estimates from tissue measures in rats.
°Bayesian calibration based on five data sets (see text for description); posterior distributions presented in this table.
Source: David et al. (2006).
1927
1928	In addition, while the BW distribution in the David et al. (2006) PBTK model used
1929	ranges from 7 to 130 kg, thus covering 6-month-old children to obese adults, there are age-
1930	dependent changes and gender-dependent differences in ventilation rates and body fat that are
1931	not explicitly included. To more accurately reflect the distribution of physiological parameters
1932	in the entire population, EPA replaced the unstructured distributions of David et al. (2006) with
1933	distributions based on available information that specifically account for population variability in
1934	age, gender, and age- and gender-specific distributions or functions for BW, QCC, alveolar
34

-------
1935	ventilation, body fat (fraction), and liver fraction (see Appendix B for more details of the
1936	evaluation of each of these parameters).
1937	The resulting set of parameter distribution characteristics, including those used as defined
1938	by David et al. (2006) are described in Table 3-9. Using this revised set of distributions,
1939	including the (revised) CYP and (published) GST activity distributions, and other distributions
1940	used as defined by David et al. (2006), the model as applied should reflect the full variability in
1941	the (U.S.) human population.
1942
35

-------
1943
Table 3-9. Parameter distributions for the human PBTK model for dichloromethane used by EPA
Distribution



(Geometric)

Lower
Upper

Parameter
Shape
mean a
SD/GSD a
bound
bound
BW
Body weight (kg)
Normal
/ (age, gender)
1st %tile
99th %tile
Flow rates






QAlvC
Alveolar ventilation (L/hour/kg°75)
Normal
f (age, gender)
f(age)
5th %tile
95th %tile
vprv
Variability in ventilation:per£usion ratio
Log-normal
1.00
0.203
0.69
1.42
QCC
Cardiac output (L/hour/kg°75)
QCCmean
=XQAlvC)
QCC
= QCCmean/vprv
Fractional flow rates (fraction of QCC)





QFC
Fat
Normal
0.05
0.0150
0.0050
0.0950
QLC
Liver
Normal
0.26
0.0910
0.010
0.533
QRC
Rapidly perfused tissues
Normal
0.50
0.10
0.20
0.80
QSC
Slow perfused tissues
Normal
0.19
0.0285
0.105
0.276
Tissue volumes (fraction BW)





VFC
Fat
Normal
/ (age, gender)
0.3 mean
0.1 mean
1.9-mean
VLC
Liver
Normal
f(age)
0.05-mean
0.85 mea
1.15-mea
VLuC
Lung
Normal
0.0115
0.00161
0.00667
0.0163
VRC
Rapidly perfused tissues
Normal
0.064
0.00640
0.0448
0.0832
VSC
Slowly perfused tissues
Normal
0.63
0.189
0.431
0.829
Partition coefficients





Section or source
B-4.3; NHANES IV
B-4.4; mean: Clewell et al. (2004);
SD: Arcus-Arth and Blaisdell (2007)
VPR/VPRmean of David et al. (2006)
B-4.5; Clewell et al. (2004) (mean)
David et al. (2006); after sampling from
these distributions, normalize:
QC-QiC
Q' =

mean: B-4.7 (Clewell et al., 2004);
otherwise, David et al. (2006); after
sampling from these distributions,
normalize:
0.9215 -BW-ViC
Vi=-

PB
PF
PL, PLu,
& PR
PS
Blood:air
Fat:blood
Livenblood, lung:arterial blood, and
rapidly perfused tissue :blood
Log-normal
Log-normal
Log-normal
Slowly perfused tissue (muscle) :blood Log-normal
Metabolism parameters (based on Monte Carte calibration from five
Maximum metabolism rate
9.7
11.9
1.43
0.80
human data sets)
1.1
7.16
13.0
1.34
4.92
28.7
1.22
0.790
2.59
1.22
0.444
1.46
Km
(mg/hr/kgXvmax)
Affinity (mg/L)
Lognormal 9.34
Log-normal	0.41
1.73
1.39
(none) (none)
0.154
1.10
Geometric mean & GSD values listed here,
converted from arithmetic mean and SD
values of David et al. (2006)
B-3 mean: David et al. (2006);
GSD & bounds: Lipscomb et al. (2003);
Xvmax = 0.88 for age <18;.
Xvmax = 0.70 for age > 18.
Geometric mean & GSD values listed here,
36

-------
Table 3-9. Parameter distributions for the human PBTK model for dichloromethane used by EPA
Distribution
(Geometric)
Parameter
Shape
mean
SD/GSD
Lower Upper
bound bound
FracR
A1	Ratio of lung VMax to liver VMax
A2	Ratio of lung KF to liver KF
Fractional MFO capacity in rapidly
perfused tissue
First order metabolism rate (|hour/kg!, 3| ')
Homozygous (-/-)
kfC	Heterozygous (+/-)
Homozygous (+/+)
Log-normal
Log-normal
Log-normal
Normal
Normal
Normal
0.00092
0.0083
0.0152
0
0.676
1.31
1.47
1.92
2.0
0
0.123
0.167
0.000291
0.00116
0
0.00
0.00
0.00292
0.0580
0.00190 0.122
0
1.291
2.145
Section or source
converted from arithmetic mean and SD
values of David et al. (2006)
David et al. (2006)
aArithmetic mean and SD listed for normal distributions; geometric mean and geometric SD (GSD) listed for log-normal distributions.
1944
37

-------
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
3.5.3. Evaluation of Rat PBTK Dichloromethane Models
Several deterministic PBTK rat models have been reported in the scientific literature
(Sweeney et al., 2004; Andersen et al., 1991, 1987; Reitz , 1991; Reitz et al.,1988a, b; U.S. EPA
1988b, 1987a, b; Gargas et al., 1986). Unlike the mouse (Marino et al., 2006) and human (David
et al., 2006), no hierarchical population model for dichloromethane in the rat exists in which
parameter uncertainty is quantitatively integrated into model calibration. Rat data are not
available that would allow for Bayesian calibration of individual metabolic parameters for the
CYP or GST pathways. Thus, EPA assessed modified versions of deterministic rat PBTK
models to select the most appropriate model for use in extrapolating internal dosimetry from rats
to humans, for example in the determination of RfDs and RfCs based on effects seen in the rat.
This work is described in detail in Appendix C and is based on evaluation of blood levels of
dichloromethane and the percent saturation of hemoglobin as COHb (%COHb) and expired
dichloromethane following intravenous injection (Angelo et al., 1986b), closed chamber gas
uptake (Gargas et al., 1986), and dichloromethane and %COHb blood levels from a 4-hour
inhalation exposure (Andersen et. al., 1991, 1987). Based on this work, the basic model
structure of Andersen et al. (1991) was chosen, with the inclusion of lung dichloromethane
metabolism via CYP (4% of liver metabolite production) and GST (14% of liver metabolite
production) pathways (estimated from Reitz et al., 1989) (Figure 3-5) with metabolic parameters
recalibrated against data of Andersen et al. (1991), based on prediction agreement of the various
parameters with the available rat data sets. Table 3-10 presents the parameter distribution data
for this model.
GST
CYP 	
CO Sub
Model
Endogenous
Production
CYP
GST
Lung
Blood
Gas
Exchange
Liver
Richly
Perfused
Slowly
Perfused
Fat
Alveolar
Air
Figure 3-5. Schematic of rat PBTK model used in current assessment.
38

-------
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
Table 3-10. Parameter values for the rat PBTK model for
dichloromethane used by EPA
Parameter	Mean
Flow rates
QCC (L/hour/kg0 74)	15.9
VPR	0.94
Fractional flow rates (percent of QCC)
Fat	9
Liver	20
Rapidly perfused tissues	56
Slowly perfused tissues	15
Tissue volumes (percent BW)
Fat	7
Liver	4
Lung (scaled as BW099)	1.15
Rapidly perfused tissues	5
Slowly perfused tissues	75
Partition coefficients
Blood:air	19.4
Fat:blood	6.19
Livenblood	0.732
Lung:arterial blood	0.46
Rapidly perfused tissue:blood	0.732
Slowly perfused tissue (muscle):blood	0.408
Metabolism parameters
Maximum metabolism rate (mg/hour/kg°7)	3.93
Affinity (mg/L)	0.524
Ratio of lung VMax to liver VMax	0.04
Ratio of lung KF to liver KF	0.14
1 st order metabolism rate (liver KF) ([hour/kg0 3]_1)	2.46
Oral absorption constant, ka (1/hr)	1.80
3.5.4. Comparison of Mouse, Rat and Human PBTK Models
The comparison of various parameters across species (Table 3-11) primarily shows the
modest inter-species differences that are known to occur in physiological parameters, also
including the approximately 2-fold differences in partition coefficients which occur because of
differences in rodent versus human blood lipid content. The 2.5-fold lower Vmaxc (CYP activity)
in rats versus mice is also typical. The most striking difference is the variation in A1 and A2.
Those values, however, reflect the in vitro differences originally quantified by Lorenz et al.
(1984) and used in the dichloromethane PBPK modeling of Anderson et al. (1987). Thus these
differences are based on independent measurements of tissue-specific metabolic capacity, and
39

-------
1978	while the specific values for mouse and human were refined through Bayesian analysis, the
1979	ultimate (posterior) values used are within a reasonable range of the in vitro measurements and
1980	so do not appear to be artifactual. (Since in vivo kinetics often indicate some differences from
1981	what would be predicted without adjustment from in vitro, it is not surprising that such
1982	differences occur here.) These differences do explain why lung-specific metrics in particular
1983	lead to lower internal dose and hence risk predictions in humans compared to whole-body
1984	metrics.
40

-------
Table 3-11. Parameters in the mouse, rat, and human PBTK model for dichloromethane used by the EPA

Mousea
Ratb

Humanc

Parameter
Mean
Value
Mean
CV/GSD (Shape, bounds)
Sources
Fractional flow rates (fraction of cardiac output) b




David et al. (2006); then
QFC Fat
0.05
0.09
0.05
0.3 (N, 0.1-1.9)
normalized:
QLC Liver
0.24
0.20
0.26
0.35 (N, 0.0385-2.05)
n QC-QiC
QRC Rapidly perfused tissues
0.52
0.56
0.50
0.2 (N, 0.4-1.6)
Q'= ^
V OiC
QSC Slowly perfused tissues
0.19
0.15
0.19
0.15 (N, 0.553-1.453)

Fractional tissue volumes (fraction of body weight) b




Fat mean: §2.2.3.6;
VFC Fat
0.04
0.07
/(age, gender)
0.3 (N, 0.1-1.9)
Liver mean: §2.2.3.7;
VLC Liver
0.04
0.04
/(age)
0.05 (N, 0.85-1.15)
otherwise David et al.
VLuC Lung
0.0115
0.0115
0.0115
0.14 (N, 0.58-1.42)
(2006); then normalized:
VRC Rapidly perfused tissues
0.05
0.05
0.064
0.1 (N, 0.7-1.3)
T/; 0.9215-BW-ViC
Vl —
VSC Slowly perfused tissues
0.78
0.75
0.63
0.3 (N, 0.684-1.32)

Partition coefficients0





PB Blood/air
23.0
19.4
9.7
1.1 (LN, 0.738-1.34)
Geometric mean (GM) &
PF Fat/blood
5.1
6.19
11.9
1.34 (LN, 0.413-2.41)
GSD/GM values
PL Liver/blood
1.6
0.73
1.43
1.22 (LN, 0.552-1.81)
converted from arithmetic
PLu Lung/blood
0.46
0.46
1.43
II
mean & SDs of David et
PR Rapidly perfused/blood
0.52
0.73
1.43
II
al. (2006)
PS Slowly perfused/blood
0.44
0.41
0.80
1.22 (LN, 0.555-1.83)

Flow rates




QCC: §2.2.3.5;
QCC Cardiac output (L/hr/kg074)
24.2
14.99
QCCmean=/QAlvC)
QCC = QCCmean/vprV
vprv = VP R/VP Rmean:
VPR ventilation/perfusion ratio
1.45
0.94
(variable)
(varies) (LN, 0.69-1.42)
David et al. (2006);
QAlvC
QCC/VPR
QCC/VPR
/ (age, gender)
f(age)(N, S^^/o)
QAlvC: §2.2.3.4;
Metabolism parameters





Vmaxc Maximum CYP metabolic rate (mg/hr/kgXvmax)
9.27
3.93
9.34
1.73 (LN, [unbounded])
Vmaxc: §2.2.2;
Xvmax CYP allometric scaling power
0.7
0.7
0.88 forage <18;
;.39(LN, 0.376-2.68)
others: David et al.



0.7 for age

(2006) (GM & GSD/GM
Km CYP affinity (mg/L)
0.574
0.524
0.41
-/-: NA
values converted from
kfC First-order GST metabolic rate constant (kg0 3/hr)
1.41
2.46
o (-/-r
+/-: 0.182 (N, 0-1.91)
arithmetic mean & SDs)



0.676 (+/-)e
+/+: 0.128 (N, 0-1.64)



1.31 (+/+)e
1.47 (LN, 0.316-3.17)

Al Ratio of lung Vmaxc to liver Vmaxc
0.207
0.04
0.00092
1.92 (LN, 0.140-6.99)

A2 Ratio of lung kfC to liver kfC
0.196
0.14
0.0083


a Based on Marino et al. (2006) (source for all mouse parameters)
b Based on Andersen et al. (1991), with the addition of lung metabolism of dichloromethane via the CYP (4% of liver metabolite production) and GST (14% of liver
41

-------
metabolite production) pathways. Physiological parameters and partition coefficients are from Andersen et al. (1991). The values for dichloromethane metabolism
in the lung (as a fractional yield of liver metabolism for each pathway) were estimated from the in vitro ratios of enzyme activity (nmol/min/mg protein) in lung and
liver cytosolic (GST) and microsomal (CYP) tissue fractions (Reitz et al., 1989). Metabolic parameters were re-optimized against the inhalation data of Andersen et
al. (1991) using a heteroscedasticity parameter value of 2, which uses relative error for the model fitting algorithm. See Appendix C for further details.
0 Based on David et al. (2006), with changes as noted. Additional sources include Clewell et al. (2004), Arcus-Arth and Blaisdell (2007), and Lipscomb et al. (2003).
See identified sections for details. Distribution values (mean and a measure of dispersion) are provided with the CV (mean/SD) presented for normal (N) distributions
and the GSD (italicized) presented for log-normal (LN) distributions. Distributions were truncated, bounds are (upper-lower bound)/mean.
"Values for the homozygous (-/-), heterozygous (+/-), and homozygous (+/+) GST-T1 genotypes, respectively.
42

-------
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS
4.1.1.	Introduction—Case Reports, Epidemiologic, and Clinical Studies
There has been considerable interest in the influence of occupational exposure to
dichloromethane in relation to a variety of conditions. The recognition that dichloromethane can
be metabolized and bound to hemoglobin to form COHb, resulting in a reduction in the oxygen
carrying capacity of the blood (Stewart et al., 1972b), prompted investigations into risk of
ischemic heart disease and other cardiovascular effects. Reports of neurological effects from
acute, high-exposure situations contributed to concern about neurological effects of chronic
exposure to lower levels of dichloromethane. A general interest in potential cancer risk became
more focused on lung and liver cancer because of the observation of these specific tumors in the
NTP (1986) experiments in mice. Details of the studies pertaining to the experimental and
epidemiologic studies of noncancer outcomes (e.g., cardiac, neurologic, hepatic, reproductive)
are presented in section 4.1.2, and studies of cancer risk are presented in section 4.1.3.
4.1.2.	Noncancer Studies
4.1.2.1. Case Reports of Acute, High-dose Exposures
Numerous case reports have been published that describe health effects resulting from
acute exposure to dichloromethane. Most of the reports describe health effects resulting from
inhalation of dichloromethane or dermal contact, but a few involve ingestion. The COHb levels
in some of these cases were relatively low (7.5-13%), so the initial toxic effects of acute
dichloromethane exposure appear to be due to its anesthetic properties as opposed to metabolic
conversion of dichloromethane to CO.
Bakinson and Jones (1985) reported on a series of 33 cases of acute inhalation exposures
to dichloromethane that occurred in the workplace over the period 1961-1980. Thirteen had lost
consciousness, and one of the workers died. Nineteen cases reported general neurological
effects, 13 reported gastrointestinal symptoms, 4 reported respiratory symptoms, and 1 reported
hepatic symptoms. Of the 19 with general neurological symptoms, all reported headache, and
dizziness was reported by 11 workers. Five workers reported one of the following symptoms:
drunkenness, confusion, lack of coordination, or paresthesia.
Rioux and Myers (1988) summarized the health effects reported for 26 cases of
dichloromethane poisoning published in the literature between 1936 and 1986. Three cases
resulted from abuse-related exposures, 2 from chronic exposures, and 21 from acute exposures.
The most common effects involved the central nervous system (CNS) (unconsciousness,
drowsiness, headache, and behavioral symptoms), pulmonary edema and dyspnea, and
dermatologic symptoms. Even severe symptoms could be reversed, but four deaths occurred.
43

-------
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
More than 10 other case reports of fatalities or poisonings have been published since the
summaries by Rioux and Myers (1988) and Bakinson and Jones (1985), and many of these
incidents involve inadequately ventilated occupational settings (Jacubovich et al., 2005; Raphael
et al., 2002; Fechner et al., 2001; Zarrabeitia et al., 2001; Goulle et al., 1999; Mahmud and
Kales, 1999; Kim et al., 1996; Tay et al., 1995; Manno et al., 1992; Leikin et al., 1990;
Shusterman et al., 1990). CNS depression and resulting narcosis, respiratory failure, and heart
failure are common features of these reports. In a survey of workers in furniture stripping shops,
10 of the 21 workers stated that they sometimes experienced dizziness, nausea, or headache
during furniture stripping operations (Hall and Rumack, 1990).
Chang et al. (1999) reported details of six patients who had ingested dichloromethane
(four in a suicide attempt and two from accidental ingestion during a state of intoxication). The
estimated amounts ingested were 350 mL or less. COHb levels, which were measured in only
two of the cases, were 8.4 and 35% (with the latter being seen in a fatal case). As in exposures
resulting from inhalation, the most common symptoms involved CNS depression, ranging from
somnolence and weakness to deep coma. Tachypnea (n = 6) and corrosive gastrointestinal tract
injury (n = 3) were also reported. Hepatic and renal failure and pancreatitis were found in the
two most severe cases.
4.1.2.2. Controlled Experiments Examining Acute Effects
Several controlled experiments were conducted in the 1970s, examining
neurophysiological effects and levels of COHb resulting from short-term (1-4 hours) exposures
to dichloromethane at levels up to 1,000 ppm or longer-term exposures at levels up to 500 ppm.
The 8-hour threshold limit value before 1975 was 500 ppm (National Institute of Occupational
Safety and Health [NIOSH], 1986). These studies are described below. With the exception of
Putz et al. (1979), there is no description in the published reports of the informed consent and
other human subjects research ethics procedures undertaken in these studies, but there is no
evidence that the conduct of the research was fundamentally unethical or significantly deficient
relative to the ethical standards prevailing at the time the research was conducted.
In 1972, Stewart et al. (1972a, b) reported results from four experiments that were
initiated because of the chance observation of an elevation in COHb saturation levels in an
individual (one of the investigators) the morning after he had spent 2 hours working with varnish
remover. Participants were medical students and faculty (including at least one of the
coauthors). A total of 11 healthy nonsmoking volunteers were placed in an exposure chamber
with mean concentrations of dichloromethane ranging from 213 to 986 ppm for 1 or 2 hours.
These experiments indicated that dichloromethane exposure at these levels resulted in COHb
saturation levels that exceeded and were more prolonged than those seen with threshold limit
value exposures to CO. The exposures also resulted in symptoms of CNS depression indicated
by visual evoked response changes and reports of light-headedness. Although return of COHb
44

-------
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
levels to background levels could take >24 hours, all of the other symptoms were reversible
within a few hours after exposure ceased.
Winneke (1974) measured auditory vigilance, visual flicker fusion frequency, and
14 psychomotor tasks in a total of 38 women exposed to dichloromethane levels of 300-800 ppm
for 4 hours in an exposure chamber. A comparison group (nine females, nine males) exposed to
100 ppm CO for 5 hours was also included. Exposure to 800 ppm dichloromethane resulted in a
statistically significant decrease in the performance of 10 of the 14 psychomotor tasks. In tests
of auditory vigilance and visual flicker fusion, depressed response was seen at 300 ppm and was
further depressed at 800 ppm. These effects were not seen with CO exposure.
Forster et al. (1974) exposed four healthy young men to dichloromethane levels ranging
from 0 to 500 ppm for 7.5 hours/day for a total of 26 days over a 6-week period to investigate
alterations in hemoglobin affinity for oxygen and altered pulmonary function. While no changes
were observed in pulmonary function, hemoglobin affinity for oxygen was increased with no
indication of adaptation to restore this affinity for oxygen to normal.
Putz et al. (1979) examined the behavioral effects seen after exposure to dichloromethane
and to CO. Twelve healthy volunteers (six men and six women) each acted as his/her own
control in separate 4-hour exposures to 70 ppm CO and 200 ppm dichloromethane. These levels
were chosen so that the COHb level would reach 5% from both the CO and dichloromethane
exposures. The experiments were conducted in a double-blind manner so that neither the
investigators nor the participant knew the exposure condition under study at any particular time.
Informed consent was obtained, and the study was reviewed by the NIOSH Human Subject
Review Board. The performance tests were dual tasks (an eye-hand coordination task in
conjunction with a tracking task), with five measures of performance assessed at six time points
over the 4-hour test period and an auditory vigilance task. Two levels of difficulty were assessed
for each task to allow assessment of whether the exposure effect was similar in low and high
difficulty tasks. The tests of eye-hand coordination, tracking tasks, and auditory vigilance
revealed significant impairment with both exposures under the more difficult task conditions.
Effects were similar or stronger in magnitude for dichloromethane compared with CO.
4.1.2.3. Observational Studies Focusing on Clinical Chemistries, Clinical Examinations, and
Symptoms
Studies in currently exposed workers
Ott et al. (1983a, c, d) evaluated several parameters of hepatic, hematopoietic, and cardiac
function in workers exposure to dichloromethane in a triacetate fiber production plant in Rock
Hill, South Carolina. Two hundred sixty-six Rock Hill workers and a comparison group of
251 workers in an acetate fiber production plant in Narrows, Virginia, were included in the
examination of urinary and blood measures. These groups included men and women, blacks and
whites, and smokers and nonsmokers. The median 8-hour TWA exposure for dichloromethane
45

-------
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
ranged from 60 to 475 ppm in Rock Hill. Acetone at levels up to over 1,000 ppm was present in
both plants, but dichloromethane and acetone exposures were inversely related.
There were differences in blood collection procedures between the two plants and in the
age, sex, race, and smoking history distribution of the study groups. The demographic and
smoking differences were accounted for in the analysis by stratification. Statistically significant
differences were seen between the workers in the two plants for COHb, serum alanine
aminotransferase (ALT), total bilirubin, and mean corpuscular hemoglobin concentration
(MCHC) (although the direction and magnitude of these differences were not reported and the
authors stated that the difference in serum ALT could be due to the differences in blood
collection procedures, which involved a sitting versus recumbent position of the subjects at the
exposed and nonexposed plants, respectively) (Ott et al., 1983c). Within the Rock Hill plant,
analyses were also conducted to examine associations between dichloromethane exposure and
the clinical parameters within specific race-sex groups by using multiple regression to control for
smoking status, age, and time of venipuncture. Positive associations were seen with COHb in all
race-sex groups (increases of 0.7-2.1% per 100 ppm increase in dichloromethane) and with total
bilirubin (increases of 0.05-0.08 mg/dL per 100 ppm increase in dichloromethane) in all groups
except nonwhite men (which was a much smaller group, n = 20, than the other groups). Red cell
count, hematocrit, hemoglobin, and aspartate aminotransferase (AST) were also positively
associated with dichloromethane exposure in white females. The increase in total bilirubin level
was not supported by parallel changes in other measures of liver function or red blood cell
turnover, suggesting that this measure was not reflecting liver damage or hemolysis.
The increased red cell count, hemoglobin, and hematocrit in women exposed to high
levels of dichloromethane (up to 475 ppm, 8-hour TWA) may indicate a compensatory
hematopoietic effect. The fact that these changes were not significant among men may be due to
higher baseline hemoglobin, which was observed when comparisons were made between
nonsmoking men and women. No such difference in the baseline values was observed among
the smoking men and women, suggesting that the compensatory advantage may be lost among
smokers.
Ott et al. (1983e) present results from a further investigation of changes in COHb,
alveolar CO, and oxygen half-saturation pressure in relation to dichloromethane exposure.
Blood samples were collected before and after shift from 136 Rock Hill and 132 Narrows
workers. For the Rock Hill workers, personal monitoring for dichloromethane exposure was
done during the shift. The TWA for dichloromethane ranged from 0-900 ppm, with a bimodal
distribution (peaks around 150 and 500 ppm) resulting from the layout of the plant. The blood
samples were used to determine blood COHb and alveolar CO levels, and the partial oxygen
pressure (P50) (that is, the pressure required to keep 50% of the blood oxygen-carrying capacity
saturated with oxygen at pH 7.4 and 37°C). Separate analyses were conducted for smokers and
nonsmokers to account for the smoking-related effects on COHb. Linear relationships were seen
46

-------
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
between dichloromethane exposure and the before-shift COHb and alveolar CO levels, reflecting
residual CO metabolism from the previous day's exposure. There were significant quadratic
relationships between dichloromethane exposure and the postshift COHb and alveolar CO levels,
indicating a partial saturation of the enzyme system metabolizing dichloromethane. The
Pso group means were lower among the exposed compared with the referents, among smokers
compared with nonsmokers, and among men compared with women. Given the relationship
between COHb and P50, an expected decrease in P50 during the shift was observed among the
exposed.
Continuous 24-hour cardiac monitoring was also evaluated in a smaller sample of
24 dichloromethane-exposed workers from the triacetate fiber production plant in Rock Hill,
South Carolina, and 26 workers from the comparison plant in Narrows, Virginia. This study (Ott
et al., 1983d) was limited to white men ages 35 or more years. Special efforts were made to
recruit men with a history of heart disease, because this group was postulated to be most likely to
demonstrate positive findings. The estimated TWA dichloromethane exposure ranged from
60 to 475 ppm in the exposed group. The evaluation examined ventricular and supraventricular
ectopic activity and S-T segment depression in the exposed and nonexposed groups.
Comparisons were also made between cardiac performance during work hours and nonwork
hours to discern possible short-term effects of recent exposure. Comparing the findings for the
24 exposed and 26 referent volunteers indicated no difference in ventricular or supraventricular
ectopic activity or S-T-segment depression. There was no difference comparing work and
nonwork hours among exposed.
Soden et al. (1996) studied all active male workers exposed to dichloromethane at a
Hoechst Celanese triacetate film production plant in Belgium. The production process was the
same as the process at the Hoechst Celanese Rock Hill plant, except the Belgium plant was
newer with better engineering controls to significantly reduce overall levels of the
dichloromethane, acetone, and methanol used in the process. The objectives of the study were to
determine the impact of varying levels of dichloromethane exposure on COHb levels, whether
successive days of dichloromethane exposure affected the COHb levels, and what impact
smoking had on COHb levels in conjunction with dichloromethane exposure. Workers were
monitored semiannually for COHb at the end of the work shift and were personally monitored
for exposure to the three solvents. Smoking status was defined based on a health assessment
questionnaire, with smokers smoking at least one cigarette per day. Among nonsmokers, a dose
response was found among COHb levels and average dichloromethane exposure levels in the
range of 7-90 ppm. The maximum COHb was 4.00% at an average exposure of 90 ppm
(correlation coefficient = 0.58, p < 0.05). Smokers' COHb levels were elevated when compared
with those of nonsmokers with similar dichloromethane air levels, but the dose-response
correlation between dichloromethane air levels and COHb levels was weaker and not statistically
significant (correlation coefficient 0.20). The maximum COHb level for smokers was 6.35% at
47

-------
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
an average dichloromethane air level of 99 ppm. The authors concluded that dichloromethane
exposures up to the levels observed did not produce COHb levels that are likely to cause cardiac
symptoms.
Cherry et al. (1983, 1981) reported the results of health evaluations of two studies of
triacetate film production workers. Cherry et al. (1981) recruited 46 of the 76 male workers at a
triacetate film factory, where workers were exposed to dichloromethane and methanol in a ratio
of 9:1 at air levels of dichloromethane ranging from 75 to 100 ppm. A small comparison group
(n = 12) of workers at this factory who worked a similar shift pattern (rapidly rotating shifts) but
who were not exposed to dichloromethane was also included. The men were asked whether they
had ever experienced cardiac symptoms (pain in the arms, chest pain sitting or lying, or chest
pain when walking or hurrying) and were asked about the presence, in the past 12 months, of
neurological disorders (frequent headaches, dizziness, loss of balance, difficulty remembering
things, numbness and tingling in the hands or feet), affective symptoms (irritability, depression,
tiredness), and stomachache (as an indicator of symptom overreporting). No difference in
response was found in history of stomachache (reported by 15% of exposed workers compared
with 17% nonexposed workers). Six of the exposed and none of the unexposed men responded
positively to the cardiac symptoms. The exposed group reported an excess of neurological
symptoms; the number (and proportion) reporting zero, one, two, and three or more symptoms
were 26 (0.56), 8 (0.17), 9 (0.20), and 3 (0.07), respectively, in exposed workers compared with
11 (0.92), 1 (0.12), 0 (0.00), and 0 (0.00), respectively, in controls (p < 0.02 for chi-square test of
linear trend). With respect to affective symptoms, the number (and proportion) reporting zero,
one, two, and three symptoms were 28 (0.61), 6 (0.13), 7 (0.15), and 5 (0.11), respectively,
among the exposed workers, and 9 (0.75), 2 (0.17), 1 (0.08), and 0 (0.0), respectively, among the
unexposed workers. The authors concluded that there was no difference between exposed and
nonexposed in reporting of affective symptoms based on a chi-square test of linear trend. There
was no discussion of the statistical power of this test or of tests of the proportion reporting a
specified number of symptoms (which may be a more appropriate test given the sample size), but
it is clear that the statistical power of this test was very low. For example, taking the simple
case of the comparison of the proportion reporting two or more symptoms and using the
approximate estimates from this study (25 and 10% in the exposed and unexposed, respectively),
approximately 75 exposed and 300 unexposed workers would be needed for a power of 0.80
(i.e., an 80% chance of rejecting the null hypothesis when the null hypothesis was false); the
actual power with the sample size of 46 and 12 is less that 0.10.
Based on these results, a follow-up study was conducted, which included a larger referent
group. This study included the symptom list described in the previous paragraph, a standardized
clinical exam (including an electrocardiograph), and neurological and psychological tests of
nerve conduction, motor speed and accuracy, intelligence, reading, and memory (Cherry et al.,
1981). Twenty-nine of the original 46 exposed workers participated in the follow-up. The men
48

-------
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
who did not participate in the follow-up were similar in age and symptoms to the men who did.
The new referent group was recruited from another plant with the same owner and a very similar
process but without dichloromethane exposure. One control, age-matched within 3 years, was
selected for each exposed worker. No differences between the groups were found in the clinical
exam, electrocardiogram, or nerve conduction tests. A statistically significant (p < 0.05) deficit
among the exposed workers was found for coarse motor speed. On two tests of overall
intelligence, the exposed group did significantly better than the referent, but, on a reading ability
test designed to assess premorbid educational level, scores for the exposed group were slightly
lower than for the referent group. (Only one of these three differences, the trail making
intelligence test, was statistically significant.) With respect to the report of neurological
symptoms in the past year, the number (and proportion) reporting zero, one, two, and three
symptoms were 17 (0.59), 4 (0.14), 6 (0.21), and 2 (0.07), respectively, among the exposed
workers, and 21 (0.72), 6 (0.21), 0 (0.0), and 2 (0.07), respectively, among the unexposed
workers, with a test of linear trend that was not statistically significant. The authors interpret the
results as indicating that the differences in neurological symptoms seen in the initial study were
due to chance and that, taken as a whole, the exposed workers had no detrimental effect
attributable to dichloromethane exposure. Again, the limitations of the statistical power of the
analysis and alternative interpretations that might have resulted from approaches taken to
improve the power were not discussed. These approaches include combining the unexposed
groups from the two analyses, using the full sample of the exposed group instead of the subset of
29 who completed the clinical exam, or using a different test (i.e., of a proportion rather than a
linear trend),
Cherry et al. (1983) compared dichloromethane-exposed workers at an acetate film
factory to nonexposed workers (from the same plant but from areas without solvent contact or
from another film production factory in which solvents were not used). The 56 exposed and
36 unexposed workers were matched to within 3 years of age. Both factories were on rapid
rotating shifts. Exposure to dichloromethane ranged from 28-173 ppm, using individual air
sampling pumps. Blood samples were taken to monitor dichloromethane levels at the beginning
and end of the shift. Study participants were asked to rate sleepiness, physical and mental
tiredness, and general health on visual analog scales with the extreme responses at either end.
Participants were also given a digit symbol substitution test and a test of simple reaction time.
No differences were seen between exposed and unexposed groups at the beginning of the shift on
the four visual analogue scales, but the exposed deteriorated more on each of the scales than did
the controls. This difference in deterioration was statistically significant (p < 0.05) during the
morning shift but was not statistically significant during the afternoon or night shifts. A
significant correlation was shown between change in mood over the course of the shift and level
of dichloromethane in the blood. No difference was seen between the exposed and referents on
the tests of reaction time or digit substitution. However, among the exposed, deterioration in the
49

-------
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
digit substitution tests at the end of the shift was significantly related to blood dichloromethane
levels.
Anundi et al. (1993) studied 12 men who worked in a graffiti-removing company. Each
worker filled out a questionnaire about previous occupational and non occupational exposure to
solvents and use of protective equipment. Half-day breathing zone samples were taken for each
of the 12 workers, and 15-minute samples were also taken for 10 workers. On the day the air
sampling was done, a structured interview pertaining to recent diseases or symptoms related to
allergies, asthma, diseases of the skin, respiratory organs, gastrointestinal tract, urinary organs,
neurological trauma and disease, and neuropsychiatric symptoms was conducted by a physician,
and blood and urine samples were collected. The results were compared with those of 233 men
from the area population. The 12 men (mean age 23 years) had worked between 3 months and
4.5 years cleaning graffiti from underground stations. No respiratory protection was used, and
the leather gloves were frequently soaked with solvent. While mixed solvent was used to do the
cleaning, dichloromethane was the predominant component, as confirmed by the air samples.
The geometric mean of the TWA calculated from the half-day samples was 127 mg/m3 (range
18-1,188 mg/m3), with half of the samples exceeding the Swedish permissible exposure limit of
120 mg/m3. The geometric mean of the 15-minute samples was 400 mg/m3 (range 6-
5,315 mg/m3), with most samples exceeding the Swedish short-time exposure limit of
300 mg/m3. Two workers had clinical laboratory data outside the normal range (urinary ai- or
p2-microglobulin, serum ALT, y-glutamyl transpeptidase), which could indicate possible kidney
and liver damage. The authors stated that in both cases factors other than the solvent exposure
(i.e., urinary tract medical condition preceding employment, history of renal stones) could have
influenced these laboratory results. The prevalence of irritation of the eyes and upper respiratory
tract (blocked nose and nasal catarrh) was much higher in the graffiti-cleaning workers compared
with the referent group (e.g., >70% of the workers compared with 18% of the comparison group
reported a blocked nose; -50% of workers and 15% of the comparison group reported eye
irritation), but there were no or much smaller differences in abnormal tiredness, headache,
nausea, or irritative cough. No acute effects on the CNS were noted.
Studies in retired workers
Lash et al. (1991) examined the hypothesis that long-term exposure to dichloromethane
produces lasting CNS effects as measured by long-term impairment on memory and attention
centers. Retired aircraft maintenance workers employed in at least 1 of 14 targeted jobs with
dichloromethane exposure for 6 or more years between 1970 and 1984 were compared to a like
group of workers without dichloromethane exposure. The unexposed workers were also retired
aircraft mechanics at the same base and held one of 10 jobs in the jet shop where little solvent
was used. The exposed group made up of painters and mechanics in the overhaul department
was chosen to maximize the exposure contrast yet minimize differences in potential confounders
50

-------
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
between exposed and nonexposed. Exposures were typically within state and federal guidelines
for dichloromethane exposure. From 1974 to 1986, when 155 measurements for
dichloromethane exposure were made, mean breathing zone TWAs ranged from 82 to 236 ppm
and averaged 225 ppm for painters and 100 ppm for mechanics.
Data collection occurred in three phases, with an initial questionnaire to all retired
members of the airline mechanics union to identify eligible workers, followed by a telephone
survey to collect medical, demographic, and general employment criteria. Subjects who
qualified were then recruited to participate in the medical evaluation. Sixty percent of the
1,758 retirees responded to the questionnaire and 259 of these retirees met the eligibility criteria.
Ninety-one men qualified for the medical evaluation based on the telephone survey; 25 retirees
exposed to solvents, and 21 unexposed retirees participated in the evaluation. All were men
between the ages of 55 and 75, without a history of alcoholism or any neurological disorder. The
25 exposed participants worked an average of 11.6 years in dichloromethane-exposed jobs
during the target period and 23.8 years in the industry.
The medical evaluation included a questionnaire about the occurrence of 33 different
symptoms in the past year, physiological measurement of odor and color vision senses, auditory
response potential, hand grip strength, and measures of reaction time (simple, choice, and
complex), short-term visual memory and visual retention, attention, and spatial ability. The only
large differences (i.e., effect size, or mean difference between groups divided by the SD of the
outcome measure, of 0.4 or greater) between the two groups were a higher score on verbal
memory tasks (effect size approximately 0.45, p = 0.11) and lower score on attention tasks
(effect size approximately -0.55, p = 0.08) and complex reaction time (effect size approximately
-0.40, p = 0.18) in the exposed compared with the control group. (Although not noted by the
authors, the power to detect a statistically significant difference between the groups, given this
sample size, was low [i.e., approximately 0.30 for an effect size of 0.40, using a two-tailed alpha
of 0.05]) (Cohen, 1987). The authors investigated the possibility of response bias, given the low
initial response to the mailed questionnaire recruiting retirees and the small number of workers
from the entire pool of eligible participants who actually participated in the medical evaluation.
Attempts were made to contact 30% of the questionnaire nonrespondents, with 46% contacted
and 31% completing the telephone interview. The only difference found between those who
responded to the mailed questionnaire and those who did not was a higher percentage of
diagnosed heart disease among the nonrespondents who were 2.5 years older and had been
retired 1.7 more years than the respondents. Those who were eligible but did not participate in
the medical evaluation were similar to the exam participants on all characteristics included in the
interview. The only difference was a higher prevalence of gout among the unexposed who did
not participate compared to the unexposed who did participate.
4.1.2.4. Observational Studies Using Workplace Medical Program Data
51

-------
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
Kolodner et al. (1990) investigated the effect of occupational exposure to
dichloromethane on six health outcomes identified in the literature or based on biological
plausibility. Participants in the study were male workers at least 19 years old at two
General Electric plastic polymer plants where dichloromethane was one of the chemicals used.
Four dichloromethane exposure categories were established based on full-shift personal air
monitoring data (8-hour TWA) collected in 1979-1985 Job titles, and industrial hygienists'
knowledge of plant operations. The mean 8-hour TWA and number of workers in each of the
four exposure groups were 49.0 ppm for the 19 workers in the highest, 10.9 ppm for the
49 workers in the intermediate, 3.3 ppm for the 56 workers in the low, and <1.0 ppm for the
772 workers in the minimal/no exposure group.
Data from 1984 annual medical exams and 1985 absence data from payroll records were
evaluated for possible health effects resulting from occupational exposure to dichloromethane.
A high percentage of workers participated in the annual medical exams, with only 5 of the
896 eligible for inclusion in the study refusing the exam completely in 1984. Six hypotheses
were specifically tested regarding dichloromethane exposure in relation to different health
outcomes: absence due to illness, hepatotoxicity (manifested by nausea, weakness and fatigue,
palpable liver, abdominal tenderness, jaundice, hepatomegaly, abnormal serum y-glutamyl
transferase, ALT, AST, or bilirubin), diabetes mellitus (manifested by weight loss, weakness and
fatigue, polydypsia, polyuria, impaired vision, excessive weight loss, elevated fasting blood
sugar, and abnormal urinary glucose or urinary acetone), CNS toxicity (manifested by headache,
lightheadedness, dizziness and vertigo, ataxia, weakness and fatigue, and abnormalities detected
in the central motor, central sensory, cranial nerve, gait, neurocoordination, or Bibinski reflex
examinations), cardiovascular abnormalities (manifested by fatigue, dyspnea, chest pain with
exertion, palpitations, or abnormalities detected in the point maximum impulse exam, blood
pressure measurements, or electrocardiogram), and neoplastic breast changes (154 women were
included in this portion of the study—manifested by painful breast, breast swelling, lump, nipple
discharge, or abnormalities detected in the breast examination).
Workers were placed in exposure categories based on their current jobs. In addition,
exposure to high noise levels occurred in both plants, and workers in each plant had exposure to
another chemical, either phenol or phosgene. The authors noted that workers tended to move
from entry-level jobs with high dichloromethane exposure to supervisory jobs with lower
dichloromethane exposure, based on the seniority system in place at both plants. Thus, current
exposure levels reported did not necessarily reflect cumulative exposure. Because of the way the
seniority system moved workers through jobs and the fact that workers were assigned to
dichloromethane exposure categories based on their current job, age was inversely related to
exposure and was controlled in the analysis of some of the continuous variables using analysis of
covariance. Age adjustment was not employed in the analysis of dichotomous variables. The
52

-------
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
mean age was 35.3, 39.7, 37.1 and 29.5 years in the minimal/no, low, medium3, and high
exposure groups, respectively. The small number of workers in the exposed groups limited the
ability to evaluate the effects of dichloromethane exposure on health outcomes related to age,
since age had to be adjusted in these analyses. The racial distribution did not differ among the
exposure groups.
The authors indicated that all the hypotheses were accepted with the exception of CNS
symptoms. However, it should be noted that the small size and younger age distribution in the
high exposure group and the lack of adjustment for age in most of the analyses make it difficult
to interpret the statistical testing that was performed. Data pertaining to neurological, hepatic,
and cardiac function are shown in Table 4-1. Among the six neurological symptoms evaluated, a
statistically significant positive exposure-effect relationship between dizziness/vertigo and
dichloromethane exposure was identified. This trend was driven most strongly by the low
frequency of this reported symptom in the minimal/no exposure group (1.2%), but there was no
linear trend across the higher levels of exposure (7.5, 2.1, and 5.3% in the low, medium, and high
exposure groups, respectively).
3 The "medium" exposure group is also referred to as the "intermediate" exposure group in Kolodner et al. (1990).
53

-------
2381
Table 4-1. Percentage of male General Electric plastic polymer workers
reporting neurologic symptoms or displaying abnormal values in measures
of neurological function, hepatic function, and cardiac function
Exposure Group"

Minimal/no
Low
Medium
High

(n = 772)
(n = 56)
(n = 49)
(n = 19)
Neurological




Headache
8.7
7.5
10.4
5.3
Lightheadedness
2.9
3.8
4.2
5.3
Dizziness/vertigo
1.2
7.5
2.1
5.3
Ataxia
0.0
1.9
0.0
0.0
Babinsky
0.0
0.0
0.0
0.0
Gait
0.0
0.0
0.0
0.0
Faintness/syncopeb
0.1
0.0
2.1
0.0
Seizures15
0.4
0.0
2.1
0.0
Paresis/paralysisb
0.7
0.0
0.0
0.0
Parasthesisb
4.0
7.5
14.6
0.0
Head trauma/concussionb
0.8
1.9
0.0
0.0
Peripheral motor examb c
0.5
0.0
0.0
0.0
Peripheral sensory examb c
1.1
2.4
5.1
0.0
Rhomberg examb'°
0.0
0.0
2.6
0.0
Hepatic




Serum gamma glutamyl transferase
8.0
16.1
12.2
5.3
Serum total bilirubin
3.0
1.8
2.0
10.0
Serum AST
1.8
3.6
4.1
0.0
Serum ALT
9.1
10.7
8.2
5.3
Cardiacd




Palpitations: percent abnormal
1.2
9.1
2.1
0.0
Electrocardiogram




borderline/abnormal
18.5
16.7
19.1
8.3
bradycardia/tachycardia abnormalities15
20.2
16.7
25.5
0.0
general rhythm abnormalities
12.0
11.1
17.0
8.3
atrial, atrioventricular, or sinus abnormalities
0.8
0.0
0.0
0.0
bundle blocks or ventricular abnormalities
3.9
5.6
10.6
8.3
axis deviations
2.6
1.9
2.1
8.3
wave abnormalities
4.0
3.7
10.6
0.0
hypertrophy
3.8
3.7
6.4
0.0
evidence of infarction
2.3
5.6
2.1
0.0
"Mean 8-hour TWA exposure was <1.0, 3.3, 0.9, and 49.0 ppm in the minimal/no, low, medium, and high groups,
respectively; mean age 35.3, 39.7, 37.1, 29.5 years in the minimal/no, low, medium, and high groups, respectively.
bThe authors considered these to be screening variables rather than hypothesis-testing variables.
°n = 629, 42, 39, and 14 in the minimal/no, low, medium, and high groups, respectively
dFor all cardiac outcomes except bradycardia/tachycardia, n = 728, 54, 47, and 12 in the minimal/no, low, medium,
and high groups, respectively. For bradycardia/tachycardia, n = 727 in the minimal/no group.
Source: Kolodner et al. (1990).
2382
2383
2384	Soden (1993) compared health-monitoring data from dichloromethane-exposed workers
2385	in the Rock Hill triacetate fiber production plant to workers from another plant making polyester
2386	fibers owned by the same company in the same geographic area. Exposed and control workers
2387	were chosen from among workers who had worked at least 10 years in their respective areas and
2388	who participated in the company's health-monitoring program between 1984 and 1986 and were
54

-------
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
m
still employed on December 31, 1986. Controls were matched by race, age, and gender to each
Rock Hill worker for a sample size of 150 and 260 in the exposed and control groups,
respectively. (The aim of the study had been 1:2 matching.) The 8-hour TWAs among the Rock
Hill workers were those reported by Lanes et al. (1990), namely 475 ppm for dichloromethane,
900 ppm for acetone, and 100 ppm for methanol. None of these exposures occurred at the
polyester plant. There was a 90% participation rate in the health-monitoring program. Six
questions in the health history portion of the health-monitoring program concerned cardiac and
neurological symptoms (chest discomfort with exercise; racing, skipping, or irregular heartbeat;
recurring severe headaches; numbness/tingling in hands or feet; loss of memory; dizziness). Part
of this program included blood samples used for standard clinical hepatic and hematologic
parameters: serum ALT, AST, total bilirubin, and hematocrit. The clinical measures were
available for 90 (60%) of the exposed and 120 (46%) of the control group because some
participants declined this part of the health-monitoring program because similar tests had been
part of recent personal medical care.
There was little difference in the frequency of reported symptoms between exposed
workers and controls: chest discomfort reported by 2.0%> of exposed and 4.0% of the controls,
irregular heartbeat reported by 5.5% of exposed and 6.0% of the controls, recurring severe
headaches reported by 3.5% of exposed and 5.5% of the controls, numbness/tingling in hands
and feet reported by 6.4% of exposed and 8.1% of the controls, loss of memory reported by 1.3%
of exposed and 0.4% of controls, and dizziness reported by 2.7% of exposed and 4.8% of
controls (Soden, 1993). The levels of the blood values were similar in the exposed and control
groups, except for a 3.1 IU/L decrease in serum AST activity (p = 0.06). The authors concluded
that this difference was not clinically significant but did not discuss the potential bias introduced
by the selective participation in this part of the study.
4.1.2.5. Studies of Ischemic Heart Disease Mortality Risk
Several studies have examined the relation between dichloromethane exposure and risk
of cardiovascular-related mortality. The methodological details of these studies are described in
section 4.1.3.2.). No evidence of increased risk of ischemic heart disease mortality was seen in
two triacetate film production cohort studies (Hearne and Pifer, 1999; Tomenson et al., 1997) or
in two triacetate fiber production cohort studies (Gibbs et al., 1996; Lanes et al., 1993).
Information on this outcome was not included in the dichloromethane analysis of civilian Air
Force base workers (Blair et al., 1998). The standardized mortality ratios (SMRs) for ischemic
heart disease mortality were <1.0 in all of the cohorts and dose groups examined (Table 4-2).
The "healthy worker effect" may have contributed to these observations. There are no case-
control studies of ischemic heart disease and dichloromethane exposure.
55

-------
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
Table 4-2. Ischemic heart disease mortality risk in four cohorts of
dichloromethane-exposed workers


Obsa
Expb
SMR
95% cr
Triacetate film production





Hearne and Pifer (1999)
Cohort 1 (men)
117
136.7
0.86
0.71-1.03

Cohort 2 (men)
122
143.3
0.85
0.71-1.02
Tomenson et al. (1997)
Men
114
123.9
0.92
0.76-1.10
Triacetate fiber production





Lanes et al. (1993)
Men and women
43
47.8
0.90
65-121
Gibbs et al. (1996)
Men





50-100 ppm
96
100.1
0.96
0.78-1.2

350-700 ppm
98
106.8
0.92
0.75-1.1

Women





50-100 ppm
32
45.8
0.70
0.48-0.99

350-700 ppm
0
3.4
-
0.0-1.1
aObs = number of observed deaths.
bExp = number of expected deaths.
°CI = confidence interval.
4.1.2.6. Studies of Suicide Risk
Suicide risk is not an outcome that was a primary hypothesis or motivation of the cohort
studies but may be relevant given the potential neuropsychological effects of dichloromethane,
as evidenced from studies of acute and chronic exposure scenarios described previously. In a
triacetate film production cohort in Rochester, New York, Hearne and Pifer (1999) reported
14 observed deaths from suicide compared with 7.8 expected, for an SMR of 1.8 (95%
confidence interval [CI] 0.98-3.0) (Table 4-3). This cohort ("Cohort 1") consisted of 1,311 men
who were first employed between 1946 and 1970 and were followed through 1994. Similar
results were seen in a different, but somewhat overlapping, cohort in this study ("Cohort 2") of
1,013 men employed between 1964 and 1970 and followed through 1994 (see section 4.1.3.3.1).
There was also evidence of increasing suicide risk with dichloromethane exposure, particularly
in the highest exposure group, in the study of triacetate fiber production workers in Maryland
(Gibbs, 1992). The triacetate fiber production cohort study in Rock Hill, South Carolina, has
published what appears to be erroneous information about suicide risk. In the 1993 paper (Lanes
et al., 1993), 4 observed and 5.21 expected cases were reported (SMR 0.77), but the SMR that
was reported with these data was 1.19 (95% CI 0.39, 2.8). This ratio would correspond to
6 observed and around 5.2 expected cases. Information on suicide was not included in the other
film and fiber cohort studies (Tomenson et al., 1997) or in the analysis of civilian Air Force base
workers (Blair et al., 1998). There are no case-control studies of suicide risk and
dichloromethane exposure.
56

-------
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
Table 4-3. Suicide risk in two cohorts of dichloromethane-exposed workers


Obsa
Expb
SMR
95% CI
Triacetate film production





Hearne and Pifer (1999)
Cohort 1
14
7.8
1.8
0.98-3.0

Cohort 2
9
5.1
1.8
0.81-3.4
Triacetate fiber production0





Gibbs (1992)
50-100 ppm
8
6.4
1.3
0.54-2.5

350-700 ppm
8
4.4
1.8
0.78-3.6
'Obs = number of observed deaths.
bExp = number of expected deaths.
°One additional study provided data on suicide risk, but some kind of error seems to be present: 4 observed and
5.21 expected cases were reported in Lanes et al. (1993), which would be an SMR of 0.77, but the SMR that was
reported with these data was 1.19 (95% CI 0.39, 2.8). This ratio would correspond to 6 observed and around
5.2 expected cases.
4.1.2.7 Studies of Infectious Disease Risk
There is limited information pertaining to infectious disease risk in relation to
dichloromethane exposure. Only one of the cohort studies (Hearne and Pifer, 1999) reported
data for the broad category of infectious and parasitic disease mortality. In Cohort 1 of this
analysis, there were no observed deaths in this category (5.6 expected), and in Cohort 2 there
were 3 observed and 4.7 expected, for an SMR of 0.64. The detailed report by Gibbs (1992) of
the cellulose triacetate fiber production cohorts in Maryland (Gibbs et al., 1996) also contained
information on the facility in South Carolina that was the site of the report by Lanes et al. (1993,
1990). Slightly elevated risks of mortality due to influenza and pneumonia were seen among the
male workers in the high exposure group in Maryland (7 observed, 5.62 expected, SMR 1.25)
and in South Carolina (3 observed, 1.33 expected, SMR 2.26). Among females, there were few
observed or expected cases (in Maryland, 1 observed, 0.23 expected, SMR 4.36; in South
Carolina, 0 observed and 0.74 expected).
4.1.2.8. Studies of Reproductive Outcomes
Pregnancy outcomes in women exposed to dichloromethane have been investigated in
two studies. Taskinen et al. (1986) studied spontaneous abortions among women employed in
eight pharmaceutical factories between 1973 and 1980. Data on pregnancy outcomes were
collected from a national hospital and clinic discharge registry in Finland from 1973 to 1981 by
matching the worker rosters to the registry. Exposure to dichloromethane was one of eight
solvents or classes of solvents included in the study. The study consisted of two parts. The first
investigated the rate of spontaneous abortions (number of spontaneous abortions divided by the
sum of spontaneous abortions and births) during, before, or after employment in the
pharmaceutical industry. One hundred and forty-two spontaneous abortions and 1,179 births
were identified among the female workers at the eight plants. Employment hire and termination
57

-------
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
dates were obtained from plant records. The spontaneous abortion rate was 10.9% during
employment compared with 10.6% before and after employment. These results compared to a
rate of 8.5% in the general population in the geographic area where the factories were located.
The rate of spontaneous abortions among workers declined over the period of the study, with a 3-
year moving average of 15% at the beginning declining to 9.5% at the end of the study. Over the
same period, the industrial hygiene allegedly improved in the plants. Ten congenital
malformations of different types were identified among the women (five among those who were
employed in the pharmaceutical industry during the pregnancy and five among those whose
pregnancies occurred before or after this employment).
The second part of the study by Taskinen et al. (1986) was a case-control study of the risk
of spontaneous abortions in relation to workplace exposures during pregnancy. The source
population consisted of women who were employed in one of the eight Finnish pharmaceutical
factories during at least 1 week of the first trimester of pregnancy during the study period. Cases
(n = 44) were selected from this population based on hospital or clinic records indicating a
spontaneous abortion, and 130 controls (women who had given birth) were age-matched
(3:1 matching, age within 2.5 years) to each case. Occupational exposure data were obtained by
questionnaires completed by the plant physician or the nursing staff, blinded to the case status of
the study member, in consultation with labor protection chiefs and department foremen. The
questionnaire requested information about job history and job tasks, exposure to eight specific
solvents or classes of solvents (aliphatic solvents, alicyclic solvents, toluene, xylene, benzene,
chloroform, dichloromethane, and other solvents), antineoplastic agents, carcinogens, hormones,
antibiotics, heavy lifting, known chronic diseases, acute diseases during pregnancy, smoking
status, and previous pregnancies. Exposure frequency to each solvent was based on the
cumulative weighted sum of the number of days/week the woman was exposed to the solvent.
While overall response to the questionnaire was 93%, less than half the questionnaires contained
information about smoking or previous pregnancies, precluding inclusion of these variables in
the analysis. The distribution of broad categories of occupations (i.e., pharmaceutical workers
and packers, laboratory assistants) was similar in both groups. However, exposure to each of the
solvents was higher in the cases compared with controls, and the results for dichloromethane
were relatively strong. For dichloromethane, the prevalence of exposure was 28.9 and 14.3% in
cases and controls, respectively, resulting in an odds ratio (OR) of 2.3 (95% CI, 1.0, 5.7). There
was also evidence of an increasing risk with higher exposure frequency, with an odds ratio (OR)
of 2.0 (95% CI 0.6, 6.6) with exposures of less than once a week and 2.8 (95% CI 0.8, 9.5) with
exposures of once a week or more. An association was also seen with exposure to four or more
solvents (OR 3.4, 95% CI 1.0, 12.5), and weaker associations were seen with other specific
solvents (e.g., chloroform, toluene).
Bell et al. (1991) investigated the relation between birth weight and maternal exposure to
airborne dichloromethane as a result of living around the triacetate film facility in Rochester,
58

-------
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
New York. For this population-based cross-sectional study, birth certificates were obtained for
all births in 1976-1987 in Monroe County, where the triacetate film facility is located. Multiple
births and births of infants weighing <750 grams were excluded. Data abstracted from the
certificate included date of birth, census tract of residence, age, race, educational level of the
mother and father, sex, gestational age, multiple births, month of the pregnancy that prenatal care
began, total previous births, total previous live births, and conditions present during the
pregnancy. An air dispersion modeling system for 250 air emissions, including
dichloromethane, predicting average annual ground level concentrations in the surrounding
community, was used to assign dichloromethane exposure levels to each birth mother. One of
four levels of exposure was assigned to each census tract based on the isopleth of exposure in
which more than half of the census tract population resided. Because of the few births among
nonwhites that occurred in areas of higher exposure, the study was restricted to whites
(n = 91,302). The number of births that occurred in each of the four exposure levels was
n = 1,085 in the high-exposure group (50 (J,g/m3 [0.014 ppm]), n = 1,795 in the moderate-
exposure group (25 |ig/m3 [0.007 ppm]), n = 6,044 in the low-exposure group (10 (J,g/m3
[0.003 ppm]), and n = 82,076 in the no-exposure group. At the levels of dichloromethane
exposure in this population no significant adverse effect on birth weight was found. There was
an 18.7 g decrease in birthweight (95% CI -51.6, 14.2) in the high- compared with the no-
exposure group, adjusting for maternal age, maternal education, parity, previous pregnancy loss,
late start of prenatal care, sex of the child, and pregnancy complications. No significant
association was found between any combination of exposure levels and birth weight. There was
no association between exposure group and risk of a low birthweight infant (i.e., <2,500 g,
OR 1.0 [95% CI 0.81, 1.2] in the high- compared with the no-exposure group). The authors
point out a number of problems with assignment of dichloromethane exposure. It is possible that
the dichloromethane exposure was overestimated using the model. Comparisons to ambient air
sampling levels collected six times/year resulted in the dichloromethane exposure derived from
the model being twice as high as the ambient air samples. There was also inaccuracy in the
assignment of dichloromethane exposure level to each birth because the exposure assignment
was made using the predominant value of the isopleth for a census tract.
Two studies have investigated the occurrence of oligospermia among men occupationally
exposed to dichloromethane exposure. Kelly (1988) studied 34 men employed in an automotive
plant as bonders, finishers, and press operators. These men were self-referred to a health center
for a variety of complaints, including neurological symptoms, musculoskeletal symptoms, and
shortness of breath. Twenty-six of the men were bonders and eight were finishers or press
operators. The job as bonder consisted of dipping hands into an open bucket of dichloromethane
and splashing it onto plastic automobile parts. The dichloromethane exposure for bonders
averaged 68 ppm with a range from 3.3 to 154.4 ppm. Eight men, all of whom were bonders,
reported symptoms of testicular and epidydimal tenderness, with confirmation on medical exam.
59

-------
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
They ranged in age from 20 to 47 years old and had been bonders for up to 2.9 years. The COHb
levels for the eight workers with genital symptoms ranged from 1.2 to 17.3%, with an average of
6.9% anywhere from 4 to 90 hours postexposure. The COHb levels for the two men who
smoked were among the highest, namely 7.3 and 17.3%. Four of the eight workers agreed to
provide semen samples; their sperm counts were 2-26 x 106/cm3. The authors stated that men
with sperm counts as low as 25 x 106/cm3 may still be fertile, but none of these men had had any
children since working with dichloromethane despite not using contraceptives. There was one
miscarriage. All four men reported dipping their hands into open buckets of dichloromethane
without any protective equipment, and two men reported feeling dizzy, giddy, and high at work.
Based on the results of the Kelly (1988) case report, Wells et al. (1989) planned to do a
study of oligospermia among 20 exposed and 20 unexposed to dichloromethane. The exposed
workers were unvasectomized men who had worked for the 3 months prior to recruitment in
furniture stripping shops. Eleven men were recruited from among 14 eligible workers at six
different shops where dichloromethane was utilized. Names of acquaintances of the exposed
were solicited as potential referents. Only one exposed man provided any names. Therefore, the
study was redirected as a case report on the 11 exposed men. The mean TWA dichloromethane
exposure was 122 ppm (range 15-366 ppm) with a mean COHb of 5.8% (range 2.2-13.5%).
The mean COHb for smokers, 10.2% (range 8.1-13.5), was higher than for nonsmokers, 3.9%
(range 2.2-5.9), and the nonsmoker levels were higher than the 2% level considered to be the
upper limit of normal in nonsmoking populations. The mean sperm count was 54 x 106/cm3
(range 23-128 x 106/cm3) compared to a population value of 47 x 106/cm3 for the same
geographic based on samples analyzed at the same laboratory. Using the standard definition for
oligospermia of 20 x 106/cm3, none of the 11 workers had oligospermia.
4.1.2.9. Summary of Noncancer Studies
The clinical and workplace studies of noncancer health effects of dichloromethane
exposure have examined markers of disease and specific clinical endpoints relating to cardiac,
neurological disease, hepatic function, and reproductive health.
Cardiac effects
The effect of dichloromethane on the formation of COHb (Stewart et al., 1972b) raised
concerns about potential risk of cardiovascular damage. To date, there is little evidence of
cardiac damage related to dichloromethane exposure in the cohort studies of dichloromethane-
exposed workers that examined ischemic heart disease mortality risk (Hearne and Pifer, 1999;
Tomenson et al., 1997; Gibbs et al., 1996; Lanes et al., 1993) or in two small cardiac monitoring
studies (Ott et al., 1983d; Cherry et al., 1981). However, limitations in these studies should be
noted, including the healthy worker effect and the absence of data pertaining to workers who
died before the establishment of the analytic cohort (Gibbs et al., 1996; Gibbs, 1992).
60

-------
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
Neurological effects
The acute effects of dichloromethane exposure on neurological function seen in
numerous case reports have also been established in experimental studies in humans (Putz et al.,
1979; Winneke, 1974; Stewart et al., 1972a, b). Relatively less is known about the long-term
effects of chronic exposures in humans. Some data from studies of workers suggest that the
effects of dichloromethane are relatively short-lived. For example, in the study by Cherry et al.
(1983) of 56 exposed and 36 unexposed workers, alterations in mood or in digit substitution test
results were seen during the course of a work shift but were not seen at the beginning of a shift.
No difference in four neurological symptoms was seen in an analysis of exposed workers
(average exposure 475 ppm, >10-year duration) and an unexposed comparison group by Soden
(1993). Other data suggest an increase in prevalence of neurological symptoms among workers
(Cherry et al., 1981) and possible detriments in attention and reaction time in complex tasks
among retired workers (Lash et al., 1991). These latter two studies are limited by the small
sample size. Thus Cherry et al. (1981) and Lash et al. (1991) have low power for detecting
statistically significant results and consequently should not be interpreted as definitive analyses
showing no effects. Rather, these analyses provide some evidence of an increased prevalence of
neurological symptoms among workers with average exposures of 75-100 ppm (Cherry et al.,
1981) and long-term effects on specific neurological measures (i.e., attention and reaction time)
in workers whose past exposures, at least for part of their work history, were in the 100-200 ppm
range (Lash et al., 1991). The increased risk of suicide (approximately a twofold increased risk)
seen in two of the worker cohort studies (Hearne and Pifer, 1999; Gibbs, 1992) is an additional
indication of potential neurological consequences of dichloromethane exposure. Adequate
studies addressing these specific issues are not available. Thus, given the suggestions from the
currently available studies, the statement that there are no long-term neurological effects of
chronic exposures to dichloromethane cannot be made with confidence.
Hepatic effects
Three studies provide data pertaining to markers of hepatic damage (i.e., serum enzymes
and bilirubin levels) (Soden, 1993; Kolodner et al., 1990; Ott et al., 1983c). Two of these studies
were based in the Rock Hill, South Carolina, cellulose triacetate fiber plant (Soden, 1993; Ott et
al., 1983c), with the most recent of the studies focusing on workers with more than 10 years
duration in a high exposure area (average exposure estimated as 475 ppm). There is some
evidence of increasing levels of serum bilirubin with increasing dichloromethane exposure in Ott
et al. (1983c) and Kolodner et al. (1990), but there are no consistent patterns with respect to the
other hepatic enzymes examined (serum y-glutamyl transferase, serum AST, serum ALT). These
studies do not provide clear evidence of hepatic damage in dichloromethane-exposed workers, to
the extent that this damage could be detected by these serologic measures; however, these data
61

-------
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
are limited and thus the absence, presence, or extent of hepatic damage is not known with
certainty.
Immune effects
Only limited, and somewhat indirect, evidence pertaining to immune-related effects of
dichloromethane in humans is available. No risk was seen in the broad category of infectious
and parasite-related mortality reported by Hearne and Pifer (1999), but there was some evidence
of an increased risk for influenza and pneumonia-related mortality at two cellulose triacetate
fiber production work sites in Maryland and South Carolina (Gibbs, 1992). In the Maryland
facility, an increased risk of cervical cancer was seen among the 938 female workers, with an
SMR of 3.0 (95% CI, 0.96, 6.9) in the 50-100 ppm group and 5.4 (95% CI 0.13, 30.1) in the
350-700 ppm group (Gibbs et al., 1996). Cervical cancer is viral mediated (human papilloma
virus), and immunosuppression is a risk factor for development of this disease, as seen by the
increased risk in immunocompromised patients and people taking immunosuppressant
medications (Leitao et al., 2008; Ognenovski et al., 2004).
Reproductive effects
Studies pertaining to various reproductive effects and dichloromethane exposure from
workplace settings (Wells et al., 1989; Kelly, 1988; Taskinen et al., 1986) or environmental
settings (Bell et al., 1991) have examined possible associations with spontaneous abortion
(Taskinen et al., 1986), low birth weight (Bell et al., 1991), and oligospermia (Wells et al., 1989;
Kelly, 1988). Of these, the data pertaining to spontaneous abortion provide the strongest
evidence of an adverse effect of dichloromethane exposure, particularly with respect to the case-
control study in which the strongest association was seen specifically with the higher frequency
category of dichloromethane exposure. It is a small study (44 cases, 130 controls), however,
with limited quantitative exposure assessment and multiple exposures (although the association
seen with dichloromethane was among the highest seen among the solvents) and so cannot be
considered to firmly establish the role of dichloromethane in induction of miscarriage. However,
the high exposure scenario, including the potential for substantial dermal exposure in the study
of Kelly (1988), also suggests the potential for adverse male reproductive effects.
4.1.3. Cancer Studies
4.1.3.1. Identification and Selection of Studies for Evaluation of Cancer Risk
Twelve epidemiologic studies of cancer risk were identified and included in this
evaluation: four cohorts for which the primary solvent exposure was to dichloromethane (two in
film production settings and two in cellulose triacetate fiber production), one large cohort of
civilian employees at a military base with exposures to a variety of solvents but that included an
assessment specifically of dichloromethane exposure, and seven case-control studies of specific
62

-------
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
cancers with data on dichloromethane exposure. One additional study (Ott et al., 1985), a cohort
of 1,919 men employed at Dow Chemical facilities, was identified but was not included in the
summary. The analysis was based on exposure to a combined group of chlorinated methanes
(e.g., carbon tetrachloride, chloroform, methyl chloride, and dichloromethane), and it was not
possible from the data presented to assess the individual effects of dichloromethane.
4.1.3.2.	Description of the Selected Studies
In this section, the study setting, methods (including exposure assessment techniques),
results pertaining to incidence or mortality from specific cancers, and a brief summary of
primary strengths and limitations are summarized for each of the 12 selected studies. When two
papers of the same cohort were available, the results from the longer period of follow-up are
emphasized in the summary. Information from earlier reports is used when these reports contain
more details regarding working conditions, study design, and exposure assessment. The
description of individual studies is followed by a summary of the evidence available from these
studies relating to specific types of cancer.
4.1.3.3.	Cellulose Triacetate Film Base Production Cohorts
4.1.3.3.1. Cellulose triacetate film base production—Rochester, New York (Eastman Kodak).
Friedlander et al. (1978) reported a cohort mortality study of workers in an Eastman Kodak
facility in Rochester, New York. This study was expanded and extended several times during
the next 20 years (Hearne and Pifer, 1999; Hearne et al., 1990, 1987). The latest analysis
provided data on two overlapping cohorts. The first cohort (Cohort 1) consisted of 1,311 male
workers employed in the roll coating division (n = 1,070) or the dope and distilling departments
(n = 241) of the Eastman Kodak facility in Rochester, New York. Men who began working in
these areas after 1945 and were employed in these areas for at least 1 year (including seasonal or
part-time work that equaled 1 full-time year equivalent) from 1946 to 1970 were included.
Follow-up time was calculated from the end of the first year of employment in the study area
through December 31, 1994. The mean duration of work in Cohort 1 was 17 years. The total
number of person-years of follow-up was 46,112, and the mean duration of follow-up was
35.2 years (range 25-49 years). The second cohort (Cohort 2) included 1,013 male workers in
the roll coating division who were employed for at least 1 year in this division between 1964 and
1970. Follow-up time was calculated from January 1, 1964, or the date of first employment in
the roll coating division for those who were employed there before 1964 and those who began in
1964 or later, respectively. Follow-up continued through December 31, 1994. The mean
duration of work in Cohort 2 was 24 years. Total follow-up time was 26,251 person-years, and
the mean duration of follow-up was 25.9 years (range 25-31 years). Cohort 2 was the focus of
previous analyses by Friedlander et al. (1978) and Hearne et al. (1990, 1987).
63

-------
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
For both cohorts, causes of death were based on the underlying causes of death recorded
on the death certificates, which were routinely obtained by the company for the processing of life
insurance claims. The expected number of deaths was calculated using appropriate age-, sex-,
calendar time-, and cause-specific death rates for men in New York State (excluding New York
City). In addition, another referent group was also used in the analysis of the second cohort.
This other referent was based on the age-, sex-, calendar time-, and cause-specific death rates of
other hourly male workers employed at the Eastman Kodak plant in Rochester, New York. (An
internal referent group was also described for Cohort 1, but data for that analysis were not
presented.)
Dichloromethane was first used in the film production process at the Eastman Kodak
facility around 1944 (Hearne et al., 1987). Cellulose triacetate was dissolved in dichloromethane
and then cast into a thin film onto revolving wheels. The film was then cured by circulating hot
air in the coating machines, and the solvent was recovered and redistilled. 1,2-Dichloropropane
and 1,2-dichloroethane were also used as solvents from the 1930s to the 1960s, but
dichloromethane was predominant (ratio 17:2:1 for dichloromethane: 1,2-dichloropropane: 1,2-
dichloroethane in general workplace air measurements) (Hearne et al., 1987).
The exposure assessment in the Rochester, New York, Eastman Kodak cohort studies
was based on employment records (start and stop dates for specific jobs in the relevant areas of
the company) in combination with air monitoring data used to estimate the exposure level for a
given job, location, and time period (Hearne et al., 1987). Air monitoring began in the 1940s,
but few data are available before 1959. In the most recent update (Hearne and Pifer, 1999), more
than 1,500 area and 2,500 breathing zone air samples were used in the exposure assessment
process. Reductions in exposures in the dope department and the distilling department began
after 1965. The highest exposure jobs were operator and maintenance workers (dope
department) and filter washing and waste operator (distilling department), with estimated 8-hour
TWA exposures of 100-520 ppm between 1946 and 1985. There was little change in estimated
exposures for jobs in the roll coating division from the 1940s through 1985, but some reduction
was seen from 1986 to 1994. The mean 8-hour TWA exposures were 39 ppm for Cohort 1 and
26 ppm for Cohort 2.
These data were used to estimate a cumulative exposure index (i.e., the summation across
all jobs held by an individual of the product of the average dichloromethane concentration as
ppm and the duration of employment in that job). The authors refer to this as a "career exposure
index." Additional adjustment in these estimates was made for respiratory protection, but the
details of this adjustment were not described. For Cohort 1, the cumulative exposure categories
used in exposure-effect analyses were <150, 150-349, 350-799, and >800 ppm. For Cohort 2,
the cumulative exposure categories were <400, 401-799, 800-1,199, and >1,200 ppm. The cut
points were chosen to produce an approximately equal number of expected total deaths in these
categories. Two different methods to calculate expected number of deaths within each exposure
64

-------
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
category were used for each cohort analysis. For Cohort 1, an internal comparison was made
based on the distribution of person-years within each exposure category, and an external
comparison was made applying New York State mortality rates. For Cohort 2, the internal
comparison using the distribution of person-years within each exposure category was also used,
but the external comparison was based on mortality rates in other hourly workers at the
Rochester, New York, Eastman Kodak work site.
There was no increased risk of mortality for all sites of cancer or for lung cancer in either
cohort analysis (Table 4-4). Data pertaining to smoking history, obtained from a survey of
workers in the New York film production facility, indicate that smoking rates were similar in the
exposed group, the internal comparison group, and the general population; therefore, it is
unlikely that differences in smoking could be masking an effect of dichloromethane.
The only specific sites for which there were increased SMRs in both cohorts were brain
and CNS cancer, Hodgkin's lymphoma, and leukemia. Pancreatic cancer mortality risk was
increased in Cohort 2 but not in Cohort 1. None of these associations were statistically
significant, and the Hodgkin's lymphoma observations were based on a total of only four cases
in both cohorts combined and so were very imprecise. Within Cohort 2, there was little
difference in results for most sites, using the different referent groups, but the point estimates for
the SMRs for brain and CNS cancer, Hodgkin's lymphoma, and leukemia were somewhat higher
using the New York State referent group compared with the internal Eastman Kodak referent
group. An attenuation of the dichloromethane association seen in the analyses using the internal
Kodak referent group would be expected if the risk of specific cancers was increased in this
comparison group, possibly because of other workplace exposures.
65

-------
Table 4-4. Mortality risk in Eastman Kodak cellulose triacetate film base production workers, Rochester,
New York
Cohort 1:
1,311 men employed 1946-1970, followed	Cohort 2:
through 1994	1,013 men employed 1964-1970, followed through 1994
New York referent group	New York referent group	Kodak referent group
Cancer type
Obsb
Expb
SMR
95% CI
Obs
Exp
SMR
95% CI
Exp
SMR
95% CI
Cancer, all sites
93
105.8
0.88
0.71-1.08
91
102.0
0.89
0.72-1.10
94.7
0.96
0.77-1.18
Liver3
1
2.4
0.42
0.01-2.36
1
2.4
0.42
0.01-2.33
1.8
0.55
0.01-3.07
Pancreas
5
5.5
0.92
0.30-2.14
8
5.3
1.51
0.65-2.98
5.1
1.55
0.67-3.06
Lung3
27
36.0
0.75
0.49-1.09
28
34.2
0.82
0.55-1.19
31.3
0.89
0.59-1.29
Brain3
6
2.8
2.16
0.79-4.69
4
2.1
1.88
0.51—4.81
2.7
1.46
0.39-3.75
Lymphatic system
5
6.6
0.75
0.24-1.78
6
5.7
1.06
0.39-2.30
5.7
1.05
0.38-2.28
Non-Hodgkin's
2
4.1
0.49
0.06-1.76
3
3.5
0.85
0.17-2.50
3.6
0.84
0.17-2.46
Hodgkin's
2
1.1
1.82
0.20-6.57
2
0.6
3.13
0.35-11.30
0.9
2.23
0.25-8.05
Multiple myeloma
1
1.5
0.68
0.01-3.79
1
1.5
0.65
0.01-3.62
1.3
0.79
0.01-4.39
Leukemia
8
3.9
2.04
0.88-4.03
6
3.5
1.73
0.63-3.76
4.4
1.37
0.50-2.98
aLiver includes liver and biliary duct; lung includes lung, trachea, and bronchus; brain includes brain and CNS.
bObs = number observed deaths, Exp = number of expected deaths.
Source: Hearne and Pifer (1999).
2762
66

-------
2763
2764
2765
2766
The authors presented the exposure-effect analysis, based on the estimated cumulative
dichloromethane exposure groups, for all sites of cancer, pancreatic cancer, lung cancer, brain
cancer, and leukemia (Table 4-5).
Table 4-5. Mortality risk by cumulative exposure in Eastman Kodak
cellulose triacetate film base production workers, Rochester, New York
Cohort, cancer,
SMRs (number of observed deaths)


Cohort 1"



Cumulative exposure





(ppm years)
<150
150-349
350-799

>800
Cancer, all sites





internal
0.81 (20)
1.02 (19)
1.10 (28)

1.07 (26)
New York
0.67
0.93
0.95

1.00
Pancreas





internal
0.74 (1)
0.00 (0)
0.77 (1)

2.34 (3)
New York
0.68
0.00
0.65

2.18
Lung1,





internal
0.78 (5)
1.07 (6)
1.25 (9)

0.90 (7)
New York
0.52
0.90
0.86

0.77
Brainb





internal
0.58 (1)
0.78(1)
1.65 (3)

0.85 (1)
New York
1.10
1.77
3.99

1.78
Leukemia





internal
0.83 (2)
0.00 (0)
0.48(1)

2.73 (5)
New York
1.61
0.00
Cohort 2C
0.98

5.79
Cumulative exposure

<400
400-799
800-1,199
>1,200
(ppm years)





Cancer, all sites





Internal

0.89(18)
0.96 (33)
1.11 (23)
1.08 (17)
New York

0.76
0.93
1.13
1.12
Pancreas





Internal

2.58 (4)
0.00 (0)
0.95 (2)
1.43 (2)
New York

2.86
0.00
1.83
2.67
Lung1,





Internal

0.95 (6)
1.15 (12)
0.94 (6)
0.82 (4)
New York

0.80
1.00
0.89
0.79
Brainb





Internal

0.00 (0)
1.13 (2)
1.37(1)
1.49(1)
New York

0.00
2.02
1.75
2.50
Leukemia





Internal

0.00 (0)
0.84 (2)
0.75 (1)
2.70 (3)
New York

0.00
1.26
1.10
4.84
aCohort 1: 1,311 men employed 1946-1970 in the roll coating division, dope department, or distilling department,
followed through 1994; mean exposure (cumulative exposure years) 66, 244, 543, and 1,782 ppm-years in the four
dose groups, respectively.
bLung includes lung, trachea, and bronchus; brain includes brain and CNS.
°Cohort 2: 1,013 men employed 1964-1970 in the roll coating division, followed through 1994; mean exposure
(cumulative exposure years) 168, 581, 981, and 1,670 ppm-years in the four dose groups, respectively.
Source: Hearne and Pifer (1999).
2767
2768
67

-------
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
There is no evidence of an exposure-effect for all site cancer mortality or lung cancer
mortality risk. The relatively sparse number of deaths for the other specific cancer types makes
it difficult to interpret the data. The patterns for pancreatic cancer differ between the two
cohorts, with increased risk at the higher dose in Cohort 1 and a U-shaped curve seen in
Cohort 2. For brain cancer mortality, a higher SMR was seen in the groups with cumulative
exposure levels of 800 ppm-years or greater compared with lower exposure groups. For
leukemia, in both cohorts, an increased mortality risk is seen in the highest exposure group
(mean approximately 1,700 ppm-years).
A strength of the Eastman Kodak cohort studies was the sampling data for
dichloromethane that allowed an assessment of each worker's exposure, using the monitoring
data and the worker's job history, making exposure-effect analyses possible. Follow-up of the
vital status of the cohort was >99% (Hearne and Pifer, 1999). There was also some information
on smoking history, too, for workers in the plant, based on a survey conducted in 1986 (Hearne
et al., 1987). A difficulty in interpreting the data, however, is that there was some overlap
between the cohorts: 707 of the men were included in both Cohort 1 and Cohort 2. Data are not
presented in a way that would allow the reader to eliminate duplicate cases and person-years so
that cases are only counted once when examining both cohorts. A strength of the Cohort
1 sampling strategy, compared with that of Cohort 2, is that Cohort 1 is limited to workers who
began work at the plant after 1945. These workers would not have had workplace exposure to
methanol and acetone, which were used at the plant in the film production process prior to that
time. Also, follow-up began with the beginning of employment in the relevant area. In contrast,
Cohort 2 was limited to workers who were employed from 1964 to 1970, so exposed workers
who left or died before 1964 were not included. The relatively small number of cases with
specific low incidence cancers (e.g., brain, leukemia) is also a limitation of the analyses of both
of the cohorts in this study. In addition, the exposure levels in both cohorts (mean 8-hour TWA
39 and 26 ppm in Cohorts 1 and 2, respectively) is relatively low compared with values seen in
other workplaces, including the cellulose triacetate fiber production cohorts described in Ott et
al. (1983a) and Gibbs et al. (1996). Also, the outcome assessment is based on mortality
(underlying cause from death certificates) rather than incidence data, and, because the Kodak
studies were limited to men, there is no information on risk of breast cancer or other female
reproductive cancers.
4.1.3.3.2. Cellulose triacetate film base production—Brantham, United Kingdom (Imperial
Chemical Industries). Tomenson et al. (1997) reported the results of a retrospective cohort
mortality study of 1,473 men who worked at a film-base production facility in Brantham,
England, anytime between 1946 and 1988 in jobs that were considered to have dichloromethane
exposure. The start of the follow-up period was not specified by the authors but is likely to have
been 1946 or the date of first employment at the plant. Follow-up of the cohort continued
68

-------
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
through December 31, 1994, and vital status was based on national records (United Kingdom
National Health Service Central Register and the Department of Social Security). Causes of
death were based on the underlying causes of death recorded on the death certificates. The
expected number of deaths was calculated using age-, sex-, calendar time-, and cause-specific
death rates for England and Wales. In addition, a comparison using mortality rates for the local
areas (Tendring and Samford) for 1968-1978 and analyses limited to workers who had been
employed for at least 3 months were also made, but the results of these analyses were not
presented. The mean duration of work in the cohort was 9 years, the total number of person-
years was 39,759, and the mean duration of follow-up was 27.0 years (7-49 years).
This facility produced cellulose diacetate film from 1950 to 1988, with other types of
films also manufactured beginning in the 1960s. Dichloromethane was the solvent used in this
process, and exposure occurred in the production of the triacetate film base and the casting of the
film into rolls. The exposure assessment was based on more than 2,700 personal or air
monitoring samples collected since 1975. An exposure matrix was constructed, assigning jobs to
1 of 20 work groups with similar exposure potential for each of four different time periods
(before 1960, 1960-1969, 1970-1979, and 1980-1988). For the 1980-1988 period, exposure
estimates for specific jobs were based on about 330 personal monitoring samples. For the earlier
time periods, information about work tasks and location was used in combination with the
information about the number of, use of, speed of, and problems with casting machines at
different times from their initial introduction in 1950. The highest exposures were estimated to
be in the casting machine operators and cleaners. Lifetime cumulative exposure to
dichloromethane was calculated as the product of the mean level of exposure for the assigned
work group and the duration of employment in that job summed across all jobs. Three categories
of cumulative exposure were used for the analysis of ever exposed workers: <400, 400-700, and
800+ ppm-years. Approximately 30% of the workers in the cohort were classified as
"unassigned" for the calculation of exposure group because sufficient information needed to
determine exposures (i.e., the location and tasks assigned to laborers and maintenance workers)
was not available. The mean 8-hour TWA exposure was estimated as 19 ppm for the cohort.
There was no increased risk of mortality for all sites of cancer (Table 4-6), and the SMRs
for most of the specific cancer sites examined (stomach, colon, rectum, liver, pancreas, lung, and
prostate) were less than 1.0. The only specific sites for which there was an increased SMR (i.e.,
1.1 or higher) were brain and CNS cancer and leukemia, and these estimates were based on few
(less than five) observed cases (Table 4-6). Tomenson et al. (1997) present the exposure-effect
analysis, based on the estimated cumulative dichloromethane exposure groups, for all sites of
cancer, pancreatic cancer, and lung cancer, and there is no evidence of an increasing effect with
increasing exposure level in these analyses. A formal exposure-effect analysis for brain cancer
or leukemia was not presented. However, the authors described two of the brain cancer cases as
having "minimal" exposure to dichloromethane (and thus presumably would have been in the
69

-------
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
<400 ppm-year cumulative exposure group). One case was estimated as having 572 ppm-years
cumulative exposure, and the other case was an electrician classified in the unassigned exposure
group. He had worked for 21 years at an exposure level "that was unlikely to have exceeded
15 ppm 8 hour TWA."
Table 4-6. Mortality risk in Imperial Chemical Industries cellulose
triacetate film base production workers, Brantham, United Kingdom:
1,473 men employed 1946-1988, followed through 1994
Cancer type
Observed
Expected"
SMR
95% CI
Cancer, all sites
68
104.6
0.65
0.51-0.82
Liver and biliary duct
0
1.5
-
-
Pancreas
3
4.4
0.68
0.14-1.99
Lung, trachea, bronchus
19
41.3
0.46
0.29-0.75
Brain and CNS system
4
2.8
1.45
0.40-3.72
Lymphatic and hematopoietic
6
7.1
0.85
0.31-1.84
Leukemia
3
2.7
1.11
0.23-1.84
aExpected, calculated from observed and SMR data reported by the authors by using the following formula:
expected = 100 x observed ^ SMR; SMRs and CIs were not calculated for categories with zero observed cases.
Source: Tomenson et al. (1997).
A strength of this study was the monitoring data available that allowed assignment of
cumulative exposure categories for use in exposure-effect analyses. However, 30% (439) of
exposed workers had insufficient work histories to determine lifetime cumulative exposure. Air
measurements were not available until 1975, and personal measures were not available until
1980. In addition, the duration of exposure was relatively low (mean, 9 years), the mean
exposure level was relatively low (mean 8-hour TWA, 19 ppm), and there were very few deaths
from specific types of cancer, which limit the statistical power of the study to examine
associations among dichloromethane and specific cancers. Other limitations, as were also noted
in the Kodak cohort studies, include the use of mortality rather than incidence to define risk, the
reliance solely on underlying causes of death from death certificates to classify specific cancer
types and the lack of information on breast cancer risk.
4.1.3.4. Cellulose Triacetate Fiber Production Cohorts
4.1.3.4.1. Cellulose triacetate fiber production—Rock Hill, South Carolina (Hoechst Celanese
Corporation). Two cohorts of cellulose triacetate fiber workers have been studied in Rock Hill,
South Carolina (Lanes et al., 1993, 1990; Ott et al., 1983a, b), and Cumberland, Maryland (Gibbs
et al., 1996; Gibbs, 1992). Workers were exposed to dichloromethane, methanol, and acetone in
both facilities.
Ott et al. (1983a, b) conducted a retrospective cohort mortality study of 1,271 acetate
fiber production workers (551 men and 720 women) employed at least 3 months from 1954 to
70

-------
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
1977 at Dow Chemical Company, Rock Hill, South Carolina. This analysis focused on ischemic
heart disease mortality risk, and there was no presentation of cancer risk. The Rock Hill cohort
study was updated twice, through September 30, 1986 (Lanes et al., 1990), and December 31,
1990 (Lanes et al., 1993), and analyses of cancer mortality risks were included in these later
reports. Causes of death information was obtained from death certificates, with coding based on
the underlying and contributing causes (Ott et al., 1983a). The referent used in the updates was
the general population of York County, South Carolina, and analyses were adjusted for age, race,
gender, and calendar period. Because the results of the mortality risk analyses were similar for
both updates, those from the 1993 paper are presented here. The mean duration of work in the
cohort was not reported, but 56% worked for fewer than 5 years (calculated from Tables 3 and
4 of Ott et al., 1983b). The mean duration of follow-up was 23.6 years in the analysis through
1986 (Lanes et al., 1990) but was not reported in the later paper (Lanes et al., 1993). The
1993 report added approximately 4.25 years of follow-up, which would result in an estimate of
approximately 28 years follow-up for this report.
The Rock Hill, South Carolina, plant began producing cellulose triacetate fiber in 1954.
Dichloromethane was used as the solvent for the initial mixing with cellulose triacetate flakes.
This mixture was then filtered and transferred to the extrusion area for drying and winding. Air
measurements taken in 1977-1978 were assumed to be representative of operations since
dichloromethane use began in 1954, based on review of processing operations. The median
8-hour TWA exposures were estimated as 140, 280, and 475 ppm in the low, moderate, and high
categories of exposure (Ott et al., 1983a). Employment records provided information on jobs
held and dates employed, and this was used in conjunction with the exposure estimates for
specific jobs and work areas to classify individual exposures. However, detailed work history
information was only available for 475 (37%) of the workers (Lanes et al., 1990), and it is not
clear how the exposure assessment was applied to workers with missing job history data.
Methanol was also used in the cellulose triacetate fiber production process, and methanol
exposure was estimated as one-tenth that of dichloromethane. Acetone exposure was used in the
production of acetate (cellulose diacetate) fiber at an adjacent part of the plant. The exposure to
acetone was inversely related to that of dichloromethane, with estimated median 8-hour TWAs
of 1,080 ppm acetone in the low dichloromethane exposure group and 110 ppm acetone in the
moderate and high dichloromethane groups in the Rock Hill plant (Ott et al., 1983a).
In the latest follow-up (Lanes et al., 1993), there was no increase in mortality risk from
cancer (all sites) or from cancer of the lung or pancreas (Table 4-7). The SMR for liver and bile
duct cancer, based on four observed cases, was 2.98 (95% CI 0.81, 7.63). This was lower than
the SMR of 5.75 (95% CI 1.82, 13.8) that was reported in the 1990 analysis, based on these same
four cases but on a shorter follow-up period (and thus lower number of expected cases). Three
of these cases were bile duct cancers. This was the first cohort study that included women, and
this study provided data on breast cancer risk. There were 3 observed breast cancer deaths
71

-------
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
compared with 5.59 expected, yielding an SMR of 0.54 (95% CI 0.11, 1.57). No data were
provided pertaining to reproductive risk factors (e.g., pregnancy history) for breast cancer among
the women in this cohort, so it is difficult to assess whether these potential confounders are likely
to have been distributed differently in the cohort compared with the referent group. Information
about brain cancer, Hodgkin's lymphoma, and leukemia (Table 4-7) was not included in this
report but was included in the report by Gibbs (1992) (see Table 11 of that report).
Table 4-7. Mortality risk in Hoechst Celanese Corporation cellulose
triacetate fiber production workers, Rock Hill, South Carolina: 1,271 men
and women employed 1954-1977, followed through 1990
Cancer type
Observed
Expected
SMRb
95% CIb'c
Cancer, all sites
39
47.7
0.82
0.58-1.52
Liver and biliary duct
4
1.34
2.98
0.81-7.63
Pancreas
2
2.42
0.83
0.10-2.99
Lung, trachea, bronchus
13
16.21
0.80
0.43-1.37
Brain and CNSa
1
1.5
0.67
0.2-3.71
Hodgkin's lymphoma3
0
0.24
-
-
Leukemia3
1
1.11
0.90
0.02-5.0
Breast cancer (women)
3
5.59
0.54
0.11-1.57
"Data for brain and CNS cancer, Hodgkin's lymphoma, and leukemia were reported in Gibbs (1992).
bSMRs and CIs were not calculated for categories with zero observed cases.
°CIs were calculated from Breslow and Day (1987, Table 2.10).
Source: Lanes et al. (1993).
There are a number of limitations of this study, including the small size of the cohort,
small number of observed cancer deaths, availability of detailed work history information for
only 37% of the workers, and use of mortality rather than incidence data. The exposure levels at
this plant were high, but the duration of exposure for most of the cohort was relatively short
(<5 years). It is the first cohort study, however, that included women and provided information
on breast cancer risk.
4.1.3.4.2. Cellulose triacetate fiber production—Cumberland, Maryland (Hoechst Celanese
Corporation). Gibbs et al. (1996) studied a cohort of 2,909 cellulose triacetate fiber production
workers (1,931 men and 978 women) at a Hoechst Celanese plant in Cumberland, Maryland.
This retrospective cohort mortality study included all workers who were employed on or after
January 1, 1970, and who worked at least 3 months. This study also included a very small
comparison group (256 men, 46 women) that was described as a "0" or "no" exposure group of
workers at the plant who worked in jobs that were not considered to have had dichloromethane
exposure, for a total of 2,187 men and 1,024 women in the exposed and nonexposed groups.
72

-------
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
The plant closed in 1981, and mortality was followed through 1989. Since 1955,
employees of this plant were exposed to dichloromethane, methanol, acetone, and finishing oils
used as lubricants. Before 1955, acetone was the only exposure. Industrial hygiene monitoring
focusing on dichloromethane, acetone, and methanol began in the late 1960s. Exposure
groupings (low, 50-100 ppm, and high, 350-700 ppm) were assigned by area in which
employees worked. The extrusion and spinning workers and jet wipers were among the high
exposure group (300-1,250 ppm 8-hour TWA). The SMR analysis that was reported used
Allegany County, Maryland, as the comparison group. Cause of death information was obtained
from death certificates, but the authors did not state whether they used underlying or underlying
and contributing cause of death information. The mean duration of work in the cohort was not
reported. The total follow-up period included 49,828 person-years (16,292 in the high exposure
group and 33,536 in the low exposure group), and the mean duration of follow-up was 17.2 years
(range 8-20 years). These data were found in Hearne and Pifer (1999, Table 7).
There was little evidence of an increase in mortality risk from cancer (all sites) or from
cancer of the liver and bile duct, pancreas, or brain in men or in women (Table 4-8). An
increasing risk with increasing exposure level was seen for prostate cancer mortality in men.
Thep-walue for the trend was not given, but the authors describe it as a "nonstatistically
significant dose-response relationship." A statistically significant SMR for prostate cancer death
was seen in the 350-700 ppm group when latency (at least 20 years since first exposure) was
included in the analysis (SMR = 2.08,/? < 0.05). Cervical cancer mortality risk was increased,
but the small number of cases in the high exposure group did not allow a precise assessment of
the pattern with respect to exposure level. There was no increased risk of breast cancer.
73

-------
2958
Table 4-8. Cancer mortality risk in Hoechst Celanese Corporation cellulose
triacetate fiber production workers, Cumberland, Maryland: 2,909 men and
women employed 1970-1981, followed through 1989
Men (n = 1,931)	Women (n = 978)
Cancer type,
exposure level3
Obsb
Expb
SMR
95% cr
Obsb
Expb
SMR
95% cr
Cancer, all sites
121



42



50-100 ppm
64
70.0
0.91
0.70-1.2
37
44.79
0.83
0.58-1.1
350-700 ppm
57
75.6
0.75
0.57-0.98
5
4.61
1.1
0.35-2.5
Liver
2



0



50-100 ppm
1
1.33
0.75
0.02-4.2
0
1.04

-
350-700 ppm
1
1.24
0.81
0.02-4.5
0
0.10

-
Pancreas
3



1



50-100 ppm
2
2.24
0.89
0.1-3.2
1
1.73
0.58
0.01-3.2
350-700 ppm
1
2.90
0.35
0.01-1.9
0
0.18

-
Lung
35



11



50-100 ppm
20
25.7
0.78
0.48-1.2
9
8.24
1.1
0.50-2.1
350-700 ppm
15
27.3
0.55
0.31-0.91
2
0.87
2.3
0.28-8.3
Brain3
2



2



50-100 ppm
1
1.88
0.53
0.01-2.96
2
0.66
3.1
0.37-10.9
350-700 ppm
1
1.94
0.52
0.01-2.87
0
0.07


Hodgkin's3








50-100 ppm
1
0.4
2.5
0.06-13.9
0
0.23


350-700 ppm
0
0.41


0
0.02


Leukemia3








50-100 ppm
4
2.14
1.9
0.51-4.8
0
1.25


350-700 ppm
1
2.28
0.44
0.01-2.4
0
0.13


Prostate
22





\nl applicable

50-100 ppm
9
6.41
1.4
0.64-2.7




350-700 ppm
1 ^
¦ :<>
1 S
0 >)5 ' 1




Cervical

\nl applicable

6



50-100 ppm




5
1.69
3.0
0.96-6.9
350-700 ppm




1
0.19
5.4
0.13-30.1
Breast3
0



10



50-100 ppm
0
0.03


9
9.8
0.92
0.42-1.7
350-700 ppm
0
0.02


1
1.07
0.93
0.02-5.2
aData for brain and CNS cancer, Hodgkin's lymphoma, leukemia, and breast cancer reported in Gibbs (1992).
bObs = number of observed deaths, Exp = number of expected deaths. Referent group = Allegany County,
Maryland. SMRs and CIs were not calculated for categories with zero observed cases.
°CIs were calculated from Breslow and Day (1987, Table 2.10).
Sources: Gibbs et al. (1996); Gibbs (1992).
2959
2960
2961	A primary limitation of this study is that workers who were exposed before 1970 but
2962	were not working at the plant in 1970 were not included in the cohort. The authors had
2963	attempted to create a cohort of all workers who were employed on or after January I, 1954, but
2964	problems with the completeness of the personnel file made it impossible to use this study design.
74

-------
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
From what the author (Gibbs, 1992) was able to determine, the records of workers who had died,
left the company, or retired before the mid to late 1960s (when a new personnel system was
developed) were not available. Additional limitations include the small size of the cohort, small
number of observed cancer deaths, and use of mortality (death certificate) data. This is
particularly problematic for cancers with relatively high survival rates (such as prostate cancer
and cervical cancer), since incidence rates are not estimated well by mortality rates in this
situation.
4.1.3.5. Solvent-Exposed Workers—Hill Air Force Base, Utah
Spirtas et al. (1991) and Blair et al. (1998) evaluated exposure to dichloromethane in
relation to mortality risk in successive retrospective cohort studies of 14,457 civilian workers
employed at Hill Air Force Base in Utah for at least 1 year from 1952-1956. The analysis was
limited to the workers that were white or who had missing data on race, resulting in a sample
size of 14,066 (10,461 men, 3,605 women). Spirtas et al. (1991) examined mortality through
1982 (3,832 deaths), and Blair et al. (1998) updated mortality through 1990 (4,195 deaths). The
underlying and contributing causes of death information from death certificates was used to
classify cause-specific mortality. SMRs were calculated by using mortality rates from the Utah
population, and an internally standardized life table method was used to adjust for age at entry
into the cohort and competing causes of death. In the Blair et al. (1998) analysis, adjusted
relative risks (rate ratios) were estimated from a Poisson regression analysis with unexposed
workers as the referent. The mean duration of work was not reported. In the analysis through
1982 (Spirtas et al., 1991), there were 22,770 person-years of follow-up in men and
3,091 person-years of follow-up in women who were classified as exposed to dichloromethane.
The total number of workers classified as exposed to dichloromethane was 1,222 (Stewart et al.,
1991), which would yield an estimated mean of approximately 21 years follow-up through 1982.
The total number of person-years included in the later report (Blair et al., 1998), with the
addition of 8 more years of follow-up, was not reported but would be expected to increase the
mean follow-up time to approximately 29 years.
Two industrial hygienists developed the exposure assessment based on walkthrough
surveys, interviews with management and labor representatives, review of historical records, job
descriptions, monitoring data and other information pertaining to chemicals used, and
organization of the work site (Blair et al., 1998; Spirtas et al., 1991). Each worker was assigned
exposure by using information on the worker's job history, which included job titles, department
codes, and dates of employment. The most detailed exposure assessment was done for
trichloroethylene, the primary focus of the study. Dichloromethane, one of 25 other exposures
analyzed, was classified as a dichotomous exposure (ever exposed, never exposed).
Blair et al. (1998) presented the mortality risk for three specific cancers in relation to
15 of the 25 chemicals classified as dichotomized exposures. The rate ratios for non-Hodgkin's
75

-------
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
lymphoma and multiple myeloma in relation to dichloromethane in men were 3.0 (95% CI 0.9,
10.0) and 3.4 (95% CI 0.9, 13.2), respectively. These rate ratios, (particularly those for multiple
myeloma), were considerably higher than the rate ratios for any of the other chemicals examined,
in which the next highest observed rate ratio was 1.8 for Freon. No cases of either of these
cancers were observed in women with dichloromethane exposure, but the rate ratio for breast
cancer in these women was 3.0 (95% CI 1.0-8.8). Associations of similar magnitude (rate ratios
of 3.0-4.0) were also seen among breast cancer and some other exposures (Freon, solder flux,
isopropyl alcohol, and trichloroethane).
This is the largest of the cohort studies that were identified that included women and
specifically reported data pertaining to breast cancer risk. The major limitation of this study is
that the exposure assessment for dichloromethane was based on a dichotomized classification. In
addition, exposure to many different types of solvents was common; thus, it is difficult to
completely separate the effects of individual exposures. Some aspects of reproductive history,
such as age at first pregnancy, are known risk factors for breast cancer. Reproductive history
was not included in this analysis, but Blair et al. (1998) note that it is unlikely that these factors
would confound the results of a few specific chemicals, since the referent group was an internal
group within the cohort (and thus would be expected to be similar in terms of socioeconomic
status) and there was no association overall between solvent exposures and breast cancer
mortality.
4.1.3.6. Case-Control Studies of Specific Cancers and Dichloromethane
Seven site-specific cancer case-control studies included dichloromethane as an exposure
of interest. These studies involve six cancer sites: brain and CNS (Cocco et al., 1999; Heineman
et al., 1994), breast (Cantor et al., 1995), kidney (Dosemeci et al., 1999), pancreas (Kernan et al.,
1999), rectum (Dumas et al., 2000), and childhood leukemia (Infante-Rivard et al., 2005). A
synopsis of cohort studies in humans is provided in Table 4-9.
4.1.3.6.1. Case-control studies of brain cancer. Heineman et al. (1994) studied the association
between astrocytic brain cancer (International Classification of Diseases 9th ed. [ICD-9] codes
191, 192, 225, and 239.7) and occupational exposure to chlorinated aliphatic hydrocarbons.
Cases were identified by using death certificates from southern Louisiana, northern New Jersey,
and the Philadelphia area. This analysis was limited to white males who died between 1978 and
1981. Controls were randomly selected from the death certificates of white males who died of
causes other than brain tumors, cerebrovascular disease, epilepsy, suicide, and homicide. The
controls were frequency matched to cases by age, year of death, and study area.
Next of kin were successfully located for interview for 654 cases and 612 controls, which
represents 88 and 83% of the identified cases and controls, respectively. Interviews were
completed for 483 cases (74%) and 386 controls (63%). There were 300 cases of astrocytic
76

-------
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
brain cancer (including astrocytoma, glioblastoma, mixed glioma with astrocytic cells). The
ascertainment of type of cancer was based on review of hospital records, which included
pathology reports for 229 cases and computerized tomography reports for 71 cases. After the
exclusion of 66 controls with a possible association between cause of death and occupational
exposure to chlorinated aliphatic hydrocarbons (some types of cancer, cirrhosis of the liver), the
final analytic sample consisted of 300 cases and 320 controls.
In the next-of-kin interviews, the work history included information about each job held
since the case (or control) was 15 years old (job title, description of tasks, name and location of
company, kinds of products, employment dates, and hours worked per week). Occupation and
industry were coded based on four-digit Standard Industrial Classification and Standard
Occupational Classification (Department of Commerce) codes. The investigators developed
matrices linked to jobs with likely exposure to dichloromethane, five other chlorinated aliphatic
hydrocarbons (carbon tetrachloride, chloroform, methyl chloroform, tetrachloroethylene, and
trichloroethylene), and organic solvents (Gomez et al., 1994). This assessment was done blinded
to case-control status. Exposure was defined as the probability of exposure to a substance (the
highest probability score for that substance among all jobs), duration of employment in the
exposed occupation and industry, specific exposure intensity categories, average intensity score
(the three-level semiquantitative exposure concentration assigned to each job multiplied by
duration of employment in the job, summed across all jobs), and cumulative exposure score
(weighted sum of years in all exposed jobs with weights based on the square of exposure
intensity [1, 2, 3] assigned to each job). Secular trends in the use of specific chemicals were
considered in the assignment of exposure potential. Exposures were lagged 10 or 20 years to
account for latency. Thus, this exposure assessment procedure was quite detailed.
Adjusting for age and study area, the OR for the association between any exposure to
dichloromethane and risk of astrocytic brain cancer was 1.3 (95% CI 0.9, 1.8). There was a
statistically significant trend (p < 0.05) with increasing probability of exposure to
dichloromethane with an OR =1.0 (95% CI, 0.7, 1.6) for low probability, OR =1.6 (95% CI 0.8,
3.0) for medium probability, and OR = 2.4 (95% CI 1.0, 5.9) for high probability compared with
the referent group of unexposed men. An increased risk with higher duration of exposure was
also observed, with OR =1.7 (95% CI 0.9, 3.6) for 21 or more years of work in exposed jobs for
all exposed workers and OR = 6.1 (95% CI 1.1, 43.8) for the combination of 21 years or more of
work in a high probability of exposure job. Similar results were seen in additional analyses,
controlling for age, study area, employment in electronics occupations and industries, and
exposure to carbon tetrachloride, tetrachloroethylene, and trichloroethylene. There was also
evidence of an association between astrocytic brain cancer risk and dichloromethane exposure,
based on the average intensity score, with an OR =1.1 (95% CI 0.7, 1.7) for the low-medium
intensity group and an OR = 2.2 (95% CI 1.1, 4.1) for the high intensity group, and trend p-value
<0.05. The combination of high intensity and high duration (21 or more years) was strongly
77

-------
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
associated with risk (OR =6.1 [95% CI 1.5, 28.3]), and a weaker association (OR =1.4 [95% CI
0.6, 3.2]) was seen for high intensity and shorter duration (2-20 years). The association between
cumulative exposure score (low, medium, and high) and astrocytic brain cancer risk was
nonlinear (ORs of 0.9, 1.9, and 1.2 in the low, medium, and high exposure categories,
respectively).
The strengths of this case-control study include a large sample size, detailed work
histories (including information not just about usual or most recent industry and occupation but
also about tasks and products for all jobs held since age 15), and comprehensive exposure
assessment and analysis along several different dimensions of exposure. The major limitations
were the lack of direct exposure information and potential inaccuracy of the descriptions of work
histories that were obtained from next-of-kin interviews. Heineman et al. (1994) acknowledge
these limitations in the report, and, in response to a letter by Norman and Boggs (1996)
criticizing the methodology and interpretation of the study, Heineman et al. (1996) noted that,
while the lack of direct exposure information must be interpreted cautiously, it does not
invalidate the results. Differential recall bias between cases and controls was unlikely because
work histories came from next-of-kin for both groups, the industrial hygienists made their
judgments blinded to disease status, and the strong associations that were seen with the exposure
measures for dichloromethane were not seen with the other solvents included in the analysis.
The relatively strong and statistically significant associations between dichloromethane and
astrocytic brain tumors were seen along multiple measures of exposure, suggesting that the
results were unlikely to be spurious. Nondifferential misclassification would, on average,
attenuate true associations and would be unlikely to result in the types of exposure-response
relationships that were observed in this study.
Norman and Boggs (1996) described an apparent inconsistency in the estimated trends in
dichloromethane and carbon tetrachloride exposure based on the methodology used in this case-
control study (described in more detail in Gomez et al. [1994]). In response, Gomez (1996)
noted that the apparent inconsistency was actually due to an error in the labeling of the lines on
one of the figures in the report rather than an inconsistency with the estimated trends. Another
point raised by Norman and Boggs (1996) was that the Heineman et al. (1994) findings were
surprising in light of the lack of brain carcinogenesis in animals. In response, Heineman et al.
(1996) pointed out that carcinogens commonly cause different cancers in animals and humans. It
can also be noted that brain tumors are exceedingly rare in animal bioassays (Sills et al., 1999).
Norman and Boggs (1996) also suggested that the results of the Heineman et al. (1994) study be
given no weight when compared with the results of the cohort studies. The authors responded by
pointing out that the cohort studies had low statistical power and large CIs around their point
estimates but were not inconsistent with an association between dichloromethane and brain
cancer (Heineman et al., 1996). This point is strengthened further by the more recent results
from the Rochester, New York, Eastman Kodak cohort (Hearne and Pifer, 1999), described
78

-------
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
previously, since an increased SMR for brain and CNS cancers was seen in the longer follow-up
period of this cohort.
In another case-control study of brain cancer and dichloromethane exposure, Cocco et al.
(1999) identified 12,980 female cases of cancer of the brain and CNS through the underlying
cause of death listing (ICD codes 191 and 192) on death certificates from 24 states from 1984 to
1992. (This collection of death certificates is a data set created by the National Center for Health
Statistics, NIOSH, and the National Cancer Institute to facilitate research on occupational
exposures and mortality risk.) The cases included 161 women with meningioma (ICD-9 codes
192.1, 192.3). Four women who died of nonmalignant diseases, excluding neurological
disorders, were chosen as controls for each case. The controls were frequency matched to the
cases by state, race, and 5-year age group. Occupation data were based on the occupation fields
in the death certificates. This job was coded based on the three-digit industry and three-digit
occupation (Department of Census) codes. The investigators developed job exposure matrices
that were applied to these industry/occupation codes. The job exposure matrices included
probability and intensity scores for 11 occupational hazards, one of which was dichloromethane,
but also included other solvents, electromagnetic fields, chlorinated aliphatic hydrocarbons,
benzene, lead, nitrosamines, insecticides, herbicides, and public contact. The investigators used
logistic regression models to estimate ORs, adjusting for each workplace exposure, marital
status, three levels of socioeconomic status (based on occupation), and age at death. For each
chemical, four levels of intensity and probability were defined (unexposed, low, medium, and
high).
A weak association between dichloromethane exposure and brain/CNS cancer was seen
(OR 1.2 [95% CI 1.1, 1.3]) (Cocco et al., 1999). There was no exposure-related trend in the
association between probability or intensity of exposure and brain cancer. A similar but more
imprecise association was seen with meningioma cancer (OR 1.2 [95% CI 0.7, 2.2]). There were
too few cases of meningioma to stratify by exposure probability and intensity.
The major limitations of this study are the use of mortality rather than incidence data and
the reliance on occupation data from death certificates. The death certificate occupation data are
based on "usual" occupation, which may be more prone to misclassification in studies of women
because of gender-related differences in work patterns (i.e., shorter duration jobs for women
compared with men). A relatively broad job exposure matrix was applied to the job information,
and typically more generic job exposure matrices result in less sensitive assessment with limited
ability to detect exposure-response trends (Teschke et al., 2002). Nondifferential
misclassification of outcome and exposure would generally result in attenuated effect estimates.
4.1.3.6.2. Case-control studies of breast cancer. Cantor et al. (1995) conducted a case-control
study of occupational exposures and breast cancer, using the 24 state (1984-1989) death
certificate data described in the previous section. Cases were women with breast cancer coded as
79

-------
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
the underlying cause of death (ICD-9 code 174). Four female controls per case were selected
from all noncancer deaths, frequency matched by age (5-year age groups) and ethnicity (black,
white). The occupation listed on the death certificate was coded based on the three-digit industry
and three-digit occupation (Department of Census) codes, and this was used with a job exposure
matrix developed by the investigators to assess 31 workplace exposures, one of which was
dichloromethane. Four exposure probability and three exposure level scores were assigned.
ORs for probability and level were calculated for each ethnic group, adjusting for age at death
and a measure of socioeconomic status (based on occupation). After excluding subjects whose
death certificate occupations were listed as homemaker, there were 29,397 white cases and
4,112 black cases (total 33,509) and 102,955 white controls and 14,839 black controls (total
117,794).
There was little evidence of an association between exposure probability and breast
cancer mortality using the probability exposure metric. The ORs were 1.05 (95% CI 0.97, 1.1)
and 0.76 (95% CI 0.3, 2.0) in probability level 3 and level 4, respectively, for white women and
1.13 (95% CI 0.9, 1.4) in probability level 3 for black women. (There were too few black
women in exposure probability level 4 for analysis.) Weak associations were seen with exposure
level. In white women, an OR of 1.17 (95% CI 1.1, 1.3) was seen with the highest exposure
level, and in black women the OR in this exposure group was 1.46 (95% CI 1.2, 1.7). In the
analysis that jointly considered exposure level and probability ratings but excluded the lowest
probability of exposure, the OR for the highest category of exposure level was 1.28 in whites
(p < 0.05) and 1.21 in blacks.
As with the Cocco et al. (1999) case-control study that used a similar methodology, the
limitations of this study include the use of an outcome defined by mortality rather than incidence,
use of usual occupation information as recorded in death certificates, and use of a very broad job
exposure matrix to classify 31 different exposures. Although information on pregnancy and
lactation history (known risk factors for breast cancer) was not available, the authors did adjust
for socioeconomic status by using the occupation data, which may have corrected for some of the
potential confounding due to reproductive history.
4.1.3.6.3. Case-control studies of pancreatic cancer. Kernan et al. (1999) conducted a case-
control study of 63,097 pancreatic cancer cases, using the 24-state (1984-1993) death certificate
data. The diagnosis of pancreatic cancer was based on underlying cause of death (ICD-9 code
157). Four controls who had died during the same time period of causes other than cancer were
selected for each case, frequency-matched by state, race, gender, and 5-year age group
(n = 252,386). Usual occupation and industry, based on the occupation data in the death
certificate, were coded by using the three-digit (Department of Census) codes. A job-exposure
matrix was used with the industry and occupation codes to evaluate exposure intensity and
probability (each categorized as high, medium, or low) for formaldehyde, dichloromethane,
80

-------
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
10 other solvents, and a combined "organic solvents" measure. Race- and gender- specific
analyses were conducted by using logistic regression to estimate ORs and 95% CIs, adjusting for
age, marital status (ever, never married), residential area (metropolitan, nonmetropolitan), and
region (east, south central, south, and west).
The point estimates for the ORs in the low, medium, and high intensity categories in the
four race-gender groups ranged from 0.8 to 1.3, with no exposure-effect trend seen in any group.
The only statistically significant OR was for high exposure intensity in white females (OR 1.3
[95% CI 1.1-1.6]), with ORs of 1.0 (95% CI 0.9, 1.1) for medium intensity and 1.1 (1.0, 1.2) for
low intensity in this group. An elevated OR was seen with high exposure probability in black
males (OR 2.2 [95% CI 1.0, 4.8]) but not in white females (OR 1.0 [95% CI 0.8, 1.4]) or white
males (OR 1.0 [05% CI 0.8, 1.3]), and the ORs were 0.9 for medium exposure probability in
these three groups. There were relatively few black females in this study, resulting in imprecise
estimates (OR 2.0 [95% CI 0.8, 5.4] for medium exposure and OR 1.5 [95% CI 0.6, 3.6] for high
exposure).
The limitations of this study, as with the other case-control studies that used the 24-state
death certificate data set, include the reliance on cause of death data from death certificates rather
than medical-record validated incidence data and the use of death certificate occupation data.
The job exposure matrix used with the occupation data was more focused than those used in
Cocco et al. (1999) and Cantor et al. (1995). Although the analysis adjusted for some
sociodemographic characteristics, it did not include measures of smoking history or diabetes,
which are known risk factors for pancreatic cancer (Lowenfels and Maisonneuve, 2005).
4.1.3.6.4. Case-control studies of renal cancer. Dosemeci et al. (1999) reported data from a
population-based case-control study of the association between occupation exposures and renal
cancer risk. The investigators identified newly diagnosed patients with histologically confirmed
renal cell carcinoma from the Minnesota Cancer Surveillance System from July 1, 1988, to
December 31, 1990. The study was limited to white cases, and age and gender-stratified controls
were ascertained by using random digit dialing (for subjects ages 20-64) and from Medicare
records (for subjects 65-85 years). Of the 796 cases and 796 controls initially identified,
438 cases (273 men, 165 women) and 687 controls (462 men, 225 women) with complete
personal interviews were included in the occupational analysis.
Data were obtained through in-person interviews that included demographic variables,
residential history, diet, smoking habits, medical history, and drug use. The occupational history
included information about the most recent and usual industry and occupation (coded using the
standard industrial and occupation codes, Department of Commerce), job activities, hire and
termination dates, and full- and part-time status. A job exposure matrix developed by the
National Cancer Institute was used with the coded job data to estimate exposure status to
dichloromethane and eight other chlorinated aliphatic hydrocarbons.
81

-------
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
ORs were adjusted for age, smoking, hypertension and use of drugs for hypertension, and
body mass index. No association between renal cell carcinoma and exposure to dichloromethane
was observed in men (OR 0.85 [95% CI 0.6, 1.2]), women (OR 0.95 [95% CI 0.4, 2.2]), or both
sexes combined (OR 0.87 [95% CI 0.6, 1.2]).
A strength of this study includes the use of incident cases of renal cancer from a defined
population area, with confirmation of the diagnosis using histology reports. The occupation
history was based on usual and most recent job, in combination with a relatively focused job
exposure matrix. In contrast to the type of exposure assessment that can be conducted in cohort
studies within a specific workplace, however, exposure measurements, based on personal or
workplace measurements, were not used, and a full lifetime job history was not obtained.
4.1.3.6.5. Case-control studies of rectal cancer. Dumas et al. (2000) reported data from a case-
control study of occupational exposures and rectal cancer conducted in Montreal, Quebec,
Canada. The investigators identified 304 newly diagnosed cases of primary rectal cancer,
confirmed on the basis of histology reports, between 1979 and 1985; 257 of these participated in
the study interview. One control group (n = 1,295) consisted of patients with other forms of
cancer (excluding lung cancer and other intestinal cancers) recruited through the same study
procedures and time period as the rectal cancer cases. A population-based control group
(n = 533), frequency matched by age strata, was drawn by using electoral lists and random digit
dialing. The occupational assessment consisted of a detailed description of each job held during
the working lifetime, including the company, products, nature of work at site, job activities, and
any additional information from the interviews that could furnish clues about exposure. The
percentage of proxy respondents was 15.2% for cases, 19.7% for other cancer controls, and
12.6%) for the population controls.
A team of industrial hygienists and chemists blinded to subjects' disease status translated
jobs into potential exposure to 294 substances with three dimensions (degree of confidence that
exposure occurred, frequency of exposure, and concentration of exposure). Each of these
exposure dimensions was categorized into none, any, or substantial exposure. Logistic
regression models adjusted for age, education, proxy versus subject responder status, cigarette
smoking, beer drinking, and body mass index. Using the cancer control group, the OR for any
exposure to dichloromethane was 1.2 (95% CI 0.5, 2.8) and the OR for substantial exposure
(confident that exposure occurred with 5 or more years of exposure at medium or high frequency
and concentration) was 3.8 (95% CI 1.1, 12.2). The results using the population-based control
group for this exposure were not presented.
The strengths of this study were the large number of incident cases, specific information
about job duties for all jobs held, and a definitive diagnosis of rectal cancer. However, the use of
the general population (rather than a known cohort of exposed workers) reduced the likelihood
that subjects were exposed to dichloromethane, resulting in relatively low statistical power for
82

-------
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
the analysis. The job exposure matrix, applied to the job information, was very broad since it
was used to evaluate 294 chemicals.
4.1.3.6.6. Case-control studies of childhood leukemia. Infante-Rivard et al. (2005) examined
the association between maternal occupational exposures, before and during pregnancy, and risk
of childhood acute lymphoblastic leukemia (ICD-9 code 204.0) by using data from a population-
based case-control study in Quebec, Canada. Incident cases diagnosed from 1980-2000 were
identified from the cancer hospitals in the province, and diagnosis was confirmed based on
clinical records from an oncologist or hematologist. Between 1980 and 1993, cases ages 0-
9 years at diagnosis were included, and from 1994 to 2000 the age range was expanded to
14 years. The number of eligible cases identified was 848, and, of these, 790 parents (93%)
participated in the study. Population-based controls, individually matched to the sex and age at
diagnosis of the cases, were identified from government registries of all children in the province
(1980-1993) and the universal health insurance files (1994-2000). The parents of 790 (86%) of
the 916 eligible controls who were identified participated in the study.
Data were collected by using a structured telephone interview. Some information (i.e.,
job title, dates, type of industry, industry name and address) was obtained for all jobs held since
age 18, and additional information (e.g., materials and machines used, typical activities) was
obtained for jobs held by the mother from 2 years before the pregnancy through the birth of the
child. Specialized exposure modules were also used to collect information about specific jobs
(e.g., nurse, waitress, hair dresser, textile dry cleaner). All of this information was reviewed by
chemists and industrial hygienists, blinded to case-control status, to classify exposure to over
300 chemicals, although the primary focus of the study was on solvents (21 individual
substances, including dichloromethane, and six mixtures). The exposure assessment included
ratings of confidence (possible, probable, and definite), frequency of exposure during a normal
workweek (<5, 5-30, or >30% of the time), and level of concentration (low = slightly above
background, high = highest possible exposure in the study population, and medium for in-
between levels).
A weak association was seen between any dichloromethane exposure during the 2 years
before pregnancy up to the birth and risk of leukemia in the child (OR 1.34 [95% CI 0.54, 3.34]),
and results were similar when limited to exposures during pregnancy. Stronger associations
were seen with probable or definite exposure (OR 3.22 [95% CI 0.88, 11.7]) compared with
possible or no exposure. The estimates for categories based on concentration and frequency
were similar but there was no evidence for an increasing risk with increasing exposure level.
4.1.3.7. Summary of Cancer Studies by Type of Cancer
The cohort and case-control studies with data relevant to the issue of dichloromethane
exposure and cancer risk are summarized in Tables 4-9 and 4-10, respectively. The strongest of
the cohort studies, in terms of design, are two of the triacetate film base production cohorts
83

-------
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
(Cohort 1 in New York and the United Kingdom cohort, reported in Hearne and Pifer [1999] and
Tomenson et al. [1997], respectively). These are the cohorts with the most extensive exposure
assessment information. The start of eligibility for cohort entrance corresponds with the
beginning of the time when the exposure potential at the work site began, and the follow-up
period is relatively long (mean >25 years). Although Cohort 2 of the New York film base
production study has similar exposure data and follow-up, this cohort was limited to workers
employed between 1964 and 1970 and therefore would have missed anyone leaving (possibly
because of illness or death) before this time. In addition, because of the overlap between
Cohort 1 and Cohort 2, including both cohorts in an evaluation would be double counting
experiences of some individuals. Several limitations of the triacetate film base production
cohorts should be noted, however. One of these limitations concerns the generalizability of the
results, given the relatively low exposure level (mean 8-hour TWA <40 ppm) compared to the
other cellulose triacetate fiber production cohorts (Gibbs et al., 1996; Ott et al., 1983a).
Exposures in small, poorly ventilated work areas are also often much higher than those seen in
these film base production cohorts (Estill and Spencer, 1996; Anundi et al., 1993). Other
limitations include the limited power to detect a risk of low-incidence cancers (including brain
and leukemia), the lack of women and thus lack of data pertaining to breast cancer, and the use
of mortality rather than incidence data. Although the exposure levels in the cohorts involved in
cellulose triacetate fiber production were much higher than those of the film production cohorts,
the duration of exposure was relatively short in the South Carolina cohort (Lanes et al., 1993),
and the majority of workers were missing job history data. In the Maryland triacetate fiber
production plant, duration of exposure was not reported and the length of follow-up was
relatively short (mean, 17 years) (Gibbs et al., 1996). Also, the cohort began in 1970, even
though production began in 1955, and the missing personnel records made it impossible to
recreate an inception cohort. The exposure assessment in the study of civilian Air Force base
workers (Blair et al., 1998) allowed for only a dichotomized classification of exposure, and there
was considerable exposure to other solvents among these workers. This Air Force base study
was the largest of the cohort studies that included women and presented data pertaining to breast
cancer.
84

-------
Table 4-9. Summary of cohort studies of cancer risk and dichloromethane exposure
Cohort
Total n, exposure level" and
duration, length of follow-up
Inclusion criteriab
Exposure assessment;
Outcome assessment
Results0
Hearne and Pifer (1999)
Cellulose triacetate film
base production;
New York
Cohort 1
Cohort 2
Tomenson et al. (1997)
Cellulose triacetate film
base production;
United Kingdom
Lanes et al. (1993)
n = 1,311 men
Mean, 39 ppm
mean duration, 17 yr
mean follow-up, 35 yr
n = 1,013 men
mean, 26 ppm
mean duration, 24 yr
mean follow-up, 26 yr
n = 1,473 men
mean, 19 ppm
mean duration, 9 yr
mean follow-up, 27 yr
n = 551 men and 720 women (total
Cellulose triacetate fiber n = 1,271); median 140, 280, and
production;
South Carolina
Gibbs et al. (1996)
Began working after
1945; worked at least
1 yr
Work history (job records) and
personal/air monitoring;
death certificate (underlying causes)
475 ppm in low, moderate, and high,
respectively; 56% <5 yr work
duration; mean follow-up, ~28 yr
n = 1,931 men and 978 women (total
Cellulose triacetate fiber n = 2,909); 50-100 ppm in low and
Employed at least 1 yr Work history (job records) and
between 1964 and 1970 personal/air monitoring;
(potential exposure death certificate (underlying causes)
began 1946)
Employed anytime Work history (job records) and
between 1946 and 1988 personal/air monitoring;
death certificate (underlying causes)
Worked at least	Job history data and personal/air
3 months in the	monitoring of specific areas (but job
preparation or extrusion history data available for 37%);
areas from 1954 to death certificate (underlying and
1977	contributing causes)
production;
Maryland
Employed on or after
January 1, 1970, for at
least 3 months
350-700 ppm in high exposure;
duration not reported; mean follow-up (potential exposure
17 yr	began 1955)
Work history (job records) and
personal/air monitoring;
death certificate (fields used not
stated)
Elevated mortality risks seen for
brain cancer, Hodgkin's disease,
and leukemia (SMRs around 2.0);
no risk for liver, lung, or pancreatic
cancer (see Table 4-4)
Elevated mortality risks seen for
brain cancer, Hodgkin's disease,
leukemia, and pancreatic cancer
(SMRs between 1.5 and 3); no risk
for liver or lung cancer (see Table 4-
4)
Elevated mortality risks seen for
brain cancer (SMR 1.45); weak
elevation for leukemia; no risk for
liver, lung, or pancreatic cancer (see
Table 4-5)
Elevated mortality risk for liver
cancer (SMR 2.98, lower than seen
in earlier study of this cohort); no
risk for lung, pancreatic, or brain
cancer (see Table 4-7)
Elevated mortality risk for prostate
cancer (men, SMRs 1.4 and 1.8)),
cervical cancer (women, SMR >
3.0), and lung cancer in women
(high exposure, SMR 2.3), but not
in men; weak risk for liver cancer,
no risk for pancreatic or brain
cancer (see Table 4-8)
85

-------
Table 4-9. Summary of cohort studies of cancer risk and dichloromethane exposure
Cohort
Total n, exposure level" and
duration, length of follow-up
Inclusion criteriab
Exposure assessment;
Outcome assessment
Results0
Blair et al. (1998)
Air Force Base, Utah
n = 10,461 men and 3,605 women
(total n = 14,066)d
dichotomized (yes, no)
exposure duration not reported
mean follow-up -29 yr	
Employed at least 1 yr
from 1952 to 1956
(potential exposure
began 1939)
Work history (job records) and
industrial hygiene assessment based
on work site review (dichotomized
exposure);
Elevated mortality risk for non-
Hodgkin's lymphoma (RR 3.0) and
multiple myeloma (RR 3.4) in men,
and breast cancer in women (RR
(underlying and contributing causes) 3.0) (see section 4.1.3.5)
a8-hour TWA.
bIf dichloromethane was used at the plant before the first date of entrance into the cohort, the year that potential exposure began is noted.
°Results are described as elevated if SMR was around 1.5 or higher. There is limited statistical power for these cause-specific analyses in these cohort studies; the
statistical significance of individual estimates is not presented in this table. RR = relative risk,
includes whites and unknown race.
3336
3337
86

-------
3338
Table 4-10. Summary of case-control studies of cancer risk and dichloromethane exposure
Location
Cancer type, n cases, n controls (source), demographic
reference	group	
Exposure assessment
Results
Brain
Heineman et al. (1994)
Brain
Cocco et al. (1999)
Breast
Cantor et al. (1995)
Louisiana, New Jersey, Philadelphia
300 cases, 320 controls (death certificates);
cancer confirmed by hospital records; white
men
24 states, U.S.
12,980 cases, 51,920 controls (death
certificates); women
24 states, U.S.
33,509 cases, 117,794 controls (death
certificates); black and white women
Job exposure matrix applied to detailed
information on all jobs held (at least
1 year) since age 15, as obtained from
next-of-kin interviews; probability,
duration, intensity, and cumulative
exposure scores; six solvents evaluated
Job exposure matrix applied to death
certificate occupation; probability, and
intensity scores; 11 exposures evaluated
Job exposure matrix applied to death
certificate job data, probability, and
exposure level; 31 substances evaluated
OR 1.3 for any exposure; increased risk
with increased probability (trend p-valuc
<0.05, OR 2.4 for high probability),
increased duration, increased intensity;
strongest effects seen in high probability
plus high duration (OR 6.1) or high
intensity and high duration (OR 6.1)
combinations; no association with
cumulative exposure score (see section
4.1.3.6.1)
Weak association overall (OR 1.2), no
trend with probability or intensity scores
(see section 4.1.3.6.1)
Little evidence of association with
exposure probability; weak association
with exposure level in whites and in
blacks (see section 4.1.3.6.2)
Pancreas
Kernan et al. (1999)
Kidney
Dosemeci et al. (1999)
24 states, U.S.
63,037 cases, 252,386 controls (death
certificates); black and white men and
women
Minnesota
438 incident cases (Minnesota cancer
registry), 687 controls (random digit dialing
and Medicare records); cancer confirmed by
histology; men and women
Job exposure matrix applied to death
certificate occupation, probability, and
intensity scores; 11 chlorinated solvents
and formaldehyde evaluated
Job exposure matrices applied to most
recent and usual job, as ascertained from
interviews; nine solvents evaluated
Little evidence of associations with
intensity or probability (see section
4.1.3.6.3)
No evidence of increased risk associated
with dichloromethane (OR 0.85 in men,
0.95 in women) (see section 4.1.3.6.4)
Rectum
Dumas et al. (2000)
Montreal, Canada
257 incident cases, 1,295 other cancer
controls from 19 hospitals; 533 population-
based controls (electoral rolls and random
digit dialing), cancer confirmed by
histology; men
Job exposure matrix applied to detailed Little evidence of an association with any
information on all jobs held, as ascertained exposure (OR 1.2), but increased risk in a
from interviews; 294 substances evaluated small, "substantial exposure" group (OR
3.8) (using cancer controls; analysis of
population controls not given for this
exposure) (see section 4.1.3.6.5)
87

-------
Table 4-10. Summary of case-control studies of cancer risk and dichloromethane exposure
3339
Location
Cancer type, n cases, n controls (source), demographic
reference	group	
Exposure assessment
Results
Childhood leukemia (acute Quebec, Canada
lymphoblastic leukemia)
Infante-Rivard et al.
(2005)
790 incident cases (hospitals—all
provinces), 790 population-based controls
(government population registries); cancer
based on oncologist or hematologist
diagnosis
ages 0-14,a both sexes
Systematic review of detailed information
on all jobs held by the mother from
2 years before pregnancy through birth of
the child; 21 individual substances and six
mixtures evaluated (mostly solvents);
confidence, frequency, and concentration
of exposure rated
Little evidence of association with any
exposure (OR 1.3), but stronger
associations (OR > 3.0, referent group =
possible/no exposure) with probable or
definite and with combinations of
frequency and concentration (see section
4.1.3.6.6)
aFrom 1980 to 1993, study was limited to diagnoses of ages 0-9, but this was expanded between 1994 and 2000 to ages 0-14.
88

-------
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
Case-control studies offer the potential for increased statistical power for assessing
associations with relatively rare cancers, such as brain cancer and leukemia. Case-control
studies are often designed to examine incidence rather than mortality, which is of particular
importance in etiologic research for diseases with relatively high survival rates and diseases in
which survival may be strongly related to factors that are difficult to adjust for without detailed
data collection (e.g., access to health care). There is a considerable range in the detail and
quality of the exposure assessment used in case-control studies, however. Case-control studies
rarely include specific measurements taken at specific work sites of individual study participants.
Although it is more difficult to determine absolute exposure levels without these individual
measurements, the exposure assessment methodology used in case-control studies can result in
useful between-group comparisons of risk if the intra-group variability is less than the inter-
group variability in potential exposure levels. Among the case-control studies with data
pertaining to cancer risk and dichloromethane exposure, the two studies with the strongest
designs are the study of brain cancer by Heineman et al. (1994) and the study of childhood
leukemia by Infante-Rivard et al. (2005). These are the studies that obtained detailed
information about all jobs held (rather than just the usual or most recent job), focused on a
relatively small number of exposures, and used medical record data to confirm the diagnosis.
Heineman et al. (1994) obtained the work history from interviews with next-of-kin, however,
which is most likely to have resulted in nondifferential misclassification of exposure, and thus
attenuation in the observed associations. The use of death certificate data to classify disease and
occupational exposures in the three studies using the large 24 state death certificate database
(brain cancer: Cocco et al. [1999]; breast cancer: Cantor et al. [1995]; pancreatic cancer: Kernan
et al. [1999]) is also likely to have resulted in nondifferential misclassification of both outcome
and exposure (and thus attenuated associations).
Considering the issues described above with respect to the strengths and limitations of the
available epidemiologic studies, a summary of the epidemiologic evidence relating to
dichloromethane exposure and specific types of cancer can be made, as described below. The
available epidemiologic data suggest an association between dichloromethane and brain cancer
and liver cancer, but not lung cancer.
4.1.3.7.1. Brain and CNS cancer. An increased risk of brain and CNS cancers was seen in the
strongest cohort studies; SMRs were 2.16 in Cohort 1 in New York (Hearne and Pifer, 1999) and
1.45 in the United Kingdom cohort (Tomenson et al., 1997). These estimates are based on a
small number of observations (six cases in New York and four in the United Kingdom) and so
are relatively imprecise. It is only in the latest follow-up of the New York film base production
cohort that an elevated SMR was observed, further suggesting that the statistical power of the
other cohort studies for examining risk of this disease may be quite low. Two case-control
studies of dichloromethane exposure and brain cancer have been conducted (Cocco et al., 1999;
89

-------
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
Heineman et al., 1994). The Heineman et al. (1994) study, which is the stronger study in terms
of exposure assessment strategy and confirmation of diagnosis, reported relatively strong trends
with increasing probability, duration, and intensity measures of exposure, but a nonlinear trend
was seen with the cumulative exposure metric. This difference could reflect a more valid
measure of relevant exposures in the brain from the intensity measure, as suggested by the study
in rats reported by Savolainen et al. (1981) in which dichloromethane levels in the brain were
much higher with a higher intensity exposure scenario compared with a constant exposure period
with an equivalent TWA (see section 3.2). The available epidemiologic studies provide some
evidence of an association between dichloromethane and brain cancer, and this area of research
represents a data gap in the understanding of the carcinogenic potential of dichloromethane.
4.1.3.7.2.	Liver and biliary duct cancer. Liver and biliary duct cancer are relatively uncommon
(age-adjusted incidence 6.2 per 100,000 person-years) (SEER website, seer.cancer.gov, accessed
April 2006), so it is difficult to study in most occupational cohorts of limited size. The cohort
study with the higher exposures, the Rock Hill, South Carolina, triacetate fiber production plant,
suggested an increased risk of liver cancer (Lanes et al., 1993, 1990). The SMR for liver and
bile duct cancer was 2.98 (95% CI 0.81, 7.63) in the latest update of this cohort. This
observation was based on four cases; two of these cases were biliary duct cancers. As the
follow-up period has increased, the strength of this association has decreased, although it is
relatively strong (albeit with wide CIs). The decrease in the SMR with increasing follow-up
reflects the increase in number of expected cases, because the four observed cases were seen
earlier in the follow-up period. No other cohort study has reported an increased risk of liver
cancer mortality, although it should be noted that there is no other inception cohort study of a
population with exposure levels similar to those of the Rock Hill plant, and no data from a case-
control study of liver cancer are available pertaining to dichloromethane exposure. The available
epidemiologic studies, with biological plausibility inferred from the results from studies in mice
and female rats (see section 4.2) (NTP, 1986; Serota et al., 1986a, b; Nitschke 1988a), provide
some evidence of an association between dichloromethane and liver and biliary duct cancer,
although it should be noted that this evidence is based on very limited epidemiologic data.
4.1.3.7.3.	Lung cancer. In the stronger cohort studies (Cohort 1 in the New York Eastman
Kodak Company triacetate film production study reported by Hearne and Pifer [1999] and the
United Kingdom triacetate film production study reported by Tomenson et al. [1997]), the SMRs
for lung cancer were well below 1.0. The New York study had also obtained data on smoking
history that indicated it was unlikely that differences in smoking could be masking an effect of
dichloromethane (Hearne et al., 1987). Lung cancer is a common cancer (age-adjusted incidence
61 per 100,000 person-years) (SEER website, seer.cancer.gov, accessed April 2006), so the
expected rates, even in small cohorts, are based on relatively robust estimates. The only group in
90

-------
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
any study that had an increased risk for lung cancer was the high-exposure women in the
triacetate fiber production cohort in Maryland (Gibbs et al., 1996). However, this was based on
only two cases and was a highly imprecise estimate (SMR 2.3 [95% CI 0.28, 8.3]). No case-
control study of dichloromethane exposure and lung cancer risk is available. The available
epidemiologic studies do not provide evidence for an association between dichloromethane and
lung cancer, although it should be noted that the studies with the best designs are limited to
relatively low exposure levels.
4.1.3.7.4.	Pancreatic cancer. An early study (Hearne et al., 1990) of Cohort 2 of the New York
triacetate film production cohort had reported 8 observed and 4.2 expected pancreatic cancer
deaths, for a twofold increased SMR (p = 0.13). This association was reduced in the subsequent
follow-up (SMR 1.5 [95% CI 0.7, 3.0]) (Hearne and Pifer, 1999) but was not seen in the more
methodologically sound Cohort 1 (SMR 0.92) or in any of the other cohorts. A meta-analysis of
the cohort studies (using the data of Hearne et al. [1990]) reported a summary association of
1.42 (95%) CI 0.80, 2.53) (Ojajarvi et al., 2001). This summary measure would be further
reduced with the updated data for Cohort 2 and the addition of Cohort 1 from Hearne and Pifer
(1999). The only case-control study of pancreatic cancer mortality risk and dichloromethane
exposure (based on death certificate data) did not report consistent patterns with respect to
intensity or exposure among the race-sex groups studied. The available epidemiologic studies do
not provide evidence for an association between dichloromethane and pancreatic cancer.
4.1.3.7.5.	Leukemia and lymphoma. Each of the individual hematopoietic cancers is relatively
uncommon, with age-adjusted incidence rates of 5 per 100,000 person-years or less (SEER
website, seer.cancer.gov, accessed April 2006). The relatively inconsistent (point estimates
ranging from 0.50 or less to 2.0 or higher) and imprecise measures of association between
dichloromethane exposure and non-Hodgkin's lymphoma, Hodgkin's lymphoma, myeloma, and
leukemia are thus expected, given the relatively small size of the available cohort studies. Only
one case-control study of any of these diseases and dichloromethane is available, and this is a
study of childhood leukemia (acute lymphoblastic leukemia) in relation to maternal occupational
history (Infante-Rivard et al., 2005). This is a large, population-based study of confirmed
incident cases of leukemia, with a detailed exposure assessment pertaining to the period before
and during pregnancy. A threefold increased risk was seen with probable or definite exposure
(OR 3.22 [95% CI 0.88, 11.7]) compared with possible or no exposure. The available
epidemiologic studies do not provide an adequate basis for the evaluation of the role of
dichloromethane in any of the specific hematopoietic cancers because of the small size of the
cohort studies and the relative lack of case-control studies pertaining to these outcomes.
91

-------
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
4.1.3.7.6. Breast cancer. Only one large cohort study included women and reported data
pertaining to breast cancer risk (Blair et al., 1998), and this is a cohort with a limited exposure
assessment (dichotomized) and multiple exposures. A relatively strong association was seen
between dichloromethane exposure and breast cancer mortality in this study (rate ratio 3.0 [95%
CI 1.0, 8.8]). Similar associations were seen with several other chemicals, and the potential
effect of confounding and misclassification of these exposures may have biased the estimate in
either direction. The only case-control study of breast cancer risk and dichloromethane exposure
used the 24-state death certificate data to classify exposure and disease. The available
epidemiologic studies do not provide an adequate basis for the evaluation of the role of
dichloromethane in breast cancer because there are currently no cohort studies with adequate
statistical power and no case-control studies with adequate exposure methodology to examine
this relationship.
4.2. SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
ANIMALS—ORAL AND INHALATION
4.2.1. Oral Exposure: Overview of Noncancer and Cancer Effects
Results from studies of animals exposed by the oral route for short-term, subchronic, and
chronic durations identify the liver and the nervous system as the most sensitive targets for
noncancer toxicity from repeated oral exposure to dichloromethane. In a 90-day exposure study,
nonneoplastic histopathologic changes in the liver were observed in F344 rats exposed to
drinking water doses of >166 mg/kg-day (males) or >209 mg/kg-day (females) (Kirschman et al.,
1986). Similar changes were seen in F344 rats in a 2-year exposure of >50 mg/kg-day (Serota et
al., 1986a).
The 2-year oral exposure study in F344 rats did not produce evidence of increasing
incidence of liver tumors across all of the dose groups in males or females (Serota et al., 1986a).
In females, however, a jagged stepped pattern of increasing incidence was observed. In a
parallel study in B6C3Fi mice (Serota et al., 1986b; Hazelton Laboratories, 1983), a clearer trend
with respect to hepatic cancer was seen in males but not females.
None of the chronic oral exposure studies included a systematic measurement of potential
neurological effects. One 14-day study focusing on neurobehavioral changes is available,
however. Changes in autonomic, neuromuscular, and sensorimotor functions were observed in
F344 rats exposed for 14 days to gavage doses >337 mg/kg-day (Moser et al., 1995) (see section
4.4.3 for more details).
No effects on reproductive parameters were observed in Charles River CD rats exposed
for 90 days to gavage doses as high as 225 mg/kg-day (General Electric Co., 1976) or in
pregnant F344 rats exposed to gavage doses of up to 450 mg/kg-day on gestation days (GDs) 6-
19 (Narotsky and Kavlock, 1995). However, no oral exposure studies examining developmental
neurobehavioral effects have been conducted (see section 4.3 for more details).
92

-------
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
4.2.1.1. Toxicity Studies of Subchronic Oral Exposures: Hepatic Effects
Kirschman et al. (1986) examined the toxicity of dichloromethane in a 90-day drinking
water study in F344 rats (20/sex/dose level). The nominal concentration of dichloromethane in
the water was 0.15, 0.45, or 1.5%. Based on BW and water consumption data, average intakes
were reported to be 0, 166, 420, or 1,200 mg/kg-day for males and 0, 209, 607, or 1,469 mg/kg-
day for females. Clinical chemistry tests (hematological and chemical variables in samples of
blood and urine) and tissue histopathology were evaluated in groups of five rats/sex/dose level
after 1 month of treatment. These endpoints were also evaluated in the remaining rats sacrificed
after 90 days of exposure.
Exposure to dichloromethane did not affect mortality or cause adverse clinical signs of
toxicity. Gross necropsy was also unremarkable. Reported changes in mean values for clinical
chemistry variables, compared with controls, included elevated serum ALT activities for all
treated males at 1 month and for the high-dose females at 3 months, elevated serum AST activity
in high-dose females at 3 months, elevated serum lactate dehydrogenase activities in mid- and
high-dose females at 3 months, and decreases in serum concentrations of fasting glucose,
cholesterol, and triglycerides in all exposed groups of both sexes at 1 and 3 months. Actual
values for clinical chemistry variables, however, were not presented in the report.
No histopathologic alterations were seen in tissues after 1 month of treatment (a detailed
description of tissues examined was not presented). In rats exposed for 3 months, exposure-
related histopathologic changes were restricted to the liver. Elevated, statistically significant,
incidences of hepatocytic vacuolation were observed in all exposed male groups and in the mid-
and high-dose female groups (see Table 4-11). The most frequently observed vacuolation was
described as generalized and occurring throughout the lobule, and Oil Red-O-staining indicated
most were lipid-containing vacuoles. The incidences of generalized vacuolation scored as mild
or moderate were higher in all of the female dose groups compared with the controls. The
authors stated that the no-observed-adverse-effect level (NOAEL) based on this study is less than
200 mg/kg-day and the lowest-observed-adverse-effect level (LOAEL) for males was
166 mg/kg-day. The authors did not explicitly provide a LOAEL for females. The results
indicate that 166 mg/kg-day and 209 mg/kg-day were the LOAELs for liver effects in male and
female rats, respectively.
93

-------
3522
Table 4-11. Incidences of histopathologic changes in livers of male and
female F344 rats exposed to dichloromethane in drinking water for 90 days
Lesion, by sex
Controls
Low dose
Mid dose
High dose
Males—n per group3
15
15
15
15
Estimated mean intake (mg/kg-day)
0
166
420
1,200
Number (%) with




Hepatocyte vacuolation (generalized, centrilobular,
1(7)
10b (67)
9b (60)
7b (47)
or periportal)




Generalized vacuolation severity:
0(0)
5b (33)
8b(53)
6b (40)
minimal
0
4
7
6
mild
0
0
1
0
moderate
0
1
0
0
Centrilobular severity:
0(0)
1(7)
0(0)
2(13)
minimal
0
1
0
0
mild
0
0
0
2
moderate
0
0
0
0
Hepatocyte degeneration
0(0)
0(0)
0(0)
2(13)
Focal granuloma
1(7)
0(0)
0(0)
1(7)
Females—n per group3
15
15
15
15
Estimated mean intake (mg/kg-day)
0
209
607
1,469
Number (%) with




Hepatocyte vacuolation (generalized, centrilobular,
6(40)
13b(87)
15b(100)
15b (100)
or periportal)




Generalized vacuolation severity:
5(33)
13b(87)
15b(100)
15b (100)
minimal
5
8
6
8
mild
0
4
5
6
moderate
0
1
4
1
Centrilobular severity:
0(0)
0(0)
1(7)
llb (28)
minimal
0
0
0
2
mild
0
0
1
4
moderate
0
0
0
3
marked
0
0
0
2
Hepatocyte degeneration
0(0)
0(0)
0(0)
12b(80)
Focal granuloma
0(0)
0(0)
4 (27)°
6b (40)
320 per group; 5 sacrificed at 1 month; these endpoints for the remaining 15 per group.
Statistical significance testing not reported by authors; Fisher's exact test for comparison with control p—
value <0.05 (two-sided).
Statistical significance testing not reported by authors; Fisher's exact test for comparison with control p -
value <0.10 (two-sided). Authors stated LOAEL = 166 mg/kg-day in males but did not explicitly provide LOAEL
for females; NOAEL is less than 200 mg/kg-day.
Source: Kirschmanetal. (1986).
3523
3524	Kirschman et al. (1986) conducted a similar 90-day study in B6C3Fi mice (20/sex/dose
3525	level). The estimated average intakes were 0, 226, 587, or 1,911 mg/kg-day for males and 231,
3526	586, or 2,030 mg/kg-day for females. Six mice (two controls, two low dose, and two mid dose)
3527	died during the study from unknown causes. Administration of dichloromethane did not cause
3528	adverse clinical signs of toxicity or affect food consumption, ophthalmology, or serum ALT
3529	activity. Gross necropsy examinations also were unremarkable.
3530	Histopathologic evaluation of tissues from mice killed after 1 month of treatment did not
3531	reveal any compound-related effects. Evaluation at 3 months showed subtle generalized or
94

-------
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
centrilobular changes in the liver (characterized as increased vacuolation with fat deposition),
which was evident in all exposed groups and most prominent in mid- and high-dose female
groups (Table 4-12). The most frequently detected change was characterized as a generalized
vacuolation. Some evidence was found for an increase in severity of the generalized vacuolation
with increasing exposure level, but the incidence of this lesion in the control mice was
substantial, especially in females (Table 4-12). Incidences for centrilobular vacuolation were
significantly increased only for the mid-dose female group. No other changes were found.
Using the results from this study to select doses for a chronic study, Kirschman et al.
(1986) expressed the opinion that the mid-dose level (587 mg/kg-day) was the LOAEL in this
study. Although incidences for generalized vacuolation were increased in the low- and mid-dose
male groups, the incidences in the high-dose groups were not significantly increased compared
with controls (Table 4-12). The study authors identified a LOAEL of 586 mg/kg-day for
centrilobular vacuolation in male B6C3Fi mice. The NOAEL for males was considered by the
investigators to be between 226 and 587 mg/kg-day.
95

-------
3547
Table 4-12. Incidences of histopathologic changes in livers of male and
female B6C3Fi mice exposed to dichloromethane in drinking water for
90 days
Lesion, by sex
Controls
Low dose
Mid dose
High dose
Males—n per group3
14
14
14
15
Estimated mean intake (mg/kg-day)
0
226
587
1,911
Number (%) with




Hepatocyte vacuolation (generalized, centrilobular,
9(64)
12 (86)
13 (93)
12 (80)
or periportal)




Generalized vacuolation, severity:
7(50)
12b (86)
13b(93)
10 (67)
minimal
4
3
9
7
mild
2
7
5
3
moderate
1
2
0
0
marked
0
0
0
0
Centrilobular severity:
2(14)
0(0)
1(7)
5(33)
minimal
2
0
0
1
mild
0
0
0
3
moderate
0
0
1
1
Females—n per group3
14
11
13
15
Estimated mean intake (mg/kg-day)
0
231
586
2,030
Number (%) with




Hepatocyte vacuolation (generalized, centrilobular,
13 (93)
11 (100)
13 (100)
13 (87)
or periportal)




Generalized vacuolation severity:
13 (93)
11 (100)
13 (100)
13 (87)
minimal
1
3
5
3
mild
8
7
6
6
moderate
4
1
2
1
marked
0
0
0
3
Centrilobular severity:
0(0)
0(0)
5° (39)
1(7)
minimal
0
0
0
0
mild
0
0
2
1
moderate
0
0
3
0
marked
0
0
0
0
320 per group; 5 sacrificed at 1 month.
Statistical significance testing not reported by authors; Fisher's exact test for comparison with controlp -
value = 0.10 for low dose group andp = 0.032 for mid-dose group (two-sided).
Statistical significance testing not reported by authors; Fisher's exact test for comparison with controlp -
value = 0.016 (two-sided). Authors say LOAEL = 587 mg/kg-day; NOAEL between 226 and 587 mg/kg-day for
males; not explicitly stated for females.
Source: Kirschmanetal. (1986).
3548
3549
3550	4.2.1.2. Toxicity Studies of Chronic Oral Exposures: Hepatic Effects and Carcinogenicity
3551	Longer-term (up to 2-year) oral exposure studies in mice and rats are summarized in
3552	Table 4-13 and described in more detail below. These studies provide additional information
3553	pertaining to hepatotoxicity and carcinogenicity.
3554
3555
3556
96

-------
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
Table 4-13. Studies of chronic oral dichloromethane exposures (up to
2 years)
Reference,
strain/species
Number per
group
Exposure information
Comments
Serota et al.
(1986a)
F344 rats
Serota et al.
(1986b);
Hazelton
Laboratories
(1983)
B6C3Fi mice
Maltoni et al.
(1988)
Sprague-Dawley
rats
Maltoni et al.
(1988)
Swiss mice
85/sex/dose +135
controls
Males 125, 200,
100, 100, 125
Females 100, 100,
50, 50, 50
50/sex/dose
50/sex/dose + 60
controls
Drinking water, 2 years, target
dose 0, 5, 50, 125, 250 mg/kg-day
Mean intake:
males 0, 6, 52, 125,
235 mg/kg-day
females 0, 6, 58, 136,
263 mg/kg-day
Drinking water, 2 years, target
dose 0, 60, 125, 185, 250 mg/kg-
day
Mean intake:
males 0,61, 124, 177,
234 mg/kg-day
females 0, 59, 118, 172,
238 mg/kg-day
Gavage, up to 64 weeks
0, 100, 500 mg/kg-day, 4-5 days
per week
Gavage, up to 64 weeks
0, 100, 500 mg/kg-day, 4-5 days
per week
Nonneoplastic liver effects
(foci/areas of alteration) in males
and females (see Table 4-14)
Jagged stepped pattern of
increasing incidence of
neoplastic nodules or
hepatcellular carcinoma in
females (i.e., increased in the 50
and 250 mg/kg-day but not 125
mg/kg-day groups) (see Table 4-
14)
Increasing trend of liver cancer
(hepatocellular adenoma or
carcinoma) in males (see Table
4-15)
High mortality in high dose
group led to termination of study
at 64 weeks; non-statistically
significant increase in malignant
mammary tumors in female rats
High mortality in high dose
group led to termination of study
at 64 weeks
4.2.1.2.1. Chronic oral exposure in F344 rats (Serota et al, 1986a). Treatment with
dichloromethane did not induce adverse clinical signs or affect survival in the F344 rats (Serota
et al., 1986a). BWs of rats in the 125 and 250 mg/kg-day groups were generally lower than in
controls throughout the study. The authors stated that the differences, although small, were
statistically significant, but the data were not shown in the published report. Water consumption
was lower throughout the study in both sexes of rats from the 125 and 250 mg/kg-day groups
relative to controls; food consumption was also lower in these groups during the first 13 weeks
of treatment. Mean hematocrit, hemoglobin, and red blood cell count were increased in both
sexes at dichloromethane levels of 50, 125, and 250 mg/kg-day for 52 and 78 weeks. Half of
these increases were reported to be statistically significant, but the report did not provide the
numerical values or specify which parameters were significant. Clinical chemistry results
showed decreases in alkaline phosphatase (AP), creatinine, blood urea nitrogen, total protein, and
cholesterol in both sexes at 250 mg/kg-day, and most of these changes were statistically
97

-------
3572	significant at one or both of the intervals evaluated. (Significant parameters not specified and the
3573	mean group values were not presented in the published report.) No significant deviations in
3574	urinary parameters were observed. Organ weights were not significantly affected by treatment
3575	with dichloromethane.
3576	No treatment-related histopathological effects were noted in the tissues examined except
3577	for the liver (Serota et al., 1986a). Examination of liver sections showed a dose-related positive
3578	trend (positive Cochran-Armitage trend test) in the incidences of foci/areas of cellular alteration
3579	in treated F344 rats (Table 4-14). Comparisons of incidences with control incidences indicated
3580	statistically significant elevations at all dose levels except 5 mg/kg-day. These liver changes
3581
Table 4-14. Incidences of nonneoplastic liver changes and liver tumors in
male and female F344 rats exposed to dichloromethane in drinking water for
2 years
Target dose (mg/kg-day)

Controls




Trend
250 with

0a
5
50
125
250
/>-valucb
recovery0
Males—n per groupd
76
34
38
35
41

15
Estimated mean intake (mg/kg-day)
0
6
52
125
235

232
Number (%) with







Liver foci/areas of alteration
52 (70)
22 (65)
35 (92)e
34 (97)e
40 (98)e
<0.0001
15 (100)e
Neoplastic nodules
9(12)
1(3)
0(0)
2(6)
1(2)
Not
2(13)






reported

Hepatocellular carcinoma
3 (4)
0(0)
0(0)
0(0)
1(2)
Not
0(0)






reported

Neoplastic nodules and
12 (16)
1(3)
0(0)
2(6)
2(5)
Not
2(13)
hepatocellular carcinoma





reported

Females—n per groupd
67
29
41
38
34

20
Estimated mean intake (mg/kg-day)
0
6
58
136
263

239
Number (%) with







Liver foci/areas of alteration
34(51)
12 (41)
30 (73)e
34 (89)e
31 (91)e
<0.0001
17 (85)e
Neoplastic nodules
0 (0)
1(3)
2(5)
1(3)
3(9)
Not
2(10)






reported

Hepatocellular carcinoma
0(0)
0(0)
2(5)
0(0)
2(6)
Not
0(0)






reported

Neoplastic nodules and
0(0)
1(3)
4 (10)f
1(3)
5 (14)f
p < 0.01
2 (10)f
hepatocellular carcinoma







3582
3583	aTwo control groups combined.
3584	bCochran- Armitage trend test was used for trend test of liver foci/areas of alteration. For tumor mortality-unadjusted
3585	analysis, a Cochran-Armitage trend test was used, and, for tumor mortality-adjusted analyses, tumor prevalence
3586	analytic method by Dinse and Lagakos (1982) was used. Similar results were seen in these two analyses.
3587	°Recovery group was exposed for 78 weeks and then had a 26-week period without dichloromethane exposure;
3588	n = 15 for nonneoplastic lesions and n = 17 for neoplastic lesions.
3589	d Number available at terminal sacrifice; starting with 135 controls (combining both control groups) and 85 per sex
3 590	per dose group except recovery group (n = 25); subtracted 5, 10, and 20 per group (except for recovery group)
3591	sacrificed at 25, 52, and 78 weeks, respectively, and subtracted unscheduled deaths, which ranged from 5 to 19 per
3592	group.
3 593	"Significantly (p < 0.05) different from control with Fisher's exact test.
3 594	Significantly (p < 0.05) different from controls with Fisher's exact test, mortality-unadjusted and mortality-adjusted
3595	analyses.
3596
3597	Source: Serota et al. (1986a).
98

-------
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
were first noted after treatment for 78 weeks and progressed until week 104. Livers of animals
treated with dichloromethane also showed an increased incidence of fatty change, but incidence
data for this lesion were not presented in the published report. The recovery group also showed
an increased incidence of areas of cellular alterations, but the fatty changes were less pronounced
than in the 250 mg/kg-day group dosed for 104 weeks. The authors indicate that 5 mg/kg-day
was a NOAEL and 50 mg/kg-day was a LOAEL for nonneoplastic liver changes in male and
female F344 rats exposed to dichloromethane in drinking water for 2 years.
Dichloromethane-exposed male rats showed no statistically significant increased
incidence of liver tumors. In females, there was a positive trend for increasing incidence of
hepatocellular carcinoma or neoplastic nodules with increasing dose (Table 4-14) (Serota et al.,
1986a). Statistically significant increases in tumor incidences were observed in the 50 and
250 mg/kg-day groups (incidence rates of 10% and 14%, respectively) but not in the 125 mg/kg-
day group (incidence rate of 3%). Incidence was also increased (10%) in a group exposed for
78 weeks followed by a 26-week period of no exposure. The characterization of malignant
potential of the nodules was not described, however, and no trend was seen in the data limited to
hepatocellular carcinomas. The incidence of hepatocellular carcinoma or neoplastic nodules in
this control group (0%) was lower than that seen in historical controls from the same laboratory
(324 female F344 rats; 4 with carcinoma, 21 with neoplastic nodules; 25/324 = 7.7%).
4.2.1.2.2. Chronic oral exposure in B6C3Fi mice (Serota et al, 1986b; Hazelton Laboratories,
1983). A 2-year drinking water study similar to the previously described study in F344 rats was
also conducted in B6C3Fi mice (Serota et al., 1986b; Hazelton Laboratories, 1983). The mice
received target doses of 0, 60, 125, 185, or 250 mg/kg-day of dichloromethane in deionized
drinking water for 24 months. Treatment groups consisted of 100 female mice in the low-dose
group and 50 in the remaining treatment groups. There were 200, 100, 100, and 125 male mice
(low- to high-dose groups) in the treated groups. One hundred females (in two groups of 50) and
125 males (in two groups of 60 and 65 mice) served as controls. (The authors do not state why
two groups of control mice were used, other than to say that the design was used due to the high
and erratic incidence of liver tumors in historical control B6C3Fi mice.) Based on water
consumption and BW measurements, mean intakes were reported to be 61, 124, 177, and
234 mg/kg-day for males and 59, 118, 172, and 238 mg/kg-day for females. Endpoints
examined included clinical signs, BW and water consumption, hematology at weeks 52 and 104,
and gross and microscopic examinations of tissues and organs at termination. All tissues from
the control and 250 mg/kg-day groups were examined microscopically, as well as the livers and
neoplasms from all groups and the eyes of all males from all groups.
Throughout the 2-year study, mice from both control and treatment groups exhibited a
high incidence of convulsions (Serota et al., 1986b; Hazelton Laboratories, 1983). The
convulsions were noted only during handling for BW determinations, and efforts to establish a
99

-------
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
basis for this response were unsuccessful. The incidence of convulsions did not correlate with an
increased mortality rate. Survival to 104 weeks was high (82% in males and 78% in females),
and no evidence for exposure-related negative effects on survival were found. Exposure had no
significant effect on BW or water consumption. Mean leukocyte count was significantly
elevated in males and females dosed with 250 mg/kg-day dichloromethane for 52 weeks, but the
authors indicated that the mean values were within the normal historical range for the laboratory.
Treatment-related nonproliferative histopathologic effects were restricted to the liver and
consisted of a marginal increase in the amount of Oil Red O-positive material in the liver of
males and females dosed with 250 mg/kg-day (group incidences for this lesion, however, were
not presented in the published report). The results indicate that 185 mg/kg-day was a NOAEL
and 250 mg/kg-day was a LOAEL for marginally increased amounts of fat in livers of male and
female B6C3Fi mice.
Incidences for proliferative hepatocellular changes in female mice were not presented in
the published reports (Serota et al., 1986b; Hazelton Laboratories, 1983), but it was reported that
exposed female mice did not show increased incidences of proliferative hepatocellular lesions.
In the male B6C3Fi mice, incidences for hepatic focal hyperplasia showed no evidence of an
exposure-related effect (Table 4-15). The authors interpret the data regarding adenomas alone or
carcinomas alone as showing no significantly elevated incidence compared with controls. The
trend tests for each of these outcomes were 0.172 and 0.147 (Hazelton Laboratories, 1983),
respectively, and none of the comparisons between individual exposure groups and the controls
was statistically significant at the chosen Bonferroni-corrected level of <0.01. However,
exposed male mice showed increased combined incidences of hepatocellular adenomas and
carcinomas, with a linear trendp-walue = 0.058 and individual /^-values of <0.05.
Table 4-15. Incidences for focal hyperplasia and tumors in the liver of male
B6C3Fi mice exposed to dichloromethane in drinking water for 2 years
Target dose (mg/kg-day)

Controls
0a
60
125
185
250
Trend
p-valueb
n per group0
125
200
100
99
125

Estimated mean intake (mg/kg-day)
0
61
124
111
234

Number (%) with






Focal hyperplasia"1
10(8)
14(7)
4 (4)
10 (10)
13 (10)
not reported
Hepatocellular adenoma
10(8)
20 (10)
14 (14)
14 (14)
15 (12)

mortality-adjusted percent and /?-valuce
(9)
(12)
(17)
(16)
(12)



p = 0.24
p = 0.064
p = 0.076
p = 0.13
0.172
Hepatocellular carcinoma
14(11)
33 (17)
18(18)
17 (17)
23 (18)

mortality-adjusted percent and /?-valuce
(13)
(19)
(21)
(19)
(21)



p = 0.082
p = 0.073
p = 0.11
p = 0.044
0.147
Hepatocellular adenoma or carcinoma
24 (19)
51 (26)
30 (30)
31 (31)
35 (28)
0.058
mortality-adjusted percent and /?-valuce
(21)
(29)
(34)
(34)
(32)



p = 0.071
p = 0.023
p = 0.019
0.036

100

-------
3661
3662
3663
3664
3665
3666
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
"Two control groups combined.
bCochran-Armitage trend test (source: Hazelton Laboratories [1983])
°Number at start of treatment.
dSome mice with hyperplasia also had hepatocellular neoplasms, but the exact number was unspecified by Serota et
al. (1986b).
"Percent calculated based on number at risk, using Kaplan-Meier estimation, taking into account mortality losses; p-
value for comparison with control group, using asymptotic normal test (source: Hazelton Laboratories [1983]).
Sources: Serota et al. (1986b); Hazelton Laboratories (1983).
Serota et al. (1986b) noted, in summary, that slight increases in proliferative
hepatocellular lesions were found in exposed male but not female B6C3Fi mice and that the
increases were not dose related and were within the range of historical control incidences. The
average incidence of hepatocellular adenomas and carcinomas in 354 control male B6C3Fi mice
in the laboratory in which the experiment was performed was 17.8% with a range of 5-40%
(Serota et al., 1986b). Serota et al. (1986b) concluded that dichloromethane "did not induce a
treatment-related carcinogenic response in B6C3Fi mice" under the conditions of this study. An
alternative conclusion, as determined by EPA, is that dichloromethane induced a carcinogenic
response in male B6C3Fi mice as evidenced by small, but statistically significant, increases in
hepatocellular adenomas and carcinomas at dose levels of 125, 185, and 250 mg/kg-day but not
at 60 mg/kg-day and by a marginally increased trend test for combined hepatocellular adenomas
and carcinomas. Results for the highest dose group (no effect on BW, histologic findings
restricted to mild histologic changes in the liver [vacuolation], and a slight, but statistically
significant, increase in incidence in liver tumors in males only) indicate that this mouse study
may not have included the maximum tolerated dose.
4.2.1.2.3. Chronic oral exposure in Sprague-Dawley rats and Swiss mice (Maltoni et al,
1988). Maltoni et al. (1988) conducted gavage carcinogenicity studies in Sprague-Dawley rats
and in Swiss mice. Groups of rats (50/sex/dose level) were gavaged with dichloromethane
(99.9% pure) in olive oil at dose levels of 0 (olive oil), 100, or 500 mg/kg-day, 4-5 days/week
for 64 weeks. This dosing regime was also used for groups of Swiss mice (50/sex/dose level
plus 60/sex as controls). Endpoints monitored included clinical signs, BW, and full necropsy at
sacrifice (when spontaneous death occurred). For each animal sacrificed, histopathologic
examinations were performed on the following organs: brain and cerebellum, zymbal glands,
interscapular brown fat, salivary glands, tongue, thymus and mediastinal lymph nodes, lungs,
liver, kidneys, adrenals, spleen, pancreas, esophagus, stomach, intestine, bladder, uterus, gonads,
and any other organs with gross lesions. High mortality was observed in male and female high-
dose rats (data not shown) and achieved significance (p < 0.01) in males. The increased
mortality became evident after 36 weeks of treatment and led to the termination of treatment at
week 64. Explanation of the mortality was not provided by the study authors. As with the rats,
101

-------
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
high mortality occurred in male and female mice from the high-dose group (p < 0.01), and the
exposure was terminated after 64 weeks.
Little information is provided regarding nonneoplastic effects (Maltoni et al., 1988).
Treatment with dichloromethane did not affect BW in the Sprague-Dawley rats. A reduction in
BW was apparent in treated mice after 36-40 weeks of treatment, but no data were shown to
determine the magnitude of the effect. The lack of reporting of nonneoplastic findings from the
histopathologic examinations precludes assigning NOAELs and LOAELs for possible
nonneoplastic effects in these studies
The Maltoni et al. (1988) studies of Sprague-Dawley rats and Swiss mice did not find
distinct exposure-related carcinogenic responses following gavage exposure to dichloromethane
at dose levels up to 500 mg/kg-day, although the early termination of the study (at 64 weeks)
limits the interpretation of this finding. Dichloromethane exposure was not related to the
percentage of either study animal bearing benign and malignant tumors or bearing malignant
tumors or to the number of total malignant tumors per 100 animals. High-dose female rats
showed an increased incidence in malignant mammary tumors, mainly due to adenocarcinomas
(8, 6, and 18% in the control, 100, and 500 mg/kg dose groups, respectively; the number of
animals examined was not provided), but the increase was not statistically significant. A dose-
related increase, although not statistically significant, in pulmonary adenomas was observed in
male mice (5, 12, and 18% in control, 100, and 500 mg/kg-day groups, respectively). When
mortality was taken into account, high-dose male mice that died in the period ranging from 52 to
78 weeks were reported to show a statistically significantly (p < 0.05) elevated incidence for
pulmonary tumors (1/14, 4/21, and 7/24 in control, 100, and 500 mg/kg-day groups,
respectively). Details of this analysis were not provided. EPA applied a Fisher's exact test to
these incidences and determined a p-value of 0.11 for the comparison of the 500 mg/kg-day
group (7/24) to the controls (1/14).
4.2.2. Inhalation Exposure: Overview of Noncancer and Cancer Effects
Inhalation dichloromethane exposure studies in rats and mice, using subchronic and
chronic durations, identify the CNS, liver, and lungs as potential toxicity targets. Data from
other studies indicate that hamsters are less susceptible to the nonneoplastic and neoplastic
effects of dichloromethane than are rats and mice.
Increased incidences of nonneoplastic liver lesions were observed in Sprague-Dawley
rats exposed to >500 ppm for 2 years (Nitschke et al., 1988a; Burek et al., 1984), F344 rats
exposed to concentrations >1,000 ppm for 2 years (Mennear et al., 1988; NTP, 1986), and
B6C3Fi mice exposed to >2,000 ppm for 2 years (Mennear et al., 1988; NTP, 1986).
Two-year inhalation exposure studies at concentrations of 2,000 or 4,000 ppm
dichloromethane produced increased incidences of lung and liver tumors in B6C3Fi mice
(Mennear et al., 1988; NTP, 1986). Additional studies examining mechanistic issues regarding
102

-------
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
this effect are described in sections 4.5.2 and 4.5.3 (Maronpot et al., 1995; Foley et al., 1993;
Kari et al., 1993). Significantly increased incidences of benign mammary tumors (primarily
fibroadenomas) were also observed in male and female F344/N rats exposed by inhalation to
2,000 or 4,000 ppm for 2 years (Mennear et al., 1988; NTP, 1986). In the male rats, the
incidence of fibromas or sarcomas originating from the subcutaneous tissue around the
mammary gland was also increased, but the difference was not statistically significant. In other
studies in Sprague-Dawley rats with exposures of 50-500 ppm (Nitschke et al., 1988a) and 500-
3,500 ppm (Burek et al., 1984), the incidence of benign mammary tumors was not increased, but
in females the number of tumors per tumor-bearing rat increased at the higher dose levels.
No obvious clinical signs of neurological impairment were observed in the 2-year
bioassays involving exposure concentrations up to 2,000 ppm in F344 rats (Mennear et al., 1988;
NTP, 1986) or 3,500 ppm in Sprague-Dawley rats (Nitschke et al., 1988a; Burek et al., 1984). In
B6C3Fimice exposed to 4,000 ppm in B6C3Fis there was some evidence of hyperactivity during
the first year of the study and lethargy during the second year, with female mice appearing to be
more sensitive (Mennear et al., 1988; NTP, 1986). Studies that evaluated batteries of
neurobehavioral endpoints following subchronic or chronic inhalation exposure are restricted to
one in which no effects were observed more than 64 hours postexposure in an observational
battery, a test of hind-limb grip strength, a battery of evoked potentials, or histologic
examinations of brain, spinal cord, or peripheral nerves in F344 rats exposed to concentrations
up to 2,000 ppm for 13 weeks (Mattsson et al., 1990) (see section 4.2.3)
No effects on reproductive performance were found in a two-generation reproductive
toxicity study with F344 rats exposed to concentrations up to 1,500 ppm for 14 and 17 weeks
before mating of the F0 and F1 generations, respectively (Nitschke et al., 1988b) (described
more completely in section 4.3). Developmental effects following exposure of Long-Evans rats
to 4,500 ppm for 14 days prior to mating and during gestation (or during gestation alone)
included decreased offspring weight at birth and changed behavioral habituation of the offspring
to novel environments (Bornschein et al., 1980; Hardin and Manson, 1980) (see section 4.3 for
more details). In standard developmental toxicity studies involving exposure to 1,250 ppm on
GDs 6-15, no adverse effects on fetal development were found in Swiss-Webster mice or
Sprague-Dawley rats (Schwetz et al., 1975) (see section 4.3).
4.2.2.1. Toxicity Studies of Subchronic Inhalation Exposures: General, Renal, and Hepatic
Effects
Data pertaining to general (e.g., BW, mortality), hepatic, and renal effects from several
inhalation exposure studies in various species, with exposure periods of 3-6 months, are
described below. (Studies providing detailed neurological data are described separately in
section 4.4.3.) The earliest study involved several different species, with exposures of
5,000 ppm for up to 6 months (Heppel et al., 1944). Two 14-week studies in dogs, monkeys,
103

-------
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
rats, and mice were conducted with exposures at 0, 1,000, and 5,000 ppm (Haun et al., 1972,
1971; Weinstein et al., 1972) and at 0, 25, and 100 ppm (Haun et al., 1972). Neurological effects
and hepatic degeneration were seen at the 1,000 ppm dose. In the lower-dose portion of the
Haun et al. (1972) study in mice, decreased cytochrome P-450 levels in liver microsomes and
some histopathologic liver changes (fat stains and cytoplasmic vacuolation) were seen at
100 ppm but more obvious adverse effects were not observed. Leuschner et al. (1984) reported
data from a high exposure (10,000 ppm) 90-day study of rats; beagle dogs were also included in
this study, at an exposure level of 5,000 ppm. No evidence of toxicity was reported by the
authors of this study. In a 13-week exposure study conducted by NTP (1986), decreased BWs
and increased incidence of foreign body pneumonia were seen at 8,400 ppm in F344 rats, and
histologic changes in the liver in B6C3F1 mice were seen at 4,200 ppm.
The first experimental study of dichloromethane exposure included dogs, rabbits, guinea
pigs, and rats, with an exposure of approximately 5,000 ppm for 7 hours/day, 5 days per week
for up to 6 months (Heppel et al., 1944). The strains of the animals, the comparability between
exposed and unexposed group (in terms of sex distribution and other attributes), and process by
which animals were chosen for histologic examination are not clearly described in the report.
Exposed animals included adult dogs (1 male and 5 females), juvenile dogs (1 male and 1 female
born in the exposure chamber and exposed daily from birth), adult rabbits (2 males and
2 females), guinea pigs (14 males), and rats (15 males and 6 females). The nonexposed control
group included 14 guinea pigs, 28 rats, 4 rabbits, and an unspecified number of dogs. Exposure
produced no significant effects on BWs except in the guinea pigs; after 131 exposures, average
BWs were 0.820 and 1.025 kg for exposed and control guinea pigs, respectively. Three exposed
guinea pigs died after 35, 90, and 96 exposures. No other deaths occurred, except for one
exposed female rat that died after 22 exposures and giving birth to a litter. Autopsy showed
thrombi in the renal vessels associated with marked cortical infarction. No adverse clinical signs
of toxicity (such as decreased activity or incoordination) were observed in exposed animals
during the study. Urinalysis, hematology tests, and tests of liver function performed on dogs
during the study showed no treatment-related effects. At termination, gross and microscopic
examination of the major organs showed no pathological changes after exposure to 5,000 ppm
dichloromethane, with the exception that two of the exposed guinea pigs that died showed
extensive pneumonia associated with moderate centrilobular fatty degeneration of the liver. The
results indicate that 5,000 ppm was a NOAEL for nonneoplastic systemic effects in dogs, rabbits,
and rats exposed 7 hours/day, 5 days/week for up to 6 months. The findings of three deaths (two
with pulmonary congestion and centrilobular fatty degeneration) and 20% decreased average
BW among the 14 exposed guinea pigs indicates that 5,000 ppm was a LOAEL in this species.
Haun et al. (1972, 1971) and Weinstein et al. (1972) reported results from studies in
which groups of 8 female beagle dogs, 4 female rhesus monkeys, 20 male Sprague-Dawley rats,
and 380 female ICR mice were continuously exposed to 0, 1,000, or 5,000 ppm dichloromethane
104

-------
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
for up to 14 weeks in whole-body exposure chambers. Gross and histopathologic examinations
were scheduled to be made on animals that died or were sacrificed during or at termination of the
study. At 5,000 ppm, obvious nervous system effects (e.g., incoordination, lethargy) were
observed in dogs, monkeys, and mice. At 1,000 ppm, these effects were most apparent in dogs
and monkeys (Haun et al., 1971). Food consumption was reduced in all species at 5,000 ppm
and in dogs and monkeys at 1,000 ppm. All exposed animals either lost weight or showed
markedly decreased BW gains compared with controls. For example, rats exposed to 1,000 or
5,000 ppm for 14 weeks showed average BWs that were roughly 10 and 20% lower than control
values. Significant numbers of dogs (4) and mice (123), as well as 1 monkey, died within the
first 3 weeks of exposure to 5,000 ppm. Because of this high mortality, all surviving 5,000 ppm
animals were sacrificed at 4 weeks of exposure, except for one half (10) of the rats that went on
to survive the 14-week exposure period. At 1,000 ppm, six of eight dogs died by 7 weeks, at
which time the remaining two were sacrificed. Monkeys, rats, and all but a few mice survived
exposure to 1,000 ppm for 14 weeks.
Gross examination of tissues showed yellow, fatty livers in dogs that died during
exposure to 1,000 or 5,000 ppm, "borderline" liver changes in 3 monkeys exposed to 5,000 ppm,
and mottled liver changes in 4/10 rats exposed to 5,000 ppm for 14 weeks (Haun et al., 1971).
Comprehensive reporting of the histologic findings from this study were not available, but Haun
et al. (1972) reported that the primary target organ was the liver and that in some exposed
animals the kidney was also affected. Light and electron microscopy of liver sections from
groups of 4-10 mice sacrificed after 1, 4, 8, and 12 hours and 1, 2, 3, 4, 6, and 7 days of
exposure to 5,000 ppm showed hepatocytes with balloon degeneration (dissociation of
polyribosomes and swelling of rough endoplasmic reticulum) as early as 12 hours of exposure
(Weinstein et al., 1972). The degeneration peaked in severity after 2 days of exposure and,
subsequently, partially reversed in severity. Information on possible histopathologic changes in
mice exposed to 1,000 ppm was not provided.
The results from this study demonstrate that dogs and mice were more sensitive than
were rats and monkeys to lethal effects, nervous system depression, and possibly liver effects
from continuous exposure to 1,000 or 5,000 ppm. The results indicate that continuous exposure
to 1,000 ppm was an adverse effect level for mortality and effects on the nervous system and
liver in dogs (exposed for up to 4 weeks) and for BW changes in rats (exposed for 14 weeks).
The 5,000 ppm level induced mortality in beagle dogs, ICR mice, and rhesus monkeys (but not
in Sprague-Dawley rats); obvious nervous system effects in dogs, mice, monkeys, and rats; and
gross liver changes in dogs, mice, monkeys, and rats.
Haun et al. (1972) also conducted studies with groups of 20 mice, 20 rats, 16 dogs, and
4 monkeys exposed continuously to 0, 25, or 100 ppm dichloromethane for 100 days (14 weeks).
The animals presumably were of the same strains and sexes as those used in the studies involving
exposure to 1,000 or 5,000 ppm dichloromethane (Haun et al., 1972, 1971; Weinstein et al.,
105

-------
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
1972). All animals underwent necropsy and histopathologic evaluation at termination of the
exposure, but a list of the tissues examined and incidence or severity data were not presented in
the report. Hematology and clinical chemistry variables (including COHb levels) were measured
in blood samples collected from dogs and monkeys at biweekly or monthly intervals during
exposure. COHb levels were elevated in a dose-related manner in monkeys and peaked at about
5% (approximately 0.8% pre-exposure) after 6 weeks of exposure. COHb levels in dogs were
unaffected by the 25 ppm exposure level and rose to about 2% (from about 0.6%) from
week 4 on in high-dose dogs. Additional groups of mice were included for assessment of
hexobarbital sleep times at monthly intervals; levels of cytochromes P-450, P-420, and bs in liver
microsomes at monthly intervals; and spontaneous physical activity at several intervals during
the study.
No clinical signs of toxicity or alterations in weight gain were seen in any of the species
examined. In dogs and monkeys, hematology and clinical chemistry results throughout the study
and at termination were unremarkable, as were the results of the gross and histopathologic
examinations. In mice exposed to 100 ppm, CYP levels in liver microsomes were significantly
decreased (compared with control values) after 30, 60, and 90 days of exposure to 100 ppm,
whereas levels of cytochrome bs and P-420 decreased after 30 days and increased after 90 days
of exposure. At 25 ppm, no significant differences from control were seen in mouse liver levels
of cytochromes. Mice exposed to 25 ppm showed no histopathologic changes, while histologic
changes in mice at 100 ppm were restricted to positive fat stains and some cytoplasmic
vacuolation in the liver. In rats at both exposure levels, the livers showed positive staining for
increased fat, and the kidneys showed evidence of nonspecific tubular degenerative and
regenerative changes. Haun et al. (1972) indicate that no distinctively adverse effects were
found in monkeys, dogs, rats, or mice continuously exposed to 25 or 100 ppm for up to
14 weeks. Decreased CYP levels in liver microsomes and some histopathologic liver changes
(fat stains and cytoplasmic vacuolation) were seen at the 100 ppm dose.
Leuschner et al. (1984) exposed Sprague-Dawley rats (20/sex/dose level) to 0 or
10,000 ppm and beagle dogs (3/sex/dose level) to 0 or 5,000 ppm dichloromethane in whole-
body exposure chambers. Exposure periods were 6 hours/day for 90 consecutive days.
Endpoints evaluated in both species included clinical signs, food and water consumption, BW,
hematology, clinical chemistry, urinalysis, and gross and microscopic evaluation of 27 organs at
termination. Electrocardiography and blood pressure measurements were also done in dogs.
The only significant effect observed in rats was a slight redness of the conjunctiva
1-10 hours after each exposure. In dogs, compound-related effects were restricted to slight
sedation throughout the exposure period and slight erythema lasting up to 10 hours after
exposure. In this 90-day study involving daily 6-hour exposures, 10,000 and 5,000 ppm were
NOAELs for behavioral, clinical chemistry, hematologic, and histologic signs of toxicity in
Sprague-Dawley rats and beagle dogs, respectively.
106

-------
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
NTP (1986) exposed groups of F344 rats and B6C3Fi mice (10/sex/dose level) to target
concentrations of 0, 525, 1,050, 2,100, 4,200, or 8,400 ppm dichloromethane, 6 hours/day,
5 days/week for 13 weeks in whole-body exposure chambers. Endpoints monitored included
clinical signs, BW, and necropsy at termination. Comprehensive sets of tissues and organs in
control and high-dose animals were histologically examined; tissues from the lower dose groups
were examined to determine the no-observed-effect level. One male and one female rat from the
8,400 ppm exposure group died before the end of the study, but the cause of death was not
discussed. The final mean BWs of 8,400 ppm male and female rats were reduced by 23 and
11%, respectively, relative to controls. Foreign-body pneumonia was present in 4/10 male and
6/10 female rats exposed to 8,400 ppm and in 1/10 female rat from the 4,200 ppm exposure
group. The liver lipid/liver weight ratios for 8,400 ppm rats of both sexes and 4,200 ppm female
rats were significantly lower than in controls. In mice, 4/10 males and 2/10 females exposed to
8,400 ppm died before the end of the study, and these deaths were considered treatment related.
Histologic changes in exposed mice consisted of hepatic centrilobular hydropic degeneration (of
minimal to mild severity) in 3/10 males and 8/10 females at 8,400 ppm and in 9/10 females from
the 4,200 ppm exposure group. Histologic changes in the 2,100 ppm mouse group were not
mentioned. The liver lipid/liver weight ratio for the high-dose female mice was significantly
lower than in controls. In this 13-week study involving 6-hour exposure periods for
5 days/week, 4,200 ppm was a NOAEL and 8,400 ppm was a LOAEL for decreased BWs and
increased incidence of foreign-body pneumonia in F344 rats. In B6C3Fi mice, 2,100 ppm was a
NOAEL and 4,200 ppm was a LOAEL for histologic changes in the liver.
4.2.2.2. Toxicity Studies from Chronic Inhalation Exposures
Chronic inhalation exposure studies are summarized in Table 4-16. Details of each study
are described below, with the results pertaining to nonneoplastic and neoplastic effects
summarized in the following sections.
107

-------
3920
Table 4-16. Studies of chronic inhalation dichloromethane exposures
Reference,
Number per


strain/species
group
Exposure information
Comments
Mennear et al. (1988);
50/sex/dose
2 years, 6 hours/day,
Nonneoplastic liver effects and
NTP (1986)

5 days/week
hemosiderosis in males and females (see
F344 rats

0, 1,000, 2,000, 4,000 ppm
Table 4-17)



Weak trend for neoplastic nodule or



hepatocellular carcinoma in females,



benign mammary tumors in males and



females (see Table 4-18)
Mennear et al. (1988);
50/sex/dose
2 years, 6 hours/day,
Varied nonneoplastic effects (see Table
NTP (1986)

5 days/week
4-19)
B6C3Fi mice

0, 2,000, 4,000 ppm
Liver and lung tumors (adenomas or



carcinomas) in males and females (see



Table 4-20)
Burek et al. (1984)
95/sex/dose
2 years, 6 hours/day,
Decreased mortality
Syrian hamsters

5 days/week
Increased CoHb at 500 ppm (see section


0, 500, 1,500, 3,500 ppm
4.2.2.2.3)
Burek et al. (1984)
92-97/sex/dose
2 years, 6 hours/day,
Nonneoplastic liver effects in males and
Sprague-Dawley rats

5 days/week
females (see Table 4-21)


0, 500, 1,500, 3,500 ppm
Increased CoHb at 500 ppm



Increased number of benign mammary



tumors per tumor bearing rat (females)



(see Table 4-21)
Nitschke et al. (1988a)
90/dose/sex
2 years, 6 hours/day,
Nonneoplastic liver effects in males and
Sprague-Dawley rats

5 days/week
females (statistically significant in


0, 50, 200, 500 ppm
females) (seeTable 4-22)



Increased CoHb at 50 ppm



Increased number of benign mammary



tumors per animal in females (see Table



4-23)
Maltoni et al. (1988)
54-60/dose
2 years, 4 hours/day,
No effects seen on total number of benign


5 days/week for 7 weeks;
or malignant cancers
Sprague-Dawley rats,

7 hours/day, 5 days/week

female

for 97 weeks



0, 100 ppm

3921
3922
3923	4.2.2.2.1. Chronic inhalation exposure in F344/Nrats (Mennear et al., 1988; NTP, 1986).
3924	NTP conducted a 2-year inhalation exposure study in F344/N rats (Mennear et al., 1988; NTP,
3925	1986). The rats (50/sex/exposure level) were exposed to dichloromethane (>99% pure) by
3926	inhalation in exposure chambers, 6 hours/day, 5 days/week for 2 years. Exposure concentrations
3927	were 0, 1,000, 2,000, or 4,000 ppm. Mean daily concentrations never exceeded 110% of target
3928	and were <90% of target in only 23 of 1,476 analyses. Endpoints monitored included clinical
3929	signs, mortality, and gross and microscopic examinations of 32 tissues at study termination.
3930	Clinical examinations were conducted weekly for 3.5 months and biweekly until month 8. After
3931	8 months, the animals were clinically examined and palpated for tumors and masses monthly
3932	until the end of the study.
3933	Dichloromethane exposure did not significantly alter BW gain or terminal BWs
3934	(Mennear et al., 1988; NTP, 1986). Survival of male rats was low in all exposed groups and the
108

-------
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
control group, and no significant exposure-related differences were apparent. Most deaths
occurred during the last 16 weeks of the study. Survival at week 86 was 36/50, 39/50, 37/50, and
33/50 for the control, 1,000, 2,000, and 4,000 ppm groups, respectively. In female rats, there
was a trend towards decreased survival, and the survival of high-dose female rats was
significantly reduced, possibly due to leukemia. Survival in the females at 86 weeks was 30/50,
22/50, 22/50, and 15/50 for the control, 1,000, 2,000, and 4,000 ppm groups, respectively.
Nonneoplastic lesions with statistically significantly elevated incidences, compared with
controls, included hepatocyte cytoplasmic vacuolation and necrosis and liver hemosiderosis in
males and females; renal tubular cell degeneration in males and females; splenic fibrosis in
males; and nasal cavity squamous metaplasia in females (Table 4-17). The results indicate that
1,000 ppm (6 hours/day, 5 days/week) was a LOAEL for nonneoplastic liver changes
(hepatocyte cytoplasmic vacuolation and necrosis, hepatic hemosiderosis) in male and female
F344/N rats. A NOAEL was not established because effects were observed at the lowest dose.
109

-------
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
Table 4-17. Incidences of nonneoplastic histologic changes in male and
female F344/N rats exposed to dichloromethane by inhalation (6 hours/day,
5 days/week) for 2 years
Exposure (ppm)a

Controls




Lesion, by sex

0
1,000

2,000
4,000
Males






n per groupb
50

50
50

50
Number (%)° with






Liver changes






Hepatocyte cytoplasmic vacuolation
8
(16)
26 (53)d
22
(44)d
25 (50)d
Hepatocyte focal necrosis
7
(14)
23 (47)d
6
(12)
16 (32)d
Hepatocytomegaly
2
(4)
10 (20)
6
(12)
5 (10)
Hemosiderosis
8
(16)
29 (59)d
37
(74)d
42 (84)d
Bile duct fibrosis
8
(16)
10 (20)
17
(34)
23 (46)d
Renal tubular cell degeneration
11
(22)
13 (26)
23
(46)d
10 (20)d
Splenic fibrosis
2
(4)
6 (12)
11
(22)d
8 (16)d
Females






n per group6
50

50
50

50
Number (%)° with






Liver changes






Hepatocyte cytoplasmic vacuolation
10
(20)
43 (86)d
44
(88)d
43 (86)d
Hepatocyte focal necrosis
2
(4)
32 (64)d
19
(38)d
9 (18)d
Hepatocytomegaly
3
(6)
10 (20)d
18
(36)d
5(10)
Hemosiderosis
19
(38)
29 (58)d
38
(76)d
45 (90)d
Bile duct fibrosis
4
(8)
3 (6)
10
(20)d
3 (6)
Renal tubular cell degeneration
14
(28)
20 (40)
22
(44)
25 (51)d
Splenic fibrosis
0
(0)
2 (4)
4
(8)
4 (8)
Nasal cavity squamous metaplasia
1
(2)
2 (4)
3
(6)
9 (18)d
al,000 ppm = 3,474 mg/m3, 2,000 ppm = 6,947 mg/m3, 4,000 ppm = 13,894 mg/m3.
bNumber of male rats necropsied per group; only 49 1,000 ppm livers were examined microscopically.
Percentages were based on the number of tissues examined microscopically per group.
Statistical significance not reported in publications but significantly (p < 0.05) different from control as calculated
by Fisher's exact test.
eNumber of females necropsied per group; only 49 4,000 ppm kidneys and spleens were examined microscopically.
Sources: Mennear et al. (1988); NTP (1986, Appendix B, Tables CI and C2).
Incidences of mammary fibroadenomas were significantly increased in 4,000 ppm males
and 2,000 and 4,000 ppm females, compared with controls (Table 4-18). Similar patterns were
seen with the combination of fibroadenomas and adenomas (not shown in Table 4-18). In males,
subcutaneous tissue fibroma or sarcoma was seen in 1/50, 1/50, 2/50, and 5/50 rats in the 0,
1,000, 2,000, and 4,000 ppm groups, respectively, but these lesions were not seen in females.
Incidences of female rats with liver neoplastic nodules or carcinomas (combined) showed a
significant trend test after survival adjustment only, but the incidences at the two highest dose
levels were not significantly increased relative to the control (Table 4-18).
Incidences for mononuclear cell leukemias in mid- and high-dose female rats were
statistically significant after a survival-adjustment analysis. However, Mennear et al. (1988)
considered the relationship between exposure to dichloromethane and mononuclear cell leukemia
110

-------
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
to be equivocal, based on the fact that most male rats had leukemia (34/50, 26/50, 32/50, and
35/50 in controls, 1,000, 2,000, and 4,000 ppm rats, respectively). Other neoplasms that had
increased incidences included mesotheliomas (predominantly in the tunica vaginalis) in males
(0/50, 2/50, 5/50, and 4/50 in controls, 1,000, 2,000, and 4,000 ppm rats, respectively). This
lesion was not considered to be related to dichloromethane exposure, because the concurrent
control incidence (0/50) for this neoplasm was low relative to earlier inhalation studies
conducted at this laboratory (4/100, 4%) and in other NTP studies with male F344/N rats
(44/1,727) (mean historical percentage across NTP studies = 3 ± 2%).
NTP (1986) concluded that there was "some evidence of carcinogenicity of
dichloromethane" in male F344/N rats as shown by increased incidence of benign mammary
gland tumors and "clear evidence of carcinogenicity" of dichloromethane in female F344/N rats
as shown by increased incidence of benign mammary gland tumors. The summary of the hepatic
effects in rats in the NTP (1986) report also notes the positive trend in the incidence of
hepatocellular neoplastic nodules or carcinomas in females, which "may have been due to
dichloromethane exposure."
Ill

-------
Table 4-18. Incidences of selected neoplastic lesions in male and female F344/N rats exposed to
dichloromethane by inhalation (6 hours/day, 5 days/week) for 2 years
Exposure (ppm)a

0 (Controls)

1,000


2,000


4,000















Trend
Neoplastic lesion, by sex
n
(%)b
(%)c
n
(%)b (%)c
n
(%)b (%)c
n
(%)b
(%)c
/>-valuc
Males













n per group
50


50


50


50



Liver—Neoplastic nodule or hepatocellular carcinoma
2
(4)
(10)
3
(6)
(13)
4
(8)
(19)
1
(2)
(6)
0.55
Liver—hepatocellular carcinoma
2
(4)
(10)
1
(2)
(4)
2
(4)
(10)
1
(2)
(6)
nr
Lung—Bronchoalveolar adenoma or carcinoma
1


1
(2)

2
(4)

1
(2)


Mammary gland













Adenoma, adenocarcinoma, or carcinoma
0
(0)

0
(0)

0
(0)

1
(2)


Subcutaneous tissue fibroma or sarcoma
1
(2)
(6)
1
(2)
(6)
2
(4)
(9)
5
(10)
(23)
0.008
Fibroadenoma
0
(0)
(0)
0
(0)
(0)
2
(4)
(12)
1
(2)
(8)
<0.001
Mammary gland or subcutaneous tissue adenoma,













fibroadenoma, fibroma, or sarcoma
1
(2)
(6)
1
(2)
(6)
4
(8)
(21)
9e
(18)
(49)
<0.001
Brain (carcinoma, not otherwise specified, invasive)
0
(0)

1
(2)

0
(0)

0
(0)


Females













n per group
50


50


50


50



Liver—Neoplastic nodule or hepatocellular carcinoma
2
(4)
(7)
1
(2)
(2)
4
(8)
(14)
5
(10)
(20)
0.08
Liver—hepatocellular carcinoma
0
(0)
(0)
0
(0)
(0)
1
(2)
(4)
0
(0)
(0)
nr
Lung—Bronchoalveolar adenoma or carcinoma
1
(2)

1
(2)

0
(0)

0
(0)


Mammary gland













Adenocarcinoma or carcinoma
1
(2)

2
(4)

2
(4)

0
(0)


Adenoma, adenocarcinoma, or carcinoma
1
(2)

2
(4)

2
(4)

1
(2)


Fibroadenoma
5
(10)
(16)
lle
(22)
(41)
13e
(26)
(44)
IT
(44)
(79)
<0.001
Mammary gland adenoma, fibroadenoma, or
6
(12)
(18)
13
(26)
(44)
14e
(28)
(45)
23e
(46)
(86)
<0.001
adenocarcinoma













Brain (carcinoma, not otherwise specified, invasive, and
1
(2)

0
(0)

2
(4)

0
(0)


oligodendroglioma/	
al,000 ppm = 3,474 mg/m3, 2,000 ppm = 6,947 mg/m3, 4,000 ppm = 13,894 mg/m3.
Percentages based on the number of tissues examined microscopically per group; for males, 49 livers and lungs were examined microscopically in the
1,000 ppm groups, and only 49 brains were examined microscopically in the 4,000 ppm group. For comparison, incidence in historical controls reported in
NTP (1986) were 1% for female liver tumors and 16% for female mammary fibroadenomas.
°Mortality-adjusted percentage.
dLife-table trend test, as reported by NTP (1986). nr = not reported.
eLife-table test comparison dose group with control < 0.05, as reported by NTP (1986).
fThe oligodendroglioma occurred in the 2,000 ppm group.
Sources: Mennear et al. (1988); NTP (1986, Appendix A and Appendix E, Tables El and E2)
112

-------
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4.2.2.2.2. Chronic inhalation exposure in B6C3Fj mice (Mennear et al., 1988; NTP, 1986). A
2-year inhalation exposure study in B6C3Fi mice, similar to that in F344/N rats, was also
conducted by NTP (Mennear et al., 1988; NTP, 1986). The mice (50/sex/exposure level) were
exposed to dichloromethane (>99% pure) by inhalation at concentrations of 0, 2,000, or
4,000 ppm in exposure chambers 6 hours/day, 5 days/week for 2 years. As with the study in rats,
mean daily concentrations in the mice never exceeded 110% of target and were <90% of target in
only 23 of 1,476 analyses. Endpoints monitored included clinical signs, mortality, and gross and
microscopic examinations of 32 tissues at study termination. Clinical examinations were
conducted weekly for 3.5 months and biweekly until month 8. After 8 months, the animals were
clinically examined and palpated monthly for tumors and masses until the end of the study.
The BW of 4,000 ppm males was comparable to controls until week 90 and 8-11% below
controls thereafter. The BW of 4,000 ppm females was 0—8% lower than that of controls from
week 51 to 95 and 17% lower at study termination. No information was provided regarding food
consumption during the study. Male and female mice from the high-dose groups (4,000 ppm)
were hyperactive during the first year of the study; during the second year, high-dose females
appeared lethargic. Exposure was associated with decreased survivability of both male and
female mice (males: 39/50, 24/50, and 11/50 and females: 25/50, 25/50, and 8/50 in controls,
2,000 ppm, and 4,000 ppm at 104 weeks, respectively). In 4,000 ppm mice, statistically
significant incidences of nonneoplastic lesions were found in the liver (cytologic degeneration),
testes (atrophy), ovary and uterus (atrophy), kidneys (tubule casts in males only), stomach
(dilatation), and spleen (splenic follicles in males only) (Table 4-19). In 2,000 ppm mice, the
only nonneoplastic lesions showing statistically significantly elevated incidences were ovarian
atrophy, renal tubule casts, and hepatocyte degeneration in female mice (Table 4-19). The
results indicate that 2,000 ppm, the lowest exposure level, was a LOAEL for nonneoplastic
changes in the ovaries, kidneys, and livers of female B6C3Fi mice. A NOAEL was not
established because effects occurred at the lowest exposure level.
113

-------
4008
Table 4-19. Incidences of nonneoplastic histologic changes in B6C3Fi
mice exposed to dichloromethane by inhalation (6 hours/day,
5 days/week) for 2 years
Exposure (ppm)a

Controls


Lesion, by sex
0
2,000
4,000
Males, n per groupb
50
50
50
Number (%)° with



Liver changes



Hepatocyte cytoplasmic vacuolation
Not reported
Not reported
Not reported
Hepatocyte focal necrosis
0 (0)
0 (0)
2 (4)
Cytologic degeneration
0 (0)
0 (0)
22 (45)d
Testicular atrophy
0 (0)
4 (8)
31 (62)d
Renal tubule casts
6 (12)
11 (22)
20 (40)d
Stomach dilatation
3 (6)
7 (15)
9 (18)d
Splenic follicular atrophy
0 (0)
3 (6)
7 (15)d
Females , n per group6
50
50
50
Number (%)° with



Liver changes



Hepatocyte cytoplasmic vacuolation
Not reported
Not reported
Not reported
Hepatocyte focal necrosis
Not reported
Not reported
Not reported
Cytologic degeneration
0 (0)
23 (48)d
21 (44)d
Ovarian atrophy
6 (12)
28 (60)d
32 (74)d
Uterus atrophy
0 (0)
1 (2)
8 (17)d
Renal tubule casts
8 (16)
23 (48)d
23 (49)d
Glandular stomach dilatation
1 (2)
2 (4)
10 (20)d
Splenic follicular atrophy
0 (0)
0 (0)
1 (2)
a2,000 ppm = 6,947 mg/m3, 4,000 ppm = 13,894 mg/m3.
bNumber of male mice necropsied per group. The number biopsied in the 0, 2,000, and 4,000 ppm
dose groups was 50, 49, and 49 for liver; 50, 49, and 50 for renal tubules; 49, 47, and 49 for
stomach; and 49, 49, and 48 for spleen.
Percentages were based on the number of tissues examined microscopically per group.
Statistical significance not reported in publications but significantly different (p < 0.05) from
control as calculated by EPA using Fisher's exact test.
"Number of females necropsied per group. The number biopsied in the 0, 2,000, and 4,000 ppm
dose groups was 50, 48, and 48 for liver; 50, 47, and 43 for ovaries; 50, 48, and 47 for uterus; 49,
48, and 47 for renal tubule; 49, 47, and 48 for stomach; and 49, 48, and 47 for spleen.
Sources: Mennear et al. (1988); NTP (1986, Appendix C, Tables D1 and D2).
4009
4010
4011	At both exposure levels, statistically significantly elevated incidences were found for
4012	hepatocellular adenomas (males only); hepatocellular carcinomas; hepatocellular adenomas and
4013	carcinomas, combined; bronchoalveolar adenomas; bronchoalveolar carcinomas; and
4014	bronchoalveolar adenomas and carcinomas (Table 4-20). Statistically significant positive trend
4015	tests were found for each of these tumor types in female mice. The trend tests were significant
4016	for the liver tumors in male mice after life-table adjustment for reduced survival. The only other
4017	statistically significant carcinogenic response was for increased incidence of hemangiosarcomas
4018	or combined hemangiomas and hemangiosarcomas in male mice exposed to 4,000 ppm. NTP
114

-------
4019	(1986) concluded that the elevated incidences of liver and lung tumors provided clear evidence
4020	of carcinogenicity in male and female B6C3Fi mice.
4021
Table 4-20. Incidences of neoplastic lesions in male and female B6C3Fi mice
exposed to dichloromethane by inhalation (6 hours/day, 5 days/week) for 2 years
	Exposure (ppm)a	
0 (Controls)	2,000	4,000	
Trend
	Neoplastic lesion, by sex	n (%)b (%)° n (%)b (%)°	n (%)b (%)° p-valued
Males
Liver
Hepatocellular adenoma	10
Hepatocellular	13
Hepatocellular adenoma or carcinoma 22
Lung
Bronchoalveolar adenoma	3
Bronchoalveolar carcinoma	2
Bronchoalveolar adenoma or carcinoma 5
Mammary adenocarcinomaf
Hemangioma or hemangiosarcoma,	2
combined	
Females
Liver
Hepatocellular adenoma	2
Hepatocellular carcinoma	1
Hepatocellular adenoma or carcinoma 3
Lung
Bronchoalveolar adenoma	2
Bronchoalveolar carcinoma	1
Bronchoalveolar adenoma or carcinoma 3
Mammary adenocarcinoma	2
Hemangioma or hemangiosarcoma,
combinedf
a2,000 ppm = 6,947 mg/m3, 4,000 ppm = 13,894 mg/m3.
bPercentages based on the number of tissues examined microscopically per group; for males, 49 livers were examined in
the 2,000 and 4,000 ppm groups; for females, only 48 livers and lungs and 49 mammary glands were microscopically
examined in the 2,000 and 4,000 ppm groups. For comparison, incidence in historical controls reported in NTP (1986)
were 28% for male liver tumors, 31% for male lung tumors, 5% for female liver tumors, and 10% for female lung
tumors.
°Mortality-adjusted percentage.
dLife-table trend test, as reported by NTP (1986).
eLife-table test comparison dose group with control < 0.05, as reported by NTP (1986).
fData not reported.
Sources: Mennear et al. (1988); NTP, (1986, Appendix E, Tables E3 and E4).
4022
4023
4024	4.2.2.2.3. Chronic inhalation exposure in Syrian hamsters (Burek et al, 1984). Burek et al.
4025	(1984) conducted a chronic toxicity and carcinogenicity study in rats and hamsters. In the
4026	hamster study, groups of 95 Syrian golden hamsters of each sex were exposed to 0 (filtered air),
4027	500, 1,500, or 3,500 ppm dichloromethane (>99% pure) under dynamic airflow conditions in
4028	whole-body exposure chambers 6 hours/day, 5 days/week for 2 years. Exposure started when the
(20)
(23)
14
(29)
(47)
14
(29) (68)
0.19
(26)
(30)
15
(30)
(44)
26e
(53) (76)
0.004
(44)
(48)
24
(49)
(67)
33e
(67) (93)
0.013
(6)
(8)
19e
(38)
(56)
24e
(48) (79)
<0.001
(4)
(5)
10e
(20)
(34)
28e
(56) (93)
<0.001
(10)
(12)
27e
(54)
(74)
40e
(80) (100)
<0.001
(4)
(5)
2
(4)
(8)
6
(12) (26)
0.08
(4)
(7)
6
(13)
(21)
22e
(46) (83)
<0.001
(1)
(4)
11
(23)
(34)
32e
(67) (97)
<0.001
(6)
(10)
16e
(33)
(48)
40e
(83) (100)
<0.001
(4)
(7)
23e
(48)
(58)
28e
(58) (91)
<0.001
(1)
(4)
13e
(27)
(46)
29e
(60) (92)
<0.001
(6)
(11)
30e
(63)
(83)
41e
(85) (100)
<0.001
(4)
(8)
3
(6)
(10)
0
(0) (0)
0.21
115

-------
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
animals were approximately 8 weeks of age. Interim sacrifices were conducted at 6, 12, and
18 months. The hamsters were observed daily during exposure days and were palpated monthly
for palpable masses starting the third month of the study. BWs were monitored weekly for the
first 8 weeks of the study and monthly thereafter. Hematologic determinations included packed
cell volume, total erythrocyte counts, total red blood cells, differential leukocyte counts, and
hemoglobin concentration. The mean corpuscular volume, mean corpuscular hemoglobin, and
MCHC were also determined. A reticulocyte count was also performed on all animals at the 18-
month kill and on 10 animals/sex/dose at 24 months. Clinical chemistry determinations included
serum AP and ALT activities, blood urea nitrogen levels, and total protein and albumin. Urinary
parameters measured were specific gravity, pH, glucose, ketones, bilirubin, occult blood, protein,
and urobilinogen. Hematology, clinical chemistries, and urinalysis were performed at interim
sacrifices and at termination. COHb was measured after a single 6-hour exposure and following
22 months of exposure. Gross and microscopic examinations were conducted on all tissues. In
addition, the weights of the brain, heart, liver, kidneys, and testes were recorded.
In the study using Syrian hamsters (Burek et al., 1984), hamsters were exposed to
analytical concentrations of dichloromethane of 510 ± 27, 1,510 ± 62, and 3,472 ± 144 ppm for
the target concentrations of 500, 1,500, and 3,500 ppm, respectively. No exposure-related
clinical signs were observed in the hamsters throughout the study. Significantly decreased
mortality was observed in females exposed to 3,500 ppm from the 13th through the 24th month
and from the 20th to the 24th month in females exposed to 1,500 ppm. Exposure to
dichloromethane had no significant effect on BW or on mean organ weights. Regarding
hematology parameters (actual data were not shown), Burek et al. (1984) stated that a few
statistically significant changes occurred, but no obvious pattern could be discerned and most
values were within the expected range for the animals. There were no exposure-related
alterations in clinical chemistry or urinalysis values. Male and female hamsters in all dose
groups had significantly elevated COHb values after a single 6-hour exposure and after
22 months of exposure, but at both time points there was no dose-response relationship above the
first dose level and no apparent significant differences in the magnitude of the changes between
the two time points. For example, mean values (± SD) for percentage COHb in male hamsters
after 22 months of exposure were 3.3 (± 3.5), 28.4 (± 5.9), 27.8 (± 2.9), and 30.2 (± 4.9), for the
control through 3,500 ppm groups, respectively. Similar values were obtained for females at
22 months and for males and females after the first day of exposure. Pathological evaluation of
hamsters showed a lack of evidence of definite target organ toxicity. Specific observations
mentioned by the authors included a trend of increasing hemosiderin in the liver of male
hamsters at 6 and 12 months; decreased amyloid deposit in organs, such as the liver, kidneys,
adrenal, and thyroid glands in exposed animals; and fewer biliary cysts in the liver. Increased
hepatic hemosiderin at the 12 month sacrifice was observed in 1/5, 1/5, 3/5, and 5/5 male
hamsters in the control through 3,500 ppm groups, respectively. No exposure-related increased
116

-------
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
incidences of hepatic hemosiderin, or other liver effects, were reported for the terminal sacrifice.
The exposure-related decreases in geriatric changes (i.e., amyloid deposits and biliary cysts)
were more prominent in females and were associated with the increased survivability in the
exposed female hamsters compared with controls. The results indicate that 3,500 ppm was a
NOAEL for adverse changes in clinical chemistry and hematological variables, as well as for
nonneoplastic histologic changes in tissues, in male and female Syrian golden hamsters. A
LOAEL was not established, based on the lack of adverse changes in clinical chemistry and
hematological variables as well as the absence of nonneoplastic histologic changes in tissues, in
male and female Syrian golden hamsters.
Evaluation of the total number of hamsters with a tumor, the number with a benign
tumor, or the number with a malignant tumor revealed no exposure-related differences in male
hamsters. In the high-dose female group, there was a statistically significant increase in the total
number of benign tumors at any tissue site (the report did not specify which sites), but this was
considered to be secondary to the increased survival of this group. Incidences of male or female
hamsters with tumors in specific tissues were not statistically significantly elevated in exposed
groups compared with control incidences. The results indicate that no statistically significant,
exposure-related carcinogenic responses occurred in male or female Syrian golden hamsters
exposed (6 hours/day, 5 days/week) to up to 3,500 ppm dichloromethane for 2 years.
4.2.2.2.4. Chronic inhalation exposure in Sprague-Dawley rats (Burek et al., 1984). In the rat
study, groups of 92-97 Sprague-Dawley rats of each sex were exposed (similar to the hamster
study described in the previous section) to 0, 500, 1,500, or 3,500 ppm dichloromethane
6 hours/day, 5 days/week for 2 years (Burek et al., 1984). Rats were approximately 8 weeks old
when exposure started. Interim sacrifices were conducted at 6, 12, 15, and 18 months.
Endpoints monitored in rats were the same as in hamsters except that total protein and albumin in
blood were not determined in rats. In addition to measurement at scheduled sacrifices, serum
ALT activity was also measured after 30 days of exposure. COHb was measured after 6, 11, 18,
and 21 months of exposure. Bone marrow cells were collected for cytogenetic studies from five
rats/sex/dose after 6 months of exposure. The scope of the pathological examinations of the rats
was the same as in the hamster study.
No significant exposure-related signs of toxicity were observed in the rats during the
study. A significant increase in mortality was seen in high-dose female rats from the 18th to the
24th month of exposure, and this appeared to be exposure-related. Exposure to dichloromethane
had no significant effect on BW gain in either males or females. The only exposure-related
alterations in organ weights was a significant increase in both absolute and relative liver weight
in high-dose males at the 18-month interim kill and a significant increase in relative liver weight
in high-dose females also at 18 months. Statistically significant changes in hematologic
parameters were restricted to increased mean corpuscular volume and mean corpuscular
117

-------
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
hemoglobin values at 15 months in males. The clinical chemistry tests revealed no significant
exposure-related effects. Male and female rats in all exposed groups had significantly elevated
COHb values at all time points, but no dose-response relationship was apparent. For example,
mean (± SD) values for percentage COHb after 21 months of exposure were 0.4 (± 0.7),
12.8 (± 2.6), 14.8 (± 4.4), and 12.2 (± 5.7) for the control through 3,500 ppm female rat groups,
respectively. Exposure-related statistically significant increases in incidences of nonneoplastic
lesions were restricted to the liver (Table 4-21). The incidences of males or females with
hepatocellular vacuolation consistent with fatty change increased as the exposure concentration
increased. Hepatocellular necrosis occurred at elevated incidences in male rats exposed to
1,500 or 3,500 ppm, compared with controls, but this endpoint was not reported in the female
data. Liver lesions were initially observed after 12 months of treatment. There was some
evidence that exposure at the two highest levels provided some inhibition of the age-related
glomerulonephropathy observed in the control rats at termination. The results indicate that the
lowest exposure level, 500 ppm, was a LOAEL for fatty changes in the liver of male and female
Sprague-Dawley rats and that exposure to >1,500 ppm induced hepatocellular necrosis in males.
118

-------
4121
Table 4-21. Incidences of selected neoplastic histologic changes in male and
female Sprague-Dawley rats exposed to dichloromethane by inhalation
(6 hours/day, 5 days/week) for 2 years


Exposure (ppm)a


Controls



Lesion, by sex
0
500
1,500
3,500
Males—n per group
92
95
95
97
Number (%) with




Liver changes




Hepatocellular necrosis
2b (2)
8 (8)
10 (10)c
11 (ll)c
Coagulation necrosis
d
--
--
--
Hepatic vacuolation (fatty change)
16b (17)
36 (38)°
43 (45)°
52 (54)°
Foci of altered hepatocytes
--
--
--
--
Foci of altered hepatocytes, basophilic
--
--
--
--
Area of altered hepatocytes
--
--
--
--
Multinucleated hepatocytes
--
--
--
--
Glomerulonephropathy




Severe
70b (76)
62 (65)
53 (56)°
39 (40)°
Any degree
92b'e (100)
91 (96)
93 (98)
90 (93)
Mammary changes




Rats with benign mammary tumors
7b (8)
3(3)
7(7)
14 (14)
Total number of benign mammary tumors
8
6
11
17
Number of tumors per tumor-bearing ratf
1.1
2.0
1.6
1.2
Females—n per group
96
95
96
97
Number (%) with




Liver changes




Hepatocellular necrosis
-
-
-
-
Coagulation necrosis
lb (1)
0 (0)
2 (2)
7 (7)
Hepatic vacuolation (fatty change)
33b (34)
49 (52)°
56 (58)°
63 (65)°
Foci of altered hepatocytes
35b (37)
36 (38)
27 (28)
50 (52)°
Foci of altered hepatocytes, basophilic
3b (3)
0 (0)
4 (4)
10 (10)
Area of altered hepatocytes
19b (20)
24 (25)
28 (29)
35 (36)°
Multinucleated hepatocytes
i° (7)
36 (38)°
34 (35)°
29 (30)°
Glomerulonephropathy




Severe
5 (5)
3 (3)
4 (4)
5 (5)
Any degree
62b (65)
64 (67)
59 (62)
48 (50)°
Mammary changes




Rats with benign mammary tumors
79 (82)
81 (85)
80 (83)
83 (86)
Total number of benign mammary tumors
165
218
245
287
Number of tumors per tumor-bearing ratf
2.1
2.7
3.1
3.5
a500 ppm = 1,737 mg/m3, 1,500 ppm = 5,210 mg/m3, 3,500 ppm = 12,158 mg/m3.
Significant dose-related trend—Cochran-Armitage trend testp < 0.05.
Significantly higher than control incidence by Fisher's exact test.
d- = Reported as "no exposure effect" by Burek et al. (1984); data not given.
eBurek et al. (1984) reported that 93/92 male mice had glomerulonephropathy in the kidney in the control group; the
incidence was corrected to 92/92.
Calculated by EPA.
Source: Burek et al. (1984).
4122
4123
4124	In females, an increasing trend was seen in the incidence of foci or areas of altered
4125	hepatocytes. Female rats in all exposed groups showed increased incidence of multinucleated
4126	hepatocytes in the centrilobular region, compared with controls, but there was no evidence of
119

-------
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
increasing incidence or severity with increasing exposure level (Table 4-21). The foci and areas
were apparent after 12 months and their number and size increased thereafter, but incidences for
neoplastic nodules in the liver or hepatocellular carcinomas were not increased in any exposure
group. A statistically significant increased incidence of salivary gland sarcomas was reported for
male rats exposed to 3,500 ppm. Burek et al. (1984) considered this finding unusual and
inconsistent with other existing data because the primary target organ for dichloromethane seems
to be the liver. Incidences of rats with benign mammary gland tumors were not statistically
significantly higher in exposed male or female groups compared with controls, and exposed male
and female groups showed no significantly increased incidences for malignant mammary gland
tumors. The average number of benign mammary tumors per tumor-bearing rat increased with
increasing exposure level. In females, the values were 2.1, 2.7, 3.1, and 3.5 in the control
through 3,500 ppm groups, respectively; males showed a similar response with increasing
exposure level, albeit to a lesser extent (Table 4-21). Burek et al. (1984) concluded that the
significance of this benign mammary tumor response (i.e., increase in number of tumors per
tumor-bearing rat) was unknown but speculated that the predisposition of this strain of rats
(historical control incidences of female with benign mammary tumors normally exceeded 80%)
plus the high exposure to dichloromethane may have resulted in the response.
4.2.2.2.5. Chronic inhalation exposure in Sprague-Dawley rats (Nitschke et al., 1988a).
Nitschke et al. (1988a) examined the toxicity and carcinogenicity of lower concentrations of
dichloromethane in Sprague-Dawley rats. Groups of 90 male and 90 female rats were exposed to
0, 50, 200, or 500 ppm dichloromethane (>99.5% pure) 6 hours/day, 5 days/week for 2 years.
Interim sacrifices were conducted at 6, 12, 15, and 18 months (five rats/sex/interval). An
additional group of 30 female rats was exposed to 500 ppm for 12 months and then exposed to
room air for up to an additional 12 months, and another group of 30 female rats was exposed to
room air for the first 12 months, followed by exposure to 500 ppm for the last 12 months of the
study. These latter groups were included to examine temporal relationships between exposure
and potential carcinogenic response. All groups of rats were examined daily for signs of toxicity
and all rats were examined for palpable masses prior to the initial exposure and at monthly
intervals after the first 12 months. BW was checked twice a month for the first 3 months and
monthly thereafter. Blood samples were collected at interim sacrifices and analyzed for total
bilirubin, cholesterol, triglycerides, potassium, estradiol, follicle-stimulating hormone, and
luteinizing hormone levels. In addition, COHb was determined at multiple times in blood
collected from the tail vein. DNA synthesis (incorporation of 3H-thymidine as a measure of
cellular proliferation) was measured in the liver of separate groups of female rats after exposure
to the various concentrations for 6 and 12 months (four females/exposure group per interval).
All rats were subjected to a complete necropsy, and sections from most tissues were processed
for microscopic examination.
120

-------
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
Exposure to dichloromethane at any of the exposure levels did not significantly alter
mortality rates, BWs, organ weights, clinical chemistry values, or plasma hormone levels
(Nitschke et al., 1988a). Blood COHb was elevated in a dose-related manner but not in an
exposure duration-related fashion, suggesting lack of accumulation with repeated exposures. For
example, mean (± SD) values for percentage COHb were 2.2 (± 1.3), 6.5 (± 1.1), 12.5 (± 0.8),
and 13.7 (± 0.6) for male rats in the control through 500 ppm groups, respectively, at the
terminal sacrifice and were similarly affected at the 6-month and 12-month intervals (e.g.,
respective values for males were 0.3(± 0.7), 2.8 (± 0.3), 9.6 (± 1.2), and 12.7 (± 1.6) at the
12-month sacrifice).
The results of the thymidine incorporation experiment revealed no detectable alteration in
the rate of liver DNA synthesis in the exposed groups compared with controls. Statistically
significantly increased incidences of nonneoplastic liver lesions (hepatic vacuolation and
multinucleated hepatocytes) occurred only in females in the 500 ppm group (Table 4-22). Male
rat incidence for hepatocyte vacuolation was elevated at 500 ppm but not to a statistically
significant degree. In the group of female rats exposed for only 12 months to 500 ppm,
significantly increased incidences of nonneoplastic lesions, compared with controls, were
restricted to liver cytoplasmic vacuolization (16/25 = 64%) and multinucleated hepatocytes
(9/25= 36%) in rats exposed during the first 12 months of the study; rats exposed only during the
last 12 months of the study showed no elevated incidences of the liver lesions.
Table 4-22. Incidences of selected nonneoplastic histologic changes in male
and female Sprague-Dawley rats exposed to dichloromethane by inhalation
(6 hours/day, 5 days/week) for 2 years
Exposure (ppm)"

Controls


Trend
Late
Early
Lesion, by sex
0
50
200
500 />-valucb
500c
500c
Males—n per group
Number (%) with
70
70
70
70
NAd
NA
Hepatic vacuolation (fatty change)
Multinucleated hepatocytes
22 (31)
e
—
28 (40)


Females—n per group
Number (%) with
70
70
70
70
25
25
Hepatic vacuolation (fatty change)
Multinucleated hepatocytes
41 (59)
8 (11)
42 (60)
6 (9)
41 (59)
12(17)
53 (76)f 0.01
27 (39)f <0.0001
15 (60)
3 (12)
16 (64)f
9 (36)f
a50 ppm = 174 mg/m3, 200 ppm = 695 mg/m3, 500 ppm = 1,737 mg/m3.
bCochran-Armitage trend test.
°Late 500 = no exposure for first 12 months followed by 500 ppm for last 12 months; early 500 = 500 ppm for first
12 months followed by no exposure for last 12 months.
dNA= there were no male rats in these exposure groups.
e- = Incidences not reported.
Significantly (p < 0.05) higher than control incidence by Fisher's exact test (Nitschke et al., 1988a).
Source: Nitschke et al. (1988a).
121

-------
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
A few fibrosarcomas or undifferentiated sarcomas in the mammary gland were seen in
the exposed rats, but these incidences were not statistically significant (Table 4-23).
Significantly increased incidences of rats with neoplastic lesions were restricted to benign
mammary tumors in female rats exposed for 2 years to 200 ppm compared with controls
(61/69 = 88%) (Table 4-23). However, significantly elevated incidences of this tumor type were
not observed in 500 ppm females, and the 200 ppm incidence was within the range for historical
control values for benign mammary tumors in female Sprague-Dawley rats (79-82%) from two
other chronic toxicity/carcinogenicity studies from the same laboratory. A slight, but statistically
significant, increase in the number of palpable masses in subcutaneous or mammary regions (at
23 months) per tumor-bearing rat was observed only in the 500 ppm female group. The numbers
of benign mammary tumors per tumor-bearing rat were slightly elevated in the exposed groups
compared with control groups, but no statistical analysis of this variable was performed. In
female rats exposed to 500 ppm (during the first or second 12 months of the study), slight but
statistically significant elevations were found in the number of palpable masses in subcutaneous
or mammary regions per tumor-bearing rat; the numbers of benign mammary tumors per tumor-
bearing rat were slightly elevated compared with those of controls, but statistical analysis of this
variable was not performed.
A statistically significant increased incidence of brain or CNS tumors was not observed,
but six astrocytoma or glioma (mixed glial cell) tumors were seen in the exposed groups (4 in
males, 2 in females). The authors concluded that there was no distinct exposure-related
malignant carcinogenic response in male or female Sprague-Dawley rats exposed (6 hours/day,
5 days/week) to up to 500 ppm dichloromethane for 2 years (Nitschke et al., 1988a).
122

-------
Table 4-23. Incidences of selected neoplastic histologic changes in male and
female Sprague-Dawley rats exposed to dichloromethane by inhalation
(6 hours/day, 5 days/week) for 2 years
Exposure (ppm)a

Controls



Late
Early
Lesion, by sex
0
50
200
500
500b
500b
Males—n per group
70
70
70
70
0
0
Number (%)° with






Liver tumors
0 (0)
0 (0)
0 (0)
0 (0)


Lung tumors
0 (0)
0 (0)
0 (0)
0 (0)


Mammary gland tumors






Adenocarcinoma or carcinoma
0 (0)
0 (0)
0 (0)
0 (0)


Fibroadenoma
2 (4)
0 (0)
2 (3)
2 (3)


Fibroma
6 (11)
1 (6)
6 (11)
10 (16)


Fibrosarcoma
0 (0)
1 (6)
1 (6)
0 (0)


Undifferentiated sarcoma
0 (0)
2 (4)
0 (0)
0 (0)


Fibroma, fibrosarcoma, or undifferentiated
6 (11)
4 (6)
7 (12)
10 (16)


sarcomad






Brain tumors






Astrocytoma or glial cell
0 (0)
1(1)
2 (3)
1 (1)


Granular cell
0 (0)
0 (0)
0 (0)
1 (1)


Females—n per group
70
70
70
70
25
25
Number (%)°with






Liver tumors






Neoplastic nodule(s)
4 (6)
4 (6)
3 (4)
4 (6)
0 (0)
1(4)
Hepatocellular carcinoma
1 (1)
0 (0)
2 (3)
1 (1)
0 (0)
0 (0)
Lung tumors
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
Mammary gland tumors






Adenocarcinoma or carcinoma
6 (9)
5 (7)
4 (6)
4 (6)
3 (12)
2 (8)
Adenoma
1 (1)
1 (1)
2 (3)
1 (1)
2 (8)
0 (0)
Fibroadenoma
51 (74)
57 (83)
60 (87)
55 (80)
22 (88)
23 (92)
Fibroma
0 (0)
1 (1)
0 (0)
1 (1)
1 (4)
1 (1)
Fibrosarcoma
1 (1)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
Number with palpable masses in subcutaneous
55 (78)
56 (81)
60 (87)
59 (86)
22 (88)
23 (92)
or mammary region






Number of palpable masses in subcutaneous or
1.8
2.1
2.0
2.T
2.3e
2.T
mammary region per tumor-bearing rat






Number with benign tumors
52 (75)
58 (84)
61f (88)
55 (80)
23 (92)
23 (92)
Number of benign tumors per tumor-bearing rat
2.0
2.3
2.2
2.7
2.2
2.6
Brain tumors






Astrocytoma or glial cell
0 (0)
0 (0)
0 (0)
2 (3)
0 (0)
0 (0)
Granular cell
1 (1)
0 (0)
0 (0)
1 (1)
0 (0)
0 (0)
'50 ppm = 174 mg/m3, 200 ppm = 695 mg/m3, 500 ppm = 1,737 mg/m3
bLate 500 = no exposure for first 12 months followed by 500 ppm for last 12 months; early 500 = 500 ppm for first
12 months followed by no exposure for last 12 months. No males were included in these exposure groups.
Percentages were based on the number of tissues examined microscopically per group. In males, 69 lungs were
examined microscopically in the 50 ppm groups, and only 57, 65, 59, and 64 mammary glands were examined in
the control, 50, 200, and 500 ppm groups, respectively. In females, 69 mammary glands were examined
microscopically in the control, 50, 200, and 500 ppm groups.
dEPA summed across these three tumors, assuming no overlap.
"Significantly (p < 0.05) higher than control by Haseman's test (Nitschke et al., 1988a).
Significantly (p < 0.05) higher than control incidence by Fisher's exact test (Nitschke et al., 1988a).
Source: Nitschke et al. (1988a).
4210
123

-------
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4.2.2.2.6. Chronic inhalation exposure in Sprague-Dawley rats (Maltoni et al., 1988). Maltoni
et al. (1988) conducted an inhalation exposure study in Sprague-Dawley rats. Two groups of
female rats (54-60/dose) were exposed to 0 or 100 ppm dichloromethane for 104 weeks. The
exposure period was 4 hours/day, 4 days/week for 7 weeks and then 7 hours/day, 5 days/week
for 97 weeks. Endpoints monitored included clinical signs, BW, and full necropsy at sacrifice
(when spontaneous death occurred). For each animal sacrificed, histopathologic examinations
were performed on the following organs: brain and cerebellum, zymbal glands, interscapular
brown fat, salivary glands, tongue, thymus and mediastinal lymph nodes, lungs, liver, kidneys,
adrenals, spleen, pancreas, esophagus, stomach, intestine, bladder, uterus, gonads, and any other
organs with gross lesions.
There was no evidence of increased mortality in the exposed group, and there was no
effect on BW (Maltoni et al., 1988). Little information was provided regarding nonneoplastic
effects, precluding assignment of NOAELs and LOAELs for possible nonneoplastic effects in
this study. Dichloromethane exposure was not related to the percentage of rats with benign
tumors and malignant tumors, malignant tumors, or the number of total malignant tumors per
100 animals. The percentage of rats with benign mammary tumors was 40.0% in controls and
64.8% in the exposed group, and the percentage of malignant mammary tumors was 3.3 and
5.5%) in controls and exposed, respectively. Neither of these differences was statistically
significant.
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL AND INHALATION
Reproductive and development studies of dichloromethane exposure are summarized in
Table 4-24 and described in detail below. No effects on reproductive performance were
observed in a 90-day gavage study in Charles River CD rats with doses up to 225 mg/kg-day
(General Electric Co., 1976) or in a two-generation reproductive toxicity study with F344 rats
exposed to concentrations up to 1,500 ppm for 14 or 17 weeks before mating of the F0 and
F1 generations, respectively, as well as during the F1 gestational period (GDs 0-21) (Nitschke et
al., 1988b). Reproductive parameters (e.g., number of litters, implants/litter, live fetuses/litter,
percent dead/litter, percent resorbed/litter, or fertility index4) were also examined in a study in
male Swiss-Webster mice administered dichloromethane (250 or 500 mg/kg) by subcutaneous
injection three times/week for 4 weeks, and in a similar study involving inhalation exposure to 0,
100, 150, or 200 ppm dichloromethane; no statistically significant effects were seen in either
protocol, although some evidence of a decrease in fertility index was seen in the 150 and
200 ppm groups (Raje et al., 1988).
fertility index defined as number of females impregnated divided by total number of females mated times 100.
124

-------
4246
Table 4-24. Summary of studies of reproductive and developmental effects of dichloromethane exposure in
animals
Species and n
Exposure dose
Exposure period
Results
Reference
Oral and Gavage
Charles River rats
(males and females),
10 per sex per dose
group
Swiss-Webster mice
(males), 20 per group
F344 rats (females),
17-21 per dose group
F344 rats (males and
females, two
generation), 30 per
sex per dose group
(F0 and Fl)
Swiss-Webster mice
(males), 20 per group
Long-Evans rats
(female), 16-21 per
dose group
Long-Evans rats
(female), 16-21 per
dose group
0, 25, 75, 225 mg/kg
(gavage)
0, 250, 500 mg/kg
(subcutaneous
injection), 3x per
week
0, 337.5,
450 mg/kg-day
(gavage)
0, 100, 500,
1,500 ppm,
6 hours/day
0, 100, 150,
200 ppm,
2 hours/day
0, 4,500 ppm
0, 4,500 ppm
90 days before mating
(10 days between last
exposure and mating
period)
4 weeks prior to mating
(1 week between last
exposure and mating
period)
GDs 6-19
No effects on fertility index, number of pups per
litter, pup survival, or Fl BW, hematology, and
clinical chemistry tests (up to 90 days of age)
No effects on fertility index, number of litters,
implants per litter, live fetuses per litter, resorption
rate; no testicular effects
Decreased maternal weight gain; no effect on
resorption rate, number of live litters, implants,
live pups, or pup weight
Inhalation
14 weeks prior to
mating (F0), GDs 0-21,
and 17 weeks prior to
mating, beginning
PND 4, (Fl)
6 weeks, prior to
mating (2 days between
last exposure and
mating period)
12-14 days before
mating and/or GDs 1-
17
12-14 days before
mating and/or GDs 1-
17
No effect on fertility index, litter size, neonatal
survival, growth rates, or histopathologic lesions
Fertility index decreased in 150 and 200 ppm
group (statistical significance depends on test
used); no effects on number of litters, implants per
litter, live fetuses per litter, resorption rate; no
testicular effects.
Gestational exposure resulted in increased absolute
and relative maternal liver weight, decreased fetal
BW
Altered rate of behavioral habituation to novel
environment (at 4 days of age). No effect on
crawling (at 10 days), movement in photocell cage
(15 days), use of running wheel (45-108 days), and
shock avoidance (4 months).
General
Electric Co.
(1976)
Raje et al.
(1988)
Narotsky and
Kavlock
(1995)
Nitschke et al.
(1988b)
Raje et al.
(1988)
Hardin and
Manson
(1980)
Bornschein et
al. (1980)
125

-------
Table 4-24. Summary of studies of reproductive and developmental effects of dichloromethane exposure in
animals
Species and n
Exposure dose
Exposure period
Results
Reference
Swiss-Webster mice
0, 1,250 ppm,
GDs 6-15
Increased incidence of extra center of ossification
Schwetz et al.
(females), 30-40 per
group
7 hours/day

in sternum, increased (-10%) maternal blood
COHb, increased maternal weight, increased
maternal absolute liver weight
(1975)
Sprague-Dawley rats
0, 1,250 ppm,
GDs 6-15
Decreased incidence of lumbar ribs or spurs,
Schwetz et al.
(females), 20-35 per
group
7 hours/day

increased incidence of delayed ossification of
sternebrae, increased (-10%) maternal blood
COHb, increased maternal absolute liver weight
(1975)
4247
4248
4249
4250
126

-------
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
Following exposure of pregnant F344 rats to gavage doses of up to 450 mg/kg-day on
GDs 6-19, maternal weight gain was decreased, but no effects were found on the number of
resorption sites, pup survivability, or pup weights at postnatal days (PNDs) 1 or 6 (Narotsky and
Kavlock, 1995). The developmental effects following exposure of Long-Evans rats to
4,500 ppm for 14 days prior to mating and during gestation (or during gestation alone) were
decreased offspring weight at birth and changed behavioral habituation of the offspring to novel
environments (Bornschein et al., 1980; Hardin and Manson, 1980) (see section 4.3.2 for more
details). In standard developmental toxicity studies involving exposure to 1,250 ppm on GDs 6-
15, no adverse effects on fetal development were found in Swiss-Webster mice or Sprague-
Dawley rats, but the incidence of minor skeletal variants (e.g., delayed ossification of sternebrae)
was increased. (Schwetz et al., 1975) (see section 4.3.2).
4.3.1. Reproductive Toxicity Studies
4.3.1.1. Oral (Gavage) Studies
In a study sponsored by the General Electric Co. (1976), Charles River CD rats
(10/sex/dose level) were administered 0, 25, 75, or 225 mg/kg-day dichloromethane by gavage in
water for 90 days. The test material was dichloromethane (of unspecified purity) purchased from
Dow Chemical Company. At approximately 100 days of age, the rats were mated 1 to 1 to
produce the F1 generation. F1 rats (15/sex/dose level) received the same treatment as F0 for
90 days, at which time they were sacrificed and necropsied. Comprehensive sets of 24 tissues
from 10 male and 10 female F1 rats from the control and 225 mg/kg-day groups were examined
microscopically after embedding, sectioning, and staining. F1 rats were monitored for clinical
signs, BW effects, and food consumption. Reproductive parameters examined were fertility
index, number of pups per litter, and pup survival. F1 rats also underwent hematology and
clinical chemistry tests and urinalysis at 1, 2, and 3 months of the study and ophthalmoscopic
examination at 3 months. There were no significant compound-related alterations in any of the
endpoints monitored.
Raje et al. (1988) administered dichloromethane (250 or 500 mg/kg) by subcutaneous
injection three times per week for 4 weeks to male Swiss-Webster mice (20/group). Mating with
unexposed females started 1 week after the last exposure and continued for 2 weeks. After the
mating period, the males were sacrificed, and the testes were examined microscopically. On
GD 17, the females were sacrificed and the uterine horns examined for live, dead, or resorbed
fetuses. The authors reported that exposure to dichloromethane had no statistically significant
effects on number of litters, implants/litter, live fetuses/litter, percent dead/litter, percent
resorbed/litter, or fertility index. Examination of the testes showed no significant alterations
compared with controls.
127

-------
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4.3.1.2. Inhalation Studies
Nitschke et al. (1988b) conducted a two-generation reproductive toxicity study in rats.
Groups of F344 rats (30/sex/dose level) were exposed by inhalation in whole-body chambers to
0, 100, 500, or 1,500 ppm dichloromethane (99.86% pure) 6 hours/day, 5 days/week for
14 weeks and then mated to produce the F1 generation. Exposure of dams continued after
mating on GDs 0-21 but was interrupted until PND 4. After weaning 30 randomly selected F1
pups/sex/dose level were exposed as the parental generation for 17 weeks and subsequently
mated to produce the F2 generation. The results showed no statistically significant exposure-
related changes in reproductive performance indices (fertility, litter size), neonatal survival,
growth rates, or histopathologic lesions in F1 (Table 4-25) or F2 weanlings sacrificed at time of
weaning. According to the authors none of the values in Table 4-25 was significantly different
from control values (a = 0.05).
Table 4-25. Reproductive outcomes in F344 rats exposed to
dichloromethane by inhalation for 14 weeks prior to mating and from
GDs 0-21

0
Exposure (ppm)a
100 500
1,500
Fertility indexb
77%
77%
63%
87%
Gestation index0
100%
100%
100%
100%
Gestation survival indexd
99.6%
100%
100%
96.6%
4-day survival index6
91.0%
95.2%
98.5%
98.6%
28-day survival indexf
99.4%
99.4%
100%
99.5%
Sex ratio on day 1 (M:F)
48:52
50:50
50:50
52:48
Litter size




Day 0
11 ± 2
10 ±2
10 ±3
11 ± 2
Day 28
7 ± 2
7 ± 2
7 ± 2
8 ± 2
Pup BWs, g




Day 1
5.2 ±0.4
5.3 ±0.5
5.3 ±0.4
5.2 ±0.4
Day 4
7.4 ±0.7
7.5 ± 1.1
7.7 ±0.7
7.3 ±0.7
Day 28, male
44.6 ±5.8
45.9 ±5.0
47.0 ±5.4
45.0 ±5.9
Day 28, female
43.2 ±4.3
43.8 ±4.5
44.4 ±5.7
43.0 ±4.8
a100 ppm = 347 mg/m3, 500 ppm = 1,737 mg/m3, 1,500 ppm = 5,210 mg/m3,
bNumber of females delivering a litter expressed as a percentage of females placed with a male.
°Number of females delivering a live litter expressed as a percentage of the number of females delivering a
litter.
Percentage of newborn pups that were alive at birth.
"Percentage of pups surviving to day 4.
Percentage of pups alive on day 4 and surviving to day 28.
Source: Nitschke et al. (1988b).
128

-------
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
Raje et al. (1988) exposed groups of male Swiss-Webster mice (20/group) to 0, 100, 150,
or 200 ppm dichloromethane (HPLC grade, JT Baker Chemical Co.) in inhalation chambers for
2 hours/day, 5 days/week for 6 weeks. Mating with unexposed females started 2 days after the
last exposure. As in the subcutaneous injection protocol described in the previous section, after
the 2-week mating period, the males were sacrificed and the females were sacrificed on GD 17.
Exposure of the male mice to dichloromethane had no statistically significant effects on number
of litters, implants/litter, live fetuses/litter, percent dead/litter, or percent resorbed/litter, and no
significant alterations in the testes were noted. The fertility index was 95%, 95%, 80%, and 80%
in the control, 100, 150 and 200 ppm groups, respectively. This decrease was not statistically
significant as reported by the authors. Details of the statistical analyses were not provided. The
overall Chi-squarep-value was 0.27. Using a Cochran-Armitage exact trend test on these data,
EPA calculated a one-sided p-v alue of 0.059. Individual /^-values for the comparison of each
group with the control group were 0.97, 0.17, and 0.17 for the 100, 150 and 200 ppm groups,
respectively. The results for the combined 150 and 200 ppm groups were statistically different
from the combined controls and 100 ppm group (Fisher's exact test, one-sided/;-value = 0.048),
suggesting a NOAEL of 100 ppm and LOAEL of 150 ppm.
4.3.2. Developmental Toxicity Studies
The metabolism of dichloromethane into CO by CYP2E1 raises concerns pertaining to
developmental neurotoxicity. Gestational exposure to CO results in developmental toxicity and
there are reports indicating that exposures as low as 75 ppm CO can result in significant
neurological effects in offspring (Giustino et al., 1999). Neurobehavioral deficits in offspring
include impaired avoidance behavior (De Salvia et al., 1995) and memory (Giustino et al., 1999).
Neurochemical changes, such as abnormal dopaminergic function (Cagiano et al., 1998) and
disruption of neuronal proliferation (Fechter, 1987), have also been observed. Oral and
inhalation dichloromethane exposure studies have demonstrated increased blood CO levels (see
section 3.3). In addition, increased blood CO levels were seen in rat fetuses exposed through
maternal inhalation to 500 ppm dichloromethane on GD 21 (Anders and Sunram, 1982), and
placental transfer of dichloromethane also occurs (Withey and Karpinski, 1985; Anders and
Sunram, 1982)
4.3.2.1. Oral (Gavage) Studies and Culture Studies
Narotsky and Kavlock (1995) evaluated developmental effects of dichloromethane
(99.9% pure) in F344 rats (17-21/dose group) treated with 0, 337.5, or 450 mg/kg-day
dichloromethane by gavage in corn oil on GDs 6-19. Dams were weighed on GDs 6, 8, 10, 13,
16, and 20 and allowed to deliver naturally. They were sacrificed on PND 6 to count uterine
implantation sites. Pups were grossly examined for developmental abnormalities and weighed
on PNDs 1, 3, and 6. Dead pups or pups with no gross abnormalities were sacrificed and
129

-------
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
examined for soft tissue abnormalities. Maternal weight gain during pregnancy was significantly
reduced in high-dose dams (by 33%, as estimated from Figure 5 of the paper); this group also
exhibited rales and nasal congestion. Treatment with dichloromethane did not induce resorptions
or alter the number of live litters on PND 1 or 6, the number of implants, the number of live pups
on PND 1 or 6, or pup weight per litter. No gross or soft tissue abnormalities were observed.
Rat embryos in culture medium were exposed to 0, 3.46, 6.54, 9.79, or 11.88 [j,mol/mL
dichloromethane for 40 hours. At the end of the exposure, embryos were observed for
development of yolk sac vasculature, crown-rump length, total embryonic protein content, and
number of somite pairs. A concentration of dichloromethane of 6.54 [j,mol/mL of culture
medium resulted in decreased crown-rump length, decreased somite number, and decreased
amount of protein per embryo, whereas no effects were seen at 3.46 [j,mol/mL (Brown-Woodman
et al., 1998). A time-course experiment conducted with a concentration of dichloromethane of
9.22 [j.mol/mL showed that marked differences in growth and development from controls were
not significant until about 8 hours of culture. Brown-Woodman et al. (1998) noted that the
concentrations that caused embryotoxicity in this study were much higher than those found in
individuals studied under controlled exposure conditions and comparable to those found in
postmortem blood after fatal inhalation.
4.3.2.2. Inhalation Studies
Schwetz et al. (1975) exposed pregnant Swiss-Webster mice (30-40/group) and Sprague-
Dawley rats (20-35/group) by inhalation in whole-body chambers to 0 or 1,250 ppm
dichloromethane (97.86% pure) 7 hours/day on GDs 6-15. Maternal BWs were recorded on
GDs 6, 10, and 16 and on the day of sacrifice (GD 18 for mice, GD 21 for rats). At sacrifice,
uterine horns were excised and examined for fetal position and number of live, dead, or absorbed
fetuses. Fetuses were observed for gross, soft tissue, and skeletal abnormalities. The only
effects seen on developing fetuses were changes in the incidence of minor skeletal variants. In
rats, the incidence of lumbar ribs or spurs was significantly decreased compared with controls,
whereas the incidence of delayed ossification of sternebrae was significantly greater than in
controls. In mice, a significant number of litters contained pups with a single extra center of
ossification in the sternum. Exposure to dichloromethane produced significantly elevated blood
COHb content in dams of both species (approximately 9—10% after 10 exposures versus 1-2% in
controls). BWs in exposed mouse dams were significantly increased (11—15%) compared with
those in controls but were not affected in exposed rat dams. Mean absolute liver weights of
exposed dams of both species were significantly elevated compared with controls, but mean
relative liver weights were not affected. The results indicate that 1,250 ppm was a LOAEL for
minimal maternal effects (increased COHb and increased absolute liver weight) and a LOAEL
for adverse effects on the fetuses.
130

-------
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
Hardin and Manson (1980) conducted a study in female Long-Evans rats to determine
whether exposure before and during gestation is more detrimental to reproductive outcome than
exposure either before or during gestation alone. Four groups of 16-21 rats were formed in
which the rats were exposed by inhalation in whole-body chambers to 4,500 ppm
dichloromethane (technical grade, >97% pure) 6 hours/day for 12-14 days before breeding
and/or on GDs 1-17 or were exposed to filtered air. Maternal BWs were measured every 4 days.
Dams were euthanized on GD 21 and livers and uteri removed. Livers were weighed, and
uterine horns were examined for fetal position and number of live, dead, or absorbed fetuses.
Fetuses were observed for gross, soft-tissue, and skeletal abnormalities. Exposure during
gestation (with or without pre-gestation exposure) significantly increased maternal liver weight
(absolute and relative) by about 10-12 and 9-12%, respectively, and decreased fetal BW by
about 9-10%) relative to those exposed to filtered air during gestation. None of the groups
showed significant alterations in the incidence of gross, external, skeletal, or soft-tissue
anomalies. Using the same study design and exposure level, Bornschein et al. (1980) observed
behavioral activities at various ages. Assessed activities included head movement/pivoting when
placed in a novel environment (4 days of age), limited crawling (10 days), movement in a
photocell cage (15 days), use of running wheel (45-108 days), and shock avoidance (4 months).
Exposure during gestation (with or without pre-gestation exposure) caused altered rates of
behavioral habituation to novel environments in the pups tested as early as 10 days of age that
were still present at 150 days of age. Growth, food and water consumption, wheel running
activity, and avoidance learning were not significantly affected by exposure to dichloromethane.
The results indicate that 4,500 ppm was a LOAEL for maternal effects (10%> increased absolute
and relative liver weight) and for effects on the fetuses (10%> decreased fetal BW and altered
behavioral habituation to novel environments).
In a study of early-life (including gestational) exposures, Maltoni et al. (1988) exposed
54 pregnant Sprague-Dawley rats to 100 ppm dichloromethane via inhalation 4 hours/day,
5 days/week for 7 weeks, followed by 7 hours/day, 5 days/week for 97 weeks. Exposure
apparently started on GD 12. Groups of 60 male and 69 female newborns continued to be
exposed after birth to 60 ppm dichloromethane 4 hours/day, 5 days/week for 7 weeks, followed
by exposure 7 hours/day, 5 days/week for 97 weeks. Additional groups of 60 male and
70 female newborn were exposed after birth to 60 ppm dichloromethane 4 hours/day,
5 days/week for 7 weeks and then for 7 hours/day, 5 days/week for 8 weeks. BWs were
measured every 2 weeks during exposure and every 8 weeks thereafter. At the end of exposure,
animals were sacrificed and histologic examinations were performed on 20 tissue types.
Early life exposures of Sprague-Dawley rats to dichloromethane (Maltoni et al., 1988)
did not affect mortality or BW in any group. Also, there was no significant effect of exposure to
dichloromethane on the percentage of animals with benign and malignant tumors and malignant
tumors, the number of malignant tumors per 100 animals, or the percentage of animals with
131

-------
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
4437
4438
4439
4440
4441
4442
4443
4444
4445
4446
4447
4448
4449
4450
4451
4452
4453
4454
benign mammary tumors, malignant mammary tumors, leukemias, pheochromocytomas, and
pheochromoblastomas. The results provide no evidence that gestational exposure to 100 ppm
dichloromethane during early life stages of development increases the susceptibility of Sprague-
Dawley rats to the potential carcinogenicity of dichloromethane, but further conclusions from
these results are precluded because the study included only one exposure level that was below
the maximum tolerated dose for adult Sprague-Dawley rats. Experiments comparing cancer
responses from early-life exposures with adult exposures are not available for F344 rats or
B6C3Fi mice, the strains of animals in which carcinogenic responses to dichloromethane have
been observed.
In summary, the potential for gestational exposure to CO, resulting from maternal
dichloromethane exposure via oral and inhalation routes, raises concerns regarding
neurodevelopmental effects. In addition, dichloromethane transfer across the placenta has also
been seen in inhalation exposure studies in rats (Withey and Karpinski, 1985; Anders and
Sunram, 1982). Although few developmental effects were observed at high exposures of
dichloromethane (Bornschein et al., 1980; Schwetz et al., 1975), there are no studies that have
thoroughly evaluated neurobehavioral and neurochemical changes resulting from gestational
dichloromethane exposure. The available data identify changes of behavior habituation at
4,500 ppm (Bornschein et al., 1980) and increases in COHb at 1,250 ppm (Schwetz et al., 1975).
The behavioral changes observed at 4,500 ppm indicate developmental neurotoxic effects. No
other neurological endpoints have been evaluated in the available developmental studies of
dichloromethane, but increases in blood COHb strongly suggest that dichloromethane is being
metabolized to CO. Gestational exposure to CO can result in significant neurological effects in
offspring, including neurobehavioral deficits (De Salvia et al., 1995), memory effects (Giustino
et al., 1999), and neurochemical changes (Cagiano et al., 1998; Fechter, 1987). As a result, it is
unknown if developmental neurotoxicity could occur at lower exposures to dichloromethane.
4.4. OTHER DURATION- OR ENDPOINT-SPECIFIC STUDIES
4.4.1. Short-term (2-Week) Studies of General and Hepatic Effects in Animals
Two short-term (2-week) studies examined hepatic and renal effects of dichloromethane
exposure in F344 rats (Berman et al., 1995) and CD-I mice (Condie et al., 1983). Berman et al.
(1995) administered dichloromethane by gavage in corn oil for up to 14 days to groups of eight
female F344 rats at dose levels of 0, 34, 101, 337, or 1,012 mg/kg-day. Starting at day 4, deaths
occurred in the 1,012 mg/kg-day exposure group, with seven of eight rats dying before the end of
the 14-day exposure period. In the dose groups that did not experience this high mortality,
incidences of increased necrotic hepatocytes were 0/8, 0/8, 0/8, and 3/8 for the 0, 34, 101, and
337 mg/kg-day groups, respectively. The increase in liver lesions was not accompanied by
increases in serum activities of ALT or AST. Kidneys, spleen, and thymus were also
histopathologically examined in this study, but none showed exposure-related lesions. The
132

-------
4455
4456
4457
4458
4459
4460
4461
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
4480
4481
4482
4483
4484
4485
4486
4487
4488
4489
4490
4491
4492
results indicate that 101 mg/kg-day was a NOAEL and 337 mg/kg-day was a LOAEL for
increased incidence of degenerative lesions in female rats exposed for 14 days. In a companion
study with groups of eight female F344 rats that were given single doses of 0, 101, 337, 1,012, or
1,889 mg/kg-day, incidences of rats with increased necrotic hepatocytes were 1/8, 0/8, 8/8, 7/8,
and 8/8, respectively (Berman et al., 1995).
Condie et al. (1983) detected exposure-related liver lesions in a 14-day gavage study in
which dichloromethane in corn oil was administered to male CD-I mice at dose levels of 0, 133,
333, or 665 mg/kg-day. Incidences of mice with minimal or slight cytoplasmic vacuolation were
1/16, 0/5, 3/5, and 4/5 for the control through high-dose groups, respectively. The kidneys were
also examined histopathologically in this study but showed no exposure-related lesions. No
other tissues were prepared for histologic examination. Blood urea nitrogen, serum creatinine,
and serum ALT activities were not significantly altered by exposure. All dose levels
significantly reduced to the same extent the active transport of p-aminohippurate into renal
cortical slices in vitro, a measure of proximal tubule function. The results most clearly identify
133 mg/kg-day as a NOAEL and 333 mg/kg-day as a LOAEL for increased incidence of
hepatocyte vacuolation in male mice.
4.4.2. Immunotoxicity Studies in Animals
Aranyi et al. (1986) studied the effects of acute inhalation exposures to 50 or 100 ppm
dichloromethane on two measures of immune response (susceptibility to respiratory infection
and mortality due to Streptoccocus zooepidemicus exposure and ability of pulmonary
macrophages to clear infection with Klebsiella pneumoniae). Female CD1 mice that were 5-
7 weeks of age at the start of the exposure portion of the experiment were used for both assays.
Up to five replicate groups of about 30 mice were challenged with viable S. zooepidemicus
during simultaneous exposure to dichloromethane or to filtered air. Deaths were recorded over a
14-day observation period. Clearance of 35S-labeled K. pneumoniae by pulmonary macrophages
was determined by measuring the ratio of the viable bacterial counts to the radioactive counts in
each animal's lungs 3 hours after infection; 18 animals were used per dose group. A single 3-
hour exposure to 100 ppm dichloromethane significantly increased the susceptibility to
respiratory infection and greater mortality following exposure to S. zooepidemicus (p < 0.01).
Twenty-six deaths occurred in 140 (18.6%) mice challenged during a 3-hour exposure to 100
ppm dichloromethane; in contrast, nine deaths occurred in 140 mice (6.4%) exposed to filtered
air. The 3-hour exposure to 100 ppm dichloromethane was associated with a statistically
significant (p < 0.001) 12% decrease in pulmonary bactericidal activity (91.6 and 19.6% of
bacteria killed in controls and 100 ppm group, respectively). No difference was seen in either
mortality rate or bactericidal activity in experiments using a single 3-hour exposure to 50 ppm or
3-hour exposures to 40 ppm dichloromethane repeated daily for 5 days compared with control
animals exposed to filtered air. These results suggest that 3-hour exposure to 50 ppm
133

-------
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
dichloromethane was a NOAEL and 100 ppm was a LOAEL for decreased immunological
competence (immunosuppression) in CD-I mice.
Aranyi et al. (1986) also conducted a similar set of experiments with 13 other chemicals
(acetaldehyde, acrolein, propylene oxide, chloroform, methyl chloroform, carbon tetrachloride,
allyl chloride, benzene, phenol, monochlorobenzene, benzyl chloride, perchloroethylene, and
ethylene trichloride). Perchloroethylene and ethylene trichloride were the only chemicals in this
group for which an increased mortality risk from streptococcal pneumonia was seen (mortality
risk 15.0 and 31.4% in controls and 50 ppm exposure groups for perchloroethylene and 13.4 and
58.1% in controls and 50 ppm exposure groups for ethylene trichloride). Decreased bactericidal
activity was also seen with acetaldehyde, acrolein, methyl chloroform, allyl chloride, benzene,
benzyl chloride, perchloroethylene, and ethylene trichloride at one or more exposures. Results
from several chemicals suggest that 5 days of exposure results in greater decrease in bactericidal
activity (i.e., acetaldehyde, acrolein, and benzene), and others (e.g., perchloroethylene) suggest
that 5 days of exposure does not result in greater suppression than a single exposure period.
There was considerable variation in both measures of immune response among the
controls in the experiments (Aranyi et al., 1986). Among the controls in the experiments with
the 13 chemicals other than dichloromethane, mortality in the streptococcal infectivity model
ranged from 5.7-22.1%, with a mean of 12.7%.5 Bactericidal activity in the klebsiella model
among controls ranged from 67.9-94.7%), with a mean of 81.8%>. The number of bacteria
deposited in the lung in an inhalation bacterial infectivity model can show considerable variation,
(i.e., between 750 to 1,500 viable streptococcus or klebsiella organisms, [Ehrlich, 1980]).
Therefore, concurrent controls are particularly import due to the variation in preparation and
aerosol administration of the bacteria in these assays.
Warbrick et al. (2003) evaluated immunocompetence in male and female Sprague-
Dawley rats by measuring the immunoglobulin M (IgM) antibody responses following
immunization with sheep red blood cells in addition to hematological parameters and
histopathology of the spleen, thymus, lungs, and liver. Groups of rats (8/sex/dose level) were
exposed to 0 or 5,000 ppm dichloromethane 6 hours/day, 5 days/week for 28 days. Rats injected
with cyclophosphamide served as positive controls. Five days before sacrifice (day 23 of
exposure) all rats were injected with sheep red blood cells. IgM levels in response to the sheep
red blood cells were comparable between dichloromethane-exposed and air-exposed rats,
indicating that dichloromethane did not produce immunosuppression in the animals under these
exposure conditions. Cyclophosphamide-treated animals had significantly lower levels of IgM
in the blood serum, indicating immunosuppression. Rats exposed to dichloromethane showed
reduced response to sound, piloerection, and hunched posture during exposures. Neither BW
gain nor the hematological parameters monitored were significantly affected by exposure to
5EPA did not include the duplicate assay of perchloroethylene in calculating this summary statistic. If this
additional assay is included, the mortality risk ranges from 5.7-45.7%, with a mean of 15.0%.
134

-------
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
dichloromethane. Relative and absolute liver weights were significantly increased in females,
but not in males. Relative spleen weight was reduced in females, and no significant changes
were seen in the weight of the thymus and lungs. Histopathology of the tissues examined was
unremarkable. Exposure to 5,000 ppm dichloromethane did not affect antibody production to the
challenge with sheep red blood cells.
In the 2-year drinking water study (Serota et al., 1986a, b) and 2-year inhalation study
(Nitschke et al., 1988a), histopathologic analyses were conducted on the lymph nodes, thymus,
and spleen among several other organs, and no significant changes were noted.
In summary, one study (Aranyi et al., 1986) demonstrated evidence of
immunosuppression, including increased risk of streptococcal-pneumonia-related mortality and
decreased clearance of klebsiella bacteria following a single dichloromethane exposure at
100 ppm for 3 hours in CD-I mice. The streptococcal and klebsiella bacterial inhalation assays
are models of respiratory infection that test for local immune effects associated with inhalation
exposure rather than systemic immunosuppression. The NOAEL identified in this study was
50 ppm. In contrast, in a functional immune assay of systemic immunosuppression conducted in
rats, Warbrick et al. (2003) did not observe changes in the antibody response to sheep red blood
cells in a 28-day inhalation exposure to 5,000 ppm dichloromethane. Histopathologic analyses
of immune system organs in chronic exposure studies for B6C3Fi mice and F344 rats (Nitschke
et al., 1988a; Serota et al., 1986a, b) revealed no changes from controls. However, no assays of
functional immunity were included in these chronic studies. The limited database for
dichloromethane does not suggest systemic immunosuppression, but the Aranyi et al. (1986)
study provides evidence of route-specific local immunosuppression from acute exposure studies
in CD1 mice. Due to the acute exposure duration used in Aryani et al. (1986), the immune
effects of short-term or chronic exposure to dichloromethane are unclear.
4.4.3. Neurotoxicology Studies in Animals
Neurological evaluations in animals during and after exposure to dichloromethane have
resulted in CNS depressant effects similar to other chlorinated solvents (e.g., trichloroethylene,
perchloroethylene) and ethanol. Overall, there are decreased motor activity, impaired memory,
and changes in responses to sensory stimuli. Neurobehavioral, neurophysiological, and
neurochemical/ neuropathological studies have been used to characterize the effects of
dichloromethane on the CNS. A brief overview of these types of studies is provided below,
followed by a detailed description of individual studies.
Neurobehavioral studies with dichloromethane used protocols to measure changes in
spontaneous motor activity, a functional observational battery (FOB) test (to evaluate gross
neurobehavioral deficits), and a task developed to assess learning and memory. The FOB
protocol includes various autonomic parameters, neuromuscular parameters, sensorimotor
parameters, excitability measures, and activity. Learning and memory changes with
135

-------
4567
4568
4569
4570
4571
4572
4573
4574
4575
4576
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
dichloromethane were studied by using a passive avoidance task. The oral and inhalation studies
that examined neurobehavioral endpoints are summarized in Table 4-26.
Neurophysiological studies with dichloromethane exposure consisted of measuring
evoked responses in response to sensory stimuli. In these studies, animals were implanted with
electrodes over the brain region that responds to the particular stimuli. For example, an electrode
would be implanted over the visual cortex in an animal presented with a visual stimulus. Once
the stimulus is presented to the animal, an evoked response is elicited from the brain region and
transmitted to the implanted electrode. During administration of a chemical, if there is a
significant change in the magnitude, shape, and latency (among other measures) in the evoked
response, then the chemical is considered to produce neurological effects. A summary of studies
examining dichloromethane exposure and neurophysiological changes is shown in Table 4-27.
In neurochemical/neuropathological studies with dichloromethane, animals were first
exposed to dichloromethane (via oral, inhalation, or injection), and then the brains were
removed. Changes in excitatory neurotransmitters, such as glutamate and acetylcholine, and the
inhibitory neurotransmitter, GABA, were measured. Additionally, dopamine and serotonin
levels, which are associated with addiction and mood, were also measured. Other parameters
that were measured included DNA/protein content and regional brain changes in the cerebellum
and hippocampus. Measurement of neurochemical changes provides mechanistic information,
and neurobehavioral and neurophysiological effects can be correlated to these results. Table
4-28 summarizes studies of neurochemical changes and dichloromethane.
136

-------
Table 4-26. Studies of neurobehavioral changes from dichloromethane, by route of exposure and type of effect
Species
Exposure(s)
Duration
Neurobehavioral effect
Reference


Oral and gavage exposure



Functional observational battery

F344 rat, female
101, 337, 1,012,
Acute—evaluated 4 and 24
FOB neuromuscular and sensorimotor
Moser et al. (1995)

1,889 mg/kg, gavage
hours after dosing
parameters significantly different from
controls at 1,012 and 1,889 mg/kg
(337 mg/kg = NOAEL)

F344 rat, female
34, 101,337, 1,012
14 day—evaluated on days 4, 9,
All FOB parameters (except activity)
Moser et al. (1995)

mg/kg-day, gavage
and 15.
significantly affected from day 4 at doses
of 337 and 1,012 mg/kg-day



Inhalation exposure



Spontaneous activity

NMRI mouse, male
400-2,500 ppm
1 hour
Initial increase in activity followed by a
pronounced decrease at exposures
600 ppm and higher
Kjellstrand et al. (1985)
Rat, male
5,000 ppm
1 hour, every other day for 10
days
Decreased spontaneous locomotor
activity
Heppel and Neal (1944)
Wistar rat, male
500 ppm
6 hours/day, 6 days
Increased preening frequency
Savolainen et al. (1977)
ICR mouse, female
5,000 ppm
Continuous, 7 days
Increased spontaneous activity in first
few hours and then decreased activity
Weinstein et al. (1972)
Sprague-Dawley rat, male
1,000, 5,000 ppm
Continuous, 14 weeks
No neurobehavioral changes
Haun et al. (1971)
ICR mouse, female
1,000, 5,000 ppm
Continuous, 14 weeks
Incoordination, lethargy
Haun et al. (1971)
Beagle dog, female
1,000, 5,000 ppm
Continuous, 14 weeks
Incoordination, lethargy
Haun et al. (1971)
Rhesus monkey, female
1,000, 5,000 ppm
Continuous, 14 weeks
Incoordination, lethargy
Haun et al. (1971)
ICR mouse, female
25, 100 ppm
Continuous, 14 weeks
Increased spontaneous activity at 25 ppm
Thomas et al. (1972)


Functional observational battery

F344 rat, male and female
50, 200, 2,000 ppm
6 hours/day, 5 days/week, 13
weeks + 65 hours exposure free
No effects observed on FOB, grip
strength
Mattsson et al. (1990)


Learning and memory

Swiss-Webster mouse,
47,000 ppm
Approximately 20 seconds +
Significant decrease in learning and
Alexeef and Kilgore (1983)
male

1 hour exposure free before
recall ability

training; retested at days 1, 2,
and 4
4587
4588
137

-------
4589
Table 4-27. Studies of neurophysiological changes as measured by evoked potentials resulting from
dichloromethane, by route of exposure
Species
Exposure(s)
Duration
SEPsa measured
Effect
Reference
Long-Evans rat, male
F344 rat, male
57.5, 115,230,
460 mg/kg, i.p."
5,000, 10,000,
15,000 ppm
F344 rat, male and female 50, 200,
2,000 ppm
Intraperitoneal
Acute; tested at	FEPa
15 minutes, 1 hour, and
5 hours after dosing
Inhalation Exposure
Acute, 1 hour; tested Electroencephalogram,
during exposure	BAERa, CAEPa, FEP,
SEP
Subchronic, 6 hour/day, FEP, CAEP, BAER,
5 days/week, 13 weeks; SEP
tested 65 hours after
last exposure
Significant changes in FEPs were noted Herr and Boyes
in animals dosed 115 mg/kg and higher; (1997)
FEP changes time and dose dependent
Significant changes in SEP, FEP,
BAER, and CAEP responses at all
exposures; slight recovery noted at
1 hour after exposure
No significant changes noted in any
evoked potential measurements
Rebert et al. (1989)
Mattsson et al.
(1990)
4590
4591
4592
4593
"SEP = somatosensory-evoked potential; FEP = flash-evoked potential; BAER = brainstem-auditory-evoked response; CAEP = cortical-auditory-evoked
potential; i.p. = intraperitoneal.
138

-------
Table 4-28. Studies of neurochemical changes from dichloromethane, by route of exposure
Species and sex
Exposure
Duration
Regions
Effect3
Reference
Sprague-Dawley rat, male 534 mg/kg
Wistar rat, male
Wistar rat, male
Wistar rat, male
Wistar rat, male
1,000 ppm TWA (basal
exposure of 100 ppm +
2,800 ppm, 1 hour peak
exposures at hours 1 and 4)
1,000 ppm TWA
1,000 ppm
1,000 ppm
Sprague-Dawley rat, male 70, 300, 1,000 ppm
Mongolian gerbil, male and 210, 350 ppm
female
Mongolian gerbil, male and 210 ppm
female
Mongolian gerbil, male and 210 ppm
female
Oral exposure
Acute, single dose; evaluated Hippocampus,
2 hours after dosed	medulla, midbrain,
hypothalamus
Inhalation exposure
6 hours/day, 5 days/week, Cerebrum,
2 weeks	cerebellum
6 hours/day, 5 days/week,	Cerebrum,
2 weeks + 7 days exposure	cerebellum
free
6 hours/day, 5 days/week,	Cerebrum,
2 weeks	cerebellum
6 hours/day, 5 days/week,
2 weeks + 7 days exposure
free
6 hours/day, 3 days
Continuous (24 hours/day),
3 months + 4 months
exposure free
Continuous (24 hours/day),
3 months
Continuous (24 hours/day),
3 months + 4 months
exposure free
Cerebrum
Caudate nucleus-
medial
Hippocampus,
cerebellum
cerebral cortex
Frontal cortex,
cerebellum
Hippocampus,
olfactory bulbs,
cerebral cortex
t acetylcholine in hippocampus	Kanada et al.
t dopamine and serotonin in medulla (1994)
I norepinephrine in midbrain
I norepinephrine and serotonin in
hypothalamus
t NADPH diaphorase, succinate	Savolainen et al.
dehydrogenase in cerebrum	(1981)
t cerebral RNA
I succinate dehydrogenase in
cerebellum
I succinate dehydrogenase in both Savolainen et al.
regions	(1981)
t acid proteinase	Savolainen et al.
I succinate dehydrogenase in	(1981)
cerebellum
I cerebral RNA	Savolainen et al.
(1981)
t catecholamine levels (70 ppm)	Fuxe et al. (1984)
I catecholamine levels (300 and 1,000
ppm)
No effect on luteinizing hormone
release
I DNA concentration per wet weight in Rosengren et al.
hippocampus (210, 350 ppm) and (1986)
cerebellar hemispheres (350 ppm)
t astroglial proteins in frontal and
sensory motor cerebral cortex
I glutamate, GABA,	Briving et al.
phosphoethanolamine in frontal cortex (1986)
t glutamate, GABA in posterior
cerebellar vermis
I DNA concentration per wet weight in Karlsson et al.
hippocampus only	(1987)
4594
4595
111 = increase; j = decrease.
139

-------
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4.4.3.1. Neurotoxicology Studies—Oral Exposures
Three studies evaluated the neurotoxic potential of dichloromethane by either
administering the solvent orally or by injection; two of these studies (Herr and Boyes, 1997;
Kanada et al., 1994) only evaluated acute effects (2-5 hours) from single-dose exposures.
Observed neurological effects included decreased spontaneous activity (Moser et al., 1995),
changes in flash-evoked potential (FEP) measurements (Herr and Boyes, 1997), and changes in
catecholamine levels in the brain (Kanada et al., 1994).
Moser et al. (1995) conducted neurobehavioral evaluations in female F344 rats following
an acute or 14-day oral administration of dichloromethane. A FOB protocol was utilized to
determine changes in autonomic parameters (lacrimation, salivation, pupil response, urination,
defecation), neuromuscular parameters (gait, righting reflex, forelimb and hind-limb grip
strength, landing foot splay), sensorimotor parameters (tail pinch, click response, touch
response), excitability measures (handling reactivity, arousal, clonic, and/or tonic movements),
and activity (rearing, motor activity). A baseline FOB was performed on all rats prior to initial
dichloromethane administration. After dichloromethane administration, a FOB was conducted at
selected time points followed by a motor activity test in a maze. In the acute study, rats were
dosed with 0, 101, 337, 1,012, or 1,889 mg/kg dichloromethane. At 4 and 24 hours after the
administered dose, rats were tested for the neurological parameters. Significant changes in the
neuromuscular and sensorimotor parameters were observed and occurred mostly in rats
administered with the highest dose. These significant changes were only observed at the 4-hour
time point and not when measured at 24 hours. The NOAEL identified by the authors for this
study was 337 mg/kg, based on no observable changes in the FOB. In the 14-day study, rats
were administered 0, 34, 101, 337, or 1,012 mg/kg-day. FOB testing was conducted on days 4
and 9 (before the daily dose) and approximately 24 hours after the last (14th) dose. With the
exception of the activity measurements, all other neurobehavioral parameters (neuromuscular,
sensorimotor, autonomic, excitability) were significantly affected from the 4th day through the
entire 14-day exposure cycle. The NOAEL identified for the 14-day study was 101 mg/kg-day,
based on FOB changes associated with the dichloromethane exposure.
A single dose acute neurophysiology study by Herr and Boyes (1997) evaluated the effect
of dichloromethane on FEPs in adult male Long-Evans rats. Rats were implanted with epidural
electrodes over the visual cortex area. After placement in an enclosed rectangular mirror
chamber, PEPs were stimulated with a 10 [j.sec flash. Baseline FEPs were collected and rats
were injected intraperitoneally with 0 (corn oil, n = 16), 57.5 (n = 15), 115 (n = 15), 230 (n = 14),
or 460 (n = 15) mg/kg dichloromethane. Animals were retested at 15 minutes, 1 hour, and
5 hours after injection. Amplitude decreases in the early FEP components were observed. The
FEP amplitude changes were time and dose dependent with maximal effects at 15 minutes after
dichloromethane dosage. All of the waveform amplitudes returned to control levels when
measured at the 1-hour time point for all doses tested. Response latencies were still different
140

-------
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
4665
4666
4667
4668
4669
from controls when measured 5 hours after dosing, but the effect was less pronounced than at the
15-minute and 1-hour time points. In this study, 57.5 mg/kg did not produce any significant
changes in the FEP measures as compared to control and was considered this study's NOAEL.
The LOAEL was 115 mg/kg based on changes in the FEP amplitudes.
Kanada et al. (1994) examined the effect of dichloromethane on acetylcholine and
catecholamines (dopamine, norepinephrine, serotonin) and their metabolites in the midbrain,
hypothalamus, hippocampus, and medulla from male Sprague-Dawley rats (4-5/group) in a
neurochemical/neuropathology study. The rats were sacrificed 2 hours after a single gavage dose
of 0 or 534 mg/kg of undiluted dichloromethane. Administration of dichloromethane
significantly increased the concentration of acetylcholine in the hippocampus by approximately
10% and increased dopamine and serotonin levels in the medulla by approximately 75%.
Dichloromethane decreased norepinephrine levels in the midbrain and hypothalamus by 12-
15%), and serotonin levels were decreased in the hypothalamus by approximately 30%>. There
was a trend toward decreased dopamine in the hypothalamus, but the variability between the
animals was so high that the effect was not significant. (These values for the percent changes
were estimated by EPA from the figures presented in the paper.) The authors speculated that
increased acetylcholine release associated with exposure to dichloromethane and other solvents
may originate from the nerve terminals.
4.4.3.2. Neurotoxicology Studies—Inhalational Exposure
The database pertaining to neurotoxic effects from inhalation exposure to
dichloromethane is considerably larger than the oral exposure database. Acute (less than 1 day)
and short-term (1-14 days) exposures resulted in an initial increase in spontaneous activity
followed by a decrease for exposures between 500 and 2,500 ppm (Kjellstrand et al., 1985;
Savolainen et al., 1977). Higher (5,000 ppm) acute and short-term exposures resulted in
decreased spontaneous activity and lethargy (Weinstein et al., 1972; Heppel andNeal, 1944).
Longer-term exposures (up to 14 weeks) produced decreased motor activity and lethargy in
several animals at 1,000 and 5,000 ppm (Haun et al., 1971), and exposures at 25 ppm for
14 weeks produced significant increases in activity in mice, starting at week 9. CNS depression
was evidenced by decreased responses in the auditory, visual, and somatosensory regions of the
brain in a study of sensory-evoked potential effects in 12 adult male F344 rats exposed to 0,
5,000, 10,000, and 15,000 ppm for 1 hour periods (Rebert et al., 1989). Altered learning and
memory abilities were demonstrated in young (3-, 5-, and 8-week-old) male Swiss-Webster mice
exposed to 168 mg/L (-47,000 ppm) dichloromethane for approximately 20 seconds (until there
was a loss of the righting reflex) (Alexeef and Kilgore, 1983).
141

-------
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4.4.3.2.1. Inhalational exposure—neurobehavioral studies.
Spontaneous motor activity—acute and short-term studies
Heppel and Neal (1944) evaluated the neurological effects of 5,000 ppm dichloromethane
in five male rats by measuring changes in spontaneous activity during and after exposure. The
five rats were not randomly selected, since the investigators chose to pick out the most active
animals in the litter. During the 1-hour testing runs, rats were placed in a rotating drum.
Spontaneous activity was reported as the number of drum revolutions/hour. Twenty control test
runs (1 run/day) were conducted prior to dichloromethane exposure runs. After the pre-exposure
period, rats were exposed to 5,000 ppm dichloromethane every other day for 1 hour, and activity
was measured in the same manner as in the control runs. Once dichloromethane exposure was
stopped, the animals were allowed to recover for 30 minutes and a second 1-hour test run was
performed to evaluate spontaneous activity during recovery. On nonexposure days, spontaneous
activity was also measured inl-hour intervals to compare to the pre-exposure period. A total of
five dichloromethane exposures, five postexposure, and five nonexposure trials were conducted
over 10 days. Spontaneous activity significantly declined (p < 0.01, Fisher's t-test) during
exposure to 5,000 ppm dichloromethane in comparison to nonexposure days. The average
number of revolutions for all five rats over the test runs was 576 on nonexposure days and 59
revolutions during dichloromethane exposure.
Weinstein et al. (1972) continuously exposed female ICR mice to 5,000 ppm
dichloromethane for up to 7 days. Clinical behavioral observations of the mice were made
during dichloromethane exposure. Within the first few hours of exposure, spontaneous activity
increased in comparison to control animals. After 24 hours of continuous exposure, there was a
considerable decrease in spontaneous activity as noted by observation only. The mice also
appeared to be very lethargic and had a hunched posture and a rough hair coat, which are all
signs of CNS depressive effects in rodents. These effects became progressively worse until after
96 hours of exposure, where many mice resumed normal activity. After the 7-day exposure,
mice were nearly as active as the control animals but had a rougher coat and were judged to be
emaciated and dehydrated.
Male Wistar rats exposed to 500 ppm dichloromethane 6 hours/day for 6 days exhibited
an increase in preening frequency and time 1 hour after the last exposure relative to controls
(Savolainen et al., 1977). However, there were no significant changes in other types of
spontaneous activity.
In the study by Kjellstrand et al. (1985), male NMRI mice were exposed to
dichloromethane concentrations ranging from 400 to 2,500 ppm. At concentrations of 600 ppm
and higher, exposures for 1 hour produced a biphasic pattern of activity characterized by an
initial increase in activity (as high as 200% of preexposure motor activity at 2,200 ppm, as
estimated from Figure 6 in Kjellstrand et al. [1985]) during exposure followed by a decreased
that reached the lowest point 1-2 hours after the end of exposure (as low as 40% motor activity
142

-------
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
at 2,200 ppm, in comparison to preexposure, as estimated from Figure 6 in Kj ell strand et al.
[1985]). Motor activity returned to normal levels after the decreased activity observed 1-2 hours
after exposure was stopped and indicated that the effect was reversible in this study design.
Spontaneous motor activity—subchronic (14 week) studies
Haun et al. (1971) reported results from studies in which female beagle dogs, female
rhesus monkeys, male Sprague-Dawley rats, and female ICR mice were continuously exposed to
0, 1,000, or 5,000 ppm dichloromethane for up to 14 weeks in whole-body exposure chambers.
Gross and histopathologic examinations were made on animals that died or were sacrificed
during or at termination of the study. At 5,000 ppm, obvious nervous system effects (e.g.,
incoordination, lethargy) were most apparent in dogs and also observed in monkeys and mice.
Rats did not demonstrate any of these sedative effects. At 1,000 ppm, these effects were
observed to a lesser extent in monkeys and mice, but dogs still displayed prominent CNS
depressive behavior. Histopathologic analysis revealed edema of the brain in three dogs that
died during exposure to 5,000 ppm dichloromethane. No other gross brain-related changes were
reported. The results indicate that continuous exposure to 1,000 ppm was an adverse effect level
for mortality and effects on the nervous system and liver in dogs (exposed for up to 4 weeks) and
for BW changes in rats (exposed for 14 weeks). The 5,000 ppm level induced mortality in
beagle dogs, ICR mice, and rhesus monkeys (but not Sprague-Dawley rats); obvious nervous
system effects in dogs, mice, monkeys, and rats; and gross liver changes in dogs, mice, monkeys,
and rats.
In the study by Thomas et al. (1972), female ICR mice were exposed continuously to 0,
25, or 100 ppm dichloromethane for 14 weeks. Spontaneous activity of mice was evaluated by
using closed circuit television for monitoring. Mice were evaluated in daily 2-hour testing
sessions. The 25 and 100 ppm exposure groups were tested for 2 weeks prior to the onset of
dichloromethane exposure. Starting at week 9, mice exposed to 25 ppm dichloromethane
exhibited increases in spontaneous activity, but no quantitative measurements or statistical
analysis were reported. The authors stated that no significant effect was observed in the group
exposed to 100 ppm.
FOB—subchronic (13 week) study
Only one study, a 13-week inhalation study in F344 rats (Mattsson et al., 1990) has
conducted an FOB testing paradigm following a subchronic exposure to dichloromethane.
Groups of rats (12/sex/exposure level) were exposed to 0, 50, 200, or 2,000 ppm
dichloromethane 6 hours/day, 5 days/week for 13 weeks. An additional group of rats was
exposed to 135 ppm CO to induce approximately 10% COHb, approximately the level produced
by saturation of oxidative metabolism of dichloromethane. After the 13 weeks of exposure
(beginning 65 hours after the last exposure), rats were subject to an FOB to evaluate any
143

-------
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
neurobehavioral changes from the dichloromethane exposure. Autonomic parameters were first
characterized and then the rat was placed in a clear plastic box to evaluate locomotor activity and
then responsiveness to touch, sharp noise, and tail pinch. Hind-limb grip strength was also
measured by using a strain gauge. All animals were examined clinically at weekly intervals and
were tested at the end of the exposure period by FOB, grip strength, BW, temperature, and
sensory-evoked potentials. No exposure-related effects were observed on the FOB, grip
strength, or sensory-evoked potentials. No histopathologic changes were noted in brains, spinal
cords, or peripheral nerves from the high-dose dichloromethane group compared with control
animals. In the absence of changes, lower concentrations were not examined.
Learning and memory—acute study
In a study by Alexeef and Kilgore (1983), a learning and memory evaluation was
conducted following acute exposure to dichloromethane. Mice were exposed to 168 mg/L
(-47,000 ppm) dichloromethane and were tested for learning ability by using a passive-
avoidance conditioning task. Male Swiss-Webster mice (3, 5, and 8 weeks old) were used in this
study. In the passive avoidance task, mice were placed on a metal platform that extended into a
hole. If the mouse went into the hole (a darkened area, which would be the preferred area for the
mouse), it received a foot shock. Prior to the training session, mice were exposed to either air or
-47,000 ppm dichloromethane. Animals were exposed to dichloromethane until there was a loss
of the righting reflex, which would take about 20 seconds on average, and then placed back in
their home cage. One hour after exposure, animals were trained to learn the passive avoidance
task. A mouse was considered to have learned the task once it remained on the platform for at
least 30 seconds without entering the hole. Mice were then tested for recollection of the task at
either 1, 2, or 4 days after the initial training session. In the learning phase of the task, 74% of
the control mice retained the task in comparison to 59% of the dichloromethane-exposed group,
indicating the significant effect of dichloromethane on learning. There was also an age-related
effect since exposed 3-week-old mice were less likely to recall the task than five- or eight-week-
old mice. There was no difference in task recall between the 5- and 8-week-old mice.
Dichloromethane, at the exposure used in the study, was demonstrated to be non-analgesic, since
pain-response times were comparable to those in air-exposed animals in the hot-plate pain test,
and therefore the results of the passive avoidance test were not confounded by potential analgesic
effects. As a result, it is demonstrated that exposure to an acute and high concentration of
dichloromethane alters learning ability in mice.
4.4.3.2.2. Inhalational exposure—neurophysiological studies. The effect of dichloromethane
on sensory stimuli was evaluated by measuring sensory-evoked responses during an acute
exposure (Rebert et al., 1989) and following a subchronic (13 week) exposure (Mattsson et al.,
1990). Rebert et al. (1989) evaluated the effects of dichloromethane on sensory-evoked
144

-------
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
potentials (auditory, visual, and somatosensory) in F344 rats exposed to 0, 5,000, 10,000, and
15,000 ppm dichloromethane for 1 hour in a head-only exposure chamber. Twelve adult male
rats were implanted with chronic epidural electrodes placed over the visual and somatosensory
cortices. Each rat served as its own control, with a 1-week recovery period between testing
sessions. During each testing session, spontaneous electroencephalograms were recorded.
Additionally, brainstem-auditory-evoked responses (BAERs) (tone stimulus), cortical-auditory-
evoked potentials (CAEPs) (click stimulus), FEPs (flash stimulus), and somatosensory-evoked
potentials (SEPs) (tail current stimulus) were measured in response to the stimuli.
Dichloromethane decreased the SEP response to the tail current stimulus, and earlier components
of the FEP response were attenuated and eventually eliminated with increasing exposures. The
BAER response profile was also significantly altered. Dichloromethane completely abolished
the CAEP at all concentrations tested. Slight recovery of this response was noted approximately
1 hour after exposure. The collective results strongly suggest a CNS depressive profile for
dichloromethane and indicate that this chemical affects the auditory, visual, and somatosensory
regions of the brain.
In a subchronic exposure study, male and female F344 rats were exposed to
dichloromethane for 6 hours/day, 5 days/week for 13 weeks (Mattsson et al., 1990). Twelve
animals of each sex were selected for exposure to 0, 50, 200, or 2,000 ppm dichloromethane or
135 ppm CO. For electrophysiological measures, rats were surgically implanted with epidural
electrodes 10 weeks after the onset of exposure. Electrodes were placed over the somatosensory,
visual, and cerebellar region. Electrophysiological measures that were recorded included FEP
measurements, cortical flick fusion responses, CAEPs, BAERs, and SEPS recorded from the
sensory (SEP-S) and cerebellar (SEP-C) regions. None of these measures were significantly
altered by any dichloromethane or CO treatment in this study. However, it should be noted that
all of the electrophysiological measures were conducted at least 65 hours after the last
dichloromethane exposure. As a result, it can be concluded that a subchronic exposure to
dichloromethane did not result in persistent changes in any of the neurophysiological measures
that were evaluated in this study. It is not known if any neurological compensation occurred
since SEP measurements were not taken during actual dichloromethane exposure in this
subchronic study.
Based on these two studies, the significant changes noted in several SEP measures during
dichloromethane exposure were not observed after a subchronic exposure where animals were
tested at least 65 hours after the last exposure. As a result, it is difficult to ascertain if tolerance
is developed to the dichloromethane-mediated changes in sensory potentials during an acute
exposure or if these effects are still maintained during repeated exposure, since measurements
were not taken during the subchronic exposure.
145

-------
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4.4.3.2.3. Inhalational exposure—neurochemistry and neuropathology studies. The studies
evaluating specific neurochemical changes in relation to dichloromethane exposure include
studies of effects of short-term (3-day to 2-week) exposures (Fuxe et al., 1984; Savolainen et al.,
1981) and subchronic (3-month) exposures (Karlsson et al., 1987; Briving et al., 1986;
Rosengren et al., 1986).
Savolainen et al. (1981) examined three different exposure schemes in male Wistar rats.
The rats were exposed to 500, 1,000, or 1,000 ppm TWA dichloromethane for 6 hours/day,
5 days/week for 2 weeks. (Note: The abstract of this paper describes the exposures as 500, 1,000,
and 100 ppm TWA, but, based on information in the body of the paper, the abstract appears to be
incorrect.) The 1,000 ppm TWA exposure consisted of a basal 100 ppm exposure with two
2,800 ppm 1-hour peak concentrations (at 1 and 4 hours) resulting in a time-weighted exposure
of 1,000 ppm. Brains were removed from rats at the end of study and analyzed. The 1,000 ppm
TWA group displayed increases in cerebral RNA. Other changes noted for this group in the
cerebrum included significant increases in NADPH diaphorase and succinate dehydrogenase
activity. In the 1,000 ppm constant exposure group, acid proteinase activity was below the levels
observed in control animals in the first week but increased to levels above control animals in the
second week. In the cerebellum, there were no changes in RNA concentration, and there was a
decrease in succinate dehydrogenase activity in both the 1,000 and 1,000 ppm TWA groups.
After a 7-day withdrawal, RNA levels in the cerebrum were significantly lower in the 1,000 ppm
group. Succinate dehydrogenase levels remained lowered in the 1,000 ppm TWA group after the
7-day exposure-free period. No significant effects were seen at 500 ppm.
Fuxe et al. (1984) evaluated changes in brain catecholamine levels after a 3-day exposure
to dichloromethane, using male Sprague-Dawley rats. Rats were exposed to 70, 300, and
1,000 ppm dichloromethane 6 hours/day for 3 consecutive days. Additional groups of rats were
exposed to the same levels of dichloromethane and given intraperitoneal injections of the
tyrosine hydroxylase inhibitor, a-methy 1 -dl-/>tyrosine methyl ester (H44/68), 2 hours prior to
sacrifice. Brains were removed, stained, and evaluated for catecholamine changes 16-18 hours
after the last exposure. Catecholamine levels were measured in the hypothalamus, frontal cortex,
and caudate nucleus among other brain regions. At all exposures, there was a significant
decrease by approximately 10-15% of catecholamine concentrations in the posterior
periventricular region of the hypothalamus. In the medial part of the caudate nucleus, which is
involved in memory processes, catecholamine levels were significantly higher (12%) in the
70 ppm group but significantly lower in the 300 ppm (1%) and 1,000 ppm (8%) groups
compared with controls. The impact of dichloromethane was also evaluated on the
hypothalamic-pituitary gonadal axis. The hypothalamus regulates secretion of reproductive
hormones, such as follicle-stimulating hormone and luteinizing hormone. The levels of the
hormone release were not significantly changed with dichloromethane exposure. However,
when rats were dosed concurrently with H44/68 and dichloromethane, statistically significant
146

-------
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
4891
4892
4893
4894
4895
4896
inversely dose-related increases in luteinizing hormone levels were observed (330, 233, and
172% higher than controls in the 70, 300, and 1,000 ppm groups, respectively). The study
overall demonstrates significant changes in catecholamine levels in the hypothalamus and
caudate nucleus. No significant changes in catecholamine levels in the frontal cortex were
reported. Catecholamine level changes in the hypothalamus did not appear to significantly affect
hormone release; however, decreased catecholamine levels in the caudate nucleus at higher
exposures may lead to memory and learning impairment.
A series of studies were conducted in male and female Mongolian gerbils exposed
continuously to 210 ppm (Karlsson et al., 1987; Briving et al., 1986), 350 ppm, or 700 ppm
(Rosengren et al., 1986) dichloromethane for 3 months, followed by a 4-month exposure-free
period. High mortality rates occurred at 350 ppm (6/10 males and 3/10 females by 71 days) and
700 ppm (10/10 males and 9/10 females by 52 days). Rosengren et al. (1986) monitored two
astroglial proteins, S-100 and GFA, as well as DNA concentrations in the brain. Decreased
DNA concentrations were noted in the hippocampus at both the 210 and 350 ppm exposures. At
350 ppm, there was also decreased DNA concentration in the cerebellar hemispheres, indicating
a decreased cell density in these regions, probably due to cell loss. Increased astroglial proteins
were found in the frontal and sensory motor cerebral cortex, which directly correlated to the
astrogliosis that was observed in those areas. Up-regulation of these astroglial proteins is a good
indicator of neuronal injury (Rosengren et al., 1986).
Karlsson et al. (1987) measured DNA concentrations in different regions of the gerbil
brain. After the solvent-free exposure period, brains were removed and the olfactory bulbs and
cerebral cortices were dissected. Brain weights and weights of the dissected brain regions were
the same between control and dichloromethane-exposed animals. The total protein concentration
per wet weight was not significantly different between dichloromethane-exposed and control
animals. However, DNA concentrations per wet weight were significantly decreased in the
hippocampus after dichloromethane exposure. No other examined regions demonstrated
significant changes in DNA concentrations after dichloromethane exposure. This selective DNA
concentration decrease observed in the hippocampus is a sign of neurotoxicity and may possibly
explain why some studies have noted memory and learning deficits with dichloromethane
exposure. In a companion paper, in which only the 210 ppm level was tested, it was found that
exposure to dichloromethane decreased the levels of glutamate, y-aminobutyric acid, and
phosphoethanolamine in the frontal cortex, while glutamine and y-aminobutyric acid were
increased in the posterior cerebellar vermis (Briving et al., 1986). Increased levels of glutamate
in the posterior cerebellar vermis could reflect an activation of astrocytic glia, since glutamine
synthetase is localized exclusively in astrocytes. The gerbils did not have a solvent-free
exposure period as in the other two studies (Karlsson et al., 1987; Rosengren et al., 1986). The
exposure regime in these studies did not affect BW or brain weight. Furthermore, the
neurochemical changes observed in these studies were not attributed to formation of CO.
147

-------
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
Neurological changes have been investigated by measuring changes in neurotransmitter
levels and changes in neurotransmitter localization. Changes in catecholamine levels in the
caudate nucleus after an acute exposure (Fuxe et al., 1984) as well as decreased DNA content in
the hippocampus after a subchronic dichloromethane exposure (Rosengren et al., 1986) suggest
that memory functions are altered since both brain regions are associated with learning and
memory. The results from Fuxe et al. (1984) directly correlated with the finding that learning
and memory were impaired in mice after an acute (single) and very high exposure (47,000 ppm)
to dichloromethane (Alexeef and Kilgore, 1983). Additionally, changes in the hippocampus also
suggest memory effects after a long-term, continual exposure to dichloromethane, although no
conclusive evidence has been presented to date. In another subchronic, continuous exposure to
350 ppm dichloromethane for 3 months, decreased DNA concentration was observed in the
cerebellar hemispheres of Mongolian gerbils and is suggestive of cell loss (Rosengren et al.,
1986). However, in a 2-week exposure study in male Wistar rats, RNA changes were not noted
in the cerebellum, although enzyme activity was significantly decreased in this region (but was
increased in the cerebrum) (Savolainen et al., 1981). These results suggest that the cerebellum is
a target for dichloromethane. Noted neurobehavioral effects that may be linked to impaired
cerebellar function include changes in motor activity and impaired neuromuscular function
(Moser et al., 1995).
4.5. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MODE OF
ACTION
4.5.1. Genotoxicity Studies
4.5.1.1. In Vitro Genotoxicity Assays
Bacterial, Yeast, and Fungi mutagenicity assays
Numerous in vitro studies have demonstrated dichloromethane as being mutagenic in
bacterial assays, yeast, and fungi, and several studies provide evidence that the genotoxic action
of dichloromethane in bacterial systems is enhanced in the presence of GSH (e.g., Dillon et al.,
1992; Their et al., 1993; Oda et al., 1996; DeMarini et al., 1997; Pegram et al., 1997) (Table 4-
29). Considering the results are primarily dependent on the presence of GSH, activation likely
involves the GST-T1 metabolic pathway, which produces two proposed DNA-reactive
metabolites, S-(chloromethyl)glutathione and formaldehyde.
Dichloromethane induced mutations in Salmonella typhimurium strains containing GSH
(e.g. TA100, TA98). These effects were not markedly influenced by the addition of exogenous
mammalian liver fractions, suggesting that endogenous metabolism in these strains was
sufficient to activate dichloromethane (Green, 1983; Jongen et al., 1982; 1978; Gocke et al.,
1981). In support of this hypothesis, dichloromethane exposure of NG-11, a glutathione-
deficient variant of S. typhimurium strain TA100, produced twofold fewer base-pair mutations
148

-------
Table 4-29. Results from in vitro genotoxicity assays of dichloromethane with bacteria, yeast, or fungi
Results
Assay
Test system
Concentration(s)
Without metabolic
activation
-S9
With metabolic
activation
+S9
Reference
Bacteria
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
Reverse mutation
S. typhimurium
TA98a, TA1003
S. typhimurium
TA98, TA100
S. typhimurium
TA1535b, TA1537b,
TA1538b
S. typhimurium
TA100
S. typhimurium
TA100
S. typhimurium
TA100, TA1535,
TA19503,
E. coli WU3610893
S. typhimurium
TA100
S. typhimurium
TA100, NG54°
S. typhimurium
TA100, TA1535 and
TA1538 (+GSTA 1-1
and GSTP 1-1)
S. typhimurium
TA1535 (+GST 5-5),
TA1535 (wild type)
S. typhimurium
TA100, TA100/NG-lld
S. typhimurium TA100,
RSJIOO0
6-hour exposure to 0, 7,000,
and 14,000 ppm
Up to 3,600 |ig/platc
Up to 3,600 |ig/platc
6-hour exposure to 0, 7,000,
and 14,000 ppm
Up to 84,000 ppm, 3-day
exposure
10 (iL/plate
+ for TA100, TA1950,
WU361089
-forTA1535
2- and 6-hour exposures to 0,	+
2,500, 5,000, 7,500, and
10,000 ppm
0, 50, 100, and 200 (iL/plate + for TA100
- for TA1535, TA1538
0-2.0 mM/plate
0, 30, 60, 130 mM/plate
Up to 24,000 ppm
+ forTA1535 (+GST 5-5)
- for TA1535 (wild type)
++ for TA100
+ for TA100/NG-11
+ for TA100
+ for RSJIOO
++
++
Not determined
Not
determined
+
Not determined
Not determined
+ for TA100
+ for RSJIOO
Jongenetal. (1978)
Gocke et al. (1981)
Gocke et al. (1981)
Jongenetal. (1982)
Green (1983)
Osterman-Golkar et al.
(1983)
Zeiger (1990)
Dillon etal. (1992)
Not determined Simula et al. (1993)
Pegram et al. (1997); Thier
etal. (1993)
Graves et al. (1994a)
DeMarini et al. (1997)
149

-------
Table 4-29. Results from in vitro genotoxicity assays of dichloromethane with bacteria, yeast, or fungi
Results
Assay
Test system
Concentration(s)
Without metabolic
activation
-S9
With metabolic
activation
+S9
Reference
Forward mutation S. typhimurium BA13
Gene mutation
Prophage
induction
Reverse mutation
Forward mutation
Forward mutation
Mitotic
segregation
S. typhimurium
TA1535/pSK1002°
NM50040
E. coli K-39 (X)
E. coli WP2 uvra
pKMlOl
E. coli K12
E. coli Uvr+, UvrET
Aspergillus nidulans
Gene conversion Saccharomyces
and recombination cerevisiae
0-130 |imol/platc
0,2.5,5.0, 10, 20 mM
10 (iL/plate
2- and 6-hour exposures to
6,300, 12,500, 25,000, and
50,000 ppm
0, 30, 60, 130 mM/plate
20,000 ppm
+++
+ NM5004
- TA1535/ pSKlOO 2
Up to 8,000 ppm
Up to 209 mM
Fungi and yeasts
+ only at 4,000 ppm; no
dose-response relationship
established
Not determined
Not determined
Not determined
Not determined
Roldan-Aijona and Pueyo
(1993)
Odaetal. (1996)
Osterman-Golkar et al.
(1983)
Dillon etal. (1992)
Graves et al. (1994a)
Zielenska et al. (1993)
Crebelli et al. (1988)
Not determined Callen et al. (1980)
a bacterial strains that have GSH (e.g. TA100, TA 98)
b bacterial strains that do not have GSH (e.g. TA1535)
0 bacterial strains engineered to have more GSH activity than wild type
d bacterial strains engineered to have less GSH activity than wild type
150

-------
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
compared with exposure of strain TA100, which produces normal levels of GSH. Furthermore,
this difference was not apparent when the culture medium contained 1 mM
GSH (Graves et al., 1994a).
In contrast to strain TA100, S. typhimurium strains TA1535, TA1537, and TA1538
(strains deficient in GSH) did not develop base-pair mutations in response to dichloromethane
exposure (Gocke et al., 1981; Osterman-Golkar et al., 1983; Simula et al., 1993; Thier et al.,
1993; Pegram et al., 1997). However, when strain TA1535 was transfected with rat GST-T1,
dichloromethane induced base-pair reverse mutations (DeMarini et al., 1997; Pegram et al.,
1997; Thier et al., 1993). A 60-fold higher concentration of dichloromethane was needed to
induce a response (i.e., a sixfold increase over background levels in reverse mutations) in S.
typhimurium strain TA100 than in TA1535 transfected with rat GST-T1 (DeMarini et al., 1997).
This study also included several trihalomethanes; dichloromethane was several fold less
genotoxic than dibromochloromethane or bromoform, but was similar in potency to
bromodichloromethane (DeMarini et al., 1997; Pegram et al., 1997). The authors suggest that
these results support a role of GST-T1 in the mutagenicity of the trihalomethanes.
The mutagenic effects of dichloromethane have also been examined in fungi and yeasts
assays with both systems reporting positive results. Fungi assays were positive for mitotic
segregation in Asperigillus ridulans (Crebelli et al., 1988) but there was not a dose response
relationship as only the 4,000 ppm dichloromethane exposure was positive (exposure up to 8,000
ppm). A yeast assay was positive for gene conversion and recombination in Saccharomyces
cerevisiae for concentrations up to 209 mM (Callen et al., 1980).
Mammalian assays
In the in vitro mammalian system studies conducted with murine cell lines (Table 4-30),
dichloromethane was negative for producing point mutations in the mouse lymphoma L5178Y
cell line (Thilagar et al., 1984), but was positive in producing single stranded DNA breaks in
mouse Clara cells (Graves et al., 1995) and mouse hepatocytes (1994b). Given that exposure to
dichloromethane results specifically in lung and liver tumors, this pattern is not surprising.
Additionally, GST is localized in the nucleus of hepatocytes and lung cells in the mouse
(Mainwaring et al., 1996), which would also increase sensitivity of these particular cell fractions
to genotoxic effects of dichloromethane. DNA single strand breaks (SSBs) were induced at
lower concentrations in mouse hepatocytes (0.5 mM) than in rat hepatocytes (30 mM). The
extent of DNA damage was shown to be reduced to the background level seen in control (no
exposure) conditions by pretreating the cells with buthionine sulfoxime to deplete cellular levels
of GSH and thus inhibit dichloromethane metabolism via the GST pathway (Graves et al. 1995;
1994b). Similar results were seen in mouse lung Clara cells. Freshly isolated Clara cells from
the lungs of B6C3Fi mice also showed significantly increased, concentration-dependent amounts
of DNA SSBs when incubated in vitro for 2 hours in the presence of 5-60 mM dichloromethane.
151

-------
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
Pretreatment with buthionine sulphoximine before Clara-cell isolation or the presence of
buthionine sulphoximine in the culture medium decreased the amount of in vitro DNA damage
induced.
In a series of experiments with freshly isolated hepatocytes from multiple species (Table
4-30), DNA-protein cross-links were detected in hepatocytes of B6C3Fi mice but not in
hepatocytes of F344 rats, Syrian golden hamsters, or three human subjects, following 2-hour in
vitro exposure to concentrations ranging from 0.5-5 mM dichloromethane (Casanova et al.,
1997). Within the range of concentrations tested, DNA-protein cross-links in mouse hepatocytes
appeared to increase with increasing concentration of dichloromethane.
Negative results for dichloromethane were predominantly seen in in vitro test systems
that used rat or hamster cell lines with low or no GST activity (Table 4-30). Several genotoxic
endpoints, including DNA and protein synthesis (Garrett and Lewtas, 1983), chromosomal
aberrations or sister chromatid exchanges (Thilagar et al., 1984; Thilagar and Kumaroo, 1983;
Jongen et al., 1981), unscheduled DNA synthesis (Thilagar et al., 1984; Andrae and Wolff, 1983;
Jongen et al., 1981), and mutations (Thilagar et al., 1984; Jongen et al., 1981) were evaluated in
these cell lines. In contrast, positive results (DNA-protein cross links and DNA SSBs) were
observed when mouse liver cytosol was included in Chinese hamster ovary (CHO) cells (Graves
et al., 1994b; Graves et al., 1995). Dichloromethane also induced hypoxanthine-guanine
phosphoribosyl transferase (HPRT) gene mutations in CHO cells when they were incubated with
GST-competent mouse liver cytosol preparations (Graves et al., 1996).
The instability of the S-(chloromethyl)glutathione-adducts presents considerable
challenges to studies of these products (Hashmi et al., 1994). Kayser and Vuilleumier (2001),
however, demonstrated the formation of DNA adducts with radiolabeled dichloromethane in calf
thymus DNA in the presence of dichloromethane dehalogenase/GST purified from a bacterial
source (Methylophilus sp. strain DM11) and GSH (Table 4-30). The type of adduct could not be
identified because of low yield, but it was determined that guanine was more actively
incorporated than cytosine, adenine or thymine by at least 2 fold in the presence of GST-
activated dichloromethane, indicating a base specificity for these adducts. Incubation of calf
thymus DNA with formaldehyde and GSH, however, did not result in detectable DNA adduct
formation. In another study, Marsch et al. (2004) further evaluated the presence of adducts in
calf thymus DNA in the presence of dichloromethane and human (GSTT1-1), rat (GST 5-5) or
bacterial (DM11) GST (Marsch et al., 2004). This study found that all three enzymes yielded a
similar pattern of adduct formation, forming primarily with guanine and to a lesser extent with
cytosine, adenine, and thymine (2-3 fold less than guanine), consistent with the results reported
by Kayser and Vuilleumier (2001). High levels of guanosine-specific adducts were also seen
with S-(l-acetoxymethyl)glutathione, a compound that is structurally similar, but more stable,
than S-(chloromethyl)glutathione (Marsch et al., 2001). These findings indicate that the S-
(chloromethyl)glutathione intermediate formed by GSH conjugation has mutagenic potential and
152

-------
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
is likely responsible, at least in part, for the mutagenic response observed following
dichloromethane exposure.
In studies with human cell lines or isolated cells, positive results were reported for sister
chromatid exchanges and chromosomal aberrations (Thilagar et al., 1994) and in the
micronucleus test (Doherty et al., 1996). Negative results with human cells were seen in the
unscheduled DNA synthesis assays (Perocco and Prodi, 1981; Jongen et al., 1981), DNA SSBs,
and DNA-protein cross-links (Graves et al., 1995; Casanova et al., 1997).
Dichloromethane-induced DNA damage (comet assay) was examined in primary cultures
of human lung epithelial cells collected by brush biopsy from four healthy volunteers (Landi et
al., 2003). This study was designed to assess the genotoxicity of four thrihalomethanes
(chloroform, bromodichloromethane, dibromochloromethane and bromoform), with
dichloromethane included because of its known activation by GST-T1. Two of the subjects were
of the GST-T1+ genotype, and two were of the GST-T1 genotype.6 The cells had been frozen,
and GST activity was not detected in the cultured cells. DNA damage was reported to occur in
the combined GST-T1 samples (tail extent moment 7.1, 13.7 and 15.3 in the 10, 100 and 1000
|iM dichloromethane groups, respectively), but not in the combined GST-T1+ samples (tail
extent moment 8.1, 11.5 and 10.4 in the 10, 100 and 1000 |iM dichloromethane groups,
respectively). This pattern was not seen across the individual samples, however, as only one
sample exhibited a clear dose-response gradient. Given the absence of GST activity, an analysis
combining the four samples could provide a more informative picture of the dose-response
relation between dichloromethane (and the other compounds) studied and DNA damage. For
dichloromethane, values of 9.4, 7.6, 12.6, and 12.9 were seen in the 0, the 10, 100 and 1000 |iM
groups, respectively. This pattern was similar to that seen with chloroform (9.4, 6.9, 11.4, and
12.7 in the 0, the 10, 100 and 1000 |iM groups, respectively), but weaker than the pattern for
bromoform (9.4, 12.5, 15.8, and 18.2 in the 0, the 10, 100 and 1000 |iM groups, respectively),
and much weaker than for bromodichloromethane (9.4, 25.2, 28.5, and 39.1 in the 0, the 10, 100
and 1000 |iM groups, respectively).7 No dose-response gradient was seen with
dibromochloromethane (9.4, 6.5, 8.1 and 8.0 in the 0, the 10, 100 and 1000 |iM groups,
respectively). This relative pattern is also seen in the estimated slopes (beta coefficient for the
change in tail extent moment per unit increase in |iM concentration): 0.0, 0.003, 0.004., 0.006
and 0.02 for dibromochloromethane, dichloromethane, chloroform, bromoform, and
bromodichloromethane, respectively (statistical significance not reported).
6	Landi et al. (2003) did not clearly describe their treatment of GST-T1+" heterozygote genotypes; the EPA believes
it is most likely they were included in the pool from which the GST-T1+ samples were drawn. In addition, there is a
discrepancy in the paper regarding this the coding of the GST-T1 genotypes. Samples A and C are noted to be the
GST-Tr samples in one part of the paper, and C and D are described as the GST-T1 samples in another part of the
paper.
7	These values are based on the mean of the GST-T1+ and the GST-T1 samples from Table 1 of Landi et al. (2003)
153

-------
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
A stronger and more consistent response was seen under the same experimental
conditions with bromodichloromethane, but dibromochloromethane resulted in no increase in
DNA damage in any of the donor cells at any concentration tested.
Several studies have examined patterns of mutations or DNA damage with
dichloromethane and formaldehyde to assess the relative role of S-(chloromethyl)glutathione and
formaldehyde in the observed genotoxicity. In a study in CHO cells incubated with
dichoromethane (0.3% plus mouse liver cytosol), 2.5-fold increases in DNA-protein cross-links
that are indicative of formaldehyde exposure were observed, compared with a 25-fold increase
when 1 mM formaldehyde was added directly to cultures. Both treatments induced a comparable
degree of DNA SSBs (Graves and Green, 1996). In a subsequent study, Graves et al. (1996)
compared the mutational spectra induced by dichloromethane to that induced by direct addition
of formaldehyde or 1,2-dibromoethane (a chemical known to act through a glutathionyl
conjugate metabolite) at the HPRT locus in CHO cells. The mutations induced by
dichloromethane and 1,2-dibromoethane were predominantly GC to AT transitions, while all six
formaldehyde-induced mutants sequenced were single base transversions. This provided further
evidence that the S-(chloromethyl)glutathione intermediate may be primarily responsible for
dichloromethane genotoxicity. In contrast, Hu et al. (2006) found evidence of significant
amounts of formaldehyde formation following dichloromethane exposure in the cytosol of V79
(hamster) cells transfected with the murine GSTT1 gene compared to the parent cell line. In
accordance with this, they observed concentration-dependent increases in DNA-protein
crosslinks in the GSTT1 transfected cells using the comet assay with and without proteinase K
treatment that frees DNA from crosslinks and allows DNA migration. These findings are
consistent with those by Casanova et al. (1997), who performed a comparison of the amounts of
DNA-protein and RNA-formaldehyde crosslinks formed following dichloromethane exposure in
hepatocytes isolated from mice, rats, hamsters, and human GSTT1 genetic variants. Only DNA-
protein crosslinks were observed in mouse hepatocytes, but RNA-formaldehyde crosslinks were
found in all species, which were highest in the mouse hepatocytes, and were 4-, 7-, and 14-fold
higher than rats, humans, and hamsters. These results showed that human hepatocytes can
metabolize dichloromethane to formaldehyde, resulting in RNA-formaldehyde crosslinks. In
addition, the results indicate, that there is considerable variation among species, and that the
human variation in the GSTT1 gene can also affect the amount of formaldehyde produced. The
authors also noted that comparing results following ectopic addition of formaldehyde directly to
cells with results following dichloromethane metabolism in situ can be misleading, as the
formaldehyde produced internally may reside in different locations intracellularly, potentially
affecting the capability of interacting with DNA. These results show that, while most studies
indicate the importance of the S-(chloromethyl)glutathione intermediate in mediating genotoxic
damage following dichloromethane exposure, DNA damage resulting from formaldehyde
formation should also be considered.
154

-------
Table 4-30. Results from in vitro genotoxicity assays of dichloromethane with mammalian systems, by type
of test
Assay
Test system
Concentrations
Results
Reference
Point mutation
DNASSBsby
alkaline elution
DNASSBsby
alkaline elution
DNA-protein
cross-links
Unscheduled
DNA synthesis
Unscheduled
DNA synthesis
DNASSBsby
alkaline elution
DNA-protein
cross-links
DNA-protein
cross-links
Mouse lymphoma
L5178Y cells
Mouse hepatocytes
Mouse Clara cells
Mouse hepatocytes
Rat hepatocytes
Rat hepatocytes
Rat hepatocytes
Rat hepatocytes
Not provided
0, 0.4, 3.0, 5.5 mM
0, 5, 10, 30, 60 mM
0.5-5 mM
Up to 16 mM
(measured); 30 mM
(nominal)
Not provided
0, 30, 90, 90 mM
0.5-5 mM
Mouse
Negative
Positive at 0.4 mM
Thilagar et al. (1984)
Graves et al. (1994b)
Positive, but DNA damage was reduced by incubating Graves et al. (1995)
in the presence of GSH depletory
Positive
Rat
Negative
Marginally positive
Positive at 30 mM
Negative
Casanova et al. (1997)
Andrae and Wolff (1983)
Thilagar et al. (1984)
Graves et al. (1994b)
Casanova et al. (1997)
Hamster with GST activity from mouse
Chinese hamster ovary 60 mM
cells
HPRTa mutation Chinese hamster ovary 2,500 ppm
analysis	cells
DNASSBsand Chinese hamster ovary 3,000 ppm (0.3%,
DNA-protein cells	volume per volume
cross-links	[v/v]) and 5,000 ppm
(0.5%, v/v)
Positive with mouse liver cytosol (negative without) Graves et al. (1994b)
at much higher concentrations of dichloromethane (60
mM) than formaldehyde (0.5-4 mM).
Positive with mouse liver cytosol	Graves et al. (1996)
Positive at concentration of 0.5% (v/v) for SSBs in Graves and Green (1996)
presence of mouse liver cytosol, but increase in DNA-
protein cross-links marginal; formaldehyde (in
absence of mouse liver cytosol) was positive at 0.5
mM for both DNA SSBs and DNA-protein cross-
links; Chinese hamster ovary cell cultures were
suspended.
155

-------
Table 4-30. Results from in vitro genotoxicity assays of dichloromethane with mammalian systems, by type
of test
Assay
Test system
Concentrations
Results
Reference
DNA-protein
cross-links
Comet Assay
Syrian golden hamster 0.5-5 mM
hepatocytes
V79 hamster cells
transfected with mouse
GSTT1
Negative
2.5, 5, 10 mM
Forward mutation
Unscheduled
DNA synthesis
Sister chromatid
exchange
Chromosomal
aberrations
Sister chromatid
exchange
DNA and protein
synthesis
DNA SSBs by
alkaline elution
Chinese hamster
epithelial cells
Chinese hamster
epithelial cells
Chinese hamster
epithelial cells
Chinese hamster ovary
cells
Chinese hamster ovary
cells
Chinese hamster ovary
cells
Hamster hepatocytes
DNA Adducts Calf thymus DNA
DNA Adducts Calf thymus DNA
Unscheduled
DNA synthesis
Unscheduled
DNA synthesis
Human peripheral
lymphocytes
Primary human
fibroblast
A significant, dose-dependent increase in DNA
damage resulting from DNA-protein crosslinks in
V79 cells transfected with mouse GSTT1 compared
to parental cells
Hamster without GST activity from mouse
5,000, 10,000, 30,000,
50,000 ppm
5,000, 10,000, 30,000,
50,000 ppm
5,000, 10,000, 20,000,
30,000, and 40,000 ppm
Not provided
Not provided
Up to 1,000 (ig/mL
0.4-90 mM
50 mM
Negative
Negative
Casanova et al. (1997)
Hu et al. (2006)
Weak positive with or without rat-liver microsomal
system
Positive, independent of rat liver S9
Negative with or without rat liver S9
Negative
Negative
Calf
Positive in the presence of bacterial GST DM11 and
dichloromethane dehalogenase. Adducts primarily
formed with the guanine residues.
0 -8.0 umol (0 - 60 mM) Positive in the presence of bacterial GST DM11, rat
GST5-5, and human GSTT1-1. Adducts primarily
formed with the guanine residues.
Human
250, 500, 1,000 ppm Negative with or without rat liver S9
5,000, 10,000, 30,000, Negative
50,000 ppm
Jongen et al. (1981)
Jongen et al. (1981)
Jongen et al. (1981)
Thilagar and Kumaroo
(1983)
Thilagar and Kumaroo
(1983)
Garrett and Lewtas (1983)
Graves et al. (1995)
Kayser and Vuilleumier
(2001)
Marsch et al. (2004)
Perocco and Prodi (1981)
Jongen et al. (1981)
156

-------
Table 4-30. Results from in vitro genotoxicity assays of dichloromethane with mammalian systems, by type
of test
Assay
Test system
Concentrations
Results
Reference
Sister chromatid
exchange
Human peripheral
lymphocytes
Not provided
Weak positive
Thilagar et al. (1984)
Chromosomal
aberrations
Human peripheral
lymphocytes
Not provided
Positive
Thilagar et al. (1984)
DNASSBsby
alkaline elution
Human hepatocytes
Up to 120 mM
Negative at concentrations between 5 and 120 mM
Graves et al. (1995)
Micronucleus test
Human AHH-1,
MCL-5, h2El cell lines
Up to 10 mM
Positive in all three cell lines
Doherty et al. (1996)
DNA-protein
cross-links
Mouse, rat, hamster,
human hepatocytes
0.5-5 mM
Negative
Casanova et al. (1997)
DNA damage by
comet assay
Primary human lung
epithelial cells
10, 100, 1,000 \M
Weak trend, independent of GST activity (GST
enzymatic activity not present in the cultured cells)
Landi et al. (2003)
'HPRT = hypoxanthine-guanine phosphoribosyl transferase
5080
157

-------
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
4.5.1.2. In Vivo Genotoxicity Assays
Genotoxicity findings in Drosophila melanogaster assays are mixed (Table 4-31). A
study of gene mutation in D. melanogaster showed a marginal increase in sex-linked recessive
deaths following oral exposure (Gocke et al., 1981). An additional feeding study (Rodriguez-
Arnaiz, 1998) reported a positive response in the somatic w/w+ assay. A third study of
D. melanogaster (Kramers et al., 1991) found no evidence of increased sex-linked recessive
deaths, somatic mutations, or recombinations following exposure to airborne dichloromethane.
Table 4-31. Results from in vivo genotoxicity assays of dichloromethane in
insects
Assay
Test system
Doses
Result
Reference
Gene mutation (sex- Drosophila
linked recessive lethal)
Gene mutation (sex- Drosophila
linked recessive lethal,
somatic mutation and
recombination)
Somatic w/w+ assay Drosophila
125, 620 mM
6 hours—1,850, 5,500 ppm
1	week—2,360, 4,660 ppm
2	weeks—1,370, 2,360 ppm
(all approximate)
50, 100, 250, 500 mM
Positive (feeding Gocke et al. (1981)
exposure)
Negative (inhalation Kramers et al. (1991)
exposure)
Positive (feeding
exposure)
Rodriguez-Arnaiz (1998)
Some in vivo studies investigating certain genotoxic endpoints in mice exposed to
dichloromethane produced negative results (Table 4-32). Unscheduled DNA synthesis was not
induced in hepatocytes from mice (and rats) after 2- or 6-hour inhalation exposure to
concentrations that were carcinogenic in the NTP (1986) mouse bioassay (Trueman and Ashby,
1987) or other exposure routes (Lefevre and Ashby, 1989). Although positive results were not
observed in the unscheduled DNA synthesis studies, it is generally recognized that this assay is
not sensitive for detecting genotoxic chemicals (Eastmond et al., 2009; Madle et al., 1994).
Distinct, unequivocal cytogenetic effects (e.g., induction of micronuclei, sister chromatid
exchanges, or chromosome aberrations) were not consistently found in bone marrow or
erythrocytes in several studies of mice after acute oral exposures (Sheldon et al., 1987) or
parenteral exposures (Westbrook-Collins et al., 1990; Gocke et al., 1981). However,
tumorigenic effects in mice are generally localized to the liver and lung (due to high GST
activity) and therefore it is not surprising that genotoxic effects were, for the most part, not
observed in the bone marrow or erythrocytes (cell types with minimal GST activity). Crebelli et
al. (1999) stated that genotoxic effects induced by halogenated hydrocarbons (such as
dichloromethane) are not very effective in inducing micronucleus formation in mouse bone
marrow and a negative bone marrow micronucleus assay should not offset the consistently
positive in vitro results (Dearfield and Moore, 2005)
When genotoxic endpoints were examined in the cancer target tissues (liver and lung) in
mice exposed to dichloromethane, positive results were consistently reported (Table 4-32).
158

-------
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
These findings provide supporting evidence that GST-pathway metabolites may be key actors in
the genotoxic effects and carcinogenic mode of action for dichloromethane. Increased sister
chromatid exchanges were found in lung cells and peripheral lymphocytes from mice exposed by
inhalation for 2 weeks to 8,000 ppm or for 12 weeks to 2,000 ppm (Allen et al., 1990). Under
the same exposure conditions, increased chromosomal aberrations in lung and bone cells and
micronuclei in peripheral red blood cells also were found (Allen et al., 1990). DNA-protein
cross-links were detected in mouse hepatocytes, but not in lung cells, after a 3-day inhalation
exposure to 4000 ppm (Casanova et al., 1992) and between 500 and 4000 ppm (Casanova et
al.,1996). DNA damage, detected as increased DNA SSBs, was observed in liver and lung tissue
of B6C3Fi mice immediately following 3-hour exposures (Graves et al., 1995). The DNA
damage was not detectable 2 hours after in vivo exposure, indicating that DNA repair occurs
rapidly. Pretreatment of mice with buthionine sulphoximine, a GSH depletor, caused a decrease,
to levels seen in controls, in the amount of DNA damage detected immediately after in vivo
exposure in liver and lung tissue, indicating GSH involvement in the genotoxic process. DNA
damage (detected by the comet assay) was also reported in liver and lung tissues from male CD-
1 mice sacrificed 24 hours after administration of a single oral dose of 1,720 mg/kg of
dichloromethane (Sasaki et al., 1998). In this study, DNA damage in lung and liver was not
detected 3 hours after dose administration, and no DNA damage occurred at either time point in
several other tissues in which a carcinogenic response was not seen in chronic animal cancer
bioassays (e.g., stomach, kidney, bone marrow).
Formation of DNA adducts was evaluated in male and female B6C3Fi mice as well as in
male F344 rats (Watanabe et al., 2007). Animals were administered 5 mg/kg, i.p., of
radiolabeled dichloromethane and sacrificed at 1 or 8 hours after administration. The kidneys
and livers were removed and the DNA was isolated from these tissues to evaluate formation of
DNA adducts. At the administered dose, DNA adducts were not detected.
Other studies in mice have looked for mutations in specific oncogenes (K-ras or H-ra.s)
(Devereux et al., 1993) or in a tumor suppressor gene (p53) (Hegi et al., 1993) in liver or lung
tumors from dichloromethane-exposed mice. These studies have not demonstrated exposure-
related patterns of mutations in these genes, although it should be noted that the statistical power
of this analysis for the lung tumors is limited (discussed further in sections 4.5.2 and 4.5.3).
159

-------
Table 4-32. Results from in vivo genotoxicity assays of dichloromethane in mice
Assay
Test system
Route and dose
Duration
Results
Reference
Micronucleus test
Micronucleus test
Micronucleus test
Micronucleus test
DNA synthesis
Unscheduled DNA synthesis
Sister chromatid exchange
Sister chromatid exchange
Sister chromatid exchange
Sister chromatid exchange
Chromosome aberrations
Chromosome aberrations
Mouse bone marrow 425, 850, or 1,700 mg/kg
Mouse bone marrow
Mouse peripheral red
blood cells
Mouse peripheral red
blood cells
Mouse liver
Mouse hepatocytes
Mouse bone marrow
Mouse bone marrow
Mouse lung cells and
peripheral lymphocytes
Mouse lung cells
Mouse bone marrow
Mouse bone marrow
Gavage,
1,250, 2,500, and
4,000 mg/kg
Inhalation 6 hr/day, 5 d/wk,
0, 4,000, 8,000 ppm
Inhalation, 6 hr/day, 5 d/wk,
0, 2,000 ppm
Gavage, 1,000 mg/kg;
inhalation, 4,000 ppm
Inhalation, 2,000 and
4,000 ppm.
Intraperitoneal, 100, 1,000,
1,500, 2,000 mg/kg
Subcutaneous, 0, 2,500,
5,000 mg/kg
Inhalation 6 hr/day, 5 d/wk,
0, 4,000, 8,000 ppm
Inhalation 6 hr/day, 5 d/wk,
0, 2,000 ppm
Intraperitoneal, 100, 1,000,
1,500, 2,000 mg/kg
Subcutaneous, 0, 2,500,
5,000 mg/kg
2 doses
single dose
2 wk
12 weeks
single dose;
2 hours
2 or 6 hours
single dose
single dose
2 weeks
12 weeks
single dose
single dose
Negative at all doses
Negative at all doses
Positive at 4,000 and
8,000 ppm
Positive at 2,000 ppm
Negative in both oral and
inhalation studies
Negative
Negative
Negative at all doses
Positive at 8,000 ppm
Positive at 2,000 ppm
Negative
Negative
Gocke et al.
(1981)
Sheldon et al.
(1987)
Allen et al.
(1990)
Allen et al.
(1990)
Lefevre and
Ashby (1989)
Trueman and
Ashby (1987)
Westbrook-
Collins et al.
(1990)
Allen et al.
(1990)
Allen et al.
(1990)
Allen et al.
(1990)
Westbrook-
Collins et al.
(1990)
Allen et al.
(1990)
160

-------
Table 4-32. Results from in vivo genotoxicity assays of dichloromethane in mice
Assay
Test system
Route and dose
Duration
Results
Reference
Chromosome aberrations
Mouse lung and bone
Inhalation, 6 hr/day, 5 d/wk,
2 weeks
Positive at 8,000 ppm
Allen et al.

marrow cells
0, 4,000, 8,000 ppm


(1990)
DNA-protein cross-links
Mouse liver and lung
Inhalation, 6 hr/day, 3 days,
3 days
Positive in mouse liver cells
Casanova et al.

cells
4,000 ppm

at 4,000 ppm; negative in
(1992)




mouse lung cells

DNA-protein cross-links
Mouse liver and lung
Inhalation, 6 hr/day, 150,
3 days
Positive in mouse liver cells
Casanova et al.

cells
500, 1,500, 3,000,

at 500-4,000 ppm; negative
(1996)


4,000 ppm

in mouse lung cells

DNA single strand breaks by
Mouse hepatocytes
Inhalation, 2,000 and
3 or 6 hours
Positive at 4,000 ppm at
Graves et al.
alkaline elution

4,000 ppm

3 and 6 hours
(1994b)
DNA single strand breaks by
Mouse liver and lung
Liver: inhalation, 2,000,
3 hours
Liver: positive at 4,000-
Graves et al.
alkaline elution
homogenate
4,000, 6,000, 8,000 ppm

8,000 ppm
(1995)


Lung: inhalation, 1,000,
3 hours
Lung: positive at 2,000-



2,000, 4,000, 6,000 ppm

4,000 ppm

DNA damage by comet assay
Mouse liver and lung
Gavage, 1,720 mg/kg;
single dose
Positive only at 24 hours
Sasaki et al.

cells
organs harvested at 0

after dosing
(1998)


(control), 3, and 24 hours



DNA damage by comet assay
Mouse stomach, urinary
Gavage, 1,720 mg/kg;
single dose
Negative 3 or 24 hr after
Sasaki et al.

bladder, kidney, brain,
organs harvested at 0

dosing
(1998)

bone marrow
(control), 3, and 24 hours



DNA adducts
Mouse liver and kidney
Intraperitoneal, 5 mg/kg
Single dose
Negative
Watanabe et al.

cells



(2007)
Kras and Hras oncogenes
Mouse liver and lung
0, 2000 ppm
Up to 104 weeks
No difference in mutation
Devereux et al.,

tumors


profile between control and
1993




dichloromethane-induced





liver tumors; number of





spontaneous lung tumors





(n=4) limits comparison at





this site

p53 tumor suppressor gene
Mouse liver and lung
0, 2000 ppm
Up to 104 weeks
Loss of heterozygocity
Hegi et al., 1993

tumors


infrequently seen

161

-------
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
Results from in vivo studies in other mammals (i.e., rats and hamsters) of hepatocyte
sensitivity to dichloromethane induction of DNA SSBs (Table 4-33) are consistent with
interspecies differences in the induction of liver tumors in the inhalation cancer bioassays. A
gavage study in rats reported the presence of DNA SSBs with a dose of 1,275 mg/kg (Kitchin
and Brown, 1989). The other available studies, however, did not find any genotoxicity following
dichloromethane exposure. No increase in unscheduled DNA synthesis in rat hepatocytes was
seen following inhalation dichloromethane exposure of 2 to 6 hours at 2000 or 4000 ppm
(Trueman and Ashby, 1987), exposure by gavage up to 1000 mg/kg (Trueman and Ashby, 1987),
or intraperitoneal exposure of 400 mg/kg (Mirsalis et al., 1989). DNA adducts were not detected
in the livers and kidneys of male F344 rats dosed with 5 mg/kg dichloromethane, i.p. (Watanabe
et al., 2007). DNA SSBs were significantly increased in hepatocytes isolated from B6C3Fi mice
exposed to 4,831 ppm (4,000 ppm nominal) for 6 hours but were not increased in hepatocytes
from Sprague-Dawley rats exposed to 4,527 ppm (4,000 ppm nominal) for 6 hours (Graves et al.,
1994b). Results from in vivo interspecies comparisons of dichloromethane induction of DNA-
protein cross-links in hepatocytes (expected products of the GSH pathway) are also consistent
with the hypothesis that the mouse is more sensitive than other mammalian species due to greater
activity of the GST pathway. DNA-protein cross-links were formed in the liver of mice, but not
hamsters, following in vivo exposure to air concentrations ranging from 500 to 4,000 ppm,
6 hours/day for 3 days (Casanova et al., 1996). The absence of a genotoxic response in the rat
and hamster is consistent with considerably lower GST activity and therefore, these mammalian
systems would be expected to be less sensitive at detecting genotoxic effects than the studies
conducted in mice.
Table 4-34 compares results from studies of mice and rats in which comparable tissue-
specific endpoints were examined in in vivo genotoxicity assays. Several of the endpoints that
were positive in mice (e.g., sister chromatid exchange, DNA-protein cross-links, comet assay)
have not been examined in the rat. Unscheduled DNA synthesis has been demonstrated in
mouse, but not in rat hepatocytes. In contrast to the positive results seen in mouse inhalation
exposure studies, DNA SSB induction was not seen in rat inhalation studies but was seen in an
oral gavage study.
In summary, the available data provide evidence for mutagenicity of dichloromethane.
Most of the in vitro bacterial assays with GST activity showed positive results when there was
GST activity. Non-positive results were reported only in bacterial assays with low GST activity;
in experiments where GST was added, positive results were then observed. Evaluation of the in
vitro mammalian studies also demonstrates consistency of the requirement for GST for
observation of genotoxic effects. In rat and hamster cell lines where GST activity is significantly
less than mouse, primarily negative results were reported following dichloromethane exposure.
However, when mouse liver cytosol or transfected mouse GST were included in these same cell
lines, mutagenic effects were reported after dichloromethane exposure. In mouse cell lines,
162

-------
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
positive results were obtained in Clara cells but no effects were observed in a mouse lymphoma
cell line, which is consistent with the absence of tumors in this site for mice. The results of in
vivo mutagenicity in mice also provide support for the site-specificity of the observed tumors.
Assays using mouse bone marrow were all negative. However, micronuclei and sister chromatid
exchange tests in peripheral blood produced a positive response at high doses. With the
exception of one study of unscheduled DNA synthesis in hepatocytes, numerous site-specific
studies in either the liver or lung were also positive at various doses. These liver and lung
studies included chromosomal aberrations, SSBs and sister chromatid exchanges, and DNA-
protein cross-links and correspond to genotoxic and mutagenic effects associated with
metabolites from the GST pathway.
163

-------
5190
Table 4-33. Results from in vivo genotoxicity assays of dichloromethane in rats and hamsters
Assay
Test system
Route and dose
Duration
Results
Reference
Unscheduled DNA synthesis
Rat hepatocytes
Gavage, 100, 500,
Liver harvested 4 and
Negative 4 or 12 hours after
Trueman and


1,000 mg/kg
12 hours after dosing
dosing
Ashby (1987)
Unscheduled DNA synthesis
Rat hepatocytes
Inhalation, 2 or 6 hours,
2 or 6 hours
Negative at both
Trueman and


2,000 and 4,000 ppm

concentrations and exposure
Ashby (1987)




durations

Unscheduled DNA synthesis
Rat hepatocytes
Intraperitoneal, single dose,
Single dose
Negative 48 hours after
Mirsalis et al.


400 mg/kg

dosing
(1989)
DNA SSBs by alkaline elution
Rat hepatocytes
Inhalation, 3 or 6 hours,
3 or 6 hours
Negative at all
Graves et al.


2,000 and 4,000 ppm

concentrations and time
(1994b)




points

DNA SSBs by alkaline elution
Rat liver homogenate
Gavage, 2 doses, 425 mg/kg
4 or 21 hours (time
Positive at 1,275 mg/kg
Kitchin and


and 1,275 mg/kg,
between dosing and

Brown (1989)


administered 4 and 21 hours
liver harvesting)




before liver harvesting



DNA SSBs by alkaline elution
Rat liver and lung
Liver: inhalation, 4,000,
3 hours
Negative for both liver and
Graves et al.

homogenate
5,000 ppm

lung at all concentrations
(1995)


Lung: inhalation, 4,000 ppm
3 hours


DNA-protein cross-links
Hamster liver and lung Inhalation, 6 hr/day, 500,
3 days
Negative at all
Casanova et al.

cells
1,500, 4,000 ppm

concentrations
(1996)
DNA adducts
Rat liver and kidney
Intraperitoneal, 5 mg/kg
Single dose
Negative
Watanabe et al.

cells



(2007)
164

-------
Table 4-34. Comparison of in vivo dichloromethane genotoxicity assays targeted to lung or liver cells, by species
Assay

Studies in mice


Studies in rats

Test system
Route, dose (duration)
Results
Reference
Test system Route, dose (duration) Results
Reference
DNA
Liver
Gavage, 1,000 mg/kg;
Negative in oral and
Lefevre and

No studies
synthesis

inhalation, 4,000 ppm
inhalation studies
Ashby




(2 hours)

(1989)


Unscheduled
Hepatocytes
Inhalation, 2,000 and
Negative
Trueman
Hepatocytes Inhalation, 2,000 and Negative
Trueman and
DNA

4,000 ppm.

and Ashby
4,000 ppm (2 or
Ashby (1987)
synthesis

(2 or 6 hours)

(1987)
6 hours)

Unscheduled




Hepatocytes Intraperitoneal, Negative
Mirsalis et al.
DNA




400 mg/kg
(1989)
synthesis






Sister
Lung cells
Inhalation 6 hr/day,
Positive at
Allen et al.

No studies
chromatid

5 days/wk, 0, 4,000, 8,000
8,000 ppm
(1990)


exchange

ppm (2 weeks)






Inhalation 6 hr/day,
Positive at





5 days/wk, 0, 2,000 ppm
2,000 ppm





(12 weeks)




Chromosome
Lung cells
Inhalation, 6 hr/day,
Positive at
Allen et al.

No studies
aberrations

5 days/wk, 0, 4,000, 8,000
8,000 ppm
(1990)




ppm (2 weeks)




DNA-protein
Liver and
Inhalation, 6 hr/day,
Positive in liver
Casanova et

No studies
cross-links
lung cells
3 days, 4,000 ppm (3 days)
4,000 ppm
al. (1992)




Inhalation, 6 hr/day, 150,
Positive in liver at





500, 1,500, 3,000,
500-4,000 ppm;





4,000 ppm (3 days)
both studies negative






in lung



DNA SSBs
Hepatocytes
Inhalation, 2,000 and
Positive at
Graves et
Hepatocytes Inhalation, 3 or 6 hours, Negative at all
Graves et al.
by alkaline

4,000 ppm (3 or 6 hours)
4,000 ppm
al. (1994b)
2,000 and 4,000 ppm concentrations
(1994b)
elution




and time points

DNA SSBs
Liver and
Liver: inhalation, 2,000,
Liver: Positive at
Graves et
Liver and Liver: inhalation, 4,000, Negative in
Graves et al.
by alkaline
lung
4,000, 6,000, 8,000 ppm (3
4,000-8,000 ppm
al. (1995)
lung 5,000 ppm liver and lung at (1995)
elution
homogenate
hours)


homogenate Lung: inhalation, all



Lung: inhalation, 1,000,
Lung: Positive at

4,000 ppm concentrations



2,000, 4,000, 6,000 ppm (3
2,000-4,000 ppm

and time points



hours)




DNA SSBs




Liver Gavage, 425 mg/kg and Positive at
Kitchin and
by alkaline




homogenate 1,275 mg/kg 1,275 mg/kg
Brown (1989)
elution






165

-------
Table 4-34. Comparison of in vivo dichloromethane genotoxicity assays targeted to lung or liver cells, by species

Studies in mice


Studies in rats

Assay Test system
Route, dose (duration)
Results
Reference
Test system Route, dose (duration) Results
Reference
DNA damage Liver and
by comet lung cells
assay
Gavage, 1,720 mg/kg;
organs harvested at 0
(control), 3, and 24 hours
Positive only at 24
hours after dosing
Sasaki et al.
(1998)

No studies
DNAadducts Liver and
kidney cells
Intraperitoneal, 5 mg/kg
Negative
Watanabe et
al. (2007)
Liver and Intraperitoneal, 5 mg/kg Negative
kidney cells
Watanabe et
al. (2007)
5191
166

-------
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
4.5.2. Mechanistic Studies of Liver Effects
One of the major target organs from dichloromethane exposure is the liver, and several
studies have focused on examining the potential mechanisms producing liver tumors. This
section summarizes the primary mechanistic studies that were conducted in order to examine the
hepatic tumors produced by dichloromethane in mice. A parallel set of studies, discussed in the
next section, focus on potential mechanisms that produce lung tumors. Briefly,
dichloromethane-induced liver tumors first appeared in mice after 52 weeks of exposure
(Maronpot et al., 1995; Kari et al., 1993), which was when tumors began to appear in control
mice, indicating a similar time course in tumor formation between treated and untreated groups.
Onset of liver tumor formation is not preceded by liver cell proliferation (Casanova et al., 1996;
Foley et al., 1993). Further mechanistic studies were conducted to assay the tumor for
significant changes in proto-oncogene activation and tumor suppressor gene inactivation
(Maronpot et al., 1995; Devereux et al., 1993; Hegi et al., 1993). A second subset of mechanistic
studies was conducted to elucidate the reason that mice are the most sensitive species to liver
tumors and if other species exhibited changes in liver function (Thier et al., 1998; Reitz et al.,
1989). It was found that mice have the highest level of GST-T1 catalytic activity but that
humans, rats, and hamsters among other species also metabolize dichloromethane in the liver to a
GST conjugate. In contrast, there has been little research focusing on the mechanisms through
which nonneoplastic hepatic effects (seen most strongly in the rat) are produced, and the role of
the parent material, metabolites of the CYP2E1 pathway, metabolites of the GST pathway, or
some combination of parent material and metabolites is not known.
Liver tumor characterization studies
Several studies have examined the time course of appearance of liver tumors in B6C3Fi
mice exposed to 2,000 or 4,000 ppm and possible links between hepatic nonneoplastic
cytotoxicity, enhanced hepatic cell proliferation, and the development of liver tumors (Casanova
et al., 1996; Maronpot et al., 1995; Foley et al., 1993; Kari et al., 1993). The studies provide no
clear evidence for a sustained liver cell proliferation response to dichloromethane that can be
linked to the development of dichloromethane-induced liver tumors. Additionally, a few studies
have examined if dichloromethane-induced liver tumors are the result of proto-oncogene
activation and tumor suppressor gene inactivation (Maronpot et al., 1995; Devereux et al., 1993;
Hegi et al., 1993).
Kari et al. (1993) (also summarized by Maronpot et al. [1995]) reported data from
6 groups of 68 female B6C3Fi mice exposed to six "stop-exposure" protocols of differing
durations and sequences, with each exposure concentration standardized at 2,000 ppm for
6 hours per day, 5 days per week. The six stop-exposure protocols were 26 weeks of exposure
followed by 78 weeks without exposure, 78 weeks without exposure followed by 26 weeks of
exposure, 52 weeks without exposure followed by 52 weeks with exposure, 52 weeks exposed
167

-------
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
followed by 52 weeks without exposure, 78 weeks exposed followed by 26 weeks without
exposure, and 26 weeks without exposure followed by 78 weeks with exposure. A control group
(no exposure, 104 weeks duration) and a maximum exposure (104 weeks duration) group were
also included. Exposure for 26 weeks did not result in an increased incidence of liver tumors
(adenomas or carcinomas). Respective percentages of animals with liver tumors were 27%
(18/67), 40% (27/67), and 34% (23/67) for the controls, early 26-week exposure and late 26-
week exposure groups, respectively. Exposure to 2,000 ppm for 52 weeks (followed by no
exposure until 104 weeks), 78 weeks (either early or late exposure periods), or 104 weeks
produced increased incidence of mice with liver tumors (p < 0.05), but this increase was not seen
in the 52-week late exposure group. Respective percentages of animals with liver tumors
(adenomas or carcinomas combined) were 44% (28/64), 31% (21/67), 62% (42/68), 48% (32/67)
and 69%) (47/68) for the 52 (early exposure), 52 (late exposure), 78 (early exposure), 78 (late
exposure), and 104 week exposure periods, respectively. With the 78 week exposures, the
difference in the liver tumor incidence between the early and late exposure periods was
statistically significant (p < 0.01). A greater increase in multiplicity of liver tumors was also
seen with the early 78-week exposure period. These data suggest that 52 weeks of exposure was
required to increase the incidence of liver tumors in mice, that early exposure was more effective
than late exposure, and that the increased risk continued after cessation of exposure.
Histopathologic examination of liver tissue at interim killings at eight time periods (13,
26, 52, 68, 75, 78, 83, or 91 weeks) of exposure to 2,000 ppm (n = 20 mice per killing) found no
evidence of nonneoplastic cytotoxicity that preceded the appearance of proliferative neoplastic
liver lesions. Incidences of mice with liver adenomas or carcinomas were elevated (between 40-
60%) at five of the six interim killings after 52 weeks. The incidence rates at each time period
were 0/20 (0%) at 13 weeks, 1/20 (5%) at 26 weeks, 8/20 (40%) at 52 weeks, 4/26 (15%) at
68 weeks, 13/20 (65%) at 75 weeks, 12/19 (63%) at 78 weeks, 8/20 (40%) at 83 weeks, and
20/30 (66%) at 91 weeks. The collected liver lesion data identify no exposure-related increased
incidence of nonneoplastic liver lesions that could be temporally linked to liver tumor
development. Liver tumors first appeared at about the same time in control and exposed animals
(52 weeks).
Foley et al. (1993) examined indices of cell proliferation in livers of female B6C3Fi mice
exposed to 1,000, 2,000, 4,000, or 8,000 ppm dichloromethane (6 hours/day, 5 days/week) for 1,
2, 3, or 4 weeks or to 2,000 ppm for 13, 26, 52, or 78 weeks but found no evidence for sustained
cell proliferation with prolonged exposure to dichloromethane. To label liver cells in S phase,
tritiated thymidine (1- to 4-week exposure protocols) or bromodeoxyuridine (13- to 78-week
protocols) was administered subcutaneously via an osmotic mini-pump for 6 days prior to
killing. Labeled hepatocytes in liver sections (from 10 mice in each exposure/duration group)
were counted to assess the number of cells in S-phase per 1,000 cells. S-phase labeling indices
in livers of exposed mice at most killings were equivalent to or less than those in control mice.
168

-------
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
5282
5283
5284
5285
5286
5287
5288
5289
5290
5291
5292
5293
5294
5295
5296
5297
5298
5299
5300
5301
5302
5303
5304
5305
A transient increase in S-phase labeling index of about two- to fivefold over controls was
observed at the 2-week killing of mice exposed to 1,000, 4,000, or 8,000 ppm. Because of the
transient nature and small magnitude of the response, it is not expected to be of significance to
the promotion of liver tumors in chronically exposed mice. Foley et al. (1993) also compared
cell proliferation labeling indices in foci of cellular alteration and non-affected liver regions in
control and exposed mice but found no significant difference between control and exposed mice.
S-phase labeling was accomplished by immunohistologic staining for proliferating cell nuclear
antigen in liver sections from 24 control mice and 15 exposed mice, with livers showing foci of
cellular alteration. In both control and exposed livers, the labeling index was about four- to
fivefold higher in foci of cellular alteration than in surrounding unaffected liver tissue.
In mice exposed to 2,000 ppm for 13-78 weeks, relative liver weights were statistically
significantly elevated compared with controls; about 10% increased at 13 and 26 weeks and
about 30-40% increased at 52 and 78 weeks. Histologic changes in liver sections of 2,000 ppm
mice exposed for 13-78 weeks were restricted to hepatocellular hypertrophy (observed at all
killing intervals) and preneoplastic (foci of cellular alteration) and neoplastic (adenoma and
carcinoma) lesions. No signs of liver tissue degeneration were found. Adenoma and focus of
alteration were first detected at 26 weeks (2/10 versus 0/10 in controls). At 52 weeks, 4/10
exposed mice had proliferative lesions (1 focus, 1 adenoma, and 2 carcinomas), compared with
1/10 in controls (1 adenoma). At 78 weeks, 7/10 exposed mice had proliferative lesions (2 foci,
3 adenomas, 6 carcinomas) compared with 1/10 in controls (1 adenoma). In summary, the
results indicate that inhalation exposure to 2,000 ppm dichloromethane produced an increase
incidence of liver tumors in female B6C3Fi mice. No evidence was found for sustained cell
proliferation or liver tissue degeneration in response to dichloromethane exposure, but exposure
was associated with relative liver weight increases and hepatocellular hypertrophy.
Casanova et al. (1996) found no clear evidence of increased cell proliferation in the livers
of male B6C3Fi mice exposed to dichloromethane concentrations >1,500 ppm 6 hours/day for
3 days. Three or four groups of three mice were exposed to 146, 498, 1,553, or 3,923 ppm
unlabeled dichloromethane for 2 days and then exposed to [14C]-labeled dichloromethane for
6 hours on the third day. Radiolabel incorporated into liver DNA deoxyribonucleosides was
measured as an index of cell proliferation. Radiolabel incorporated into liver DNA
deoxyribonucleosides increased approximately fivefold from 146 to 1,553 ppm, but further
increases were not apparent at 3,923 ppm. (In contrast, as described in section 4.5.3, radiolabel
incorporation into lung DNA deoxyribonucleosides displayed a 27-fold increase over this
concentration range.) The small magnitude of the increase in radiolabel incorporation into liver
DNA deoxyribonucleosides with increasing exposure concentration suggests that little, if any,
enhanced cell proliferation occurred in the liver in response to dichloromethane exposure.
Devereux et al. (1993) (also reported in Maronpot et al. [1995]) analyzed liver tumors in
female B6C3Fi mice for the presence of activated H-ra.s oncogenes. Fifty dichloromethane-
169

-------
5306
5307
5308
5309
5310
5311
5312
5313
5314
5315
5316
5317
5318
5319
5320
5321
5322
5323
5324
5325
5326
5327
5328
5329
5330
5331
5332
5333
5334
5335
5336
5337
5338
5339
5340
5341
5342
5343
induced and 49 spontaneous liver tumors were screened for H-ras mutations. There was a
relatively high frequency of activated H-ras among the nonexposed B6C3Fi mice: 67% of the
spontaneous tumors and 76% of the dichloromethane-induced tumors contained mutations in the
H-ras gene. Overall, the mutation profile of the dichloromethane-induced tumors did not
significantly differ from the spontaneous tumors.
Similarly, Hegi et al. (1993) analyzed the liver tumors from female B6C3Fi mice for
inactivation of the tumor suppressor genes, p53 and Rb-1. Half of the liver tumors used for this
study had an activated H-ras oncogene. Twenty liver tumors (15 carcinomas and 5 adenomas)
were screened for loss of heterozygosity (LOH) on chromosome 11 and 14, which is associated
with malignant conversion of the p53 gene (chromosome 11) and the Rb-1 gene (chromosome
14). Only one tumor out of 20 contained a LOH at chromosome 14 and no dichloromethane-
induced liver tumors contained a LOH at chromosome 11.
Liver metabolic studies
As described in detail in section 3.3, GST-T1 enzymatic activity and distribution is
variable among species, and there is also considerable intraspecies variability among humans. In
summary, Reitz et al. (1989) demonstrated a greater metabolic activity with respect to
dichloromethane in livers of B6C3Fi mice compare with F344 rats, Syrian golden hamsters, and
humans. The rates of in vitro metabolism by the GST pathway were about 4-, 12-, and 20-fold
greater in B6C3Fi mouse liver samples than in F344 rat, human, and Syrian golden hamster
samples, respectively (Reitz et al., 1989). A more recent study characterized the
dichloromethane metabolic capacity specifically of hepatic GST-T1 (Thier et al., 1998).
Enzymatic activities of GST-T1 in liver from F344 rats, B6C3Fi mice, Syrian golden hamsters,
and humans with three different GST-T1 phenotypes (nonconjugators, low conjugators, high
conjugators) showed the following order with dichloromethane as a substrate: mouse » rat >
human high conjugators>human low conjugators > hamster > human nonconjugators.
4.5.3. Mechanistic Studies of Lung Effects
The finding of increased lung tumors in B6C3Fi mice exposed to dichloromethane
(Mennear et al., 1988; NTP, 1986) has stimulated a number of studies designed to examine the
mechanism for dichloromethane-induced lung tumors in this animal. The lung tumor mechanism
studies were conducted with B6C3Fi mice, and the frequency of lung tumors in control animals
was very low. Time-course studies for lung tumor development were conducted, and it was
found that the onset of lung tumor development was much shorter than liver tumors (Kari et al.,
1993) (reported in Maronpot et al. [1995]). As a result, it is hypothesized that a potential
common mechanism independent of liver metabolism is producing tumors in the lung. As with
the liver tumors, there were no significant increases in mutations for the K-ras oncogene
(Devereux et al., 1993) or thep53 and Rb-1 tumor suppressor genes (Hegi et al., 1993).
170

-------
5344
5345
5346
5347
5348
5349
5350
5351
5352
5353
5354
5355
5356
5357
5358
5359
5360
5361
5362
5363
5364
5365
5366
5367
5368
5369
5370
5371
5372
5373
5374
5375
5376
5377
5378
5379
5380
5381
Additionally, the Clara cells, which are non-ciliary secretory cells found in the primary
bronchioles of the lung, are selectively targeted after dichloromethane exposure. Acute
dichloromethane exposure produces Clara cell vacuolization, which is not sustained with long-
term exposure (Foster et al., 1992). There is a correlation between the acute effects on the Clara
cell and the lung tumors from chronic exposure to dichloromethane (Kari et al., 1993).
However, the exact mechanism for producing these lung effects is not completely understood at
this point. Provided below is a summary of the studies examining the potential mechanisms for
producing lung tumors resulting from dichloromethane exposure.
Lung tumor characterization studies
Kari et al. (1993) (also summarized in Maronpot et al. [1995]) demonstrated that only 26
weeks of exposure to 2,000 ppm was necessary to produce significantly increased incidence of
female B6C3Fi mice with lung tumors. In the six "stop-exposure" protocol experiment
(26 weeks exposure followed by 78 weeks without exposure, 78 weeks without exposure
followed by 26 weeks exposure, 52 weeks without exposure followed by 52 weeks with
exposure, 52 weeks exposed followed by 52 weeks without exposure, 78 weeks exposed
followed by 26 weeks without exposure, and 26 weeks without exposure followed by 78 weeks
with exposure), early but not late exposure for 26 or 52 weeks resulted in an increased incidence
of animals with lung tumors (adenoma or carcinomas). Respective percentages of animals with
lung tumors were 7.5% (5/67), 31% (21/68), 4% (3/67), 63% (40/63), and 15% (10/67) for the
controls, early 26-, late 26-, early 52-, and late 52-week exposure groups, respectively. With the
78-week exposures, both the early and late exposure regimens produced an increased incidence
of lung tumors compared with controls (56% [38/68] and 19% [13/68], respectively), compared
with the incidence of 63% (42/67) seen in the group exposed for the full 104 weeks. Thus a
plateauing of risk was seen, with similar incidence rates seen with the early 52-week, early
78-week, and 104-week exposure periods. The difference in the lung tumor incidence between
the early and late exposure periods of similar duration was statistically significant (p < 0.01) for
the 26-, 52-, and 78-week duration protocols. A greater increase in multiplicity of lung tumors
was also seen with the early 78-week exposure period. As with the liver tumor data from the
same series of experiments, these data suggest that early exposure was more effective than late
exposure and that the increased risk continued after cessation of exposure.
Histopathologic examination of lung tissue from mice killed at 13, 26, 52, 68, 75, 78, 83,
or 91 weeks of exposure to 2,000 ppm (n = 20 mice per killing) found no evidence of
nonneoplastic cytotoxicity that preceded the appearance of proliferative neoplastic lung lesions.
In contrast, incidences of mice with lung adenomas or carcinomas (combined) were elevated at
interim killings >52 weeks; incidences for the interim killings of mice exposed to 2,000 ppm
(6 hours per day, 5 days per week) between 13 and 91 weeks were 0/20 (0%) at 13 weeks,
0/20 (9%) at 26 weeks, 6/20 (30%) at 52 weeks, 6/26 (23%) at 68 weeks, 8/20 (40%) at
171

-------
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
75 weeks, 9/19 (47%) at 78 weeks, 10/20 (50%) at 83 weeks, and 14/30 (47%) at 91 weeks,
respectively. Lung hyperplasia was found at an increased incidence only at 91 weeks, well after
the 26- and 52-week periods that induced increased incidences of mice with lung tumors.
Kanno et al. (1993) found no evidence for histologic changes or increased cell
proliferation in lung tissue of female B6C3Fi mice exposed to 2,000 or 8,000 ppm
dichloromethane for 1, 2, 3, or 4 weeks, compared with control mice, or in mice exposed to
2,000 ppm for 13 or 26 weeks. Osmotic mini-pumps were used to deliver tritiated thymidine and
label cells undergoing replicative DNA synthesis over 6-day periods before killing. Labeled
cells undergoing rapid DNA synthesis and cell proliferation were assessed in sections of
proximal and terminal bronchioles and alveoli of lungs from groups of 5 mice exposed for 1-
4 weeks or 10 mice exposed for 13 or 26 weeks. There were no exposure-related histopathologic
or labeling index changes in the alveoli, but lower labeling indices were found in the bronchiolar
epithelium of exposed mice compared with controls.
The combined results from the Kari et al. (1993) and Kanno et al. (1993) studies are
consistent with the hypothesis that dichloromethane-induced lung tumors in B6C3Fi mice are not
preceded by overt cytotoxicity, enhanced and sustained cell proliferation, or hyperplasia in the
lung. Two other studies (Casanova et al., 1996; Foster et al., 1992), however, have reported
evidence for enhanced cell proliferation in lungs of B6C3Fi mice exposed for acute durations to
airborne dichloromethane. Only one of these studies (Foster et al., 1992), however, looked for
sustained cell proliferation in the lung with prolonged exposure. In agreement with the results
from Kanno et al. (1993), no evidence was found for sustained cell proliferation in lungs with
prolonged exposure to dichloromethane at concentrations demonstrated to induce lung tumors in
mice.
Casanova et al. (1996) detected evidence of increased cell proliferation in the lungs of
male B6C3Fi mice exposed to dichloromethane concentrations >1,500 ppm 6 hours/day for
3 days. Three or four groups of three mice were exposed to 146, 498, 1,553, or 3,923 ppm
unlabeled dichloromethane for 2 days and then exposed to [14C]-labeled dichloromethane for
6 hours on the third day. Radiolabel incorporated into lung DNA deoxyribonucleosides (after
removal of DNA-protein cross links containing radiolabeled formaldehyde) was measured as an
index of cell proliferation. Radiolabel incorporation into lung DNA deoxyribonucleosides
increased with increasing exposure concentration, with the amount increasing by about 27-fold
between 146 and 3,923 ppm. In hamsters that did not develop tumors in response to chronic
inhalation exposure to 3,500 ppm dichloromethane (Burek et al., 1984), no evidence for
enhanced radiolabel incorporation into lung DNA deoxyribonucleosides was found following
acute exposure (Casanova et al., 1996).
Devereux et al. (1993) (also summarized in Maronpot et al. [1995]) analyzed lung tumors
in female B6C3Fi mice for the presence of activated K-ra.s oncogenes. Fifty-four
dichloromethane-induced and 17 spontaneous lung tumors (7 from theNTP [1986] study and 10
172

-------
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
from a study in C57BL/6 x C34F1 mice reported by Candrian et al. [1991]) were screened for K-
ras mutations. Twenty percent of the dichloromethane-induced tumors and 24% of the
spontaneous tumors contained mutations in the K-ras gene. Devereux et al. (1993) stated that
there may be a significant difference in the incidence of K-ras activation between spontaneous
and dichloromethane-induced tumors. However, the small number of the spontaneous tumors
that were available for the study limits the conclusions that can be made from the results.
Hegi et al. (1993) analyzed the lung tumors from female B6C3Fi mice for inactivation of
the tumor suppressor genes, p53 and Rb-1. Forty-nine dichloromethane-induced lung
carcinomas, five lung adenomas, and seven spontaneous lung carcinomas were screened for
LOH on mouse chromosome 11 and 14, which is associated with malignant conversion of the
p53 gene (chromosome 11) and the Rb-1 gene (chromosome 14). Fourteen percent (n = 7) of the
dichloromethane-induced lung carcinomas exhibited LOH at chromosome 11. No p53 mutations
were detected in the seven spontaneous lung tumors or the five dichloromethane-induced lung
adenomas. Only three dichloromethane-induced tumors exhibited LOH at chromosome 14. The
authors noted that inactivation of the p53 and Rb-1 tumor suppressor genes infrequently occur in
lung and liver tumors.
Clara cell studies
Foster et al. (1992) found enhanced cell proliferation in bronchiolar cells and, to a lesser
degree, alveolar cells in the lungs of male B6C3Fi mice exposed for acute durations (2, 5, 8, or
9 days) to 4,000 ppm dichloromethane (6 hours/day, 5 days/week), but the response was less
distinct after subchronic durations of exposure (89, 92, or 93 days). To measure cell
proliferation, mice (n = 5 per exposure-duration group) were given subcutaneous doses of
tritiated thymidine for five consecutive days prior to killing. Labeled cells in bronchiolar or
alveolar epithelium in lung sections were counted to assess the number of cells in S phase per
1,000 cells. Counts of bronchiolar epithelium cells in S phase in exposed mice sacrificed on
days 2, 5, 8, and 9 were approximately 2-, 15-, 3-, and 5-fold higher, respectively, than those of
unexposed mice at day 0 of the experiment. In exposed mice sacrificed on days 89, 92, and 93,
less than twofold increases in bronchiolar epithelium cell labeling were observed. Increased cell
labeling was found in alveolar epithelium only on day 8 (about seven- to eightfold increase) and
day 9 (about fourfold increase). Vacuolation of the Clara cells of the bronchiolar epithelium was
observed on day 2 (scored as ++, majority of cells affected), day 9 (+++, virtually all the cells
affected), and day 44 (+, moderate effect to cells) but was not apparent on days 5, 8, 40, 43, 89,
92, or 93. No hyperplasia of the bronchiolar epithelium or changes to Type II alveolar epithelial
cells were observed in the lungs of any of the exposed mice at any time point. The appearance
and disappearance of the Clara cell vacuolation were generally correlated with the appearance
and disappearance of enhanced cell proliferation in the bronchiolar epithelium; enhanced cell
proliferation was observed on days 2, 5, 8, and 9 (along with appearance of Clara cell
173

-------
5458
5459
5460
5461
5462
5463
5464
5465
5466
5467
5468
5469
5470
5471
5472
5473
5474
5475
5476
5477
5478
5479
5480
5481
5482
5483
5484
5485
5486
5487
5488
5489
5490
5491
5492
5493
5494
5495
vacuolation on days 2 and 9) but was not observed on days 89, 92, and 93 when Clara cell
lesions also were not observed. This suggests that cell proliferation was enhanced in response to
Clara cell damage but was not sustained with repeated exposure to dichloromethane.
Currently, a mechanistic connection has not been established between the acute effects of
dichloromethane on Clara cells in the bronchiolar epithelium and the development of lung
tumors in B6C3Fi mice exposed by inhalation to concentrations >2,000 ppm dichloromethane
for 2 years (NTP, 1986) or for 26 weeks followed by no exposure through 2 years (Maronpot et
al., 1995; Kari et al., 1993). There appears to be a concordance between exposure concentrations
inducing acute Clara cell vacuolation (>2,000 ppm) and those inducing lung tumors
(>2,000 ppm). However, transient acute Clara cell vacuolation does not appear to progress to
necrosis or lead to sustained cell proliferation (which could promote the growth of tumor-
initiated cells) and appears to be dependent on CYP metabolism of dichloromethane (see the
following paragraphs discussing pertinent findings reported by Foster et al. [1994, 1992]). In
contrast, there is consistent and specific evidence for an association between the formation of
DNA-reactive GST-pathway metabolites and the formation of lung and liver tumors from
inhalation exposure (see sections 4.5.2 and 4.7.3).
Foster et al. (1992) noted that the appearance and disappearance of Clara cell vacuolation
in mouse lungs showed concordance with temporal patterns for immunologic staining for
CYP2B1 and 2B2 levels in lung sections. A similar temporal pattern was reported for CYP2B1
and 2B2 monooxygenase activities (ethoxycoumarin O-dealkylation or aldrin epoxidation)
assayed in lung microsomes. When there was a marked decrease in CYP2B1 and 2B2 staining
(e.g., on day 5) or monooxygenase activities, the lesion was not present. Similarly, the
appearance of the lesion was preceded (the day before) by the recovery of monooxygenase
activities or immunologic staining close to control levels. These patterns suggested to Foster et
al. (1992) that Clara cells may have developed tolerance to dichloromethane due to inactivation
of a CYP isozyme.
In subsequent studies, increased percentages of vacuolated bronchiolar epithelium cells
were noted in mice exposed to 2,000 ppm (26.3 ± 6.7%) or 4,000 ppm (64.8 ± 12.8%), but
vacuolated cells were not observed in bronchiolar epithelium of lung sections from control mice
or mice exposed to 125, 250, 500, or 1,000 ppm (Foster et al., 1994). Pretreatment with the CYP
inhibitor, piperonyl butoxide, counteracted the 2,000 ppm effect (2.4 ± 3.6% vacuolated cells),
whereas GSH-depleted mice showed no statistically significant change in percentage of
vacuolated cells (32.7 ± 16.9%) compared with the mean percentage in mice exposed to
2,000 ppm without pretreatment. No consistent, statistically significant, exposure-related
changes were found in cytosolic GST metabolic activities (with dichloromethane as substrate) or
microsomal CYP monooxygenase activities (ethoxycoumarin O-dealkylation), but mean
cytosolic levels of nonprotein sulfhydryl compounds were elevated in lungs of mice exposed to
1,000 and 2,000 ppm (134.6 ± 17.1 and 146.4 ± 6.7 nmol/mg protein, respectively) compared
174

-------
5496
5497
5498
5499
5500
5501
5502
5503
5504
5505
5506
5507
5508
5509
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
5523
5524
5525
5526
5527
5528
5529
5530
5531
5532
5533
with control levels (109.5 ± 7.6 nmol/mg protein). Increased cell proliferation was found in
cultured Clara cells isolated from 4,000 ppm exposed mice compared with nonexposed mice;
respective values for percentage of cells in S phase were 18.97 ±1.18 and 2.02 ± 0.86% (Foster
etal., 1994).
Results from the studies by Foster et al. (1994, 1992) indicate that 6-hour exposures of
B6C3Fi mice to dichloromethane concentrations >2,000 ppm caused transient Clara cell
vacuolation in the bronchiolar epithelium, which was not consistently observed following
repeated exposures. With repeated exposure to 4,000 ppm, the Clara cell vacuolation did not
progress to necrosis, and no hyperplasia of the bronchiolar epithelium was found after up to
13 weeks of exposure. The transient Clara cell vacuolation was decreased by CYP inhibition
with piperonyl butoxide and was unaffected by GSH depletion, indicating that a CYP metabolite
was involved. Clara cell vacuolation was not found after five consecutive daily 6-hour
exposures to 4,000 ppm but reappeared after 2 days without exposure followed by two additional
consecutive daily exposures (day 9). With repeated exposure, the lesion was detected at a
diminished severity on day 44 (but was not found on day 40 or 43) and on day 93 (but was not
found on day 89 or 92). The temporal pattern of Clara cell vacuolation with repeated exposure
was reflected in the occurrence of transiently decreased CYP metabolic activity after the
appearance of vacuolation. Foster et al. (1994, 1992) proposed that the diminishment of severity,
or the disappearance, of the Clara cell vacuolation with repeated exposure was due to the
development of a tolerance to dichloromethane, linked with a decrease of CYP metabolism of
di chl oromethane.
4.5.4. Mechanistic Studies of Neurological Effects
Several neurobehavioral studies (see section 4.4.3 for a complete summary) have
demonstrated that dichloromethane exposure results in decreased spontaneous motor activity
with pronounced lethargy at high concentrations (1,000 ppm or greater). These effects,
combined with the observation that dichloromethane impairs learning and memory (Alexeef and
Kilgore, 1983) and affects production of evoked responses to sensory stimuli (Herr and Boyes,
1997; Rebert et al., 1989), indicate that dichloromethane produces significant neurological
effects. The mechanisms behind producing these changes have been examined by measuring
changes in neurotransmitter levels and changes in neurotransmitter localization. Specific brain
regions (e.g., hippocampus, caudate nucleus, cerebellum) were analyzed to determine if
dichloromethane-induced behavioral effects, such as learning and memory (hippocampus,
caudate nucleus) and movement (cerebellum), are resulting from pathological changes in these
regions. Changes in neurotransmitter levels were also monitored to see if there was any
correlation in behavior and neurochemical changes. Summaries of these studies are provided
below. It is not yet known if dichloromethane directly interacts with neuronal receptors, as has
been demonstrated for toluene and ethanol, two other solvents with neurobehavioral and
175

-------
5534
5535
5536
5537
5538
5539
5540
5541
5542
5543
5544
5545
5546
5547
5548
5549
5550
5551
5552
5553
5554
5555
5556
5557
5558
5559
5560
5561
5562
5563
5564
5565
5566
5567
5568
5569
5570
5571
neurophysiological profiles that are similar to those of dichloromethane (for a review see Bowen
et al. [2006]).
Kanada et al. (1994) examined the effect of dichloromethane on acetylcholine and
catecholamines (dopamine, norepinephrine, serotonin) and their metabolites in the midbrain,
hypothalamus, hippocampus, and medulla from male Sprague-Dawley rats (four to five per
group). The rats were sacrificed 2 hours after a single gavage dose of 0 or 534 mg/kg of
undiluted dichloromethane. Administration of dichloromethane significantly increased the
concentration of acetylcholine in the hippocampus and increased dopamine and serotonin levels
in the medulla. Dichloromethane decreased norepinephrine levels in the midbrain, and
hypothalamus and serotonin levels were decreased in the hypothalamus. There was a trend
toward decreased dopamine in the hypothalamus, but the variability between the animals was so
high that the effect was not significant. The authors speculated that increased acetylcholine
release from dichloromethane administration may be due to decreased acetylcholine release from
the nerve terminals. It is unclear as to how these neurochemical changes could be correlated to
the neurobehavioral changes observed after dichloromethane exposure.
In a 2-week exposure study, male Wistar rats were exposed to dichloromethane at 500 or
1,000 ppm (6 hours/day, 5 days/week for 1 or 2 weeks) or 1,000 ppm TWA (1 hour at 100 ppm,
1	hour peak at 2,800 ppm, 1 hour at 100 ppm, repeated immediately, 5 days/week for 1 or
2	weeks) (Savolainen et al., 1981). Brains were removed from rats at the end of the study and
analyzed. The 1,000 ppm TWA group displayed increases in cerebral RNA. Other changes
noted for this group in the cerebrum included significant increases in NADPH-diaphorase and
succinate dehydrogenase activity. These changes suggest increased neural activity to possibly
offset the overall inhibitory effect of dichloromethane in the CNS. It could also possibly explain
why lethargic effects decrease with continued dichloromethane exposure, and this result
demonstrates a neuroprotective mechanism resulting from dichloromethane exposure. After a
7-day withdrawal, RNA levels in the cerebrum were significantly lower in the 1,000 ppm group.
Succinate dehydrogenase levels remained lowered in the 1,000 ppm TWA group after the 7-day
exposure-free period.
Changes in brain catecholamine levels after a subacute exposure to dichloromethane were
evaluated using male Sprague-Dawley rats (Fuxe et al., 1984). Rats were exposed to 70, 300,
and 1,000 ppm dichloromethane, 6 hours/day for 3 consecutive days. At all exposures, there was
a significant decrease of catecholamine concentrations in the posterior periventricular region of
the hypothalamus. The impact of dichloromethane was also evaluated on the hypothalamic-
pituitary gonadal axis. The hypothalamus regulates secretion of reproductive hormones, such as
follicle-stimulating hormone and luteinizing hormone. The levels of the hormone release were
not significantly changed with dichloromethane exposure. In the caudate nucleus, which is
involved in memory processes, the catecholamine level initially increased (at 70 ppm) and then
was lower (1,000 ppm) in comparison to the control. The study overall demonstrates significant
176

-------
5572
5573
5574
5575
5576
5577
5578
5579
5580
5581
5582
5583
5584
5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
changes in catecholamine levels in the hypothalamus and caudate nucleus. Catecholamine level
changes in the hypothalamus did not have any significant effect on hormonal release and
decreased catecholamine levels in the caudate nucleus at higher exposures may lead to memory
and learning impairment.
A series of studies were conducted in male and female Mongolian gerbils exposed
continuously to >210 ppm dichloromethane for 3 months, followed by a 4-month exposure-free
period (Karlsson et al., 1987; Briving et al., 1986; Rosengren et al., 1986). Decreased DNA
concentrations were noted in the hippocampus at both the 210 and 350 ppm exposures. At
350 ppm, there was also decreased DNA concentration in the cerebellar hemispheres, indicating
a decreased cell density in these regions probably due to cell loss (Rosengren et al., 1986).
These findings indicate that the cerebellum, which is the section of the brain that regulates motor
control, is a target for dichloromethane. In the same study, increased astroglial proteins were
found in the frontal and sensory motor cerebral cortex, which directly correlated to the
astrogliosis that was observed in those areas. Up-regulation of these astroglial proteins is a good
indicator of neuronal injury (Rosengren et al., 1986).
Karlsson et al. (1987) measured DNA concentrations in different regions of the gerbil
brain. The total brain protein concentration per wet weight was not significantly different
between dichloromethane-exposed and control animals. However, DNA concentrations per wet
weight were significantly decreased in the hippocampus after dichloromethane exposure. No
other examined regions demonstrated significant changes in DNA concentrations after
dichloromethane exposure. Therefore, this result indicates that the hippocampus, which plays a
role in the formation of new memories, is another target for dichloromethane in the CNS. This
selective DNA concentration decrease observed in the hippocampus is a sign of neurotoxicity as
noted by the authors and may possibly explain why some studies have noted memory and
learning deficits with dichloromethane exposure.
At a 210 ppm exposure, Briving et al. (1986) observed that dichloromethane decreased
glutamate, y-aminobutyric acid, and phosphoethanolamine levels in the frontal cortex, while
glutamate and y-aminobutyric acid were increased in the posterior cerebellar vermis. Increased
levels of glutamate in the posterior cerebellar vermis could reflect an activation of astrocytic glia,
since glutamine synthetase is localized exclusively in astrocytes.
4.6. SYNTHESIS OF MAJOR NONCANCER EFFECTS
4.6.1. Oral Exposures
4.6.1.1. Summary of Human Data
Information on noncancer effects in humans exposed orally to dichloromethane are
restricted to case reports of neurological impairment, liver and kidney effects (as severe as organ
failure), and gastrointestinal irritation in individuals who ingested amounts ranging from about
25 to 300 mL (Chang et al., 1999; Hughes and Tracey, 1993). Neurological effects with these
177

-------
5610
5611
5612
5613
5614
5615
5616
5617
5618
5619
5620
5621
5622
5623
5624
5625
5626
5627
5628
5629
5630
5631
5632
5633
5634
5635
5636
5637
5638
5639
5640
5641
5642
5643
5644
5645
5646
5647
individuals consisted of general CNS depressive symptoms, such as drowsiness, confusion,
headache, and dizziness. Hemoglobinuria has been noted as a kidney effect associated with
ingestions.
4.6.1.2. Summary of Animal Data
Acute oral or intraperitoneal administration of dichloromethane in animals has resulted in
several significant effects. General activity and function were affected as evidenced by
decreased neuromuscular activity (Moser et al., 1995). Additionally, decreased sensorimotor
function was detected through measurement of evoked potentials (Herr and Boyes, 1997) and by
using the FOB (Moser et al., 1995). Neurochemical changes (e.g. acetylcholine, dopamine,
norepinephrine, serotonin) were detected 2 hours after oral dosage of dichloromethane within
specific parts of the brain. It should be noted that all the acute effects that were observed after
oral or intraperitoneal administration occurred within 5 hours after dosage. No other significant
organ effects were noted after a single acute oral exposure, but in oral pharmacokinetic studies it
is known that dichloromethane is primarily distributed to the liver, lungs, and kidneys (Angelo et
al., 1986a).
Results from short-term, subchronic, and chronic oral toxicity studies in laboratory
animals are summarized in Table 4-35. The data indicate that rats may be more sensitive than
mice to nonneoplastic liver effects from orally administered dichloromethane, as evidenced by
lower NOAELs and LOAELs, with more severe liver effects in rats. The most frequently
observed liver effect was hepatocyte vacuolation, seen with drinking water exposure (90 days) in
F344 rats at >166 mg/kg-day and B6C3Fi mice at 586 mg/kg-day (Kirschman et al., 1986) and
with gavage exposure (14 days) in CD-I mice at 333 mg/kg-day (Condie et al., 1983).
Hepatocyte degeneration or necrosis was observed in female F344 rats exposed in drinking water
for 90 days to 1,469 mg/kg-day (Kirschman et al., 1986) and in female F344 rats exposed by
gavage for 14 days to 337 mg/kg-day (Berman et al., 1995), but was not seen in a 90-day
drinking water study in B6C3Fi mice exposed to doses as high as 2,030 mg/kg-day (Kirschman
et al., 1986). In the chronic-duration (2-year) study, liver effects were described as
nonneoplastic foci and areas of alteration in F344 rats exposed to drinking water doses between
50 and 250 mg/kg-day; an increased incidence of fatty changes in the liver was also noted but the
incidence of the latter was not provided (Serota et al., 1986a). These effects were considered to
be nonneoplastic for several reasons. Serota et al., (1986b) observed a dose-related increased
incidence of 0, 65, 92, 97, 98 and 100% in male rats and 51, 41, 73, 89, 91 and 85% in female
rats for the 0, 5, 50, 125, 250 and 250 with recovery groups, respectively. Evidence for liver
tumors has been reported in female rats only. Specifically, evidence for liver tumors in rats
includes a small number of hepatocellular carcinomas observed in female rats at 50 and 250
mg/kg-day, which reached statistical significance (for trend and for individual pairwise
comparisons) only with the combined grouping of neoplastic nodules and hepatocellular
178

-------
5648
5649
5650
5651
5652
5653
5654
5655
5656
5657
5658
5659
5660
5661
5662
5663
5664
5665
5666
5667
5668
carcinomas. In male rats, only one hepatocellular carcinoma was observed in all of the exposure
groups (compared to 4 in the controls), and the incidence of neoplastic nodules and
hepatocellular carcinomas was higher in controls (16%) than in any exposure group (16, 3, 0, 6,
5 and 13% for the 0, 5, 50,125, 250 mg/kg-day and 250 with recovery groups, respectively). The
authors (Serota et al., 1986a) did not elaborate on the characterization of the altered foci.
However, the characterization of altered foci could range from a focal change in fat distribution
(nonneoplastic effect) to enzyme altered foci which are generally considered a precursor to
tumor formation (Goodman et al., 1994). Serota et al (1986a) reported an increased incidence of
fatty change in the liver at doses of 50 mg/kg-day and higher, but the incidence was not reported.
In addition, a 90-day study (Kirschman et al., 1986) demonstrated that increased fatty deposits
were present in the hepatocyte vacuoles. Therefore, the altered foci (i.e. change in fat
distribution) observed by Serota et al., (1986b) may represent a precursor to fatty liver changes
which is considered a nonneoplastic effect. Taken together, the data support the conclusion that
the altered foci were nonneoplastic.
The NOAEL and LOAEL, 101 and 337 mg/kg-day, for altered neurological functions in
female F344 rats (as reported by Moser et al. [1995]) were identical to those reported by Berman
et al. (1995) for hepatocyte necrosis in female F344 rats. In the 90-day (Kirschman et al., 1986)
and 104-week (Serota et al., 1986a, b) drinking water studies, no obvious clinical signs of
neurological impairment were observed in rats or mice at exposure levels that induced liver
effects (see Table 4-35), but this study did not include a standardized neurological testing
battery.
179

-------
Table 4-35. NOAELs and LOAELs in selected animal studies involving oral exposure to dichloromethane
for short-term, subchronic, or chronic durations
Type of effect and
exposure, reference
Species and exposure details
Results
NOAEL LOAEL
(mg/kg-day)
Hepatic, 14-day gavage
Bermanetal. (1995)
Condie et al. (1983)
F344 rat, female, 8/dose group
0, 34, 101, 337, 1,012 mg/kg-day
CD-I mouse, male, 5/group for
histological examinations; 8/group for
blood urea nitrogen, serum creatinine,
and serum glutamate-pyruvate
transaminase; 0, 133, 333, 665 mg/kg-
day
Hepatic, 90-day drinking water
Kirschman et al. (1986) F344 rat, male and female; 15/sex/group;
males 0, 166, 420, 1,200 mg/kg-day
females 0, 209, 607, 1,469 mg/kg-day
Kirschman et al. (1986) B6C3Fi mouse, male and female,
males 0, 226, 587, 1,911 mg/kg-day
females 0, 231, 586, 2,030 mg/kg-day
Hepatic, 104-week drinking water
Serota et al. (1986a)
Serota et al. (1986b)
Hazelton Laboratories
(1983)
Neurologic, 14 day
Moseretal. (1995)
Reproductive
General Electric Co.
(1976)
F344 rat, male and female,
0, 5, 50, 125, 250 mg/kg-day
B6C3Fi mouse, male and female,
0, 60, 125, 185, 250 mg/kg-day
F344 rat, female,
0, 34, 101, 337, 1,012 mg/kg-day
Charles River CD rat, male and female,
gavage for 90 d before mating (10 days
between last exposure and mating
period); 0, 25, 75, 225 mg/kg-day; F1
offspring received same treatment as
parents for 90 days
Hepatocyte necrosis
Hepatocyte vacuolation (minimal to mild in 3/5)
101
133
Hepatic vacuolation (generalized, centrilobular, or
periportal, at lowest dose, in 10/15 males and 13/15 females
compared with 1/15 males and 6/15 females in controls)
Hepatic vacuolation (increased severity of centrilobular fatty
change in mid- and high-dose groups compared with
controls)
Liver foci/areas of alteration (considered nonneoplastic
histologic changes); fatty liver changes also seen at same
doses but incidence data not reported; no evidence that
increased altered foci progresses to liver tumor formation
Some evidence of fatty liver; marginal increase in the Oil
Red-O-positive material in the liver
FOB 24 hours postexposure: altered autonomic,
neuromuscular, and sensorimotor and excitability measures
Reproductive performance of F0 and histologic examination
of tissues from F1 offspring
231
101
337
333
Not	166
identified
586
50
185	250
337
225	Not
identified
180

-------
Table 4-35. NOAELs and LOAELs in selected animal studies involving oral exposure to dichloromethane
for short-term, subchronic, or chronic durations
Type of effect and
exposure, reference
Species and exposure details
Results
NOAEL LOAEL
(mg/kg-day)
Raje et al. (1988)
Developmental
Narotsky and Kavlock
(1995)
Other developmental
No studies
Swiss-Webster mouse, male, 0, 250,
500 mg/kg (subcutaneous injection),
3x per week, 4 weeks prior to mating
with nonexposed females (1 week
between last exposure and mating period)
No statistically significant effects on testes, number of
litters, live fetuses/litter, percent dead fetuses/litter, percent
resorbed/litter, or fertility index
F344 rat, pregnant female, gavage on Maternal: weight gain depression
GDs 6-19; 0, 338, 450 mg/kg-day
Fetal: no effects on pup survival, resorptions, pup weight
200	Not
identified
338
450
450	Not
identified
181

-------
5669
5670
5671
5672
5673
5674
5675
5676
5677
5678
5679
5680
5681
5682
5683
5684
5685
5686
5687
5688
5689
5690
5691
5692
5693
5694
5695
5696
5697
5698
5699
5700
5701
5702
5703
5704
5705
5706
Results from a limited number of studies do not provide evidence for effects on
reproductive or developmental endpoints (Table 4-35). No effects on pup survival, resorptions,
or pup weight were found following exposure of pregnant F344 rats to doses as high as
450 mg/kg-day on GDs 6-19, a dose that depressed maternal weight gain (Narotsky and
Kavlock, 1995), and no effects on reproductive performance endpoints (fertility index, number
of pups per litter, pup survival) were found in Charles River CD rats exposed for 90 days before
mating to doses as high as 225 mg/kg-day. There are no oral exposure studies focusing on
neurobehavioral effects or other developmental outcomes.
4.6.2. Inhalation Exposures
4.6.2.1. Summary of Human Data
As discussed in sections 4.1.3.1 and 4.1.3.2, acute inhalation exposure of humans to
dichloromethane has been associated with cardiovascular impairments due to decreased oxygen
availability from COHb formation and neurological impairment from interaction of
dichloromethane with nervous system membranes. Results from studies of acutely exposed
human subjects indicate that acute neurobehavioral deficits, measured, for example, by
psychomotor tasks, tests of hand-eye coordination, visual evoked response changes, and auditory
vigilance, may occur at concentrations >200 ppm with 4-8 hours of exposure (Bos et al., 2006;
American Conference of Governmental Industrial Hygienists [ACGIH], 2001; ATSDR, 2000;
Cherry et al., 1983; Putz et al., 1979; Gamberale et al., 1975; Winneke, 1974).
The clinical and workplace studies of noncancer health effects of chronic
dichloromethane exposure have examined markers of disease and specific clinical endpoints
relating to cardiac, neurological disease, hepatic function, and reproductive health. As
summarized in section 4.1.2.9, the limited available data do not provide evidence of cardiac
damage related to dichloromethane exposure in occupationally exposed workers (Hearne and
Pifer, 1999; Tomenson et al., 1997; Gibbs et al., 1996; Lanes et al., 1993; Ott et al., 1983d;
Cherry et al., 1981). Relatively little is known about the long-term neurological effects of
chronic exposures, although there are studies that provide some evidence of an increased
prevalence of neurological symptoms among workers with average exposures of 75-100 ppm
(Cherry et al., 1981), long-term effects on some neurological measures (i.e., possible detriments
in attention and reaction time in complex tasks) in workers whose past exposures were in the
100-200 ppm range (Lash et al., 1991), and an increased risk of suicide in worker cohort studies
(Hearne and Pifer, 1999; Gibbs, 1992). Given the suggestions from these studies and their
limitations (particularly with respect to sample size and power considerations), the statement that
there are no long-term neurological effects of chronic exposures to dichloromethane cannot be
made with confidence. With respect to markers of hepatic damage, three studies measured
serum enzyme and bilirubin levels in workers exposed to dichloromethane (Soden, 1993;
Kolodner et al., 1990; Ott et al., 1983c). There is some evidence of increasing levels of serum
182

-------
5707
5708
5709
5710
5711
5712
5713
5714
5715
5716
5717
5718
5719
5720
5721
5722
5723
5724
5725
5726
5727
5728
5729
5730
5731
5732
5733
5734
5735
5736
5737
5738
5739
5740
5741
5742
5743
5744
bilirubin with increasing dichloromethane exposure (Kolodner et al., 1990; Ott et al., 1983c), but
there are no consistent patterns with respect to the other hepatic enzymes examined (serum y-
glutamyl transferase, serum AST, serum ALT). Thus these studies do not provide clear evidence
of hepatic damage in dichloromethane exposed workers to the extent that this damage could be
detected by these serologic measures.
Only limited, and somewhat indirect, evidence pertaining to immune-related effects of
dichloromethane in humans is available. No risk of the broad category of infection- and parasite-
related mortality was reported by Hearne and Pifer (1999), but there was some evidence of an
increased risk of influenza and pneumonia-related mortality at two cellulose triacetate fiber
production work sites in Maryland and South Carolina (Gibbs, 1992).
Few studies have been conducted pertaining to reproductive effects (i.e., spontaneous
abortion, low birth weight, or oligospermia) of dichloromethane exposure from workplace
settings (Wells et al., 1989; Kelly, 1988; Taskinen et al., 1986) or environmental settings (Bell et
al., 1991). Of these, the data pertaining to spontaneous abortion provide the strongest evidence
of an adverse effect of dichloromethane exposure. The limitations of the only study (Taskinen et
al., 1986) pertaining to this outcome, however, do not allow firm conclusions to be made
regarding dichloromethane and risk of spontaneous abortion in humans.
4.6.2.2. Summary of Animal Studies
Acute and short-term (up to 7 days) inhalational exposure to dichloromethane has
resulted in neurological and hepatocellular changes. Several neurological-mediated parameters
were reported, including decreased activity (Kjellstrand et al., 1985; Weinstein et al., 1972;
Heppel and Neal, 1944), impairment of learning and memory (Alexeef and Kilgore, 1983), and
changes in responses to sensory stimuli (Rebert et al., 1989). Although learning and memory
properties were impaired in one acute exposure (47,000 ppm until loss of righting reflex), it
should be noted that this effect has not been characterized by using other learning and memory
tasks nor any other exposure paradigms. In a 3-day exposure to dichloromethane (70, 300, or
1,000 ppm 6 hours/day), it was found that in the rat brain there were changes in catecholamine
(dopamine, serotonin, norepinephrine) in the hypothalamus and caudate nucleus (Fuxe et al.,
1984). The catecholamine level changes did not affect hormonal release, which is a primary
function of the hypothalamus.
Another acute exposure study examined immunological response as measured by
increased streptoccocal pneumonia-related mortality and decreased bactericidal activity of
pulmonary macrophages in CD-I mice following a single 3-hour exposure to dichloromethane at
100 ppm (Aranyi et al., 1986). No effects were seen at 50 ppm. A 4-week inhalation exposure
to 5,000 ppm dichloromethane in rats did not result in changes in immune response as measured
by the sheep red blood cell assay (Warbrick et al., 2003). These studies suggest a localized,
portal-of-entry effect within the lung, without evidence of systemic immunosuppression.
183

-------
5745	Mouse hepatocytes showed balloon degeneration (dissociation of polyribosomes and
5746	swelling of rough endoplasmic reticulum) within 12 hours of exposure to 5,000 ppm (Weinstein
5747	et al., 1972). A subacute exposure in Wistar rats to 500 ppm dichloromethane 6 hours/day for 6
5748	days resulted in increased hemochrome content in liver microsomal CYP (Savolainen et al.,
5749	1977).
5750	Results pertaining to liver, lung, and neurological effects from longer (>7 days)
5751	subchronic and chronic inhalation toxicity studies in laboratory animals are summarized in
5752	Table 4-36; reproductive and developmental studies are summarized in Table 4-37.
184

-------
Table 4-36. NOAELs and LOAELs in animal studies involving inhalation exposure to dichloromethane for
subchronic or chronic durations, hepatic, pulmonary, and neurologic effects
Type of effect and
exposure period,
reference
NOAEL
LOAEL
Species and exposure details
Results
ppm
Hepatic, subchronic (13-14 weeks)
Haun et al. (1971)	Beagle, female (n = 8);
Haunetal. (1972)
Leuscher et al. (1984)
NTP (1986)
NTP (1986)
rhesus monkeys, females (n = 4);
Sprague-Dawley rats, male (n = 20),
ICR mice, females (n = 380)
0, 1,000, 5,000 ppm (continuous
exposure; 14 weeks)
Beagle (n = 16);
Rhesus monkey (n = 4);
Sprague-Dawley rats (n = 20),
ICR mice (n = 20)
0, 25, 100 ppm (continuous exposure;
14 weeks)
Sprague-Dawley rat, male and female,
(20/sex/group) - 0, 1,000 ppm (6
hours/day; 90 days)
Beagle, male and female (3/sex/group) ¦
0, 5,000 ppm
F344/N rat, male and female
(10/sex/group)
0, 525, 1,050, 2,100, 4,200, 8,400 ppm
(6 hours/day, 5 days/week, 13 weeks)
B6C3Fi mouse, male and female
(10/sex/group)
0, 525, 1,050, 2,100, 4,200, 8,400 ppm
(6 hours/day, 5 days/week, 13 weeks)
Hepatic, 2 years (6 hours/day, 5 dayAveek)
Mennear et al. (1988); F344/N rat, male and female
NTP (1986)	0, 1,000, 2,000, 4,000 ppm
Fatty liver at 1,000 ppm in dogs, "borderline" liver
changes in monkey at 5,000 ppm, mottled liver
changes in rats at 5,000 ppm
Decreased movement and lethargy at 1,000 ppm in
dogs, mice, and monkey.
CYP levels decreased in liver microsomes in mice
at 100 ppm
Increased fatty liver content at 25, 100 ppm in rats
No liver effects noted
Not identified
Not identified
1,000
Not identified
25;
Not identified
Not identified
25
Decreased lipid:liver weight ratios at
4,200 (females); 8,400 (males); decreased BWby
23% and 11% in males and females at 8,400 ppm
compared with controls; one male and one female
died at 8,400 ppm before the end of the study.
Hepatocyte centrilobular degeneration at 4,200
females) and 8,400 (males); decreased lipid:liver
weight ratios at 8,400 (females); at 8,400 ppm, 4/10
males and 2/10 females died before end of study.
Hepatocyte vacuolation and necrosis
Hemosiderosis in liver
Renal tubular degeneration
1,000
5,000
4,200
2,100
Not identified
Not identified
1,000
1,000 dog
1,000 monkey
5,000 rat
1,000 mouse
100 dog
25 monkey
25 rat
100 mouse
Not identified
Not identified
8,400
4,200
1,000
1,000
2,000
185

-------
Table 4-36. NOAELs and LOAELs in animal studies involving inhalation exposure to dichloromethane for
subchronic or chronic durations, hepatic, pulmonary, and neurologic effects
Type of effect and
exposure period,
reference
NOAEL
LOAEL
Species and exposure details
Results
ppm
Mennear et al. (1988);
NTP (1986)
Burek et al. (1984)
Burek et al. (1984)
Nitschke et al. (1988a)
B6C3Fi mouse, male and female
0, 2,000, 4,000 ppm
Syrian golden hamster, male and female
0, 500, 1,500, 3,500 ppm
Sprague-Dawley rat, male and female
0, 500, 1,500, 3,500 ppm
Sprague-Dawley rat, male and female
0, 50, 200, 500 ppm
Pulmonary, 13 weeks (6 hours/day, 5 days/week)
NTP (1986)
Foster et al. (1992)
Neurological, 14 days
Savolainen et al. (1981)
F344 rat, male and female
0, 525, 1,050, 2,100, 4,200, 8,400 ppm
B6C3Fi mouse, male and female
0, 4,000 ppm
Wistar rats, male
500, 1,000, 1,000 TWA (100 + 2,800
1-hour peaks3) ppm (6 hours/day,
5 days/week, 2 weeks)
Neurological, 13-14 weeks
Mattsson et al. (1990) F344 rat, male and female
0, 50, 200, 2,000 ppm
(6 hours/day, 5 days/week)
Haunetal. (1971)
Beagle dogs (female);
Rhesus monkeys (female);
Sprague-Dawley rats (male);
ICR mice (females)
0, 1,000, 5,000 ppm
(continuous exposure)
Hepatocyte degeneration
Renal tubule casts
No effects on histologic, clinical chemistry,
urinalytic, and hematologic variables no obvious
clinical signs of toxicity
Hepatocyte vacuolation (M and F)
Hepatocyte necrosis (M only), no obvious clinical
signs of toxicity)
Hepatocyte vacuolation significantly increased in
females; non-significant increase in males at 500
ppm (31% in controls and 40% in 500 ppm group).
Foreign body pneumonia
Clara cell vacuolation
Not identified 2,000
Not identified 2,000
3,500
Not identified
500
200
4,200
Increased RNA in cerebrum at 1,000 ppm;
increased enzymatic activities'3 in cerebrum and
cerebellum at 1,000 ppm TWA
500
No exposure-related effects on an observational 2,000
battery, hind-limb grip strength, a battery of evoked
potentials, or histology of brain, spinal cord,
peripheral nerves; measured 64 hours postexposure
CNS depression most evident in dogs	Not identified
Not identified
1,000
Not identified
Not identified
500
1,500
500
;,400
Not identified 4,000
1,000 for brain
RNA
concentration;
1,000 TWA for
brain enzymatic
activity
Not identified
1,000 dog
1,000 monkey
5,000 rat
1,000 mouse
186

-------
Table 4-36. NOAELs and LOAELs in animal studies involving inhalation exposure to dichloromethane for
subchronic or chronic durations, hepatic, pulmonary, and neurologic effects
Type of effect and
exposure period,
reference
NOAEL
LOAEL
Species and exposure details
Results
ppm
Karlsson et al. (1987)
Briving et al. (1986)
Rosengren et al. (1986)
Thomas et al. (1972)
CoHb, 13-14 weeks
Haun et al. (1972)
Mongolian gerbils, male and female
210, 350, 700 ppm (continuous exposure,
followed by 4 month exposure-free
period)
ICR mice, female
0, 25, 100 ppm, continuous
Beagles (n = 16);
Rhesus monkeys (n = 4);
Sprague-Dawley rats (n = 20),
ICR mice (n = 20)
0, 25, 100 ppm (continuous exposure;
14 weeks)
COHb, 2 years (6 hours/day, 5 day/week)
Syrian golden hamster, male and female
0, 500, 1,500, 3,500 ppm
Sprague-Dawley rat, male and female
0, 500, 1,500, 3,500 ppm
Sprague-Dawley rat, male and female
0, 50, 200, 500 ppm
Burek et al. (1984)
Burek et al. (1984)
Nitschke et al. (1988a)
Astrogliosis in frontal and sensory motor cerebral Not identified 210
cortex suggested by increases in astroglial proteins;
cell loss in cerebellar regions; decreased DNA in
hippocampus; neurochemical changes observed at
all exposures
Increased spontaneous activity observed at 25 ppm Not identified 25
but not 100 ppm
CoHb levels significantly higher at 25, 100 ppm Not identified 25
for monkeys and 100 ppm for beagles
About 30% COHb in each exposed group
About 12-14% COHb in each exposed group
COHb values at 2 years: about 2, 7, 13, 14%
Equivalent to 1,000 ppm TWA.
''Decreased GSH, y-aminobutyric acid, and phosphoethanolamine in frontal cortex; GSH and y-aminobutyric acid increased in posterior cerebellar vermis.
5753
5754
5755
5756
187

-------
5757
Table 4-37. NOAELs and LOAELs in selected animal studies involving inhalation exposure to
dichloromethane, reproductive and developmental effects
Type of effect and exposure
period, reference
NOAEL
LOAEL
Species and exposure details
Results
ppm
Nitschke et al. (1988b)
Mennear et al. (1988); NTP
(1986)
Raje et al. (1988)
Schwetz et al. (1975)
Schwetz et al. (1975)
Reproductive
F344 rat, male and female, FO: 6 hr/d,
5	d/wk for 14 wk before mating and
GDs 0 to 21; Fl: 6 hr/d, 5 d/wk,
beginning PND 4 for 17 wk before
mating; 0, 100, 500, 1,500 ppm
B6C3Fi mouse; 0, 2,000 or 4,000 ppm,
6	hours/day, 5 days/week for 2 years
Swiss-Webster mouse, male, 2 hr/d,
5 d/wk for 6 wk before mating with
nonexposed females; 0, 100, 150,
200 ppm
No statistically significant effects on fertility
or litter size, neonatal survival, growth rates,
or histopathologic lesions in Fl or
F2 weanlings
Testicular atrophy
Ovarian atrophy (considered secondary to
hepatic effects)
No statistically significant effects on testes,
number of litters, live fetuses/litter, percent
dead fetuses/litter, percent resorbed/litter
Swiss-Webster mouse, pregnant female,
7 hr/d, GDs 6-15; 0, 1,250 ppm
Sprague-Dawley rat, pregnant female,
7 hr/d, GDs 6-15; 0, 1,250 ppm
Maternal effects: 9-10% COHb; increased
absolute, not relative, liver weight, increased
maternal weight (11-15%).
Fetal effects: increased litters with extra
center of ossification in sternum
Maternal: 9-10% COHb; increased absolute,
not relative, liver weight
Fetal: increased incidence of delayed
ossification of sternebrae
Other developmental
Bornschein et al. (1980); Hardin Long-Evans rat, female, 6 hr/d for 12-
and Manson (1980)	14 d before breeding and GDs 1-17;
6 hr/d on GDs 1-15; 0, 4,500 ppm
1,500
Not identified
2,000
Not identified
200
Fertility index was lower in 150 and 200 ppm
groups (80%) compared with controls and
100 ppm groups (95%) (statistical
significance depends on test used).
Developmental
100
4,000
2,000
Not identified
150
Not identified 1,250
1,250
Not identified
Not identified 1,250
1,250
Not identified
Maternal (both protocols): increased absolute
and relative liver weight (-10%)
Fetal/offspring: decreased fetal BW (-10%);
changed behavioral habituation to novel
environments; no changes in gross, skeletal,
or soft-tissue anomalies
Not identified 4,500
Not identified 4,500
188

-------
5758
5759
5760
5761
5762
5763
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
5790
5791
5792
5793
5794
5795
Hepatic centrilobular degeneration was observed in several studies containing different
species and inhalational exposures. This effect was observed in guinea pigs exposed to
5,000 ppm (7 hours/day) for 6 months (Heppel et al., 1944). Monkeys, rats, and mice
continuously exposed (24 hours/day) to 5,000 ppm dichloromethane for 14 weeks also had
increased centrilobular degeneration (Haun et al., 1972, 1971). This effect was also observed at
lower exposures when mice were exposed to 4,200 ppm for 6 hours/day for 13 weeks (NTP,
1986) and in dogs exposed to 1,000 ppm for 24 hours/day for up to 14 weeks (Haun et al., 1972,
1971).
Increased incidences of histologic hepatic lesions were not found in F344 rats exposed to
4,200 or 8,400 ppm 6 hours/day for 13 weeks (NTP, 1986) or in Sprague-Dawley rats exposed to
10,000 ppm 6 hours/day for 90 days (Leuschner et al., 1984). Hepatic lesions were also not
observed in beagle dogs exposed to 5,000 ppm 6 hours/day for 90 days (Leuschner et al., 1984)
or in dogs, monkeys, rats, and mice exposed to 25 or 100 ppm for 24 hours/day for up to
14 weeks (Haun et al., 1972). Heppel et al. (1944) also demonstrated absence of hepatic lesions
in unspecified strains of monkeys, rabbits, and rats exposed to 10,000 ppm 4 hours/day for up to
8 weeks and in unspecified strains of dogs, rabbits, and rats exposed to 5,000 ppm 7 hours/day
for up to 6 months.
Gross neurological impairments were observed in several laboratory species with
repeated exposure to 10,000 ppm for 4 hours/day for 8 weeks (Heppel et al., 1944) or to 1,000 or
5,000 ppm for 24 hours/day for 14 weeks (Haun et al., 1972, 1971). Dogs exposed to 5,000 ppm
6	hours/day for 90 days showed slight sedation during exposures, but Sprague-Dawley rats
exposed to 10,000 ppm for 90 days did not (Leuschner et al., 1984). In F344 rats exposed to
concentrations up to 2,000 ppm 6 hours/day for 13 weeks, no effects were observed on an
observational battery, hind-limb grip strength, a battery of evoked potentials, or histology of the
brain, spinal cord, or peripheral nerves; these tests were conducted beginning 65 hours or more
after the last exposure (Mattsson et al., 1990).
Exposure-related nonneoplastic effects on the lungs reported in the subchronic studies
were restricted to foreign body pneumonia in rats exposed to 8,400 ppm 6 hours/day for
13 weeks (NTP, 1986), Clara cell vacuolation in mice exposed to 4,000 ppm 6 hours/day for
13 weeks (Foster et al., 1992), and pulmonary congestion in guinea pigs exposed to 5,000 ppm
7	hours/day for 6 months (Heppel et al., 1944).
The chronic duration inhalation studies were conducted at lower exposure levels than the
short-term and subchronic studies and provide results indicating that the liver is the most
sensitive target for noncancer toxicity in rats and mice (Table 4-36). Life-time exposure was
associated with hepatocyte vacuolation and necrosis in F344 rats exposed to 1,000 ppm
6 hours/day (Mennear et al., 1988; NTP, 1986), hepatocyte vacuolation in Sprague-Dawley rats
exposed to 500 ppm 6 hours/day (Nitschke et al., 1988a; Burek et al., 1984), and hepatocyte
degeneration in B6C3Fi mice exposed to 2,000 ppm 6 hours/day (lower concentrations were not
189

-------
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
5814
5815
5816
5817
5818
5819
5820
5821
5822
5823
5824
5825
5826
5827
5828
5829
5830
5831
5832
5833
tested in mice) (Mennear et al., 1988; NTP, 1986). As shown in Tables 4-36 and 4-37, other
effects observed include renal tubular degenerations in F344 rats and B6C3Fi mice at 2,000 ppm,
testicular atrophy in B6C3Fi mice at 4,000 ppm, and ovarian atrophy in B6C3Fi mice at
2,000 ppm (considered secondary to hepatic effects). No exposure-related increased incidences
of nonneoplastic lung lesions were found in any of the chronic studies (Table 4-36).
In comparison to rats and mice, Syrian golden hamsters are less sensitive to the chronic
inhalation toxicity of dichloromethane. No exposure-related changes were found in
comprehensive sets of histologic, clinical chemistry, urinalytic, and hematologic variables
measured in hamsters exposed for 2 years to 500, 1,500, or 3,500 ppm for 6 hours/day, with the
exception that mean COHb percentages were about 30% in each of these groups compared with
a mean value of about 3% for the controls (Burek et al., 1984).
The reproductive and developmental studies are limiting in terms of the exposure
regimen used, with two of the developmental studies using only a single, relatively high daily
exposure over the gestational period (1,250 ppm, GD 6-15 in Schwetz et al. [1975] and 4,500
ppm, GD 1-17 in Hardin and Manson [1980] and Bornschein [1980]). No significant effects on
reproductive performance variables were found in a two-generation reproduction assay with
F344 rats exposed to concentrations as high as 1,500 ppm (Nitschke et al., 1988b). No effects on
most of the measures of reproductive performance were observed in male mice exposed to
200 ppm for 2 hours/day for 6 weeks before mating to nonexposed females. Fertility index was
reduced in the 150 and 200 ppm groups, but the statistical significance of this effect varied
considerably depending on the statistical test used in this analysis (Raje et al., 1988). No adverse
effects on fetal development were found following exposure of pregnant Swiss-Webster mice or
Sprague-Dawley rats to 1,250 ppm 6 hours/day on GDs 6-15 (Schwetz et al., 1975). Following
exposure of female Long-Evans rats to 4,500 ppm (6 hours/day) for 14 days before breeding plus
during gestation or during gestation alone, a 10% decrease in fetal BW and changed behavioral
habituation of the offspring to novel environments were seen (Bornschein et al., 1980; Hardin
and Manson, 1980). No exposure-related changes in gross, skeletal, or soft-tissue anomalies
were found.
4.6.3. Mode of Action Information
4.6.3.1. Mode of Action for Nonneoplastic Liver Effects
Studies of chronically exposed rats, both by the oral route and the inhalation route,
identified liver changes as the most sensitive exposure-related noncancer effect associated with
exposure to dichloromethane (Tables 4-35 to 4-37). The liver changes included increased
incidence of liver foci/areas of alteration and hepatocyte vacuolation in rats and degenerative
liver effects in rats, guinea pigs, monkeys, and mice.
The mode of action by which dichloromethane induces these nonneoplastic hepatic
effects is unknown. The determination of whether or not these effects are due to the parent
190

-------
5834
5835
5836
5837
5838
5839
5840
5841
5842
5843
5844
5845
5846
5847
5848
5849
5850
5851
5852
5853
5854
5855
5856
5857
5858
5859
5860
5861
5862
5863
5864
5865
5866
5867
5868
5869
5870
5871
material, metabolites of the CYP2E1 pathway, metabolites of the GST pathway, or some
combination of parent material and metabolites has not been elucidated. The available data
indicate that rats may be more sensitive than mice to the noncancer hepatotoxicity, but a
mechanistic explanation of this possible interspecies difference is not currently available.
4.6.3.2.	Mode of Action for Nonneoplastic Lung Effects
Single 6-hour inhalation exposures to concentrations >2,000 ppm dichloromethane
produced a transient vacuolation of Clara cells in the bronchiolar epithelium of B6C3Fi mice.
Vacuolization of the Clara cells disappeared or was diminished with repeated exposure and was
correlated with subsequent transient diminishment of CYP metabolic activity. CYP inhibition
with piperonyl butoxide counteracted the vacuolation observed in the Clara cells (Foster et al.,
1994, 1992). With repeated exposure to 4,000 ppm (up to 13 weeks), the Clara cell vacuolation
did not appear to progress to necrosis, and no hyperplasia of the bronchiolar epithelium was
found. Foster et al. (1994, 1992) proposed that the diminished severity or disappearance of Clara
cell vacuolation with repeated exposure was due to the development of tolerance to
dichloromethane, linked with a transient decrease of CYP metabolism of dichloromethane. The
available data suggests that CYP metabolism of dichloromethane may be involved in the mode
of action for the acute effects of dichloromethane on the bronchiolar epithelium of mice.
Mode of action research attention on lung effects from chronic exposure to
dichloromethane has focused on neoplastic effect; nonneoplastic lung effects have received
relatively little attention. No exposure-related increased incidences of nonneoplastic lung lesions
(including epithelial hyperplasia) were found in any of the chronic studies listed in Table 4-36,
but chronic inhalation exposure of B6C3Fi mice to concentrations >2,000 ppm has consistently
been shown to induce lung tumors in several studies (Kari et al., 1993; NTP, 1986). In a study
that included interim sacrifices at 13, 26, 52, 68, 75, 78, 83, and 91 weeks of B6C3Fi mice
exposed to 2,000 ppm, hyperplasia of lung epithelium (the only nonneoplastic lung lesion found)
was found in only three of the eight interim sacrifices (68, 78, and 91 weeks) and was only
statistically significantly elevated at 91 weeks (5/30 versus 0/15 in controls) (Kari et al., 1993).
4.6.3.3.	Mode of Action for Neurological Effects
Results from studies of acutely exposed human subjects indicate that mild
neurobehavioral deficits may occur at air concentrations >200 ppm with 4-8 hours of exposure
(Bos et al., 2006; ACGIH, 2001; AT SDR, 2000; Cherry et al., 1983; Putz et al., 1979; Gamberale
et al., 1975; Winneke, 1974). Acute high-dose exposures also resulted in gross neurological
impairments in several laboratory species (Haun et al., 1972, 1971; Heppel et al., 1944).
Exposure of F344 rats to concentrations up to 2,000 ppm 6 hours/day for 13 weeks produced no
effects on an observational battery, hind-limb grip strength, a battery of evoked potentials, or
histology of the brain, spinal cord, or peripheral nerves (Mattsson et al., 1990). However, oral
191

-------
5872
5873
5874
5875
5876
5877
5878
5879
5880
5881
5882
5883
5884
5885
5886
5887
5888
5889
5890
5891
5892
5893
5894
5895
5896
5897
5898
5899
5900
5901
5902
5903
5904
5905
5906
5907
5908
exposures have been shown to alter autonomic, neuromuscular, and sensorimotor functions have
been observed in F344 rats exposed to gavage doses >337 mg/kg-day for 14 days (Moser et al.,
1995).
Dichloromethane may be metabolized by the CYP2E1 enzyme to CO (Guengerich, 1997;
Hashmi et al., 1994; Gargas et al., 1986). Many of the acute human exposure studies evaluated if
CO was the primary metabolite responsible for producing the CNS depressant effects observed
during dichloromethane exposure. Overall, at lower exposures and acute durations, it appears
that CO is the primary mediator of the neurobehavioral effects. Putz et al. (1979) demonstrated
that similar neurobehavioral deficits were present when an equivalent COHb blood level (and
CO exposure) was achieved between CO and dichloromethane exposures. Incidentally, after a
longer duration, neurobehavioral deficits are more pronounced with dichloromethane exposure in
comparison to CO exposure alone. This additional increase in the CNS depressive effects is
most likely due to the saturation of the CYP2E1 metabolic pathway. In humans, saturation of the
CYP2E1 metabolic pathway was seen at approximately 400-500 ppm after a 1-hour exposure
(Ott et al., 1983e). CYP2E1 pathway saturation with dichloromethane has also been noted in
hamsters (Burek et al., 1984) and in rats (McKenna et al, 1982; Nitschke et al., 1988a). It is
highly probable that initially CYP2E1 is metabolizing dichloromethane to CO, which results in
the neurological effects. However, at higher concentrations (greater than 500 ppm) and for
longer durations (greater than 3 hours), the CYP2E1 pathway is most likely saturated. As a
result, either the remaining dichloromethane could be metabolized by the GST pathway or the
parent compound is producing the effects itself.
Once the CYP2E1 enzyme is saturated, it is unknown whether dichloromethane or a
GST-T1 pathway metabolite (e.g., formaldehyde) mediates the resulting neurological effects.
Based on the available literature on other solvents, such as toluene and perchloroethylene (for a
review see Bowen et al. [2006]), it can be hypothesized that once the CYP2E1 enzyme is
saturated dichloromethane or a GST metabolite may interact directly with excitatory and
inhibitory receptors, such as the NMD A, GAB A, dopamine, and serotonin receptors among other
targets, to produce the resulting neurobehavioral effects. This hypothesis is supported by the
evidence that changes in relation to dichloromethane exposures in glutamate, GAB A, dopamine,
serotonin, acetylcholine, and other neurotransmitters are found in the brain (Kanada et al., 1994;
Briving et al., 1986; Fuxe et al., 1984). Additionally, several neurobehavioral effects, such as
decreased spontaneous motor activity, deficits in learning and memory, and deficits in FOB
parameters are similar to other more characterized solvents such as toluene. However, more
comprehensive studies specifically designed to determine the mode of action for
dichloromethane-induced impairment of neurological functions have not been conducted.
4.6.3.4. Mode of Action for Reproductive and Developmental Effects
192

-------
5909
5910
5911
5912
5913
5914
5915
5916
5917
5918
5919
5920
5921
5922
5923
5924
5925
5926
5927
5928
5929
5930
5931
5932
5933
5934
5935
5936
5937
5938
5939
5940
5941
5942
5943
5944
5945
5946
No significant effects on reproductive performance variables were found in a two-
generation reproduction assay with F344 rats exposed to concentrations as high as 1,500 ppm
(Nitschke et al., 1988b), and no effects were seen on most of the measures of reproductive
performance examined in a study of male mice exposed to 200 ppm for 2 hours/day for 6 weeks
before mating to nonexposed females (Raje et al., 1988). In the mouse study, fertility index
(number of females impregnated divided by total number of females mated times 100) was
reduced in the 150 and 200 ppm groups (Raje et al., 1988), but the statistical significance of this
effect varied considerably depending on the statistical test used in the analysis. Mechanistic
studies of dichloromethane or its metabolites that would provide mode of action information on
reproductive effects in the male are not available.
The mode of action for developmental effects can be hypothesized to involve the
CYP2E1 pathway and, specifically, the production of CO. CO is a known developmental
neurotoxicant. Demonstrated effects include neurobehavioral deficits and neurochemical
changes (Giustino et al., 1999; Cagiano et al., 1998; De Salvia et al., 1995; Fechter, 1987). In
addition, placental transfer of dichloromethane has been demonstrated with inhalation exposure
(Withey and Karpinski, 1985; Anders and Sunram, 1982). Pups exposed in utero to high
concentrations of dichloromethane (4,500 ppm) demonstrated neurobehavioral-related changes
in comparison to air-exposed animals (Bornschein et al., 1980). This observed effect coupled
with the known developmental neurotoxicological effects produced by CO suggests that the
CYP2E1 metabolic pathway is involved in producing observed and suspected
neurodevelopmental effects. In humans, CYP2E1 activity in the brain occurs earlier in gestation
than it does in the liver, with activity in the brain seen in the first trimester (Johnsrud et al., 2003;
Brzezinski et al., 1999). Thus, the direct effects of dichloromethane in fetal circulation, as well
as the effects of CO and the effects of the CYP2E1-related metabolism in the fetal liver and the
fetal brain, may be relevant to the risk of developmental effects in humans. Mechanistic studies
of dichloromethane or its metabolites that would provide mode of action information on other
noted developmental effects such as delayed ossification (Schwetz et al., 1975) are not available.
4.6.3.5. Mode of Action for Immunotoxicity
Evidence of a localized immunosuppressive effect in the lung resulting from inhalation
dichloromethane exposure was seen in an acute exposure (3 hours, 100 ppm) study in CD-I mice
(Aranyi et al., 1986). The lung infectivity assay used in this study examined response to
bacterial challenges (i.e., risk of streptococcal-pneumonia-related mortality and clearance of
Klebsiella bacteria). The innate immune response plays an important role in limiting the initial
lung burden of bacteria through the activity of macrophages, neutrophils, and dendritic cells, and
alveolar macrophages are particularly important in the response to respiratory infections
(Marriott and Dockrell, 2007). The adaptive response develops from several days up to several
weeks following infection, so that an effective immune response in a lung infectivity assay
193

-------
5947
5948
5949
5950
5951
5952
5953
5954
5955
5956
5957
5958
5959
5960
5961
5962
5963
5964
5965
5966
5967
5968
5969
5970
5971
5972
5973
5974
5975
5976
5977
5978
5979
5980
5981
5982
5983
5984
requires multiple immune mechanisms and in particular cooperation of macrophages,
neutrophils, and T cells along with the appropriate cytokines (Selgrade and Gilmour, 2006).
Although immunosuppression in the Streptococcal and Klebsiella infectivity models has been
reported in the acute exposure scenarios tested in Aranyi et al. (1986), mechanistic studies of
dichloromethane or its metabolites that would provide mode of action information on the
immune system cells or function have not been performed.
4.7. EVALUATION OF CARCINOGENICITY
4.7.1. Summary of Overall Weight of Evidence
Following U.S. EPA (2005a) Guidelines for Carcinogen Risk Assessment,
dichloromethane is "likely to be carcinogenic in humans" by the inhalation and oral routes of
exposure, based predominantly on evidence of carcinogenicity at two sites in 2-year bioassays in
B6C3Fi mice (liver and lung tumors with inhalation exposure in both sexes, liver tumors with
drinking water exposure in males only). In addition, evidence of a trend for increased risk of
liver tumors (described as neoplastic nodule or hepatocellular carcinoma) was seen in female
F344 rats exposed via drinking water (p < 0.01) (Serota et al., 1986a) or inhalation (p = 0.08)
(NTP, 1986). However, the potential malignant characterization of the nodules was not
described, and no trend was seen in the data limited to hepatocellular carcinomas. Additional
evidence of the tumorigenic potential of dichloromethane comes from the observation of an
increase in benign mammary tumors following inhalation exposure (NTP, 1986; Burek et al.,
1986b; Nitschke et al. 1988a). An inhalation study (exposures of 0, 50, 200, and 500 ppm) also
reported the presence of another relatively rare tumor in rats, astrocytoma or glioma (mixed glial
cell) tumors (Nitschke et al., 1988a). This collection of studies in the rat does not provide
evidence for a carcinogenic response that is as strong as that seen in the mouse. Taken together,
however, the rat data provide supporting evidence of carcinogenicity. Studies in humans found
some evidence linking occupational exposure to dichloromethane and increased risk for some
specific cancers, including brain cancer (Hearne and Pifer, 1999; Tomenson et al., 1997;
Heineman et al., 1994) and liver cancer (Lanes et al., 1993, 1990).
The proposed mode of action for dichloromethane-induced liver tumors is through a
mutagenic mode of carcinogenic action. Mode of action data indicate that dichloromethane-
induced DNA damage in cancer target tissues of mice involves DNA-reactive metabolites
produced via a metabolic pathway initially catalyzed by GST-T1. Evidence of mutagenicity
includes in vitro bacterial and mammalian assays as well as in vivo mammalian system assays,
although mutational events in critical genes (tumor suppressor genes, oncogenes) leading to
tumor initiation and tumor promotion have not been established. This metabolic pathway has
been found in human tissues, albeit at lower activities than in mouse tissues; therefore, the cancer
results in animals are considered relevant to humans.
194

-------
5985
5986
5987
5988
5989
5990
5991
5992
5993
5994
5995
5996
5997
5998
5999
6000
6001
6002
6003
6004
6005
6006
6007
6008
6009
6010
6011
6012
6013
6014
6015
6016
6017
6018
6019
6020
6021
4.7.2. Synthesis of Human, Animal, and Other Supporting Evidence
Section 4.1.2 reviewed the results, strengths, and limitations of epidemiological research
of dichloromethane and cancer, including cohort and case-control studies. The available
epidemiologic studies provide some evidence of an association between dichloromethane and
brain cancer and liver cancer, but the available data are limited.
Two small cohort studies with relatively good exposure metrics and relatively long
follow-up periods (mean over 25 years) reported an increased risk of brain cancer, with SMRs of
1.45 (95% CI 0.40-3.72) in Tomenson et al. (1997) and 2.2 (95% CI 0.79-4.69) in Cohort 1 of
Hearne and Pifer (1999). Cohort 1 is an inception cohort, following workers from the beginning
of their employment, which is methodologically more robust than Cohort 2, which only included
workers who were working between 1964 and 1970. These observations are supported by the
data from a case-control study of brain cancer that reported relatively strong trends with
increasing probability, duration, and intensity measures of exposure but not with a cumulative
exposure measure. This difference could reflect a relatively more valid measure of relevant
exposures in the brain from the intensity measure, as suggested by the study in rats reported by
Savolainen et al. (1981) in which dichloromethane levels in the brain were much higher with a
higher intensity exposure scenario compared with a constant exposure period with an equivalent
TWA (see section 3.2). A statistically significant increased incidence of brain or CNS tumors
has not been observed in any of the animal cancer bioassays, but a 2-year study using relatively
low exposure levels (0, 50, 200, and 500 ppm) in Sprague-Dawley rats observed a total of six
astrocytoma or glioma (mixed glial cell) tumors in the exposed groups (in females, the incidence
was 0, 0, 0, and 2 in the 0, 50, 200, and 500 ppm exposure groups; in males, the incidence was 0,
1, 2, and 1 in the 0, 50, 200, and 500 ppm exposure groups; sample size of each group was
70 rats). These tumors are exceedingly rare in rats, and there are few examples of statistically
significant trends in animal bioassays (Sills et al., 1999). These cancers were not seen in two
other studies in rats, both involving higher doses (1,000-4,000 ppm) (NTP, 1986; Burek et al.,
1984), or in a high dose (2,000-4,000 ppm) study in mice (NTP, 1986).
With respect to epidemiologic studies of liver and biliary duct cancer, the highest
exposure cohort, based in the Rock Hill, South Carolina, triacetate fiber production plant,
suggested an increased risk of liver cancer, with an SMR of 2.98 (95% CI 0.81, 7.63) in the latest
study update (Lanes et al., 1993). This observation was based on four cases; an earlier analysis
in this cohort reported an SMR of 5.75 (95% CI 1.82, 13.8), based on these same four cases but
with a shorter follow-up period (and thus a lower number of expected cases) (Lanes et al., 1990).
No other cohort study has reported an increased risk of liver cancer mortality, although it should
be noted that there is no other inception cohort study of a population with exposure levels similar
to those of the Rock Hill plant, and no data from a case-control study of liver cancer are
available pertaining to dichloromethane exposure.
195

-------
6022
6023
6024
6025
6026
6027
6028
6029
6030
6031
6032
6033
6034
6035
6036
6037
6038
6039
6040
6041
6042
6043
6044
6045
6046
6047
6048
6049
6050
6051
The primary limitation of all of the available dichloromethane cohort studies is the
limited statistical power for the estimation of effects relating to relatively rare cancers (such as
brain cancer, liver cancer, and leukemia). Limitations with respect to studies of other cancers
can also be noted. With respect to breast cancer, the only cohort that included a significant
percentage of women had limited exposure information (analysis was based on a dichotomous
exposure variable) and co-exposure to other solvents (Blair et al., 1998). The only breast cancer
case-control study available used death certificate data to classify disease and occupational
exposure (Cantor et al., 1995), which is likely to result in significant misclassification; exposure
misclassification in particular would be expected to result in an attenuated measure of
association (Rothman and Greenland, 1998). No studies of adult leukemia and dichloromethane
exposure and only one study of childhood leukemia (acute lymphoblastic leukemia) in relation to
maternal occupational dichloromethane exposure were found. Thus, EPA views the
epidemiologic data pertaining to breast cancer and leukemia as inadequate to assess carcinogenic
potential.
In addition to the epidemiologic studies, several dichloromethane cancer bioassays in
animals are available. In the only oral exposure cancer bioassay involving lifetime exposure,
increases in incidence of liver adenomas and carcinomas were observed in male (trend /?-value =
0.058) but not female B6C3Fi mice exposed for 2 years (Table 4-38) (Serota et al., 1986b;
Hazelton Laboratories, 1983). The authors concluded that these increases were "within the
normal fluctuation of this type of tumor incidence," noting that there was no dose-related trend
and that most of the individual group paired tests were not statistically significant after use of a
Bonferroni correction factor. [The trend /;-value and pairwise test ^-values were not given in the
Serota et al. (1986b) paper but can be found in the full report (Hazelton Laboratories, 1983)].
However, the trend />value for these results is of borderline statistical significance and it may not
be reasonable to apply a correction for multiple comparisons given the lack of independence of
the groups and given a specific focus on the liver as a target organ. In Syrian golden hamsters
exposed to 500, 1,500, or 3,500 ppm for 2 years, no statistically significantly increased
incidences of tumors were found in any tissues (Burek et al., 1984).
196

-------
6052
6053
6054
6055
6056
6057
6058
6059
6060
6061
6062
6063
6064
6065
6066
Table 4-38. Incidence of liver tumors in male B6C3Fi mice exposed
to dichloromethane in a 2-year oral exposure (drinking water) study3
Estimated mean intake
(mg/kg-day)a
Controls
0
61 124 177 234
Trend
Number of male miceb
125c
200 100 99 125
/j-valucd
Number of cancers (%)
Hepatocellular adenoma or
carcinoma
Mortality-adjusted percent"
Mortality-adjusted /?-valuce
24 (19)
(22)
51 (26) 30 (30) 31 (31) 35 (28)
(29) (34) (35) (32)
P = 0.071 p = 0.023 p = 0.019 /? = 0.036
0.058
"Target doses were 60, 125, 185, and 250 mg/kg-day from the lowest dose group (excluding controls) to
the highest dose group, respectively.
bNo significant increases in females were found, but incidence data were not reported.
°Two control groups combined. The incidence in control groups 1 and 2 were 20 and 23%, respectively.
dCochran-Armitage trend test (source: Hazelton Laboratories [1983]).
"Mortality-adjusted percent calculated, based on number at risk, using Kaplan-Meier estimation, taking into
account mortality losses; p-valuc for comparison with control group, using asymptotic normal test
(source: Hazelton Laboratories [1983]).
Sources: Serota et al. (1986b); Hazelton Laboratories (1983).
In a similar study in F344 rats (Serota et al., 1986a), no increased incidence of liver
tumors was seen in male rats, and the pattern in female rats was characterized by a jagged
stepped pattern of increasing incidence of hepatocellular carcinoma or neoplastic nodule
(Table 4-39). Information was not provided which would allow characterization of the nodules
as benign or malignant. Statistically significant increases in incidences were observed in the 50
and 250 mg/kg-day groups (incidence rates of 0, 3, 10, 3, and 14%, respectively, for the 0, 5, 50,
125, and 250 mg/kg-day groups) and in the group exposed to 250 mg/kg-day for 78 weeks
followed by a 26-week period of no exposure (incidence rate 10%). A similar pattern, but with
more sparse data, was seen for hepatocellular carcinomas, with 2 incidences in the 50 mg/kg-day
and 2 in the 250 mg/kg-day groups. The authors concluded that dichloromethane exposure did
not result in an increased incidence of liver tumors, because the increase was based on a low rate
(0%) in the controls and because of a lack of monotonicity.
197

-------
6067
6068
6069
6070
6071
6072
6073
6074
6075
6076
6077
Table 4-39. Incidences of liver tumors in male and female F344 rats exposed
to dichloromethane in drinking water for 2 years
Target dose (mg/kg-day)
Controls	Trend 250 with
	0a 5 50 125 250 /;-valueh recovery0
Males
Estimated mean intake (mg/kg-day)
0
6
52
125
235

232
n per groupd
76
34
38
35
41

15
Number (%) with neoplastic nodules
9(12)
1(3)
0(0)
2(6)
1(2)
Not
reported
2(13)
Number (%) with hepatocellular
carcinoma
3(4)
0(0)
0(0)
0(0)
1(2)
Not
reported
0(0)
Number (%) with neoplastic nodules
and hepatocellular carcinoma
12 (16)
1(3)
0(0)
2(6)
2(5)
Not
reported
2(13)
-emalcs







Estimated mean intake (mg/kg-day)
0
6
58
136
263

239
n per groupd
67
29
41
38
34

20
Number (%) with neoplastic nodules
0(0)
1(3)
2(5)
1(3)
3(9)

2 (10)e
Number (%) with hepatocellular
carcinoma
0(0)
0(0)
2(5)
0(0)
2(6)
Not
reported
0(0)
Number (%) with neoplastic nodules
and hepatocellular carcinoma
0(0)
1(3)
4 (10)e
1(3)
5 (14)e
p < 0.01
2 (10)e
"Two control groups combined.
bCochran-Armitage trend test was used for trend test of liver foci/areas of alteration. For tumor mortality-
unadjusted analyses, a Cochran-Armitage trend test was used, and, for tumor mortality-adjusted analyses, a tumor
prevalence analytic method by Dinse and Lagakos (1982) was used. Similar results were seen in these two
analyses.
°Recovery group was exposed for 78 weeks and then had a 26-week period without dichloromethane exposure;
n = 17 for neoplastic lesions.
dn available at terminal sacrifice; starting with 135 controls (combining both control groups) and 85 per sex per
dose group except recovery group (n = 25); subtracted 5, 10, and 20 per group (except for recovery group)
sacrificed at 25, 52, and 78 weeks, respectively, and subtracted unscheduled deaths, which ranged from 5 to 19 per
group.
"Significantly (p < 0.05) different from controls with Fisher's exact test, mortality-unadjusted and mortality-
adjusted analyses.
Source: Serota et al. (1986a).
Another oral (gavage) exposure study in Sprague-Dawley rats and in Swiss mice provides
limited data concerning cancer incidence because the study was terminated early (at 64 weeks)
due to high treatment-related mortality (Maltoni et al., 1988). Exposure groups included controls
(olive oil), 100, or 500 mg/kg-day 4-5 days/week. High-dose female rats showed an increased
incidence of malignant mammary tumors, mainly adenocarcinomas (8, 6, and 18% in the control,
100, and 500 mg/kg dose groups, respectively), but the increase was not statistically significant.
Data were not provided to allow an analysis accounting for differing mortality rates. A dose-
related increase, although not statistically significant, in pulmonary adenomas was observed in
198

-------
6078
6079
6080
6081
6082
6083
6084
6085
6086
6087
6088
6089
6090
6091
6092
6093
6094
6095
6096
male mice (5, 12, and 18% in control, 100, and 500 mg/kg-day groups, respectively). When
mortality was taken into account, high-dose male mice that died in the period ranging from 52 to
78 weeks were reported to show a statistically significantly (p < 0.05) elevated incidence for
pulmonary tumors (1/14, 4/21, and 7/24 in control, 100, and 500 mg/kg-day groups,
respectively). Details of this analysis were not provided. EPA applied a Fisher's exact test to
these incidences and determined a p-value of 0.11 for the comparison of the 500 mg/kg-day
group (7/24) to the controls (1/14).
As discussed in section 4.2, repeated inhalation exposure to concentrations of 2,000 or
4,000 ppm dichloromethane produced increased incidences of lung and liver tumors in B6C3Fi
mice (Mennear et al., 1988; NTP, 1986). The incidence of mortality-adjusted liver tumors across
dose groups (0, 2,000, and 4,000 ppm) increased from 48 to 67 and 93%, respectively, in male
mice (trendp-value = 0.013) and from 10 to 48 and 100% female mice (trend/^-values <0.001)
(Table 4-40). For lung tumors, the mortality-adjusted incidence was 12, 74, and 100% in males
and 11, 83, and 100% in females in the 0, 2,000, and 4,000 ppm groups, respectively (trendp-
values <0.001). Elevated incidences of lung and liver tumors in B6C3Fi mice were observed
with 52 weeks of exposure to 2,000 ppm, and lung tumors were also elevated by week 104 in
mice exposed for only 26 weeks to 2,000 ppm, followed by 78 weeks without exposure
(Maronpot et al., 1995; Kari et al., 1993).
199

-------
6099
6100
6101
6102
6103
6104
6105
6106
6107
6108
6109
6110
6111
6112
6113
6114
Table 4-40. Incidences of selected neoplastic lesions in B6C3Fi mice exposed
to dichloromethane by inhalation (6 hours/day, 5 days/week) for 2 years
Exposure (ppm)a
0 (Controls)	2,000	4,000	
Trend
Sex and neoplastic lesion	n (%)b (%)° n (%)b (%)° n (%)b (%)° />-valuc
Males
Liver—hepatocellular adenoma or 22 (44) (48) 24 (49) (67) 33e (67) (93) 0.013
carcinoma
Lung—bronchoalveolar adenoma or 5 (10) (12) 27e (54) (74) 40e (80) (100) <0.001
carcinoma
Females
Liver—hepatocellular adenoma or	3 (6) (10) 16e (33) (48) 40e (83) (100) <0.001
carcinoma
Lung—bronchoalveolar adenoma or 3 (6) (11) 30e (63) (83) 41e (85) (100) <0.001
carcinoma
a2,000 ppm = 6,947 mg/m3, 4,000 ppm = 13,894 mg/m3.
bTotal sample size was 50 per sex and dose group. Percentages based on the number of tissues examined
microscopically per group; for male mice, 49 livers were examined in the 2,000 and 4,000 ppm groups; for female
mice, 48 liver and lungs were examined. For comparison, incidence in historical controls reported in NTP (1986)
were 28% for male liver tumors, 31% for male lung tumors, 5% for female liver tumors, and 10% for female lung
tumors.
°Mortality-adjusted percentage.
dLife-table trend test, as reported by NTP (1986).
eLife-table test comparison dose group with control <0.05, as reported by NTP (1986).
Sources: Mennear et al. (1988); NTP (1986).
Liver tumors are relatively rare in F344 rats, and a moderate trend of increasing incidence
of what was described as neoplastic nodules or hepatocellular carcinoma was seen in the females
(trendp-value = 0.08) but not the males in the NTP (1986) study (Table 4-41). As with the rat
oral exposure study by Serota et al., 91986a), these nodules were not characterized as benign or
malignant. There was no evidence of an increasing trend in incidence when hepatocellular
carcinomas only were considered.
Female F344 rats exposed by inhalation to 2,000 or 4,000 ppm showed significantly
increased incidences of benign mammary tumors (adenomas or fibroadenomas) (Table 4-41); the
number of benign mammary tumors per animal also increased with dichloromethane exposure in
studies in Sprague-Dawley rats at levels of 50-500 ppm (Nitschke et al., 1988a) and 500-3,500
ppm (Burek et al., 1984) (Table 4-42). Male rats in two of these studies (Nitscke et al., 1988a ;
NTP, 1986) also exhibited a low rate of sarcoma or fibrosarcoma in mammary gland or
subcutaneous tissue around the mammary gland.
200

-------
Table 4-41. Incidences of selected neoplastic lesions in F344/N rats exposed to dichloromethane by inhalation
(6 hours/day, 5 days/week) for 2 years
Exposure (ppm)a
0 (Controls)	1,000	2,000	4,000	
Trend
Sex and neoplastic lesion	n (%)b (%)° n (%)b (%)° n (%)b (%)° n (%)b (%)° />-valued
Males
Liver—Neoplastic nodule or hepatocellular carcinoma
Liver—hepatocellular carcinoma
Lung—bronchoalveolar adenoma or carcinoma
Mammary gland
Adenoma, adenocarcinoma, or carcinoma
Subcutaneous tissue fibroma or sarcoma
Fibroadenoma
Mammary gland or subcutaneous tissue adenoma,
fibroadenoma, fibroma, or sarcoma
Females
Liver—neoplastic nodule or hepatocellular carcinoma
Liver—hepatocellular carcinoma
Lung—bronchoalveolar adenoma or carcinoma
Mammary gland
Adenocarcinoma or carcinoma
Adenoma, adenocarcinoma, or carcinoma
Fibroadenoma
Mammary gland adenoma, fibroadenoma, or
adenocarcinoma
al,000 ppm = 3,474 mg/m3, 2,000 ppm = 6,947 mg/m3, 4,000 ppm = 13,894 mg/m3.
bTotal sample size was 50 per sex and dose group. Percentages based on the number of tissues examined microscopically per group; for male rats, 49 livers
were examined in the 2,000 and 4,000 ppm groups; for females, only 48 liver and lungs and 49 mammary glands were microscopically examined in the 2,000
and 4,000 ppm groups. For comparison, incidence in historical controls reported in NTP (1986) were 1% for female liver tumors and 16% for female
mammary fibroadenomas.
°Mortality-adjusted percentage.
dLife-table trend test, as reported by NTP (1986). nr = not reported.
eLife-table test comparison dose group with control <0.05, as reported by NTP (1986).
Sources: Mennear et al. (1988); NTP (1986).
6115
6116
2
(4)
(10)
3
(6)
(13)
4
(8)
(19)
1
(2)
(6)
0.55
2
(4)
(10)
1
(2)
(4)
2
(4)
(10)
1
(2)
(6)
nr
1


1
(2)

2
(4)

1
(2)


0
(0)

0
(0)

0
(0)

1
(2)


1
(2)
(6)
1
(2)
(6)
2
(4)
(9)
5
(10)
(23)
0.008
0
(0)
(0)
0
(0)
(0)
2
(4)
(12)
1
(2)
(8)
<0.001
1
(2)
(6)
1
(2)
(6)
4
(8)
(21)
9d
(18)
(49)
<0.001
2
(4)
(7)
1
(2)
(2)
4
(8)
(14)
5
(10)
(20)
0.08
0
(0)
(0)
0
(0)
(0)
1
(2)
(4)
0
(0)
(0)
nr
1
(2)

1
(2)

0
(0)

0
(0)


1
(2)

2
(4)

2
(4)

0
(0)


1
(2)

2
(4)

2
(4)

1
(2)


5
(10)
(16)
lld
(22)
(41)
13d
(26)
(44)
22d
(44)
(79)
<0.001
6
(12)
(18)
13
(26)
(44)
14d
(28)
(45)
23e
(46)
(86)
<0.001
201

-------
Table 4-42. Incidences of mammary gland tumors in two studies of male and
female Sprague-Dawley rats exposed to dichloromethane by inhalation
(6 hours/day, 5 days/week) for 2 years
Exposure (ppm)a

Controls



Late
Early
Study, lesion
0
50
200
500
500b
500b 1,500 3,500

Nitschke et al. (1988a)



Males—n per group
57
65
59
64
C
C C C
Number (%) with






Mammary gland tumors






Adenocarcinoma or carcinoma
0 (0)
0 (0)
0 (0)
0 (0)


Fibroadenoma
2 (4)
0 (0)
2 (3)
2 (3)


Fibroma
6 (11)
1 (6)
6 (11)
10(16)


Fibrosarcoma
0 (0)
1 (6)
1 (6)
0 (0)


Undifferentiated sarcoma
0 (0)
2 (4)
0 (0)
0 (0)


Fibroma, fibrosarcoma, or
6 (11)
4 (6)
7
10 (16)


undifferentiated sarcomad


(12)



Females—n per group
69
69
69
69
25
25
Number (%) with mammary gland






Mammary gland tumors






Adenocarcinoma or carcinoma
6 (9)
5 (7)
4 (6)
4 (6)
3 (12)
2 (8)
Adenoma
1 (1)
1 (1)
2 (3)
1 (1)
2 (8)
0 (0)
Fibroadenoma
51 (74)
57 (83)
60 (87)
55 (80) 22 (88)
23 (92)
Fibroma
0 (0)
1 (1)
0 (0)
1 (1)
1 (4)
1 (1)
Fibrosarcoma
1 (1)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
Number with benign tumors6
52 (74)
58 (83)
61(87)f
55 (79)
23 (92)
23 (92)
Number of benign tumors per tumor-
2.0
2.3
2.2
2.7
2.2
2.6
bearing raf







Burek etaL (1984)





Males







n per group
92
C
95
C
c
96
97
Number (%) with benign tumors
7(8)

3 (3)


7(7)
14(14)
Total number of benign tumors
8

6


11
17
Number of tumors per tumor-bearing
rat8
1.1

2.0


1.6
1.2
Females







n per group
96
C
95
C
c
96
97
Number (%) with benign tumors
79 (82)

81 (85)


80 (83)
83 (86)
Total number of benign tumors
165

218


245
287
Number of tumors per tumor-bearing
ratf
2.1

2.7


3.1
3.5
a50 ppm = 174 mg/m3, 200 ppm = 695 mg/m3, 500 ppm = 1,737 mg/m3, 1,500 ppm = 5,210 mg/m3, 3,500 ppm =
12,158 mg/m3.
'"Late 500 = no exposure for first 12 months followed by 500 ppm for last 12 months; early 500 = 500 ppm for first 12
months followed by no exposure for last 12 months.
°No data for this exposure level in this study.
dEPA summed across these tumor types, assuming no overlap.
eIn historical controls, percent with benign tumors reported as 79-82% and number per tumor-bearing rat was 2.1.
Significantly (p < 0.05) higher than control incidence by Fisher's exact test (Nitschke et al., 1988a).
Calculated by EPA.
Sources: Nitschke et al. (1988a); Burek et al. (1984b).
6117
202

-------
6118
6119
6120
6121
6122
6123
6124
6125
6126
6127
6128
6129
6130
6131
6132
6133
6134
6135
6136
6137
6138
6139
6140
6141
6142
6143
6144
6145
6146
6147
6148
6149
6150
6151
6152
6153
6154
6155
6156
Supporting evidence for the carcinogenicity of dichloromethane comes from the results
of genotoxicity and mode of action studies discussed in section 4.5. A mutagenic mode of
carcinogenic action for dichloromethane involves metabolic activation by GST, as evidenced by
several observations, including the enhancement of dichloromethane mutagenic activity in
normally unresponsive S. typhimurium strain TA1535 after it is transfected with the gene for rat
GST-T1 (DeMarini et al., 1997; Thier et al., 1993); increased HPRT gene mutations and DNA
damage (DNA SSBs) in CHO cells when they are incubated with dichloromethane in the
presence of mouse liver cytosol preparations rich in GST enzymatic activities (Graves and
Green, 1996; Graves et al., 1996, 1994b); the detection of DNA damage (DNA SSBs) in liver
and lung tissue of B6C3Fi mice immediately following 6-hour inhalation exposure to
dichloromethane (2,000-8,000 ppm); and a suppression of the DNA damage when mice were
pretreated with buthionine sulphoximine, a GSH depletor (Graves et al., 1995).
Additional data from several studies indicate that dichloromethane genotoxicity is
expressed in cancer target tissues in mice following in vivo exposure. Increased sister chromatid
exchanges were observed in lung cells of B6C3Fi mice after 90 days of inhalation exposure to
2,000 ppm or 10 days of exposure to 4,000 or 8,000 ppm (Allen et al., 1990). DNA damage
(comet assay) was detected in liver and lung tissue (but not stomach, kidney, brain, or bone
marrow) 24 hours after oral administration of 1,720 mg/kg dichloromethane to CD-I mice
(Sasaki et al., 1998). DNA-protein cross-links were observed in the liver of B6C3Fi mice but
not hamsters, following inhalation exposure to concentrations ranging from 500 to 4,000 ppm
6 hours/day for 3 days (Casanova et al., 1996, 1992). Much less is known about genotoxicity in
the liver in rats. Studies of single-strand DNA breaks in rat hepatocytes or liver homogenate
were negative, with inhalation exposures up to 5,000 ppm for 3 hours (Graves et al., 1995,
1994b), but positive results were seen in a high-dose gavage study (1,275 mg/kg) (Kitchin and
Brown, 1989). Few other specific types of genotoxicity endpoints (e.g., sister chromatid
exchange, DNA-protein cross-links) have been studied in the rat liver.
Since there are limited data on mutagenic events following oral exposure, EPA conducted
a pharmacokinetic analysis to evaluate how comparable the internal doses to the liver in the oral
bioassay (Serota et al., 1986b; Hazelton Laboratories, 1983) were to the internal doses to the
liver in the inhalation bioassay (Mennear et al., 1988; NTP, 1986). The PBTK model of Marino
et al. (2006) predicted that the average daily amount of dichloromethane metabolized via GST
per liter of liver was about 14-fold lower in mice exposed to the highest dose of 244 mg/kg-day
in the drinking water bioassay than in mice exposed to the lowest inhalation exposure of 2,000
ppm, inducing liver tumors (Table 4-43). Thus, the lower incidence of liver tumors induced by
oral doses of 244 mg/kg-day, compared with the higher incidence induced by inhalation
exposure to 2,000 ppm, is consistent with the predicted lower liver dose of GST metabolites (and
hence lower probability of DNA modification) with oral exposure-
203

-------
6157
6158
6159
6160
6161
6162
6163
6164
6165
6166
6167
6168
6169
6170
6171
6172
6173
6174
6175
6176
6177
6178
Table 4-43. Comparison of internal dose metrics in
inhalation and oral exposure scenarios, in male mice and
rats
Internal exposure in liver (mg metabolized through
GST pathway/L liver tissue/day)3
Male
External dose	Mouse	Rat
Inhalation (ppm)
2,000
2,364
1,502
4,000
4,972
3,111
Oral (mg/kg-day)b


61
17.5
77.0
124
63.3
233.5
174
112.0
385.4
234
169.5
589.8
a Mouse values derived by EPA from the PBTK model of Marino et al. (2006); rat values
derived from EPA based on the modified PBTK model of Andersen et al. (1991) (see
Appendix C for model details).
b Actual doses administered to mice (Serota et al., 1986a); BWs not given for males and
females, so simulation results only provided for one gender.
4.7.3. Mode of Action Information
4.7.3.1. Hypothesized Mode of Action
The hypothesized mode of action for dichloromethane-induced tumors is through a
mutagenic mode of carcinogenic action. Specifically, the data indicate that dichloromethane is
metabolized by GST to reactive metabolites that induce mutations in DNA leading to
carcinogenicity. Much of the experimental mode of action research has focused on the liver and
lung, the sites of tumor formation in chronic bioassays (Mennear et al., 1988; NTP, 1986; Serota
et al., 1986b, Hazleton Laboratories, 1983). The mode of action is potentially relevant to other
sites, particularly those in which GST-T1 is expressed, such as mammary tissue (Lehmann and
Wagner, 2008) and the brain (Juronen et al., 1996).
Support for the importance of GST in the hypothesized mutagenic mode of action has
been demonstrated in in vitro bacterial and mammalian assays as well as in vivo mammalian
system assays. Dichloromethane is consistently mutagenic in S. typhimurium strains with GST
capability, but did not produce mutagenic effects in non-GST S. typhimurium strains
(summarized in Section 4.5.1.1 and Table 4-29). In vitro mammalian cell studies (see Table 4-
30) have consistently demonstrated genotoxic effects in CHO cell lines when a mouse liver
cytosol fraction was exogenously added and in mouse Clara cells; positive responses were seen
in studies measuring DNA-protein crosslinks, HPRT mutation analysis, and DNA SSBs. Other
204

-------
6179
6180
6181
6182
6183
6184
6185
6186
6187
6188
6189
6190
6191
6192
6193
6194
6195
6196
6197
6198
6199
6200
6201
6202
6203
6204
6205
6206
6207
6208
6209
6210
6211
6212
6213
6214
6215
6216
studies have demonstrated DNA adducts with dichloromethane exposure in calf thymus DNA in
the presence of bacterial GST DM11. Negative results were seen in most of the other in vitro cell
studies using rat hepatocytes or CHO cells without mouse liver cytosol incubation. These
studies were conducted in cell lines where GST activity is considerably lower than in mouse cell
lines and therefore these results are not unexpected.
In studies with human cell lines or isolated cells, positive results were reported for sister
chromatid exchanges, chromosomal aberrations, and in the micronucleus test. In vivo studies in
mice (Section 4.5.1.2 and Table 4-32) consistently showed genotoxic effects following
dichloromethane exposure in the liver and lung, where tumors are observed. Other organs in the
mouse were evaluated and mutagenic changes were not consistenly observed. The specificity of
the observed effects support the hypothesized mode of action since these positive mutagenic
responses are seen in organs where tumor formation occurs (i.e., liver and lung) rather than in
areas that were not the site of tumors in the mouse bioassays (e.g., stomach, bladder, kidney). In
vivo genotoxicity studies in rats and hamsters (the other test systems used, see Table 4-33) were
predominantly non-positive. However, rats and hamsters have considerably lower GST activity
than the mouse and may be less sensitive to dichloromethane-induced genotoxic effects.
In vivo binding of S-(chloromethyl)glutathione, dichloromethane's reactive GST
metabolite,to DNA was not demonstrated in one study in rats and mice using a relatively low
dose (5 mg/kg). The reactivity of the postulated DNA-reactive species and the instability of the
derived adducts presents considerable challenges to the ability to provide direct evidence of
adduct formation. Thus this lack of in vivo evidence of S-(chloromethyl)glutathione binding to
DNA does not in itself represent a basis for invalidating the proposed mode of action.
4.7.3.1.1. Experimental support for the hypothesized mode of action
Strength, consistency, and specificity of association
It is hypothesized that mutagenic events lead to the development of liver and lung tumors
following dichloromethane exposure. Several observations from experimental studies support the
mutagenicity of dichloromethane and the key role of GST metabolism and the formation of
DNA-reactive GST-pathway metabolites. The GST pathway produces two metabolites of
dichloromethane, S-(chloromethyl)glutathione and formaldehyde, which are potentially reactive
with DNA and other cell macromolecules. Enhanced dichloromethane genotoxicity in bacterial
and mammalian in vitro assays with the introduction of GST metabolic capacity provides support
that GST metabolism and metabolites are involved (DeMarini et al., 1997; Graves and Green,
1996; Graves et al., 1996, 1995, 1994b; Thier et al., 1993).
In bacterial strains where GST activity was not present (e.g., TA1535, TA1538),
mutagenic effects were not reported following dichloromethane exposure (Gocke et al., 1981;
Osterman-Golkar et al., 1983; Simula et al., 1993; Oda et al., 1996). Further tests of GST-
dependent mutagenicity were evaluated by transfecting GST into non-GST bacterial strains or
205

-------
6217
6218
6219
6220
6221
6222
6223
6224
6225
6226
6227
6228
6229
6230
6231
6232
6233
6234
6235
6236
6237
6238
6239
6240
6241
6242
6243
6244
6245
6246
6247
6248
6249
6250
6251
6252
6253
6254
decreasing GST activity in GST bacterial strains (e.g. TA100). When GSTT1-1 was cloned into
bacterial strain TA1535, dichloromethane treatment resulted in reverse mutations in this new
GST+ TA1535 strain and these mutations were independent of rat S9 metabolic activation (Their
et al., 1993; Pegram et al., 1997; DeMarini et al., 1997). Similarly, TA100/NG-11, a bacterial
strain with decreased GST activity in comparison to the wild-type TA100 strain, showed
significantly decreased mutagenicity (reverse mutations) following dichloromethane treatment
(Graves et al., 1994a).
In vitro mammalian genotoxicity studies also support the importance of the GST pathway
in relation to the positive effects observed following dichloromethane exposure. Positive results
in the in vitro assays were limited to experiments with the presence of GST in the cell system.
When mouse liver cytosol was added to hamster cell lines, dichloromethane induced DNA-
protein crosslinks, DNA SSBs, and HPRT gene mutations (Graves et al., 1994b; Graves et al.,
1996; Graves and Green, 1996). Additionally, in mouse Clara cells (GST is localized in the lung
cells of mice), DNA SSBs were reported following dichloromethane treatment and the extent of
DNA damage was significantly decreased when the cells were pretreated with a glutathione
depletor (Graves et al., 1995). Other studies evaluating similar genotoxic endpoints in rat or
CHO cells without modification of the low GST activity in the test system generally reported no
evidence of genotoxic events (Graves et al., 1995; Andrae and Wolff, 1983; Jongen et al., 1981;
Garrett and Lewis, 1983; Thilagar and Kumaroo, 1983). A study evaluating the genotoxic
effects of dichloromethane (up to 6 mM) in freshly isolated mouse, rat, hamster, and human
hepatocytes provides additional supporting evidence of the influence of GST activity on
mutagenicity (Casanova et al., 1997). Positive results were only observed in hepatocytes from
B6C3Fi mice; the interspecies variability in effects correlated proportionally with the enhanced
GST metabolic capacity in mice (Reitz et al., 1989). In studies with human cell lines or isolated
cells, positive results were reported for sister chromatid exchanges, chromosomal aberrations,
DNA damage, and in the micronucleus test. Negative results were obtained with human cells in
unscheduled DNA synthesis assays (Perocco and Prodi, 1981; Jongen et al., 1981) and
dichloromethane was not demonstrated to be genotoxic in studies of human hepatocytes (Graves
et al., 1995; Casanova et al., 1997).
Two of three in vivo genotoxicity studies in insects reported positive results.
Genotoxicity was observed in Drosophila for the gene mutation assay (Gocke et al., 1981) and
the somatic assay (Rodriguez-Arnaiz, 1998) when dichloromethane was administered through
the food. When Drosophila were exposed to dichloromethane via inhalation, genotoxic effects
were negative as measured through gene mutation assays (sex-linked recessive lethal, somatic
mutation and recombination) (Kramers et al., 1991).
In vivo genotoxicity studies reported DNA-protein cross links, DNA SSBs, chromosomal
aberrations, and sister chromatid exchanges in liver cells of B6C3Fi mice following acute
inhalation exposure to concentrations producing liver tumors with chronic exposure (Casanova et
206

-------
6255
6256
6257
6258
6259
6260
6261
6262
6263
6264
6265
6266
6267
6268
6269
6270
6271
6272
6273
6274
6275
6276
6277
6278
6279
6280
6281
6282
6283
6284
6285
6286
6287
6288
6289
6290
6291
6292
al., 1996, 1992; Graves et al., 1995, 1994b). The formation of DNA SSBs was suppressed when
the mice were pretreated with a GSH depletor (Graves et al., 1995), providing additional support
for the involvement of GST metabolism. Increased sister chromatid exchanges and
chromosomal aberrations were found in the lungs in mice exposed to dichloromethane for
2 weeks to 8,000 ppm or for 12 weeks to 2,000 ppm. In this study, however, there was evidence
of damage at other sites, too: sister chromatid exchanges were also seen in peripheral
lymphocytes, chromosomal aberrations were seen in bone marrow and micronuclei were seen in
in peripheral red blood cells under the same exposure protocol (Allen et al., 1990). As was seen
in the liver, DNA SSBs were seen in lungs of B6C3Fi mice following acute inhalation exposure
to concentrations producing lung tumors with chronic exposure, and this effect was suppressed
with pretreatment with a GSH depletor, buthionine sulfoximine (Graves et al., 1995). Other
studies of sister chromatid exchange (Allen et al., 1990) or DNA damage detected by the comet
assay (Sasaki et al., 1998) also provide evidence of genotoxic effects specifically in lung cells in
mice. These in vivo mammaliam genotoxicity studies demonstrate site-specific effects
correlating to the dichloromethane-induced tumors in animals. Additional evidence for site
specificity comes from a study in which DNA damage (detected by the comet assay) was
enhanced in liver tissue, but not stomach, kidney, brain, or bone marrow, 24 hours after oral
administration of 1,720 mg/kg dichloromethane to CD-I mice (Sasaki et al., 1998).
DNA reaction products (e.g., DNA adducts) produced by GST metabolites, such as S-
(chloromethyl)glutathione, have not been identified in in vivo studies (Watanabet et al., 2007).
The authors speculated that these results are due to the instability of the reaction products
(Hashmi et al., 1994). However, adducts with nucleosides have been observed in in vitro
studies, when DNA was treated with S-(l-acetoxymethyl)glutathione, a compound structurally
similar to S-(chloromethyl)glutathione and in calf thymus DNA in the presence of
dichloromethane, but not formaldehyde, and GST (Marsch et al., 2004; Kayser and Vuilleumier,
2001). These findings indicate that the S-(chloromethyl)glutathione intermediate formed by
GSH conjugation has mutagenic potential and is likely responsible, at least in part, for the
mutagenic response observed following dichloromethane exposure. However, other studies (Hu
et al., 2006; Casanova et al., 1996) provide evidence of formaldehyde-related DNA-protein
cross-links in relation to dichloromethane exposure. These results show that, while most studies
indicate the importance of the S-(chloromethyl)glutathione intermediate in mediating genotoxic
damage following dichloromethane exposure, DNA damage resulting from formaldehyde
formation should also be considered.
Mutagenic data in critical genes leading to the initiation of dichloromethane-induced liver
or lung tumors are not available. In vivo assays evaluating mutations in tumor suppressor genes
and oncogenes reported similar frequencies of activated H-ras genes and inactivation of the
tumor suppressor genes, p53 and Rb-1 in the liver tumors seen in the nonexposed and
dichloromethane-exposed B6C3Fi mice (Devereaux et al., 1993; Hegi et al., 1993). There were
207

-------
6293
6294
6295
6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306
6307
6308
6309
6310
6311
6312
6313
6314
6315
6316
6317
6318
6319
6320
6321
6322
6323
6324
6325
6326
6327
6328
6329
6330
too few lung tumors (n=4) in controls to provide a conclusive comparison of mutation patterns
between exposed and nonexposed tumors.
Dose-response concordance
Statistically significant increases in liver tumor incidences in male and female (2,000 and
4,000 ppm) mice were observed in the inhalation bioassay in B6C3Fi mice (NTP, 1986).
Several studies provide evidence of an association between mutagenic events mediated by GST-
pathway metabolites with the exposure levels inducing liver tumors in B6C3Fi in this study, and
concentration dependent increases in genotoxicity have been observed in in vitro and in vivo
assays.
In vitro mammalian genotoxicity studies were positive and demonstrated a dose-response
relationship for DNA protein crosslinks, DNA single stranded breaks, and DNA damage as
measured by the comet assay at concentrations ranging from 2.5 to 60 mM when mouse liver
cytosol was added or if mouse GSTT1-1 was transfected into hamster cell lines (Graves et al.,
1994b; Graves et al., 1996; Hu et al., 2006). In mouse hepatocytes, DNA protein crosslinks were
observed following dichloromethane exposures ranging between 0.5 - 6.0 mM (Casanova et al.,
1997). DNA-protein cross-links were detected in mouse hepatocytes incubated with 1.9 mM
dichloromethane (Casanova et al., 1997), a concentration chosen based on its correspondence to
the TWA liver concentration of dichloromethane that was predicted by the Andersen et al.
(1987) PBTK model for mice exposed by inhalation to 4,000 ppm for 6 hours (a dose that
resulted in increased liver tumor incidence in the two-year bioassay reported by NTP, 1986).
Consistent with the relative lack of liver tumor responses in Syrian golden hamsters (Burek et al.,
1984) and F344 rats (NTP, 1986) with chronic exposure to 3,500 or 4,000 ppm, hepatocytes
from these strains of animals did not form detectable DNA-protein cross-links when incubated
with 1.9 mM dichloromethane (Casanova et al., 1997)	
DNA-protein cross-links were not detected in livers of mice exposed to 146 ppm
6 hours/day for 3 days, but a concentration-dependent increase in DNA-protein cross-links was
observed in DNA from livers of mice exposed to several concentrations between 500 and
4,000 ppm (Casanova et al., 1996). Following exposure under similar conditions (concentrations
of 498, 1,553, or 3,923 ppm), DNA-protein cross-links were not detected in the livers of Syrian
golden hamsters, a species that did not develop tumors after chronic inhalation exposure to
dichloromethane (Casanova et al., 1996, 1992). Increased DNA SSBs were detected in liver
tissue of B6C3Fi mice immediately following a 6-hour inhalation exposure to dichloromethane
at concentrations ranging from 2,000 to 8,000 ppm (Graves et al., 1995), and in mouse
hepatocytes after a 3-hour exposure to 4000 (but not 2000) ppm (Graves et al., 1994b).
Statistically significant increases in the incidence of lung tumors were observed in the
inhalation chronic bioassay in male and female B6C3Fi mice exposed to 2,000 or 4,000 ppm
dichloromethane (Mennear et al., 1988; NTP, 1986). Evidence of mutagenicity at these exposure
208

-------
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
levels comes from two inhalation studies (Graves et al., 1995; Allen et al., 1990). Increased
DNA SSBs were detected in lung tissue of B6C3Fi mice immediately following a 6-hour
inhalation exposure to dichloromethane at concentrations ranging from 2,000 to 8,000 ppm
(Graves et al., 1995). In the study by Allen et al. (1990), increased presence of sister chromatid
exchanges was observed in mouse lung cells following a 12-week exposures at 2000 ppm;
shorter durations of exposure (2 weeks) were positive for measures of sister chromatid exchange
and chromosome aberrations at 8000 ppm, but not at 2000 or 4000 ppm.
DNA adducts were observed and increased with dose in an in vitro preparation of calf
thymus DNA when treated with dichloromethane (5 - 60 mM) and bacterial, rat, or human GST
(Marsch et al., 2004).
Temporal relationship
Dichloromethane-induced liver and lung tumors first appeared in mice after 52 weeks of
exposure (Maronpot et al., 1995; Kari et al., 1993). The detection of DNA-protein cross-links in
the livers of B6C3Fi mice following short-term inhalation exposures to dichloromethane
concentrations that induced tumors with chronic exposure (Casanova et al., 1996, 1992) provides
temporal support for the proposed mutagenic mode of action. Additional supporting evidence
comes from observations that increased levels of DNA SSBs were detected in the liver and lungs
of B6C3Fi mice immediately following 3-hour inhalation exposure to 2,000-8,000 ppm
dichloromethane (Graves et al., 1995; 1994b). Single dose and inhalation exposure studies of 6
hours or less did not detect an effect on DNA synthesis (Lefevre and Ashby, 1989) or
unscheduled DNA synthesis (Trueman and Ashby, 1987) in mouse liver cells.
Biological plausibility and coherence
Bioactivation of a parent compound into a mutagenic metabolite resulting in cancer is a
plausible mode of action carcinogenicity in humans and is a generally accepted mode of action.
Dichloromethane-induced carcinogenicity is hypothesized to be due to metabolism of the parent
compound by the GST pathway (GST-T1) to a metabolite that is tumorigenic. The GST
metabolite, S-(chloromethyl)glutathione, formed from dichloromethane, has been characterized
as labile and highly reactive through in vitro evaluation of dichloromethane metabolism in
hepatocytes using 13C-NMR techniques (Hashmi et al., 1994) and through an enzyme digestion
assay using calf thymus DNA and GST-T1 enzyme (Marsch et al., 2004). The hypothesis that
the formation of a mutagenic metabolite is a preliminary step resulting in carcinogenicity is
based on evidence that malignant tumors are primarily located in areas where dichloromethane is
highly metabolized by GST-T1, such as the liver and the lung, and on mutagenicity studies
indicating the importance of the GST pathway and that the lung and liver are more prone to
mutagenic effects of dichloromethane (Sasaki et al., 1998; Casanova et al., 1996, 1992; Graves
et al., 1995, 1994b). The site selectivity of the mutagenicity in liver and lung tissue as evidenced
209

-------
6369
6370
6371
6372
6373
6374
6375
6376
6377
6378
6379
6380
6381
6382
6383
6384
6385
6386
6387
6388
6389
6390
6391
6392
6393
6394
6395
6396
6397
6398
6399
6400
6401
6402
6403
6404
6405
by several studies suggests that the GST reactive metabolite remains in the tissue where it is
formed. Collectively, the studies support the hypothesis that dichloromethane-mediated
carcinogenicity results from a GST metabolite that produces selective DNA damage in the
tissues where the metabolite is formed, but this hypothesis is based, in part, on assumptions
regarding metabolite clearance and reactivity. DNA damage in the liver and lung as well as the
increased incidence of tumor formation resulting from dichloromethane exposure indicates
coherence of the mutagenic and carcinogenic effects and is evidence supporting a mutagenic
mode of action.
Differences in GST activity in mice compared with other species, and the interspecies
variability in genotoxic effects corresponding to interspecies variability in tumor response
support the mode of action hypothesis. DNA SSBs were not detected in liver or lung cells in rats
exposed to similar inhalation exposures that induce strand breaks in mice (Graves et al., 1995;
Graves et al., 1994b), and were detected at much lower in vitro concentrations in isolated
hepatocytes from B6C3Fi mice (0.4 mM) than in hepatocytes from Alpk:APfSD rats (30 mM)
(Graves et al., 1995, figure 3). The difference in susceptibility to carcinogenic response between
mice and rats likely reflects differences in GST metabolism. Toxicokinetic studies indicate that,
with increasing exposure levels, increasing amounts of dichloromethane are metabolized via
GST metabolism.
4.7.3.1.2. Other possible modes of action for liver or lung tumors in rodents.
Data are not available to support other possible modes of action for the liver and lung
tumors in rodents. Efforts to observe sustained cell proliferation in liver following
dichloromethane exposure of B6C3Fi mice have been unsuccessful. Groups of female B6C3Fi
mice were exposed to 0 or 2,000 ppm dichloromethane 6 hours/day, 5 days/week for up to 78
weeks did not exhibit enhanced cell proliferation in the liver when assessed at various intervals
during exposure (Foley et al., 1993). .
Indices of enhanced cell proliferation have been measured in the lungs of male B6C3Fi
mice following acute duration exposure at concentrations of about 1,500, 2,500, or 4,000 ppm
dichloromethane (6 hours/day for 2 days) but not at exposure concentrations of 150 or 500 ppm
and not in lungs of Syrian golden hamsters exposed to concentrations up to 4,000 ppm
(Casanova et al., 1996). Earlier studies showed somewhat consistent findings in that the
numbers of bronchiolar cells undergoing DNA synthesis (thymidine incorporation labeling) were
markedly increased (about 6- to 15-fold) in bronchi olar cells of B6C3Fi mice exposed to
4,000 ppm dichloromethane 6 hours/day on days 5, 8, and 9 of exposure, but no evidence for
increased cell proliferation was found after 89, 92, or 93 days of exposure (Foster et al., 1992).
The results suggest that enhanced cell proliferation is not sustained in the lung with longer-term
exposure to dichloromethane concentrations associated with lung tumor development in mice,
210

-------
6406
6407
6408
6409
6410
6411
6412
6413
6414
6415
6416
6417
6418
6419
6420
6421
6422
6423
6424
6425
6426
6427
6428
6429
6430
6431
6432
6433
6434
6435
6436
6437
6438
6439
6440
6441
6442
and that this mode of tumor promotion is not important in the development of dichloromethane-
induced lung tumors.
4.7.3.2. General Conclusions About the Mode of Action for Tumors in Rodents and
Relevance to Humans
The mode of action for dichloromethane is hypothesized to involve mutagenicity via
reactive metabolites. Mechanistic evidence indicates that dichloromethane-induced DNA
damage in cancer target tissues of mice involves DNA-reactive metabolites produced via a
metabolic pathway initially catalyzed by GST. Although mutational events in critical genes
leading to tumor initiation have not been established, evidence supporting a mutagenic mode of
action includes the identification of mutagenic response (reverse mutations) in short-term
bacterial assays (with microsomal activation) and induced DNA-protein cross-links and DNA
SSBs in mammalian cell assays. There are numerous positive in vivo genotoxicity studies
specifically examining responses in the liver and/or lung; these studies included evidence of
chromosomal aberrations, SSBs and sister chromatid exchanges, and DNA-protein cross-links.
The negative in vivo genotoxicity assays are generally those that were based on a micronucleus
test using mouse bone marrow, which is expected as halogenated hydrocarbons (such as
dichloromethane) are not very effective in this type of assay (Crebelli et al., 1999; Dearfield and
Moore, 2005).
Is the hypothesized mode of action sufficiently supported in test animals?
Consistent and specific evidence for the association between the formation of DNA-
reactive GST-pathway metabolites and the formation of liver and lung tumors from inhalation
includes (1) enhanced GST metabolic capacity in the liver and lung and enhanced localization of
GST-T1 in hepatic cell nuclei in B6C3Fi mice, compared with rats and hamsters, which do not
show strong tumor responses to chronic inhalation exposure; (2) the detection of DNA-protein
cross-links, or DNA SSBs in livers and lungs of B6C3Fi mice following acute inhalation
exposure to concentrations that produce tumors with chronic exposure; (3) suppression of the
formation of DNA SSBs in livers and lungs of B6C3Fi mice pretreated with a GSH depletor; (4)
the inability to detect DNA-protein cross-links or DNA SSBs in livers or lungs of similarly
exposed rats or hamsters (5) detection of DNA SSBs at much lower in vitro concentrations in
isolated hepatocytes from B6C3Fi mice than in hepatocytes from Alpk:APfSD rats; (6) dose-
response concordance and a temporal relationship for the formation of DNA-protein cross-links
and DNA SSBs with the formation of liver and lung tumors in B6C3Fi mice exposed to
dichloromethane; (7) the detection of increased sister chromatid exchanges in lung cells from
CD-I mice exposed by inhalation to dichloromethane; and (8) enhancement of dichloromethane
genotoxicity in bacterial and mammalian in vitro assays with the introduction of GST metabolic
211

-------
6443
6444
6445
6446
6447
6448
6449
6450
6451
6452
6453
6454
6455
6456
6457
6458
6459
6460
6461
6462
6463
6464
6465
6466
6467
6468
6469
6470
6471
6472
6473
6474
6475
6476
6477
6478
6479
6480
capacity. However, mutations in critical genes linked to initiation of tumor cells have not been
identified.
The much weaker carcinogenic response in the liver of rats and mice to chronic drinking
water exposure (Serota et al., 1986a, b) than that noted in mice exposed by inhalation (Kari et al.,
1993; NTP, 1986) is correlated with much smaller amounts of GST metabolites produced in the
liver under the exposure conditions of the oral bioassay than in the inhalation bioassay (Andersen
et al., 1987).
In conclusion, there is sufficient evidence supporting a mutagenic mode of action and to
establish the involvement of GST metabolism in the lung and liver carcinogenicity of
dichloromethane in mice.
Is the hypothesized mode of action relevant to humans?
The postulated mode of action, that dichloromethane is metabolized by GST to reactive
metabolites that induce mutations in DNA leading to carcinogenicity is possible in humans.
Mutagenicity as a mode of action for carcinogenicity in humans is generally accepted and is a
biologically plausible mechanism for tumor induction. The toxicokinetic and toxicodynamic
processes that would enable reactive metabolites to produce mutations in animal models are
biologically plausible in humans. Furthermore, the detection of the GST pathway in human
tissues indicates that the hypothesized mode of action involving reactive metabolites from this
pathway, S-(chloromethyl)glutathione and formaldehyde, is relevant to humans.
Another factor that may play a role in the apparent species differences in carcinogenicity
resulting from dichloromethane exposure is species differences in intracellular localization of
GST-T1 (Sherratt et al., 2002; Mainwaring et al., 1996). In mouse liver tissue, GST-T1 appears
to be localized in the nuclei of hepatocytes and bile-duct epithelium, but rat liver does not show
preferential nuclear localization of GST-T1. In human liver tissue, some hepatocytes show
nuclear localization of GST-T1 and others show localization in cytoplasm. Nuclear production
of S-(chloromethyl)glutathione catalyzed by GST-T1 in the nucleus is more likely than
cytoplasmic production to lead to DNA alkylation. The finding of some nuclear localization of
GST-T1 in human liver tissue supports the relevance of the hypothesized mode of action to
humans.
Comparisons in mice, rats, humans, and hamsters of GST enzyme activity in liver and
lung tissues have indicated the following rank order: mice > rats > or ~ humans > hamsters
(Reitz et al., 1989; Thier et al., 1998). This relative ranking in GST activity corresponds to the
rank order of the strength of the association between inhalation exposure to dichloromethane and
liver tumors in long-term cancer bioassays with mice, rats, and hamsters. This relative ranking
does not preclude the relevance of the hypothesized mode of action to humans, however.
212

-------
6481
6482
6483
6484
6485
6486
6487
6488
6489
6490
6491
6492
6493
6494
6495
6496
6497
6498
6499
6500
6501
6502
6503
6504
6505
6506
6507
6508
6509
6510
6511
6512
6513
6514
6515
6516
6517
6518
Which populations or lifestages can be particularly susceptible to the hypothesized mode of
actionl
As discussed in section 3.3, a polymorphism of the GST-T1 gene is present in humans.
People with two functional copies of the gene (+/+) readily conjugate GSH to dichloromethane.
Individuals having only one working copy of the gene (+/-) display relatively decreased
conjugation ability. Individuals with no functional copy of the gene (-/-) do not express active
GST-T1 protein and do not metabolize dichloromethane via a GST-related pathway (Thier et al.,
1998). Thus the GST-T1+/+ (wild-type) genotype would be considered to be the more "at risk"
population; this subgroup represents approximately 30% of the U.S. population (Haber et al.,
2002) but would be expected to be more common among Caucasians and African-Americans
than among Asians (Raimondi et al., 2006; Garte et al., 2001; Nelson et al., 1995) (see Table
3-3).
According to the Supplemental Guidance for Assessing Susceptibility from Early-Life
Exposure to Carcinogens (U.S. EPA, 2005b), children exposed to carcinogens with a mutagenic
mode of action are assumed to have increased early-life susceptibility. The Supplemental
Guidance (U.S. EPA, 2005b) recommends the application of age-dependent adjustment factors
(ADAFs) for carcinogens that act through a mutagenic mode of action. Although the database is
lacking in vivo evidence of specific mutagenic events following chronic exposure to
dichloromethane, the weight of the available evidence indicates that dichloromethane is acting
through a mutagenic mode of carcinogenic action. Application of ADAFs is recommended for
both the oral and inhalation routes of exposure when risks are assessed that are associated with
early-life exposure.
4.8. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.8.1. Possible Childhood Susceptibility
In humans, hepatic CYP2E1 begins to be expressed in the second trimester (Johnsrud et
al., 2003), increases significantly in the third trimester, and continues to increase during the first
year of life (Hines, 2007; Johnsrud et al., 2003; Treluyer et al., 1996; Vieira et al., 1996). In the
fetal brain, however, CYP2E1 activity is seen as early as GD 50, with increasing levels seen until
at least the end of the first trimester (Brzezinski et al., 1999). Neurobehavioral effects of
dichloromethane are seen with acute exposures in adults, and the available data regarding
neurological symptom history and standardized testing suggest the possibility of longer-term
effects. The relatively high activity of CYP2E1 in the brain compared to the liver of the
developing human fetus raises the potential for neurodevelopmental effects from
dichloromethane exposure. Results from a developmental toxicity study in rats also raise
concern for possible neurodevelopmental effects. Decreased offspring weight at birth and
changed behavioral habituation of the offspring to novel environments were seen following
exposure of adult Long-Evans rats to 4,500 ppm for 14 days prior to mating and during gestation
213

-------
6519
6520
6521
6522
6523
6524
6525
6526
6527
6528
6529
6530
6531
6532
6533
6534
6535
6536
6537
6538
6539
6540
6541
6542
6543
6544
6545
6546
6547
6548
6549
6550
6551
6552
6553
6554
6555
6556
(or during gestation alone) (Bornschein et al., 1980; Hardin and Manson, 1980). In the only
other animal study examining possible early-life susceptibility to dichloromethane toxicity,
Alexeef and Kilgore (1983) found that exposure of young male mice to approximately
47,000 ppm for about 20 seconds significantly impaired the ability to learn, using a passive-
avoidance conditioning task. Three-week-old mice were more affected than 5- or 8-week-old
mice. The broad issue of childhood susceptibility to chronic neurobehavioral effects of early life
exposure represents a data gap in the understanding of the health effects of dichloromethane.
The relatively low CYP2E1 activity in the liver of infants would tend to shift metabolism
of dichloromethane to the GST pathway. This shift could affect cancer risk, given the evidence
of genotoxicity through this metabolic pathway. However, the available data in humans are not
sufficient to address the question of whether in utero or early life exposures represent a period of
increased susceptibility to potential carcinogenic effects of dichloromethane. A threefold
increased risk of childhood leukemia (acute lymphoblastic leukemia) was seen in relation to
maternal occupational exposure in the year before and during pregnancy in one population-based
case-control study (OR 3.22 [95% CI 0.88, 11.7]) for ratings of "probable or definite" exposure
compared with possible or no exposure (Infante-Rivard et al., 2005). The estimates for
categories based on concentration and frequency were similar, but there was no evidence for an
increasing risk with increasing exposure level.
Experiments comparing cancer responses from early-life exposures with those from adult
exposures are not available for F344 rats or B6C3Fi mice, the strains of animals in which
carcinogenic responses to dichloromethane have been observed (mammary gland tumors in F344
rats and liver and lung tumors in B6C3Fi mice exposed by inhalation; liver tumors in female
F344 rats and male B6C3Fi mice exposed via drinking water). Animal data evaluating the effect
of age on the susceptibility to dichloromethane carcinogenicity are restricted to a bioassay in
which 54 pregnant Sprague-Dawley rats were exposed, starting on GD 12, to 100 ppm
dichloromethane 4 hours/day, 5 days/week for 7 weeks, followed by 7 hours/day, 5 days/week
for 97 weeks (Maltoni et al., 1988). Groups of 60 male and 69 female newborns continued to be
exposed after birth to 60 ppm dichloromethane 4 hours/day, 5 days/week for 7 weeks, followed
by exposure 7 hours/day, 5 days/week for 97 weeks. Additional groups of 60 male and
70 female newborns were exposed after birth to 60 ppm dichloromethane 4 hours/day,
5 days/week for 7 weeks and then for 7 hours/day, 5 days/week for 8 weeks. Endpoints
monitored included clinical signs, BW, and full necropsy at sacrifice (when spontaneous death
occurred). For each animal sacrificed, histopathologic examinations were performed on the
following organs: brain and cerebellum, zymbal glands, interscapular brown fat, salivary glands,
tongue, thymus and mediastinal lymph nodes, lungs, liver, kidneys, adrenals, spleen, pancreas,
esophagus, stomach, intestine, bladder, uterus, gonads, and any other organs with gross lesions.
There was no significant effect of exposure to dichloromethane on the incidence of benign or
malignant tumors among adults or the progeny. The results provide no evidence that Sprague-
214

-------
6557
6558
6559
6560
6561
6562
6563
6564
6565
6566
6567
6568
6569
6570
6571
6572
6573
6574
6575
6576
6577
6578
6579
6580
6581
6582
6583
6584
6585
6586
6587
6588
6589
6590
6591
6592
6593
6594
Dawley rats would be more sensitive to potential carcinogenic activity of dichloromethane
during early life stages. Further conclusions from these results are precluded because the study
included only one exposure level, which was below the maximum tolerated dose for adult
Sprague-Dawley rats.
4.8.2.	Possible Gender Differences
The limited data available from studies in humans do not indicate that there are large
differences by gender in sensitivity to cardiovascular, neurologic, cancer, or other effects; studies
have not been conducted specifically to examine this question and so do not provide information
pertaining to smaller or more subtle differences. The available animal studies similarly do not
establish whether either gender may be more susceptible to the toxic effects of dichloromethane.
Studies of the carcinogenic effects of dichloromethane, either by inhalation or by the oral route,
have not suggested an increased susceptibility of either male or female animals.
4.8.3.	Other Susceptible Populations
As discussed in section 3.3, a polymorphism exists within the GST-T1 gene in humans,
resulting in individuals with diminished, or a lack of, ability to conjugate GSH to
dichloromethane. While the possible effects of this polymorphism on the toxicity of
dichloromethane have not been directly demonstrated, it can be inferred from the proposed mode
of action that a decrease in the GST-T1 metabolic pathway would result in a decreased
generation of reactive metabolites and a decrease in any chronic effects mediated through those
metabolites (Jonsson and Johanson, 2001; El-Masri et al., 1999).
Interindividual variation in the ability to metabolize dichloromethane via GST-T1 is
associated with genetic polymorphisms in humans. Estimated U.S. population prevalence of
nonconjugators (-/- at the GST-T1 locus) is about 20%, but higher prevalences (47-64%) have
been reported for Asians (Raimondi et al., 2006; Haber et al., 2002; Garte et al., 2001; Nelson et
al., 1995). Although nonconjugators are expected to have negligible extra risk for
dichloromethane-induced cancer, the U.S. prevalences for low (+/- at the GST-T1 locus) and
high (+/+) conjugators have been estimated at 48 and 32%, respectively (Haber et al., 2002).
The liver and kidney are the most enriched tissues in GST-T1, but evidence is available for the
presence of GST-T1 in other tissues, including the brain and lung, at lower levels (Sherratt et al.,
2002, 1997).
Individuals may vary in their ability to metabolize dichloromethane through the CYP2E1
pathway. Individuals with decreased CYP2E1 activity may experience decreased generation of
CO and an increased level of GST-related metabolites, following exposure to dichloromethane,
which may result in increased susceptibility to the chronic effects of dichloromethane from GST-
related metabolites. Conversely, individuals with higher CYP2E1 activity may experience
relatively increased generation of CO at a given dichloromethane exposure level and, therefore,
215

-------
6595
6596
6597
6598
6599
6600
6601
6602
6603
6604
6605
may be more susceptible to the acute CO-related toxicity or other chronic effects of
dichloromethane. Several studies indicate a three- to sevenfold variability in CYP2E1 activity
among humans, as assessed by various types of measurements among "healthy" volunteers
(Sweeney et al., 2004; Haufroid et al., 2003; Lipscomb et al., 2003; Bernauer et al., 2000; Lucas
et al., 2001, 1999; Kim et al., 1995; Shimada et al., 1994). This variability is incorporated into
the PBTK models for dichloromethane. Factors that may induce or inhibit CYP2E1 activity
(e.g., obesity, alcohol use, diabetes) or co-exposures (i.e., to various solvents or medications)
(Lucas et al., 1999) may result in greater variation within segments of the population. This
variation in CYP2E1 activity may result in earlier saturation of this pathway and greater
exposure to the parent compound, which would be of particular relevance to neurological effects.
216

-------
6606
6607
6608
6609
6610
6611
6612
6613
6614
6615
6616
6617
6618
6619
6620
6621
6622
6623
6624
6625
6626
6627
6628
6629
6630
6631
6632
6633
6634
6635
6636
6637
6638
6639
6640
6641
6642
6643
5. DOSE-RESPONSE ASSESSMENTS
5.1. ORAL REFERENCE DOSE (RfD)
5.1.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
As discussed in section 4.6.1, human data for oral exposures to dichloromethane are
limited to case reports involving intentional (i.e., suicidal) or accidental, acute ingestion
exposures (Chang et al., 1999; Hughes and Tracey, 1993). Reported effects reflect frank toxicity
from very high doses, such as marked CNS depression, injury to the gastrointestinal tract, liver
and kidney failure, coma, and death. No studies of human chronic oral exposures are available.
In the absence of adequate studies evaluating possible health effects in humans repeatedly
exposed to dichloromethane via the oral route, the results from the chronic laboratory animal
studies are assumed to be relevant to humans.
The database of laboratory animal oral exposure studies includes 90-day (Kirschman et
al., 1986) and 2-year drinking water toxicity studies in F344 rats (Serota et al., 1986a) and
B6C3Fi mice (Serota et al., 1986b). A reproductive study exposed Charles River CD rats via
gavage before mating (General Electric Co., 1976), and a developmental study exposed F344 rats
via gavage during GDs 6-19 (Narotsky and Kavlock, 1995). A 14-day gavage study examined
neurotoxicity in F344 rats (Moser et al., 1995).
Hepatic effects (hepatic vacuolation, nonneoplastic liver foci) are the critical dose-
dependent noncancer effects associated with oral exposure to dichloromethane (see Table 4-35).
The 90-day drinking water toxicity study in F344 rats (Kirschman et al., 1986) reported
significant increases in hepatocyte vacuolation and necrosis in animals dosed between 166 -
1200 mg/kg-day (males) or 200 - 1469 mg/kg-day (females). These doses were used to develop
dosing levels for the 104-week drinking water study (Serota et al., 1986a). The 104-week
drinking water study of F344 rats (Serota et al., 1986a) provides adequate data to describe dose-
response relationships for nonneoplastic liver lesions from chronic oral exposure to
dichloromethane (e.g., includes four exposure levels and a control group). In this study, rats
dosed at 50 mg/kg-day or higher in both sexes had increased fatty livers, but quantitative data
were not provided by the authors. Liver lesions, described as foci or areas of cellular alteration,
were also seen in this study in the same dose groups in which the fatty changes had occurred. A
limitation of this study is that Serota et al (1986a) did not describe the evaluation of the altered
foci in detail. However, increases in altered foci did not correspond to tumor rate incidences in
either male or female rats. Instead, the altered foci correlated more closely to fatty liver
incidence changes for both sexes in the rats. Altered foci could range from a focal fatty change
(nonneoplastic) to an enzymatic altered foci change (neoplastic) (Goodman et al., 1994). Several
lines of evidence were considered in determining whether the lesions should be characterized as
nonneoplastic or neoplastic: 1) There is a congruence between the incidence of this lesion and
217

-------
6644
6645
6646
6647
6648
6649
6650
6651
6652
6653
6654
6655
6656
6657
6658
6659
6660
6661
6662
6663
6664
6665
6666
6667
6668
6669
6670
6671
6672
6673
the incidence of the fatty liver in the study by Serota et al. (1986a); 2) At higher doses,
hepatocyte vacuolation and hepatocyte necrosis were seen (Kirschman et al., 1986; Berman et
al., 1995); and 3) there is no clear indication that these altered foci progress to liver tumors since
the rate of increased foci did not correlate with liver tumor increases in either male or female
rats. Based on these observations, the altered foci were determined to be more likely to be
representative of a focal fatty change (nonneoplastic) than a neoplastic event.
The LOAELs for nonneoplastic liver lesions in rodents following repeated oral exposure
(50-586 mg/kg-day) (Table 4-35) are in the same range or below the NOAELs of 225 mg/kg-day
for reproductive performance in Charles River CD rats exposed for 90 days before mating
(General Electric Co., 1976) and 450 mg/kg-day for developmental toxicity in pregnant F344
rats exposed during gestation (Narotsky and Kavlock, 1995). The LOAEL (337 mg/kg-day) and
NOAEL (101 mg/kg-day) for mild neurological impairment in a 14-day gavage exposure study
of F344 rats (Moser et al., 1995) indicates that the threshold for neurological effects may be
similar to the threshold for liver effects. A limitation of the Moser et al. (1995) study, however,
is that the observed effects were limited to measures taken within 4 hours of exposure.
The subchronic (i.e., 90-day or less study) data were not considered in the selection of a
principal study for deriving the chronic RfD because the database contains reliable dose-response
data from a chronic study at lower doses than the 90-day study (Kirschman et al., 1986)
(conducted to provide data pertaining to relevant doses to use in the chronic study). The data
from the subchronic studies are, however, used to corroborate the findings in the chronic studies
with respect to relevant endpoints (i.e., hepatic and neurological effects). The neurotoxicity
study was not selected as the principal study due to the limited measurements to inform the
chronic exposure to dichloromethane. The rat rather than the mouse chronic bioassay (Serota et
al., 1986a) was selected as the principal study for the RfD because of the consistent evidence that
rats may be more sensitive than mice to nonneoplastic liver effects from orally administered
dichloromethane; available rat LOAELs for nonneoplastic liver lesions are lower than mouse
LOAELs (see Table 4-35). Figure 5-1 is an exposure-response array that presents NOAELs,
LOAELs, and the dose range tested, corresponding to selected health effects from the short-term
(neurotoxicological) and subchronic studies, and from the chronic, reproductive, and
developmental toxicity studies that were evaluated for use in the derivation of the RfD.
218

-------
2000
1400
1200
« 1000
ONOAEL
¦ LOAEL
The vertical lines =
range of exposures in
study.
Closed dots (•) =
exposure concentrations
used in study
£
B
9J
5/5
O
—
800
600
400
200
o
~
o

t
Fatty liver
Nonneoplastic (B6C3F1
liver foci (F344 mice, male
rat, male and and
female) - Serota female) -
etal. (1986a) Serota et
al. (1986b)
CHRONIC HEPATIC
* ?
Flepatocyte
vacuolation (F344
rat) - Kirschman et
al. (1986)
Flepatocyte
vacuolation
(B6C3F1 male
mice) -
Kirschman et
al. (1986)
SUBCHRONIC HEPATIC
Neurologic,
Functional
Observational
Battery (F344
rat, female) -
Moser et al.
(1995)
NEUROTOX
Reproductive
Performance
(CD rat, male
and female) -
General
Electric Co.
(1976)
Reproductive
organs;
performance
(male Swiss
Webster) -
Raje et al.
(1988)
Maternal weight
gain (F344 rat,
pregnant female)
- Narotsky and
Kavlock (1995)
Fetal
Toxicity
(F344 rat) -
Narotsky
and Kavlock
(1995)
REPRODUCTIVE AND DEVELOPMENTAL
6674
6675
6676
Figure 5-1. Exposure Response Array for Oral Exposure to Dichloromethane.
Figure 5-1. Exposure response array for oral exposure to dichloromethane
219

-------
6677
6678
6679
6680
6681
6682
6683
6684
6685
6686
6687
6688
6689
6690
6691
6692
6693
6694
6695
6696
6697
6698
6699
6700
6701
6702
6703
6704
6705
6706
6707
6708
6709
6710
6711
6712
6713
6714
5.1.2. Derivation Process for Noncancer Reference Values
The toxicity values (oral RfD and inhalation RfC) for noncancer endpoints were derived
by using rat and human PBTK models to calculate internal doses in rats from experimental
exposures and extrapolate points of departure to human equivalent exposures. Figure 5-2
illustrates the process of using the PBTK models for toxicity value derivation. The process for
the RfD and RfC is summarized below, using the example of a noncancer liver effect.
A deterministic PBTK model for dichloromethane in rats was first used to convert rat
drinking water or inhalation exposures to values of an internal liver dose metric (see Appendix C
for details of the rat PBTK model). Available models in EPA benchmark dose (BMD) software
(BMDS) version 2.0 were then fit to the liver lesion incidence data and internal liver dose data
for rats, and BMDi0s and their lower 95% confidence limits associated with a 10% extra risk
(BMDLio) were calculated from each of the models. Adequacy of model fit was assessed by
overall % goodness of fit (p-value >0.10) and examination of residuals, particularly in the region
of the benchmark response (BMR). The choice of best-fitting model was based on the lowest
Akaike's Information Criterion [AIC] among models with adequate fits (U.S. EPA, 2000b).
The use of a PBTK model can replace the use of the BW°75 scaling factor to account for
interspecies differences in toxicokinetics. The decision with respect to use of a scaling factor
depends on the dose metric that is used. Where PBTK models predict the concentration (in
particular, the AUC) of the proximate causative agent, a scaling factor to account for interspecies
differences is not typically used. That is, it is assumed that if the time-averaged (or steady-state)
concentration of the proximate causative agent predicted by the PBTK model in the target tissue
is the same in the test species as in humans, and the test species was exposed for an equivalent
portion of its lifetime (2 years in rats and mice being equivalent to a 70-year lifetime in humans),
then the resulting risks in the two species are the same. However, when the PBTK model
predicts the rate of production of the agent, rather than its concentration, then a BW° 75 scaling
factor may be appropriate, depending on what is known or expected regarding the rate of
clearance of the agent or metabolite of interest. Two different scenarios can be considered. If
the metabolite formed is considered to be highly reactive, then it can be assumed that the rate of
clearance (i.e., disappearance due to local reactivity) for this metabolite, per volume tissue, is
equal in rodents and humans. Thus, in that situation, as with the AUC dose metric, no BW° 75
scaling factor is necessary, although differences in tissue volume fraction in humans versus rats
(as occurs for liver) should be and are accounted for by the PBTK model. However, if the
metabolite is not highly reactive, then it is expected that interspecies differences in clearance or
removal of the toxic metabolite follow the generally assumed BW°75 scaling for rates of
metabolism and blood circulation. In this case, or in situations in which the reactivity or rate of
removal of the metabolite has not been established, it is appropriate to use a scaling factor, based
on BW ratios, to account for this difference. In the case of the noncancer liver effects of
dichloromethane, very limited information is available on the mechanism(s) involved in creating
220

-------
Benchmark Dose Analysis
Rodent Dose
Response Data
PBTK
Model
Estimates of Rodent
Internal Dose
BMD
Modeling
Monte Carlo Sampling from
Distributions of Human Model
Parameters
Probabilistic
Human PBTK Model
(What administered doses will
produce a BMDL10 in a
population?)
Multistage
£
JVLDL
BMD
Scaling
Factor
Rodent Internal BMDL
Human Internal
BMDLln
95% Lower Bound Estimate of Internal
Dose Associated with a 10% response
Distribution of Human Equivalent Administered
Doses (mg/kg) or Inhalation Concentrations (nig/1113)
(Points of Departure)
Divide by Uncertainty Factors
for Interspecies Toxicodynamic
Variability, Human
Toxicodynamic Variability and
Database Deficiencies )
Oral Reference Doses or
Inhalation References
Concentrations
Recommend lower percentile (e.g.,
1st) to protect sensitive individuals
6715
6716	Figure 5-2. Process for deriving noncancer oral RfDs and inhalation RfCs using rodent and human PBTK
6717	models.
221

-------
6718
6719
6720
6721
6722
6723
6724
6725
6726
6727
6728
6729
6730
6731
6732
6733
6734
6735
6736
6737
6738
6739
6740
6741
6742
6743
6744
6745
6746
6747
6748
6749
6750
6751
6752
6753
6754
the type of hepatic damage seen. The dose metric used in the PBTK modeling is a rate of
metabolism, rather than the concentration of putative toxic metabolites, and the clearance of
these metabolites may be slower per volume tissue in the human compared with the rat. Thus the
rat internal dose metric for noncancer effects was adjusted by dividing by a pharmacokinetic
scaling factor to obtain a human-equivalent internal BMDLio
A probabilistic PBTK model for dichloromethane in humans, adapted from the model
of David et al. (2006) as described in Appendix B, was then used to calculate distributions of
chronic exposures associated with the human equivalent internal BMDLio, based on the
responses in rats. Parameters in the human PBTK model are distributions that incorporate
information about dichloromethane toxicokinetic and physiological variability and
uncertainty among humans, incorporating information about both the CYP2E1 and GST-T1
metabolic pathways (see Table 3-9 and Appendix B). Monte Carlo sampling was performed
in which each human model parameter was defined by a value randomly drawn from each
respective parameter distribution. The model was then executed by using the human internal
BMDLio as input, and the resulting human equivalent administered dose or human equivalent
concentration (HEC) was recorded. This process was repeated for 10,000 iterations to
generate a distribution of human equivalent administered doses or concentrations.
The parameter statistics reported by David et al. (2006) include both the inter-individual
variability that would have been elucidated by the Bayesian analysis (variation between mean
values for each individual for which data were available) and uncertainty in those values. Since
EPA's objective is to account for both population variability and parameter uncertainty,
however, these statistics were primarily used as published in David et al. (2006) (exceptions
discussed in Appendix B) to define population distributions. Assuming that these parameters are
distributed independently, ignoring the covariance that was likely represented in the actual
posterior chains, will tend to over-estimate the overall range of parameters and hence distribution
of dose metrics in the population, compared to what one would obtain if the covariance were
explicitly included. Thus if the covariance (i.e., the variance-covariance matrix) for the set of
parameters had been reported by David et al., it could have been used to narrow the predicted
distribution of internal doses, or equivalent applied doses. Lacking such information the
approach used will not under-estimate risk or over-estimate lower bounds on human equivalent
exposure levels.
From these distributions of human equivalent administered doses (or concentrations),
candidate RfDs or RfCs were derived by dividing the first percentile value (point of
departure) by uncertainty factors (UFs) to account for uncertainty about potential interspecies
toxicodynamic variability, human toxicodynamic variability, and database deficiencies. The
first percentile was chosen because it allowed generation of a stable estimate for the lower
end of the distribution while being protective of the overall human population, including
222

-------
6755
6756
6757
6758
6759
6760
6761
6762
6763
6764
6765
6766
6767
6768
6769
6770
6771
6772
6773
6774
6775
6776
6777
6778
6779
6780
6781
6782
6783
6784
6785
6786
6787
6788
6789
6790
6791
6792
sensitive individuals. Choosing this lower point replaces the use of an additional UF to
account for human toxicokinetic variability.
5.1.3. Evaluation of Dose Metrics for Use in Noncancer Reference Value Derivations
There are no data to support the role of a specific metabolite in the development of
the noncancer liver lesions seen in oral and inhalation exposure studies. Four dose metrics
were examined as potential metrics for the internal dose of interest: rate of hepatic
metabolism through the CYP pathway, rate of hepatic metabolism through the GST pathway,
the combined rate of hepatic metabolism through the CYP and GST pathways, and the
concentration (area under the curve, AUC) of dichloromethane, the parent compound, in the
liver. The dose-response patterns for each of these metrics in the oral study in rats (Serota et
al., 1986a) and in two inhalation studies in rats (Nitscke et al., 1988a; Burek et al., 1984)
were examined for fit and congruence.
Using the oral exposure data, only one of the seven models, the log-logistic model,
produced an adequately fit (p >0.10) for the GST metabolism metric and the
dichloromethane AUC metrics. Adequate model fit was seen in all of the models using the
CYP dose metric with the oral data, and using the GST, CYP, and AUC dose metrics for the
inhalation data.
A limitation in using the GST metric can be observed when comparing the oral and
inhalation responses at various exposure levels. At 200 ppm, where the GST metric is
predicted by the PBTK model to be 93 mg metabolism/L liver/day, no liver effects were
seen. In contrast, liver responses were elevated at an oral dose of 50 mg/kg-day, where the
GST metric is predicted to be 60 mg metabolism/L liver/day (see Tables 5-1 and 5-5,
respectively, for the oral and inhalation internal metrics). Thus the liver GST metric
produces an inconsistency in the dose-response relationship, with very different responses
observed depending on the route of exposure. A similar inconsistency occurs with the AUC
metric. These differences are not observed, however, when using the CYP metric. At the
200 ppm inhalation exposure, where no hepatoxicity was observed, the CYP metric is
predicted to be 660 mg/L liver/day. This internal CYP metabolism metric is less than that
predicted for the oral dose for the 50 mg/kg-day group (i.e., 872 mg metabolism/L liver/day),
in which liver effects were observed. Thus, the CYP internal metric is consistent with the
observed responses seen in the oral and inhalation exposure studies.
The GST metabolism and the AUC dose metrics did not present reasonable choices
based on model fit and consistency of response across studies at comparable dose levels.
Given these results, the combination of hepatic metabolism through the GST and the CYP
pathways would not be expected to result in an improvement to a metric based only on CYP
metabolism. The CYP-metabolism dose metric is the most consistent with the data. This
metric was selected for the subsequent RfD and RfC derivations. The lack of information on
223

-------
6793
6794
6795
6796
6797
6798
6799
6800
6801
6802
6803
6804
6805
6806
6807
6808
6809
6810
6811
6812
6813
6814
mechanisms with respect to noncancer health effects represents data gaps in the
understanding of the health effects of dichloromethane.
5.1.4. Methods of Analysis—Including Models (PBTK, BMD, etc.)
PBTK models for dichloromethane in rats were described previously in section 3.5.
From the evaluation described in Appendix C, a modified model of Andersen et al. (1991)
was selected for the calculation of internal dosimetry of ingested dichloromethane in the rats
in the principal study (Serota et al., 1986a).
PBTK model simulations of the drinking water study of Serota et al. (1986a) (Table
5-1) were performed to calculate average lifetime daily internal liver doses in male and
female F344 rats. In the absence of data for group- and sex-specific BWs, reference values
were used for male and female F344 rats in chronic studies (U.S. EPA, 1988a). The mode of
action by which dichloromethane induces nonneoplastic liver effects in rodents has not
received research attention to determine the role of the parent material, metabolites of the
CYP2E1 pathway, metabolites of the GST pathway, or some combination of parent material
and metabolites. In the absence of this kind of knowledge, and considering the pattern of
response seen in the oral and inhalation studies (as described in section 5.1.3.), an internal
dose metric based on the amount of dichloromethane metabolized via the CYP pathway in
the liver (mg dichloromethane metabolized via CYP pathway per liter liver per day) was
used. Figure 5-3 shows the comparison between oral external and internal doses, using this
dose metric for the rat and for the human.
224

-------
Table 5-1. Incidence data for nonneoplastic liver lesions and internal liver
doses, based on various metrics, in male and female F344 rats exposed to
dichloromethane in drinking water for 2 years (Serota et al., 1986a)
Nominal
(actual)		Rat internal liver doseb

daily intake
Rat liver


GST and
Parent
Sex
(mg/kg-day)
lesion incidence"
CYP
GST
CYP
AUC
Male
0(0)
52/76 (68%)
0
0
0
0
(BW =
5(6)
22/34 (65%)
133.9
2.1
136.1
0.5
380 g)
50 (52)
35/38 (92%)°
872.7
58.8
931.4
13.1

125 (125)
34/35 (97%)°
1,433.1
236.0
1,669.1
52.6

250 (235)
40/41 (98%)°
1,868.6
561.5
2,430.0
125.0
Female
0(0)
34/67 (51%)
0
0
0
0
(BW =
5(6)
12/29 (41%)
134.5
2.1
136.6
0.4
229 g)
50 (58)
30/41 (73%)°
977.8
66.0
1,043.8
12.6

125 (136)
34/38 (89%)°
1577.0
258.7
1,835.7
49.5

250 (263)
31/34 (91%)°
2070.0
642.4
2,712.3
122.9
aLiver foci/areas of cellular alteration; number affected divided by total sample size,
internal doses were estimated using a rat PBTK model from simulations of actual daily doses
reported by the study authors. CYP dose is in units of mg dichloromethane metabolized via CYP
pathway/L tissue/day; GST dose is in units of mg dichloromethane metabolized via GST pathway/L
tissue/day.; GST and CYP dose is in units of mg dichloromethane metabolized via CYP and GST
pathways/L tissue/day; and Parent AUC dose is in units of mg dichloromethane*hrs)/L tissue.
Significantly (p < 0.05) different from control with Fisher's exact test.
Source: Serota etal., 1986a.
6817
6818
6819
225

-------
6820
6821
6822
6823
6824
6825
6826
6827
6828
6829
6830
6831
6832
6833
6834
6835
6836
6837
6838
6839
6840
6841
6842
6843
6844
6845
10,000
>
 o
— w
^ O
* -
I J5 1,000
O L.
,0 flj
rs *->
4-> f
E
s
u
D
os
E
100
CYP
metabolism
	
	

	
	
	
	4
	4










	|










	j









~ ~ '
~ * 1






~
~ <
~ -
~ +
	u	I
I- " j

	<
<
~
>
	¦
•
		¦
~
¦J
jjj
ij
, -
k...$	O	.5
i		i	¦L**-"
4	i	*	



~ ~ *
	 	j
i	 ¦ 	

~ <



	;
V

r



Rat

¦
«
~





T
~	





¦ numan mean
~	Human 5th %
~	Human 95th %
*












10
100
Oral dose (mg/kg/d)
Figure 5-3. PBTK model-derived internal doses (nig dichloromethane metabolized via the
CYP pathway per liter liver per day) in rats and humans and their associated external
exposures (mg/kg-day), used for the derivation of RfDs. Six simulated daily drinking water
episodes are described by Reitz et al. (1997). The human metabolism rates were estimated
using a computational sample of 1000 individuals per dose, including random samples of the
three GST-T1 polymorphisms (+/+, +/-, -/-) in the current U.S. population based on data from
Haber et al. (2002). Since a different set of samples was used for each dose, some stochasticity
is evident as the human points (values) do not fall on smooth curves.
The seven dichotomous dose-response models in BMDS version 2.0 were fit to the rat
liver lesion incidence data and PBTK model-derived internal dose data to derive a rat internal
BMD „, and corresponding BMDL H, associated with 10% extra risk (Table 5-2). The quantal
model is identical to the one-stage multistage model and so is not included in this set of models.
A BMR of 10% was selected because, in the absence of information regarding the magnitude of
change in a response that is thought to be minimally biologically significant, a BMR of 10% is
generally recommended, since it provides a consistent basis of comparison across assessments.
There are no additional data to suggest that the critical response has a greater sensitivity that
would warrant a lower BMR. The male rats exhibited a greater sensitivity compared to the
female rats (based on lower BMDLio values for all of the models examined) and thus the male
data are used as the basis for the RfD derivation. The logistic model was the best fitting model
for the male incidence data, based on AIC value among models with adequate fit (U.S.
EPA,2000b). Modeling results are shown in detail in Appendix D-l).
226

-------
6846
6847
6848
6849
6850
6851
6852
6853
6854
6855
6856
6857
Table 5-2. BMD modeling results for incidence of noncancer liver lesions
in male and female F344 rats exposed to dichloromethane in drinking
water for 2 years, based liver-specific CYP metabolism dose metric (mg
dichloromethane metabolism via CYP pathway per liter liver tissue per
day)



goodness of fit

Sex and model3
BMDio
BMDL10
p-value
AIC
Males




Gamma3
151.73
48.93
0.62
185.33
Logistic
85.17
61.78
0.75
183.61
Log-logistica
213.73
37.06
0.83
184.79
Multistage (l)a
68.62
47.58
0.71
183.74
Probit
98.87
75.49
0.69
183.81
Log-probita
197.65
77.56
0.81
184.84
Weibulf
117.29
48.39
0.57
185.49
Females




Gamma3
336.38
98.70
0.52
233.07
Logistic
169.77
134.87
0.59
231.70
Log-logistica
404.87
101.15
0.60
232.80
Multistage (l)a
123.59
91.46
0.47
232.32
Probit
179.59
146.27
0.59
231.70
Log-probita
400.95
173.57
0.60
232.80
Weibulf
283.24
97.31
0.47
233.27
aThese models in EPA BMDS version 2.0 were fit to the rat dose-response data shown in
Table 5-1 by using internal dose metrics calculated with the rat PBTK model. Details of the
models are as follows: Gamma and Weibull models restrict power >1; Log-logistic and
Log-probit models restrict to slope >1, multistage model restrict betas >0; lowest degree
polynomial with an adequate fit is reported (degree of polynomial noted in parentheses).
Bolded model is the best-fitting model in the most sensitive sex (males), which is used in
the RfD derivation.
Source: Serota et al. (1986a).
The BMDL io from the logistic model was used as the point of departure for the RfD
calculations (Table 5-3). This rat internal dose metric for noncancer effects was adjusted to
obtain a human-equivalent internal BMDLio by dividing by a pharmacokinetic scaling factor
based on a ratio of BWs (BWhuman/BWrat)0'25 = 4.09). This scaling factor was used because the
metric is a rate of metabolism, rather than the concentration of putative toxic metabolites, and the
clearance of these metabolites may be slower per volume tissue in the human compared with the
rat (that is, total rate of removal may scale as BW0'75, while tissue volume scales as BW1).
The human PBTK model (adapted from David et al. [2006], as described in Appendix B),
using Monte Carlo sampling techniques, was used to calculate quantiles of human equivalent
administered oral daily doses (in mg/kg-day) associated with the internal BMDLio values
227

-------
6858
6859
6860
6861
6862
6863
6864
6865
6866
6867
6868
6869
6870
6871
(Table 5-3), as described above in section 5.1.2. The human model used parameter values
derived from Monte Carlo sampling of probability distributions for each parameter, including
MCMC-derived distributions for the metabolic parameters for the metabolism through the
CYP2E1 pathway (Vmax and Km) and a distribution of GST metabolic rate constants that is
weighted to reflect the estimated frequency of GST-T1 genotypes (20% GST-T1 7, 48% GST-
Tl+/~, and 32% GST-T1+/+) in the current U.S. population, based on data from Haber et al.
(2002). All simulations also included a distribution of CYP activity based on data from
Lipscomb et al. (2003). The drinking water exposures were comprised of six discrete drinking
water episodes for specified times and percentages of total daily intake (Reitz et al., 1997). The
mean and two lower points on the distributions of human equivalent administered daily doses
derived from the Serota et al. (1986a) data for male rats, using the BMDLio from the logistic
model, are shown in Table 5-3.
Table 5-3. RfD for dichloromethane based on PBTK model-derived
probability distributions of human drinking water exposures
extrapolated from nonneoplastic liver lesion incidence data for male
rats exposed via drinking water for 2 years, based on liver-specific
CYP metabolism dose metric (mg dichloromethane metabolized via
CYP pathway per liter liver tissue per day)

Rat
internal
BMDL10b
Human
Human equivalent dose
(mg/kg-day)d
Human
RfD
(mg/m3)e
Model3
internal
BMDL10C
jst gth
percentile percentile ^®ean
Logistic
61.78
15.11
0.214 0.252 0.395
7 x 10-3
aBased on the best-fitting model from Table 5-2.
bRat dichloromethane PBTK model-derived internal liver dose associated with
the lower bound on 10% extra risk for developing liver foci/areas of cellular
alteration.
cHuman dichloromethane internal liver dose, derived by dividing the rat internal
BMDLio by a scaling factor of 4.09 [(BWhUman/BWrat)0 25 ] to account for potential
interspecies pharmacokinetic differences in the clearance of metabolites.
dPBTK model-derived distributions of daily average dichloromethane drinking
water doses predicted by the PBTK model to yield an internal dose in humans
equal to the dichloromethane internal BMDLio,
eHuman RfD, based on male rat data, derived by dividing the 1st percentile of
human equivalent dose value by a total UF of 30: 3 (10°5) for possible
toxicodynamic differences between species, 3 (100 5) for variability in human
toxicodynamic response, and 3 (10°5) for database deficiencies. The 1st
percentile point of departure is a stable estimate of the lower end of the
distribution. Use of this value in the lower tail replaces use of a UF for human
toxicokinetic variability.
Source: Serota etal. (1986a).
228

-------
6872
6873
6874
6875
6876
6877
6878
6879
6880
6881
6882
6883
6884
6885
6886
6887
6888
6889
6890
6891
6892
6893
6894
6895
6896
6897
6898
6899
6900
6901
6902
6903
6904
6905
6906
6907
6908
6909
5.1.5. RfD Derivation—Including Application of Uncertainty Factors (UFs)
The 1st percentile point of departure is a stable estimate of the lower end of the
distribution. Use of this value associated with a sensitive human population addresses the
uncertainty associated with human toxicokinetic variability. To derive the candidate RfD based
on data from male rats, the first percentile value of the distribution of human equivalent
administered dose associated with the male rat-derived BMDLio was divided by a composite UF
of 30 (3 [10°5] to account for uncertainty about interspecies toxicodynamic equivalence, 3 [10°5]
to account for uncertainty about toxicodynamic variability in humans, and 3 [10°5] for database
deficiencies) (Table 5-3). The resulting RfD recommended for dichloromethane is 7 x 10 3
mg/kg-day.
In deriving this RfD, factors for the following areas of uncertainty were considered:
•	Uncertainty in extrapolating from laboratory animals to humans (UFa). The use of
PBTK models to extrapolate internal doses from rats to humans reduces toxicokinetic
uncertainty in extrapolating from the rat liver lesion data but does not account for the
possibility that humans may be more sensitive than rats to dichloromethane due to
toxicodynamic differences. A UF of 3 (10°5) to account for this toxicodynamic
uncertainty was used, as shown in Table 5-3.
•	Uncertainty about variation from average humans to sensitive humans (UFh)- The
probabilistic human PBTK model used in this assessment incorporates the best available
information about variability in toxicokinetic disposition of dichloromethane in humans
but does not account for humans who may be sensitive due to toxicodynamic factors.
Thus, a UF of 3 (10°5) was applied to account for possible toxicodynamic differences in
sensitive humans.
•	Uncertainty in extrapolating from LOAELs to NOAELs (UF[). A UF for extrapolation
from a LOAEL to a NOAEL was not applied because BMD modeling was used to
determine the POD, and this factor was addressed as one of the considerations in
selecting the BMR. The BMR was selected based on the assumption that it represents a
minimum biologically significant change.
•	Uncertainty in extrapolating from subchronic to chronic durations (UFs). The derived
RfD is based on results from a chronic-duration drinking water toxicity study. No cross-
duration UF is necessary.
•	Uncertainty reflecting incompleteness of the overall database (UFd)- The oral database
for dichloromethane includes well-conducted lifetime drinking water studies in rats
229

-------
6910
6911
6912
6913
6914
6915
6916
6917
6918
6919
6920
6921
6922
6923
6924
6925
6926
6927
6928
6929
6930
6931
6932
6933
6934
6935
6936
6937
6938
6939
6940
6941
6942
6943
6944
6945
6946
6947
(Serota et al., 1986a) and mice (Serota et al., 1986b) and a supporting subchronic study in
rats and mice (Kirschman et al., 1986). These studies provided dose-response data for
the hepatic effects of dichloromethane. The database also includes one-generation oral
reproductive toxicity (General Electric Co., 1976) and developmental toxicity (Narotsky
and Kavlock, 1995) studies that found no reproductive or developmental effects at dose
levels in the range of doses associated with liver lesions. A two-generation oral exposure
study is not available; however, a two-generation inhalation exposure study by Nitschke
et al. (1988a) reported no effect on fertility index, litter size, neonatal survival, growth
rates, or histopathologic lesions at exposures of >100 ppm. However, there have been no
oral exposure studies that evaluated neurobehavioral effects in offspring. This is a
relevant endpoint given the increase in blood CO (a known developmental neurotoxicant)
that occurs through the CYP2E1 metabolic pathway for dichloromethane after oral and
inhalation exposures. There are no oral exposure studies that include functional immune
assays; however, there is a 4-week inhalation study of potential systemic immunotoxicity
that found no effect of dichloromethane exposure at concentrations up to 5,000 ppm on
the antibody response to sheep red blood cells (Warbrick et al., 2003). The Warbrick et
al. (2003) data suggest that systemic immunosuppression is not a concern for
dichloromethane exposure. Route-specific local immunosuppression from acute
inhalation exposure in CD1 mice was seen in Aryani et al. (1986). The findings from
Aryani et al. (1986) were considered to be portal-of-entry effects involving local
immunosuppression within the lung (Streptococcus and Klebsiella infectivity models)
and unlikely to be observed following oral exposure. Because of concern regarding the
adequacy of available data pertaining to possible neurodevelopmental toxicity and the
lack of a two-generation reproductive study, a UFD of 3 was applied.
5.1.6.	Previous RfD Assessment
The previous IRIS assessment derived an RfD of 0.06 mg/kg-day based on the NOAELs
of 5.85 and 6.47 mg/kg-day for nonneoplastic liver toxicity (foci/areas of cellular alteration) in
male and female rats, respectively, in a 2-year drinking water study (Serota et al., 1986a). The
LOAELs associated with these NOAELs were 52.58 and 58.32 mg/kg-day for males and
females, respectively. The RfD of 0.06 mg/kg-day was derived by dividing the average NOAEL
of 6 mg/kg-day (for male and female rats) by a UF of 100 (10 for intraspecies variability and 10
for interspecies variability).
5.1.7.	RfD Comparison Information
Use of the mean value (3.95Q/Q0 1 mg/kg-day) of the human equivalent administered
dose distribution instead of the 1st percentile, with an additional UF of 3 (10°5) to account for
human toxicokinetic variability, would yield an RfD of 4 x | | [10 3 mg/kg-day.
230

-------
6948
6949
6950
6951
6952
6953
6954
6955
6956
6957
6958
6959
6960
Additional comparisons between the derived RfD and values developed from other
endpoints or data sets using NOAEL/LOAEL methods are shown in Table 5-4 and Figure 5-4.
NOAELs were used as comparison points of departure and were not scaled allometrically. The
point of departure for three endpoints (Serota et al., 1986a; Moser et al., 1995; Narotsky and
Kavlock) are presented in Table 5-4. For Serota et al. (1986a), this is based on BMD modeling
of a 10% increase in liver lesions using internal liver dose metric (mg dichloromethane
metabolism via CYP pathway per liter liver tissue per day) derived from a rat PBTK model.
After an allometric scaling factor of 4.09 was applied, the human internal BMDLio was 15.11
mg/kg-day. A probabilistic human PBTK model adapted from David et al. (2006) was used to
generate a distribution of human equivalent doses from the human internal BMDLio and the first
percentile of this distribution was used as the point of departure. For other studies, POD is based
on the lowest non-control dose in which no effect was seen. These NOAELs that were used as
points of departure and were not scaled allometrically
231

-------
6961
Table 5-4. Potential points of departure with applied UFs and resulting candidate RfDs



Uncertainty Factors Applied
b












Candidate


PODa
POD Type and
Total


ufl


RfD

Endpoint
(mg/kg-day)
Description
UF
ufa
UFh

UFS
ufd
(mg/kg-day)
Reference
Nonneoplastic
liver foci, male
rats
61.78
BMD; 10% increase
in incidence of liver
lesion
30
3
3
1
1
3
7 x 103
Serota et al.
(1986a)


NOAEL; No effect at








Neurological
changes (FOB),
female rats
101
POD, approximate
doubling of severity
score of neuromuscular
and sensorimotor
domains
3,000
10
10
1
10
3
3.4 x 10~2
Moser et al.
(1995)
Maternal weight
gain, female rats
338
NOAEL; No effect at
POD, approximate 33%
decrease in weight gain
seen at next dose
300
10
10
1
1
3
1.1
Narotsky and
Kavlock
(1995)
aPOD = point of departure. .
bUFA = uncertainty in extrapolating from laboratory animals to humans, UFH = uncertainty about variation from average humans to sensitive humans, UFL=
uncertainty about extrapolating from LOAEL to NOAEL, UFS = uncertainty in extrapolating from subchronic to chronic durations, and UFD = uncertainty
reflecting incompleteness of the overall database. A UF for extrapolation from a LOAEL to NOAEL (UFL) was not used for any of these studies. For the
Serota et al., (1986a) study, the use of the first percentile of the human equivalent dose distribution as the point of departure replaces the use of a UFH for
human toxicokinetic variability.
Bolded value is the basis for the RfD of 7 x 10"3 mg/kg-day.
232

-------
¦B
I"
"wo
E
w
V
•/:
S
O
Point of Departure
UFa - Interspecies;
animal to human
UFh - Intraspecies;
human variability
UFs - Subchronic to
clironic exposure
duration
I I UFD-Database
^ Reference Dose
Nonneoplastic liver foci
- first percentile Human
Equivalent
Admininstered Dose
from rat;
Serotaetal. (1988a)
Neurologic, Functional
Observational Battery-
NOAEL from rats; Moser
et al. (1995)
Maternal weight gam
- NOAEL from rats,
Narotsky and
Kavlock (1995)
6962
6963
6964	Figure 5-4. Comparison of candidate RfDs derived from selected point of departures for endpoints
6965	presented in Table 5-4.
6966
233

-------
6967
6968
6969
6970
6971
6972
6973
6974
6975
6976
6977
6978
6979
6980
6981
6982
6983
6984
6985
6986
6987
6988
6989
6990
6991
6992
6993
6994
6995
5.2. INHALATION REFERENCE CONCENTRATION (RfC)
5.2.1. Choice of Principal Study and Critical Effect—with Rationale and Justification
Figure 5-5 includes exposure-response arrays from some of the human studies that were
evaluated for use in the derivation of the RfC. Several acute-duration controlled exposure
studies (section 4.1.2.2) and cross-sectional occupational studies (sections 4.1.2.3 and 4.1.2.4) in
humans are available that show neurological effects from dichloromethane exposure. These
effects include an increase in prevalence of neurological symptoms among workers (Cherry et
al., 1981) and possible detriments in attention and reaction time in complex tasks among retired
workers (Lash et al., 1991). However, these studies have inadequate power for the detection of
effects with an acceptable level of precision. In addition, the Cherry et al. (1981) study is limited
by the definition and documentation of neurological symptom history, and the Lash et al. (1991)
study has exposure measurements from 1974-1986, but the work histories of exposed workers
go back to the 1940s. Ott et al. (1983c) reported an increase in serum bilirubin among exposed
workers, but there was no association seen with respect to the other hepatic enzymes examined
(serum y-glutamyl transferase, serum AST, serum ALT), and no evidence of hepatic effects was
seen in a later study of the same cohort (Soden, 1993). Because of these limitations, these
human studies of chronic exposures do not serve as an adequate basis for RfC derivation. As
discussed in section 5.2.6, however, the quantitative measures of neurological function from
Cherry et al. (1983) were used to derive a comparative RfC.
The database of experimental animal dichloromethane inhalation studies includes
numerous 90-day and 2-year studies, with data on hepatic, pulmonary, and neurological effects,
(see Table 4-36) and reproductive and developmental studies (Table 4-37) (see summary in
Section 4.6.2). NOAELs, LOAELs, and the dose range tested corresponding to selected health
effects from the chronic studies are shown in Figure 5-5, and effects seen in subchronic,
reproductive, and developmental studies are shown in Figure 5-6. The subchronic (i.e., 90-day
or less study) data were not considered in the selection of a principal study for deriving the RfC
because the database contains reliable dose-response data from the chronic study at lower doses
than the 90-day study. The data from the subchronic studies are, however, used to corroborate
the findings with respect to relevant endpoints (i.e., hepatic and neurological effects).
234

-------
10000
1000
a
3
a
o
"-C
•-
S3
w
S3
O
u
100
10
n
}
I I
ONOAEL BLOAEL
The vertical lines = range of
exposures in study.
Closed dots (0) = exposure
concentrations used in study
J-4
n r
3 ?
Hepatocyte
necrosis,
Sprague-
Dawley rat
(Burek et
al., 1984)
3 %
Hepatocyte
vacuolation,
Sprague-
Dawley rat
(Nitschke et
al., 1988a)
Hepatocyte
vacuolation,
male and
female
Sprague-
Dawley rat
(Burek et
al., 1984)
RAT
Hepatocyte
vacuolation,
necrosis,
hemosiderosis
in liver, male
and female
F344 rat
(Mennear et al
1988; NIP,
1986)
Renal tubular
degeneration,
male and
female F344
rat (Mennear
et al., 1988;
NTP, 1986)
Hepatocyte
degeneration,
male and
female
B6C3F1 mice
(Mennear et
al., 1988;
NTP, 1986)
Renal
tubule casts,
mlae and
female
B6C3F1
mice
(Mennear et
al., 1988;
NTP, 1986)
MOUSE
Changes	Chronic	Cardiac -
inCNS	neurological	ST segment
measures, effects,	depression,
males	males	males (Ott
(Lash et	(Cherry et	et al.,
al., 1991)	al., 1983)	1983c)
HUMAN
6996
6997
6998
6999	Figure 5-5. Exposure response array for chronic (animal) or occupational (human) inhalation exposure to
7000	dichloromethane (log Y axis)
235

-------
10,00Cr
1,00a
a
a,
¦w
s
0
1
s
u
&
©
U
100-
10
H
!
o
t
o
o
ONOAEL
¦ LOAEL
The vertical lines =
range of exposures in
study.
Closed dots (•) =
exposure concentrations
used in study
u
.[lipid :liver Hepatocyte
weight centrilobular1
ratios (F334 degeneratiorJ
rat, $ and (B6C3F1
$) - NTP, mice, $ and
1986 $) - NTP,
1986
HEPATIC
Foreign	Clara cell
body	vacuolation
pneumonia	(B6C3F1
(F344 rat,	mice, $ and
(J and?)-	¥)- Foster
NTP, IS86	etal, 1992
PULMONARY
Increased	Increased IgM
Infection	response
Susceptibility,	(Sprague-
(CD1 mice,	Dawleyrat* $
$) - Aranyi et	and $) -
al., 1986	Warbrick et al.,
2003
IMMUNO
FOB, Grip
Strength,
SEPs, (F344
rat, $ and $)
- Mattsson
et al., 1990
NEURO
Adverse | Maternal
fetal effects liver weight
t	t
(Swiss Webster
mice and
Sprague-D a wley
rats) -Schwetz,
1975
t Maternal
Fetal body
Reproductive
Reproductive
liver weight;
weight and
performance;
organs;
4 fetal
histopathology growth rates;
performance
bw/altered
(Sprague-
organ
(<£ Swiss
habituation
Dawley rat, $
histopath.
Webster) -
(F344 rats) -
and $)-
(F344 rats) -
Raje et al.
Hardin and
Maltoni et al.,
Nitschke et
1988
Manson,
1988
al, 1988b

1980;


Bomschein,



1980



REPRODUCTIVE AND DEVELOPMENTAL
7001
7002
7003	Figure 5-6. Exposure response array for subacute to subehronic inhalation exposure to dichloromethane (log Y
7004	axis)
236

-------
7005
7006
7007
7008
7009
7010
7011
7012
7013
7014
7015
7016
7017
7018
7019
7020
7021
7022
7023
7024
7025
7026
7027
7028
7029
7030
7031
7032
7033
7034
7035
7036
7037
7038
7039
7040
7041
7042
Hepatic effects (hepatic vacuolation and necrosis, hemosiderosis, hepatocyte
degeneration) are the critical dose-dependent noncancer effects associated with inhalation
exposure to dichloromethane. These effects were seen in mice (Mennear et al., 1988; NTP,
1986) and rats (Mennear et al., 1988; Nitschke et al., 1988a; NTP, 1986; Burek et al., 1984), but
not in Syrian golden hamsters (Burek et al., 1984). Inhalation bioassays with Sprague-Dawley
rats identified the lowest inhalation LOAEL for nonneoplastic liver lesions in the database: 500
ppm (6 hours/day, 5 days/week for 2 years) (Nitschke et al., 1988a; Burek et al., 1984), and
Nitschke et al. (1988a) identified a NOAEL of 200 ppm in female rats. Based on the results
reviewed above, nonneoplastic liver lesions (specifically, hepatic vacuolation) in rats are
identified as the critical noncancer effect from chronic dichloromethane inhalation in animals.
Because Nitschke et al. (1988a) examined a range of exposures that included doses at the low
end of the rangecompared with the range examined in Burek et al. (1984), the former study was
selected as the principal study for derivation of a chronic inhalation RfC.
Reproductive performance (e.g., as assessed by number of litters, resorption rate, fetal
survival, and growth) was not affected in two generations of F344 rats exposed to up to
1,500 ppm for 14 or 17 weeks before mating of the F0 and F1 generations, respectively
(Nitschke et al., 1988b) or in a study of Swiss-Webster mice or Sprague-Dawley rats exposed to
1,250 ppm on GDs 6-15 (Schwetz et al., 1975). A decrease in fertility index was seen in the 150
and 200 ppm groups in a study of male Swiss-Webster mice exposed via inhalation for 6 weeks
prior to mating (Raje et al., 1988), but the statistical significance of this effect varied
considerably depending on the statistical test used in this analysis. Two types of developmental
effects, decreased offspring weight at birth and changed behavioral habituation of the offspring
to novel environments, were seen in Long-Evans rats following exposure to 4,500 ppm for
14 days prior to mating and during gestation (or during gestation alone) (Bornschein et al., 1980;
Hardin and Manson, 1980). This dose was the only exposure dose used in this study. Schwetz et
al. (1975) did not observe an adverse effect on gross development or soft tissue abnormalities in
a study involving exposure to 1,250 ppm on GD 6 in Swiss-Webster mice or Sprague-Dawley
rats, but an increase in delayed ossification of the sternebrae was seen.
Neurological impairment was not seen in lifetime rodent bioassays involving exposure to
airborne dichloromethane concentrations of <2,000 ppm in F344 rats (Mennear et al., 1988;
NTP, 1986), <3,500 ppm in Sprague-Dawley rats (Nitschke et al., 1988a; Burek et al., 1984), or
<4,000 ppm in B6C3Fi mice (Mennear et al., 1988; NTP, 1986). It should be noted, however,
that these studies did not include standardized neurological or neurobehavioral testing. The sole
subchronic or chronic study in which neurobehavioral batteries were utilized found no effects in
an observational battery, a test of hind-limb grip strength, a battery of evoked potentials, or
brain, spinal cord, or peripheral nerve histology in F344 rats exposed to concentrations up to
2,000 ppm for 13 weeks, with the tests performed beginning 65 hours after the last exposure
(Mattsson et al., 1990).
237

-------
7043
7044
7045
7046
7047
7048
7049
7050
7051
7052
7053
7054
7055
7056
7057
7058
7059
7060
7061
7062
7063
7064
7065
7066
Other effects associated with lifetime inhalation exposure to dichloromethane include
renal tubular degeneration in F344 rats exposed to >2,000 ppm, testicular atrophy in male
B6C3Fi mice exposed to 4,000 ppm, and ovarian atrophy in female B6C3Fi mice exposed to
>2,000 ppm (Mennear et al., 1988; NTP, 1986). No effects on histologic, clinical chemistry,
urinalysis, or hematologic variables were found in Syrian golden hamsters exposed to
concentrations up to 3,500 ppm for 2 years, with the exception that the mean COHb percentage
of exposed hamsters was about 30%, compared with values of about 3% in controls (Burek et al.,
1984).
5.2.2.	Derivation Process for Reference Concentration Values
The derivation process used for the RfC parallels the process described in section 5.1.2
on the RfD derivation; consideration of dose metrics was described in section 5.1.3. As was
noted in the RfD discussion, the mechanistic issues with respect to noncancer health effects
represents data gaps in the understanding of the health effects of dichloromethane.
5.2.3.	Methods of Analysis—Including Models (PBTK, BMD, etc.)
The modified rat PBTK model of Andersen et al. (1991), described in Appendix C and
also used in the derivation of the RfD (Figure 5-2), was used for calculating internal dosimetry of
inhaled dichloromethane in Sprague-Dawley rats. Simulations of 6 hours/day, 5 days/week
inhalation exposures used in the Nitschke et al. (1988a) study were performed to calculate
average daily internal liver doses (Table 5-5). In the absence of data for group- and sex-specific
BWs, reference values for male and female Sprague-Dawley rats in chronic studies were used
(U.S. EPA, 1988a).
238

-------
7067
Table 5-5. Incidence data for nonneoplastic liver lesions (hepatic vacuolation) and
internal liver doses, based on various metrics, in female Sprague-Dawley rats exposed to
dichloromethane via inhalation for 2 years (Nitschke et al., 1988a)
Rat internal liver doseb
Sex
Exposure
(ppm)
Liver lesion
incidence3
CYP
GST
GST and
CYP
Parent
AUC
Male
0
22/70 (31)





50
Not reported
Not modeled because results for middle two doses were not reported

200
Not reported





500
28/70 (40)




Female
0
41/70 (59%)
0
0
0
0
(BW =
50
42/70 (60%)
280.3
6.3
286.6
1.2
229 g)
200
41/70 (58%)
656.5
93.2
749.7
17.8

500
53/70 (76%)°
772.6
359.0
1,131.6
68.7
aNumber affected divided by total sample size.
internal doses were estimated using a rat PBTK model using exposures reported by study authors (50 ppm =
174 mg/m3, 200 ppm = 695 mg/m3, and 500 ppm = 1737 mg/m3) and are weighted-average daily values for 1
week of exposure @ 6 h/day, 5 day/week. CYP dose is in units of mg dichloromethane metabolized via CYP
pathway/L tissue/day; GST dose is in units of mg dichloromethane metabolized via GST pathway/L tissue/day.;
GST and CYP dose is in units of mg dichloromethane metabolized via CYP and GST pathways/L tissue/day;
and Parent AUC dose is in units of mg dichloromethane*hrs)/L tissue.
Significantly (p < 0.05) different from control with Fisher's exact test.
Source: Nitschke et al., 1988a.
7068
7069	As described in section 5.1.2, the internal dose metric used was based on total hepatic
7070	metabolism through the CYP2E1 pathway (mg dichloromethane metabolized via CYP
7071	pathway/L liver/day). Figure 5-7 shows the comparison between inhalation external and internal
7072	doses, using this dose metric for the rat and the human.
7073
239

-------
7074
7075
7076
7077
7078
7079
7080
7081
7082
7083
7084
7085
7086
7087
7088
7089
7090
7091
7092
7093
7094
7095
7096
7097
7098
7099
>
"U
J_
SJ
>
T3
N
©
J2
ro
SJ
£
z
0
D
01
£
10,000
	
CD
w
O
"D
"TO
C
l_
fl)
1,000
100
10
10
CMP metabolism


— Rat
¦ Human mean
~	Human 5th %
~	Human 95th %
I I I i i I I
100	1,000
Inhalation concentration (ppm)
Figure 5-7. PBTK model-derived internal doses (mg dichloromethane metabolized
via the CYP pathway/L liver/day) in rats and humans versus external exposures
(ppm). Average daily doses were calculated from simulated rat exposures of 6 hours/day, 5
days/week, while simulated human exposures were continuous. The human metabolism rates
were estimated using a computational sample of 1000 individuals per dose, including random
samples of the three GST-T1 polymorphisms (+/+. +/-, -/-) in the current U.S. population based
on data from Haber et al. (2002). Since a different set of samples was used for each dose, some
stochasticity is evident as the human points (values) do not fall on smooth curves.
The seven dichotomous dose-response models available in EPA BMDS version 2.0 were
fit to the female rat liver lesion incidence of Nitschke et al. (1988a) and PBTK model-derived
internal dose data to derive rat internal BMDio and the associated BMDLio values (Table 5-6).
The quantal model is identical to the one-stage multistage model; therefore, it is not included in
this set of models. A BMR of 10% was selected because, in the absence of information
regarding the magnitude of change in a response that is thought to be minimally biologically
significant, a BMR of 10% is generally recommended, as it provides a consistent basis of
comparison across assessments. There are no additional data to suggest that the critical response
has a greater sensitivity that would warrant a lower BMR. The log-probit model was the best
fitting model for the female incidence data, based on AIC value among models with adequate fit.
Modeling results are shown in detail in Appendix D-2).
240

-------
7100
7101
7102
7103
7104
7105
7106
7107
7108
7109
7110
7111
7112
Table 5-6. BMD modeling results for incidence of noncancer liver
lesions in female Sprague-Dawley rats exposed to dichloromethane by
inhalation for 2 years, based on liver specific CYP metabolism metric
(mg dichloromethane metabolized via CYP pathway per liter liver
tissue per day)
Model3
BMDio
BMDL10
x2
goodness of fit
/?-value
AIC
Gamma3
614.27
225.96
0.48
367.22
Logistic
274.58
150.43
0.14
369.77
Log-logistica
697.90
499.42
0.94
365.90
Multistage (3)a
506.94
153.13
0.25
368.53
Probit
275.49
152.52
0.14
369.75
Log-probita
728.96
523.94
0.98
365.82
Weibull3
706.45
487.45
0.95
365.87
aThese models in EPA BMDS version 2.0 were fit to the rat dose-response data shown in
Table 5-5 by using internal dose metrics calculated with the rat PBTK model. Gamma and
Weibull models restrict power >1; Log-logistic and Log-probit models restrict to slope >1,
multistage model restrict betas >0; lowest degree polynomial with an adequate fit reported
(degree of polynomial in parentheses).
Bolded model is the best-fitting model in the most sensitive sex (males), which is used
in the RfC derivation.
Source: Nitschke et al., (1988a).
As with the RfD derivation, the human-equivalent internal BMDLio was obtained by
dividing this rat internal dose metric by a pharmacokinetic scaling factor based on a ratio of BWs
(scaling factor = 4.09) (Table 5-7). This scaling factor was used because the metric is a rate of
metabolism rather than the concentration of putative toxic metabolites, and the clearance of these
metabolites may be slower per volume tissue in the human compared with the rat. A probabilistic
PBTK model for dichloromethane in humans, adapted from the model of David et al. (2006) as
described in Appendix B, was then used with Monte Carlo sampling to calculate distributions of
chronic HECs (in units of mg/m3) associated with the internal BMDLio, based on the responses
in female Sprague-Dawley rats. Estimated mean, first, and fifth percentiles of this distribution
are shown in Table 5-7.
241

-------
7113
Table 5-7. Inhalation RfC for dichloromethane based on PBTK model-
derived probability distributions of human inhalation exposure
extrapolated from nonneoplastic liver lesion data for female rats exposed
via inhalation for 2 years, based on liver-specific CYP metabolism dose
metric (mg dichloromethane metabolized via CYP pathway per liter liver
tissue per day)

Rat
Human
HEC (mg/m3)d
Human RfC
(mg/m3)e
Model3
internal
BMDL10b
internal
BMDL10C
jst gth
percentile percentile ^®ean
Log-probit
523.94
128.10
16.63 20.89 47.36
0.2
aBased on the best-fitting model from Table 5-6.
bRat dichloromethane PBTK model-derived internal liver dose associated with lower bound
on 10% extra risk for developing hepatocyte vacuolation.
cHuman dichloromethane internal liver dose, derived by dividing the rat internal BMDLio by
a scaling factor of 4.09 [(BWhumaii/BWrat)0'25 ] to account for potential interspecies
pharmacokinetic differences in the clearance of metabolites.
dPBTK model-derived distributions of long-term, daily average airborne dichloromethane
concentrations predicted by the PBTK model to yield an internal dose in humans equal to the
dichloromethane internal BMDLio.
eHuman candidate RfC, based on female rat data, derived by dividing the 1st percentile of
HEC values by a total UF of 100: 3 (10°5) for possible toxicodynamic differences between
species, 3 (10°5) for variability in human toxicodynamic response, and 10 for database
deficiencies. The 1st percentile point of departure is a stable estimate of the lower end of the
distribution. Use of this value in the lower tail replaces use of a UF for human toxicokinetic
variability.
242

-------
7114
7115
7116
7117
7118
7119
7120
7121
7122
7123
7124
7125
7126
7127
7128
7129
7130
7131
7132
7133
7134
7135
7136
7137
7138
7139
7140
7141
7142
7143
7144
7145
7146
7147
7148
7149
7150
7151
7152
5.2.4. RfC Derivation—Including Application of Uncertainty Factors (UFs)
The 1st percentile point of departure is a stable estimate of the lower end of the
distribution. Use of this value associated with a sensitive human population addresses the
uncertainty associated with human toxicokinetic variability. The RfC was calculated by dividing
the first percentile of the HEC distribution in Table 5-7 by a composite UF of 100 (3 [10°5] to
account for uncertainty about interspecies toxicodynamic equivalence, 3 [10°5] to account for
uncertainty about toxicodynamic variability in humans, and 10 for database deficiencies). The
resulting RfC was 0.2 mg/m3 based on liver lesions in female Sprague-Dawley rats in Nitschke et
al. (1988a). In deriving this RfC, factors for the following areas of uncertainty were considered:
•	Uncertainty in extrapolating from laboratory animals to humans (UFa)- The use of
PBTK models to extrapolate internal doses from rats to humans reduces toxicokinetic
uncertainty in extrapolating from the rat liver lesion data but does not account for the
possibility that humans may be more sensitive than rats to dichloromethane due to
toxicodynamic differences. A UF of 3 (10°5) to account for this toxicodynamic
uncertainty was applied, as shown previously in Table 5-7.
•	Uncertainty about variation in human toxicokinetics (UFH). The probabilistic human
PBTK model used in this assessment incorporates the best available information about
variability in toxicokinetic disposition of dichloromethane in humans but does not
account for humans who may be sensitive due to toxicodynamic factors. Thus, a UF of
3 (10°5) was applied to account for possible toxicodynamic differences in sensitive
humans.
Uncertainty in extrapolating from LOAELs to NOAEL (UFi). A UF for extrapolation from a
LOAEL to a NOAEL was not applied because BMD modeling was used to determine the POD,
and this factor was addressed as one of the considerations in selecting the BMR. The BMR was
selected based on the assumption that it represents a minimum biologically significant change.
•	Uncertainty in extrapolating from subchronic to chronic durations (UFs). The derived
RfD is based on results from a chronic-duration drinking water toxicity study. No cross-
duration UF is necessary.
•	Uncertainty reflecting incompleteness of the overall database (UFD). A UF of 10 was
selected to address the deficiencies in the dichloromethane toxicity database. The
inhalation database for dichloromethane includes several well-conducted chronic
inhalation studies. In these chronic exposure studies, the liver was identified as the most
sensitive noncancer target organ in rats (Nitschke et al., 1988a; NTP, 1986; Burek et al.,
243

-------
7153
7154
7155
7156
7157
7158
7159
7160
7161
7162
7163
7164
7165
7166
7167
7168
7169
7170
7171
7172
7173
7174
7175
7176
7177
7178
7179
7180
7181
7182
7183
7184
7185
7186
7187
7188
7189
7190
1984). The critical effect of hepatocyte vacuolation was corroborated in the two principal
studies (Nitschke et al., 1988a; Burek et al., 1984), which identified 500 ppm as the
lowest inhalation LOAEL for noncancer liver lesions. Gross signs of neurologic
impairment were not seen in lifetime rodent inhalation bioassays for dichloromethane at
exposure levels up to 4,000 ppm (see section 4.2.2.2 for references), and no exposure-
related effects were observed in an observational battery, a test of hind-limb grip
strength, a battery of evoked potentials, or histologic examinations of nervous tissues in
F344 rats exposed to dichloromethane concentrations as high as 2,000 ppm (Mattson et
al., 1990). A two-generation reproductive study in F344 rats reported no effect on
fertility index, litter size, neonatal survival, growth rates, or histopathologic lesions at
exposures >100 ppm dichloromethane (Nitschke et al., 1988b). Fertility index (measured
by number of unexposed females impregnated by exposed males per total number of
unexposed females mated) was reduced following inhalation exposure of male mice to
150 and 200 ppm dichloromethane for 2 hours/day for 6 weeks, but the statistical
significance of this effect varied considerably depending on the statistical test used in this
analysis. (Raje et al., 1988). The available developmental studies include single-dose
studies that use relatively high exposure concentrations (1,250 ppm in Schwetz et al.
[1975]; 4,500 ppm in Hardin and Manson [1980]; and 4,500 ppm in Bornschein et
al.[1980]). In one of the single-dose studies, decreased offspring weight at birth and
changed behavioral habituation of the offspring to novel environments were seen
following exposure of adult Long-Evans rats to 4,500 ppm for 14 days prior to mating
and during gestation (or during gestation alone) (Bornschein et al., 1980; Hardin and
Manson, 1980). CO, a known developmental neurotoxicant, is produced through the
CYP2E1 metabolic pathway for dichloromethane. Schwetz et al. (1975) reported
increased concentrations (-10% higher compared with controls) in maternal blood COHb
levels in mice and rats exposed during GDs 6-15. A chronic exposure study in F344 rats
reported a dose-related increase in blood COHb in females exposed to 50, 200, and 500
ppm, beginning with the first measure taken after 6 months of exposure (Nitschke et al.,
1988a). The increase was seen at the lowest exposure group (50 ppm). Anders and
Sunram (1982) reported elevated CO levels in maternal and fetal blood in rats following
exposure to 500 ppm for 1 hour on GD 21; levels were similar in the maternal and fetal
samples. Placental transfer of dichloromethane was also seen, although levels were lower
in the fetus. The results from the single dose developmental toxicity study in rats
(Bornschein et al., 1980; Hardin and Manson, 1980), in addition to the known increase in
CO, the placental transfer of dichloromethane, and the relatively high activity of CYP2E1
in the brain compared to the liver of the developing human fetus (Hines, 2007; Brzezinski
et al., 1999; Johnsrud et al., 2003), raise uncertainty regarding possible
neurodevelopmental toxicity from gestational exposure to inhaled dichloromethane. In
244

-------
7191
7192
7193
7194
7195
7196
7197
7198
7199
7200
7201
7202
7203
7204
7205
7206
7207
7208
7209
7210
7211
7212
7213
7214
7215
7216
7217
7218
7219
7220
7221
7222
7223
7224
7225
7226
7227
7228
addition, Aranyi et al. (1986) demonstrated evidence of immunosuppression following a
single 100 ppm dichloromethane exposure for three hours in CD-I mice. This study used
a functional immune assay that is directly relevant to humans (i.e., increased risk of
Streptococcal pneumonia-related mortality and decreased clearance of Klebsiella
bacteria). No effects were seen with 50 ppm exposure for either 1 or 5 days. Systemic
immunosuppression was not seen in a 4-week, 5,000 ppm inhalation exposure study,
measuring the antibody response to sheep red blood cells in Sprague-Dawley rats
(Warbrick et al., 2003). These studies suggest a localized, portal-of-entry effect within
the lung rather than a systemic immunosuppression. Therefore, in consideration of the
entire database for dichloromethane, a database UF of 10 was selected. This UF accounts
for the lack of neurodevelopmental toxicity studies and developmental toxicity studies at
low doses.
5.2.5.	Previous RfC Assessment
No RfC was derived in the previous IRIS assessment.
5.2.6.	RfC Comparison Information
A candidate RfC, based on a different approach to accounting for human toxicokinetic
variability is similar to thederived RfC of 0.2 mg/m3. Use of the mean value on the HEC
distribution (47.36), with an additional UF of 3 (10°5) to account for human toxicokinetic
variability, would yield an RfC of 0.2 mg/m3.
For an additional comparison, an RfC was derived based on neurological endpoints from
human occupational exposures. Cherry et al. (1983) compared 56 exposed and 36 unexposed
workers at an acetate film manufacturing plant for dichloromethane inhalation exposure, blood
levels of dichloromethane, subjective self-reporting of general health, and two objective,
quantitative measurements of neurological function (digit symbol substitution and simple
reaction time). The exposed and unexposed individuals were matched to within 3 years of age.
The measured dichloromethane concentrations from personal breathing zone sampling of the
exposed workers ranged from 28 to 173 ppm. No information on exposure duration was given,
and Cherry et al. (1983) did not indicate if the exposure measurements were indicative of
historical exposure levels. There were no significant differences between exposed and
unexposed workers in subjective or objective measurements collected at the beginning of the
work shift on a Monday (after 2 nonworking days). Exposed workers showed a slightly slower
(but not significant) score than the control workers on a reaction time test, but the scores did not
deteriorate during the shift. These findings suggest that repeated inhalation exposures in the
range of 28-173 ppm do not result in significant effects, but the actual duration of exposure of
the workers is uncertain. In the absence of data for the mean exposure levels, the exposure range
midpoint of 101 ppm serves as a NOAEL for chronic neurological effects from dichloromethane
245

-------
7229
7230
7231
7232
7233
7234
7235
7236
7237
7238
7239
7240
7241
7242
7243
7244
7245
7246
7247
7248
7249
7250
7251
7252
7253
7254
7255
7256
7257
7258
7259
7260
7261
7262
7263
7264
7265
7266
exposure. Thus, a candidate RfC of 3.5 mg/m3 was derived by dividing the NOAEL of 351
mg/m3 (101 ppm) by a composite UF of 100. A UF of 10 was applied to account for potentially
susceptible individuals in the absence of quantitative information on the variability of
neurological response to dichloromethane in the human population. A UF of 10 was applied for
database deficiencies. The duration of exposures of acetate film workers (Cherry et al., 1983)
was not reported, and a limited number of endpoints was evaluated. Further, definitive
neurological batteries were not administered in chronic-duration animal bioassays.
Another candidate RfC was developed by using the neurological data from the study by
potential long-term CNS effects in a study of retired aircraft maintenance workers (Lash et al.,
1991). Retired aircraft maintenance workers, ages 55-75 years, employed in at least one of
14 targeted jobs (e.g., paint strippers) with dichloromethane exposure for 6 or more years
between 1970 and 1984 (n = 25) were compared to a like group of workers without
dichloromethane exposure (n = 21). From 1974 to 1986, when 155 measurements for
dichloromethane exposure were made, mean breathing zone TWAs ranged from 82 to 236 ppm
and averaged 225 ppm for painters and 100 ppm for mechanics; information on exposure levels
prior to this time was not provided. The evaluation included several standard neurological tests,
including physiological measurement of odor and color vision senses, auditory response
potential, hand grip strength, measures of reaction time (simple, choice, and complex), short-
term visual memory and visual retention, attention, and spatial ability. The exposed group had a
higher score on verbal memory tasks (effect size approximately 0.45, p = 0.11) and lower score
on attention tasks (effect size approximately -0.55,p = 0.08) and complex reaction time (effect
size approximately -0.40,p = 0.18) compared with the control group. None of these differences
were statistically significant. Given the sample size, however, the power to detect a statistically
significant difference between the groups was very low (i.e., approximately 0.30 for an effect
size of 0.40 using a two-tailed alpha of 0.05) (Cohen, 1987), and these results cannot be taken as
evidence of no effect. An estimated exposure level from the study can be generated from the
midpoint value from the exposure range (82-236 ppm; mean =159 ppm), converted to 552
mg/m3. If these results are viewed as a LOAEL and this estimated mean exposure level of 552
mg/m3 was used, a composite UF of 1,000 would be applied for interspecies toxicodynamics
(10), extrapolation from a LOAEL to a NOAEL (10), and database uncertainties (10), resulting
in an RfC of 0.55 mg/m3.
The value of the candidate RfC based on the data from Cherry et al. (1983), 3.5 mg/m3, is
approximately 15-fold higher, and the value of the candidate RfC based on the data from Lash et
al. (1991), 0.55 mg/m3 is approximately three times higher than thederived RfC of 0.2 mg/m3,
based on liver lesions in rats. The animal-derived RfC is preferable to the human-derived RfC
because of the uncertainties about the exposure durations for the workers in the Cherry et al.
(1983) study and uncertainties regarding the exposures and effect sizes in Lash et al. (1991), and
because the RfC based on the rat data is more health protective.
246

-------
7267	Additional comparisons among the RfC and candidate values developed from other
7268	endpoints or data sets, using NOAEL/LOAEL methods, are shown in Table 5-8 and Figure 5-8.
247

-------
Table 5-8. Potential points of departure with applied UFs and resulting candidate RfCs
Point of

departure
POD Type and
Total




RfC (mg/m3)

Endpoint
(mg/m3)a
Description1"
UF
ufa
UFh
ufl ufs
ufd
Reference
Hepatocyte vacuolation,
female rat
523
BMD, 10% increase
in incidence of liver
lesion
100
3
3
1 1
10
0.2
Nitschke et al.
(1988a)
Renal tubular degeneration;
NOAEL, male rat
620
NOAEL
1000
3
10
1 1
10
2.07
Mennear et al.
(1988); NTP (1986)
Reproductive - fertility
index; NOAEL, male
mouse
20.7
No effect at POD,
16% decrease in
fertility index seen at
LOAEL dose
300
3
10
1 1
10
0.071
Raje et al. (1988)
Increased infection









susceptibility (mortality
15.5
NOAEL
3,000
3
10
1 10
10
0.005
Aranyi et al. (1986)
risk), female mouse









Increased IgM production,
male and female rat
17,366
NOAEL
3,000
3
10
1 10
10
1.03
Warbrick et al.
(2003)
Chronic CNS effects,
human male
351
NOAEL
100
1
10
1 1
3
3.51
Cherry et al. (1983)
CNS changes, human male
552
LOAEL
1,000
1
10
10 1
3
0.55
Lash et al. (1991)
aPOD = point of departure. For Nitschke et al. (1988a), this is based on BMD modeling of a 10% increase in liver lesions using internal liver dose metric (mg
dichloromethane metabolism via CYP pathway per liter liver tissue per day) derived from a rat PBTK model. After an allometric scaling factor of 4.09 was
applied, the human internal BMDLio was 128 mg/m3. A probabilistic human PBTK model adapted from David et al. (2006) was used to generate a distribution
of human equivalent concentrations from the human internal BMDLio and the first percentile of this distribution was used as the point of departure. For other
rodent studies, the NOAEL or LOAEL concentration, in mg/m3, was adjusted to a continuous exposure taking into account hours per day and days per week of
exposure. This adjusted exposure was then converted to an HEC by multiplying the value by a dosimetric adjustment factor (DAF). Blood:air partition
coefficients were 8.24 for humans, 19.8 for rats, and 23 for mice. Since the blood:air partition coefficients for both the mice and rats were greater than for
humans, a DAF of 1 is recommended and was used. NOAELs or LOAELs were used as points of departure in human studies since the concentrations were
already human exposures.
''Extra risk defined for incidence data as (Incidencei - Incidence0)/(1-Incidence0), where 1 = dose at observed increased and 0 = background incidence
°UFa = uncertainty in extrapolating from laboratory animals to humans, UFH = uncertainty about variation from average humans to sensitive humans, UFL=
uncertainty about extrapolating from LOAEL to NOAEL, and UFD = uncertainty reflecting incompleteness of the overall database. A UF extrapolating from
subchronic to chronic durations (UFS) was not used for any of these studies.
Bolded value is the basis of the RfC of 0.2 mg/m3.
248

-------
Hepatocyte
vacuolation,
lsl percentile
HEC from
female rat -
Nitscbke et
al. (1983a)
Renal tubular
degeneration,
adj. HEC
from rat -
Mennear et
al. (1988);
NTP (1986)
Reproductive
Performance
- Fertility
Index,, adj.
HEC mouse -
Raje et al.
(1988)
Increased
Infection
Susceptibility,
adj. HEC
mouse -
Aranyi et al.
(1986)
Increased
IgM adj.
HEC rat -
Warbrick et
al. (2003)
Chronic
CNS effects,
NOAEL
from human
males -
Cherry et al.
(1983)
CNS
changes,
LOAEL
from human
males-Lash
et al. (1991)
0 Point of Departure
U UFa- Interspecies;
animal to human
^ UFh - Intraspecies;
human variability
1! UFl- LOAEL to
NOAEL
HD UFS - Subchronic to
Chronic
Q UFd - Database
^ Reference Dose
7269
7270	Figure 5-8. Comparison of candidate RfCs derived from selected point of departures for endpoints presented in
7271	Table 5-8.
249

-------
7272
7273
7274
7275
7276
7277
7278
7279
7280
7281
7282
7283
7284
7285
7286
7287
7288
7289
7290
7291
7292
7293
7294
7295
7296
7297
7298
7299
7300
7301
7302
7303
7304
7305
7306
7307
5.3. UNCERTAINTIES IN THE ORAL REFERENCE DOSE AND INHALATION
REFERENCE CONCENTRATION
Risk assessments need to include a discussion of uncertainties associated with the derived
toxicity values. For dichloromethane, uncertainties related to inter- and intraspecies differences in
toxicodynamics and database deficiencies are treated quantitatively via the UF approach (U.S. EPA,
1994b). Uncertainties in the toxicokinetic differences of dichloromethane between species and within
humans are reduced by application of the PBTK models for rats and humans. These and other areas of
uncertainty of the derived RfD and RfC are discussed below.
Adequacy of database for derivation of RfD and RfC
As summarized in sections 4.6.1.1 and 4.6.2.1, data from the available human studies on the
health effects from occupational inhalation exposures provide some, but not conclusive, evidence of
long-term health consequences of chronic dichloromethane exposure, specifically with respect to
neurologic and hepatic damage. These data are not adequate for derivation of an RfD or RfC. However,
a broad range of animal toxicology data is available for the hazard assessment of dichloromethane, as
described in chapter 4. The database of oral (Table 4-35) and inhalation (Tables 4-36 and 4-37) toxicity
studies includes numerous chronic, subchronic, acute, reproductive, and developmental studies. Liver
toxicity in multiple rodent species is consistently identified as the most sensitive noncancer effect from
oral and inhalation exposure to dichloromethane. In addition to the oral and inhalation toxicity data,
there are numerous studies describing the toxicokinetics of dichloromethane. Consideration of the
available dose-response data to determine an estimate of oral exposure that is likely to be without an
appreciable risk of adverse noncancer health effects over a lifetime has led to the selection of noncancer
liver lesions in the 2-year drinking water study in F344 rats (Serota et al., 1986a) as the critical effect
and principal study for deriving the RfD for dichloromethane. The critical effect selected for the
derivation of the chronic RfC is also hepatic lesions; two different studies in Sprague-Dawley rats
(Nitschke et al.,1988a; Burek et al., 1984), spanning overlapping exposures, reported data on hepatic
vacuolation, and the lower exposure study was chosen as the principal study (Nitschke et al.,1988a).
A critical data uncertainty was identified for neurodevelopmental effects. Animal bioassays
have not identified gross or microscopic effects on neural tissues from long-term exposures or single
(Schwetz et al., 1975) or multigenerational (Nitschke et al., 1988b) developmental toxicity studies.
However, behavioral changes were observed in pups born to rats exposed to high levels (4,500 ppm) of
dichloromethane (Bornschein et al., 1980; Hardin and Manson, 1980); lower exposures were not
examined in this study. Uncertainty exists as to the development of neurological effects from lower
gestational exposures in animals or humans. In addition, a critical data uncertainty has been identified
that relates to potential immunotoxicity, specifically immunosuppression seen as a localized portal-of-
entry effect within the lung with an acute inhalation exposure. The lack of data on immune effects from
250

-------
7308
7309
7310
7311
7312
7313
7314
7315
7316
7317
7318
7319
7320
7321
7322
7323
7324
7325
7326
7327
7328
7329
7330
7331
7332
7333
7334
7335
7336
7337
7338
7339
7340
7341
7342
7343
longer-term exposure represents a significant data gap and is of particular importance because of the
potential importance of immunosuppression with respect to response to infections and tumor
surveillance. The weight of evidence for nonneoplastic effects in humans and animals suggests that the
development of liver lesions is the most sensitive effect, with a UF applied because of the lack of
neurodevelopmental studies and, for the RfC, the uncertainty regarding immunotoxicity.
Dose-response modeling
The selection of the BMD model(s) for the quantitation of the RfD and RfC does not lead to
significant uncertainty in estimating the point of departure. It should be noted, however, that a level of
uncertainty is inherent given the lack of data in the region of the BMR.
Interspecies extrapolation of dosimetry and risk
The extrapolation of internal dichloromethane dosimetry from nonneoplastic rat responses to
human risk was accomplished using PBTK models for dichloromethane in rats and humans.
Uncertainties in rat and human dosimetry used for RfD and RfC derivation can arise from uncertainties
in the PBTK models to accurately simulate the toxicokinetics of dichloromethane for animals under
bioassay conditions and humans experiencing relatively low, chronic environmental exposures.
There is uncertainty associated with the pharmacokinetic data used for model parameter
estimation and structure validation. The data are primarily measurements of parent dichloromethane
kinetics (e.g., blood or closed-chamber air concentrations over time), rather than measurements of
metabolite levels which can be unambiguously attributed to one of the two principal metabolic pathways
(GST and CYP). For the mouse model in particular, only parent dichloromethane data were used,
though exhaled amounts of CO2 and CO are available. Marino et al. (2006) did include data from mice
pre-treated with trans-1,2-dichloroethylene (tDCE), a specific CYP 2E1 inhibitor, but the authors
assumed without verification that 100% of the CYP 2E1 activity was eliminated by the inhibitor when
using those data. In contrast, Mathews et al. (1997) found that pretreatment of F344 rats by tDCE (100
mg/kg ip) only yielded 65% inhibition of CYP 2E1. If a significant fraction of the CYP 2E1 activity
was not eliminated in the dichloromethane experiments, then that activity is erroneously assigned to the
GST pathway in the parameter estimation Marino et al. (2006).
In addition to the possibility of incomplete inhibition of CYP 2E1 effecting the data
interpretation, the Michaelis-Menten rate equation used in all of the published PBTK models for
dichloromethane, including that of Marino et al. (2006), has in fact not been shown to accurately
describe the CYP 2El-mediated metabolism of dichloromethane in the relevant concentration range.
While Michaelis-Menten kinetics usually describe CYP-mediated oxidation data quite well, the
approach of Marion et al. (2006) implicitly assumes that any metabolism not described by the Michaelis-
Menten equation is GST-mediated. If pathway-specific metabolite data were used to define or bound
251

-------
7344
7345
7346
7347
7348
7349
7350
7351
7352
7353
7354
7355
7356
7357
7358
7359
7360
7361
7362
7363
7364
7365
7366
7367
7368
7369
7370
7371
7372
7373
7374
7375
7376
7377
7378
7379
the ratio of GST to CYP metabolism, the resulting estimates would be less sensitive to what otherwise
might be small errors in the CYP rate equation. But in the modeling of the mouse data by Marino et al.
(2006), the fraction of total metabolism assigned to the CYP pathway depends quite strongly on the
assumed form of the CYP rate equation, along with the assumption of 100% inhibition by tDCE. In
EPA's modeling of the rat in vivo PK data, using the same model structure and equations, a set of
parameter values could not be found which described both the parent dichloromethane kinetics and the
total amount of CO exhaled at both high and low exposure levels; in particular see panel C of Figure C-3
and note discrepancy between model and 50 mg/kg data. That the model does not describe well the
dose-dependent shift in metabolism shown by those CO data suggests that the dose-dependence of the
CYP Michaelis-Menten rate-equation is not adequate. As will be shown, an alternative equation for
CYP kinetics may fit the existing dichloromethane data better than Michaelis-Menten kinetics, with the
result that a higher portion of total dichloromethane metabolism would be interpreted as being CYP-
mediated. Thus there is uncertainty in the choice of equation for the CYP pathway, which leads to
uncertainty in the estimated GST:CYP metabolic ratio, upon which current risk predictions are based.
The potential error in assuming Michaelis-Menten kinetics for CYP-mediated oxidation of
dichloromethane is reinforced by examining the in vitro oxidative (i.e., CYP-specific) kinetics of
dichloromethane reported by Reitz et al. (1989). When extrapolated from in vitro to in vivo, the
apparent values of the oxidative saturation constant, Km, identified by Reitz et al. (1989) for mice, rats,
and humans are over 2 orders of magnitude greater than those obtained from in vivo PBTK modeling.
Part of the explanation for this apparent discrepancy lies in the disparate concentration ranges
investigated: Reitz et al. (1989) used much higher dichloromethane concentrations in vitro than those
observed in or predicted for the various in vivo pharmacokinetic studies. In particular, the oxidation of
dichloromethane could involve two oxidative processes, one with a high affinity (low Km)
corresponding to the nonlinearity observed in vivo and one with a low affinity (high Km) corresponding
to the nonlinearity observed in vitro. Further, the low-affinity process would have nearly linear kinetics
in the exposure range used for the in vivo dosimetry studies and hence be difficult to distinguish from
GST-mediated metabolism unless pathway-specific metabolite data are used. One can hypothesize that
this second oxidative process is not inhibited by tDCE and hence corresponds to the 35% of oxidative
metabolism which was observed to remain in rats after tDCE treatment by Mathews et al. (1997).
The data of Reitz et al (1989) could simply indicate a second CYP with low-affinity
dichloromethane activity. However that possibility is contradicted by the results of Kim and Kim (1996)
who observed that another CYP 2El-specific inhibitor, disulfiram, completely abolished
dichloromethane-induced increases on COHb in rats. Another possible explanation which would
support the findings observed in Kim and Kim (1996) as well as Reitz et al (1989) and the various in
vivo data is that a number of CYPs exhibit "atypical" kinetics, not described by the classic Michaelis-
Menten equation, consistent with the enzymes having dual binding sites as proposed by Korzekwa et al
252

-------
7380
7381
7382
7383
7384
7385
7386
7387
7388
7389
7390
7391
7392
7393
7394
7395
7396
7397
7398
7399
7400
7401
7402
(1988). (Korzekwa et al. (1988) demonstrated atypical kinetics for several CYP-isozyme/substrate
pairs, but not specifically for CYP 2E1.) Figure 5-9 shows kinetic model fits to the in vitro mouse
dichloromethane oxidation kinetic data of Reitz et al. (1989), after expressing those data on a per gram
of liver basis. Both the standard Michaelis-Menten kinetic equation (solid line) and the dual-binding
equation (dashed line) given by Korzekwa et al. (1988) are shown. In particular, the high-affinity (low)
Km for the dual-binding equation was set equal to that obtained by Marino et al. (2006) from their
PBTK modeling. This figure shows that the dual binding model is not only consistent with the apparent
high-affinity saturation obtained from in vivo PBTK modeling (Km of Marino et al. (2006)), but also
with the apparent low-affinity (high Km) data of Reitz et al. (1989), and describes those in vitro data
better than the standard Michaelis-Menten equation. (Reitz et al. (1989) used classic Lineweaver-Burk
plots to display their kinetic data; i.e., 1/reaction rate vs. 1/concentration. The systematic discrepancy
between their data and Michaelis-Menten kinetics evident in Figure 5-9 is much less obvious with that
scaling, which likely explains why they made no note of it.)

2.5
a>
>
O)
O)
E
a>
£
X
~ Reitz et al. (1989) data
	Michaelis-Menten kinetics
Q.
5 0.5
	Dual-binding CYP kinetics
0
100
200
300
400
500
[DCM] (irtg/L)
Figure 5-9. Comparison of dichloromethane oxidation rate data with alternate
kinetic models. Dichloromethane (DCM) oxidation data obtained with mouse liver
microsomes by Reitz et al. (1989) (points), expressed on a per gram of liver basis,
are shown with a fitted Michaelis-Menten equation (solid line) or a fitted dual-
binding-site equation as described by Korzekwa et al. (1988) (dashed line), where
the high affinity saturation constant of the dual-binding-site equation set equal to
the mean Km determined for mice via PBTK modeling by Marino et al. (2006). The
Km for the Michaelis-Menten equation (108 mg/L) is inconsistent with the in vivo
253

-------
7403
7404
7405
7406
7407
7408
7409
7410
7411
7412
7413
7414
7415
7416
7417
7418
7419
7420
7421
7422
7423
7424
7425
7426
7427
7428
7429
7430
7431
7432
7433
7434
7435
7436
DCM dosimetry data, while the in vitro data shown here are inconsistent with the
Km estimated in vivo (0.42 mg/L) if that equation is used.
In summary regarding model equations, the current PBTK model used the standard Michaelis-
Menten equation to describe CYP 2El-catalyzed oxidation of small volatile organic compounds.
Analysis of the dichloromethane (pharmaco)kinetic data and evaluation of the inconsistencies describe
above suggest that an alternate equation, which would impact risk predictions, may better represent CYP
2E1-induced oxidation of dichloromethane. However, this hypothesis requires further laboratory
testing, for example, by measuring dichloromethane oxidation in a bacterial expression system where
only CYP 2E1 is expressed over a concentration range sufficient to firmly distinguish between the two
kinetic forms indicated in the figure above. Until such experiments are conducted, the existing PBTK
model remains the best available science for dose- and hence risk-extrapolation from rodents to humans.
Still, this model structure uncertainty implies uncertainty in the quantitative results obtained with the
model. Analysis of the GST-mediated metabolism of dichloromethane measured by Reitz et al. (1988)
shows that those results are within a factor of three of the GST kinetic parameters used in the current
PBTK model, indicating that the any error in the GST:CYP balance is no greater than that, a reasonable
level of uncertainty.
One other component of quantitative uncertainty arises in examining the results of the Bayesian
modeling for the human PBTK model of David et al. (2006). The authors reported Bayesian posterior
statistics for the population average of each fitted parameter when calibration was performed either with
specific published data sets or the entire combined data set. While one would generally expect that the
values obtained from the combined data set should be a weighted average of the values from individual
data sets, the population mean for the liver GST activity (coefficient), KFc, was 0.852 while the values
from the individual data sets ranged from 1.92-34.0 kg°'3/h.
A clarification provided by D. Marino (personal communication)8 is that the parameter bounds
stated in the text of David et al. (2006) were only applied for the analysis of the DiVincenzo and Kaplan
(1981) and the combined data set. But according to the text and distribution prior statistics specified, the
upper bound for KFc would have been 12 kg°'3/h (mean + 2.5 standard deviations (SDs), with mean = 2
and SD = mean*CV = 2*2 = 4). The data of Andersen et al. (1991) were not used in the combined
analysis because only group average values were available from that source, rather than individual data.
Since the remaining study-specific mean KFc values were 7.95, 5.87, 34.0, and 1.92, with CVs of less
than 2, it seems unlikely that application of this upper bound would result in a value of KFc of only
0.852 kg0 3/h. Given that there had been convergence problems with the combined data set when
parameter values were unbounded, it is possible that convergence had not actually been reached after
8 Email from Dale Marino to Glinda Cooper dated April 25, 2007.
254

-------
7437
7438
7439
7440
7441
7442
7443
7444
7445
7446
7447
7448
7449
7450
7451
7452
7453
7454
7455
7456
7457
7458
7459
7460
7461
7462
7463
7464
7465
7466
7467
7468
7469
7470
7471
parameter bounds were introduced, and a higher value for KFc would have been obtained had the chain
been continued longer.
Since the numerical average of the mean KFc values for the four data sets included in the
combined data set was 12.4 and the upper bound was 12, the impact of using an intermediate value of
KFc, specifically the DiVincenzo and Kaplan value of 5.87 kg°'3/h was explored. Changing only the KFc
is not realistic since the dichloromethane data effectively define total metabolism (sum of CYP and GST
pathways) and there is naturally a negative correlation between the predicted CYP metabolic rate and
the GST metabolic rate, required to describe this total. Therefore, it would be inconsistent with the
dichloromethane data to increase KFC without adjusting the CYP metabolic rate downward, and likewise
all other parameters. The distributions for all of the fitted parameters were rescaled by the ratio of the
mean for DiVincenzo and Kaplan (1981) to the mean for the combined data set (e.g., the distribution for
KFC was multiplied by 5.87/0.852, the ratio of the two posterior means). The resulting predicted upper-
bound (95th and 99th percentile) GST metabolism rates for a fixed level of exposure (1 |ig/m3 inhaled
concentration or 1 mg/kg/day oral exposure) in the GST-T1 +/+ population increased by more than a
factor of 10. (For inhalation exposure the mean value also increased by over 10-fold, but for oral
exposure the mean increased by only 2-fold.) Since the majority of metabolism occurs via the CYP
pathway at these low levels, there is not a proportionate (i.e., over 10-fold) decrease in that rate, but
HEC and HED calculations increased by 10-30% for the mixed GST-T1 population, depending on the
route of exposure and distribution statistic compared. Thus the impact of this model uncertainty appears
to be relatively small for the noncancer assessment, but quite large for the cancer assessment.
The dose metric used in the models is the rate of metabolism to a putative toxic metabolite,
rather than the concentration average or area under the concentration curve of the metabolite, so the
model specifically fails to account for rodent-human differences in clearance or removal of the toxic
metabolite. A scaling factor based on BW ratios, was used to account for this difference.
The rat model was modified and utilized in a deterministic manner. Data were not available to
perform a hierarchical Bayesian calibration in the rat. Thus, uncertainties in the rat model predictions
had to be assessed qualitatively. To address these uncertainties, a sensitivity analysis was conducted to
determine which model parameters most influence the predictions for a given dose metric and exposure
scenario.
Sensitivity is a measure of the degree to which a given model output variable (i.e., dose metric)
is influenced by perturbation in the value of model parameters. The approach implemented was a
univariate analysis in which the value of an individual model parameter was perturbed by an amount
(A), in the forward and reverse direction (i.e., an increase and decrease from the nominal value), and the
change in the output variable was determined. Sensitivity coefficients were calculated as follows:
255

-------
7472
7473
7474
7475
7476
7477
7478
7479
7480
7481
7482
7483
7484
7485
7486
7487
7488
7489
7490
7491
7492
7493
7494
7495
7496
7497
7498
7499
/O)
fix + Ax) - f{x) x
Ax
/(*)
(Eq. 5-1)
where x is the model parameter f(x) is the output variable, Ax is the perturbation of the parameter from
the nominal value, andf'(x) is the sensitivity coefficient. In equation 5-1, the sensitivity coefficients are
scaled to the nominal value of x and f(x) to eliminate the potential effect of units of expression.
Therefore, the sensitivity coefficient is a measure of the proportional (unitless) change in the output
variable produced by proportional change in the parameter value. Parameters that have higher
sensitivity coefficients have greater influence on the output variable. They are considered more
sensitive than parameters with lower values. The results of the sensitivity analysis are useful for
assessing uncertainty in model predictions, based on the level of confidence or uncertainty in the model
parameter(s) to which the dose metric is most sensitive.
Sensitivity coefficients for the noncancer dose metric (mg dichloromethane metabolized via
CYP-mediated pathway/L liver/day), were determined for each of the model parameters; a similar
analysis was also done for a metric based on the GST-mediated pathway. Sensitivity analyses for both
oral and inhalation exposures were performed. The exposure conditions were set to be near or just
below the lowest bioassay exposure resulting in significant increases in the critical effect.
For the CYP-mediated metabolism from oral exposure, the VLC and VSC (liver volume and
slowly perfused tissue volume, respectively) parameters exert the largest influence (Figure 5-10). The
high influence of these two parameters was due to the fact that the dose metric is a tissue-specific rate of
metabolism, the majority of CYP metabolism is attributed to the liver, and that changes in liver volume
have a greater impact on the total CYP metabolism that the individual Vmax value. For inhalation
exposures VMAXC, in addition to VLC and VSC have the highest sensitivity coefficients (Figure 5-11).
The physiological parameters (VLC and VSC) are known with a high degree of confidence (Brown et
al., 1997). Vmaxc for the rat was estimated by fitting to the PK data as described in Chapter 3 and
Appendix C, subject to model structure/equation uncertainties as detailed above, and hence is known
with less certainty than the physiological parameters. That total exhaled CO, which is a specific to the
CYP pathway, is within 50% of measured levels (Fig. C-8, panel C), however, provides a similar level
of confidence in the balance between CYP and GST pathways predicted by the rat PBTK model.
256

-------
7500
7501
7502
7503
7504
7505
7506
7507
7508
7509
7510
7511
7512
7513
7514
7515
7516
Oral exposure
£ VMAXC
a)
~	CYP
~	GST
Normalized sensitivity coefficient
Figure 5-10. Sensitivity coefficients for long-term mass CYP- and GST-mediated
metabolites per liver volume from a daily drinking water concentration of 10 mg/L in rats.
KFC = GST-mediated metabolism rate; A2 = proportion of liver GST metabolism attributed to
the lung; KA = oral absorption rate from gut; VMAXC = CYP-mediated maximum rate of
metabolism; PB = blood:air partition coefficient; VSC = slowly perfused tissue volume; VLC =
liver volume; VPR = Ventilation perfusion ratio ; QCC = cardiac output constant.
VMAXC
a>
E
2
(0
a.
Inhalation exposure
B CYP
~ GST
Normalized sensitivity coefficient
Figure 5-11. Sensitivity coefficients for long-term mass CYP- and GST-mediated
metabolites per liver volume from a long-term average daily inhalation concentration of
500 ppm in rats. (KA is not included since it has no impact on inhalation dosimetry.) KFC
= GST-mediated metabolism rate; A2 = proportion of liver GST metabolism attributed to the
lung; VMAXC = CYP-mediated maximum rate of metabolism; PB = blood:air partition
coefficient; VSC = slowly perfused tissue volume; VLC = liver volume; VPR = Ventilation
perfusion ratio ; QCC = cardiac output constant.
257

-------
7517
7518
7519
7520
7521
7522
7523
7524
7525
7526
7527
7528
7529
7530
7531
7532
7533
7534
7535
7536
7537
7538
7539
7540
7541
7542
7543
7544
7545
7546
7547
7548
7549
7550
7551
7552
In summary, the uncertainties associated with use of the rat PBTK model should not markedly
affect the values of the RfD and RfC based on the metrics considered. An additional uncertainty results
from the lack of knowledge concerning the most relevant dose metric (e.g., a specific metabolite) for the
non-cancer endpoints considered. This basic research question represents a data gap. This uncertainty
was addressed by considering different dose metrics (CYP metabolism alone, GST metabolism alone,
sum of GST and CYP, and the AUC of the parent compound). The GST metabolism and the AUC dose
metrics did not present reasonable choices based on model fit and consistency of response across studies
at comparable dose levels. Given these results, the combination of hepatic metabolism through the GST
and the CYP pathways would not be expected to result in an improvement to a metric based only on
CYP metabolism. The CYP-metabolism dose metric seems to be most consistent with the data., and so
is the metric chosen for the RfD and RfC derivations.
Sensitive human populations
The potential for sensitivity to dichloromethane in a portion of the human population due to
pharmacokinetic differences was addressed quantitatively by using a human probabilistic PBTK model,
as modified by the EPA, to generate distributions of human exposures likely to result in a specified
internal BMDLio. The model and resulting distributions take into account the known non-chemical-
specific variability in human physiology as well as total variability and uncertainty in dichloromethane-
specific metabolic capability. The first percentile values of the distributions of human equivalent doses
(Table 5-3) and HECs (Tables 5-7) served as points of departure for candidate RfDs and RfCs,
respectively, to protect toxicokinetically sensitive individuals. Selection of the first percentile allows
generation of a stable estimate for the lower end of the distribution. The mean value of the human
equivalent oral dose in Table 5-3 was about two-fold higher than the corresponding first percentile
values, and the mean value of human equivalent inhalation concentration in Table 5-7 was
approximately three-fold higher than the first percentile value. The internal dose metric in the analyses
described in these tables was the mg dichloromethane metabolized via the CYP pathway per liter liver
per day, and thus the comparisons of the first percentile and mean values give estimates of the amount of
variability in the population to metabolize dichloromethane by the CYP metabolic pathways on a liver-
specific basis. The mean: 1st percentile ratios for these distributions is attributed to the dependence of
the dose metric on hepatic blood flow rate (metabolism being flow limited). This blood flow is expected
to be highly and tightly correlated with liver volume, resulting in very similar delivery of
dichloromethane per volume liver across the population. While the mean: 1st percentile ratios for the
oral distribution is less than the default intra-human toxicokinetic UF of 3, it is quite similar to that
obtained by Sweeney et al. (2003) for acrylonitrile, where an extensive sensitivity analysis indicated a
99th percentile:mean ratio of less than 2.2 among several internal dose metrics. The population-
structured distributions for physiological parameters and broadened distributions for metabolic
258

-------
7553
7554
7555
7556
7557
7558
7559
7560
7561
7562
7563
7564
7565
7566
7567
7568
7569
7570
7571
7572
7573
7574
7575
7576
7577
7578
7579
7580
parameters used here provide a good degree of confidence that the population variability has not been
under-estimated.
The internal dose metric used in the RfD and RfC derivations was based on the rate of CYP
metabolism. GST-T1 polymorphisms could affect this rate, as the GST-T1 null genotype would be
expected to result in an increase in the metabolism through the CYP pathway, resulting in a greater
sensitivity to a CYP-related effect. The effect of GST variability on the RfD and RfC values was
examined by comparing results obtained specifically for the GST-T1 null genotype to those obtained for
the population of mixed genotypes. The values for human equivalent doses and HECs were very similar
for these two groups (e.g., mean HEC 47.36 and 47.49 for the mixed and the GST-T1"" null genotypes,
respectively; 1st percentile HEC 16.63 and 16.69 for the mixed and the GST-T1"" null genotypes,
respectively), and use of this population would not result in a change in the recommended RfD or RfC.
As a further level of sensitivity analysis, we compared model predictions of the human
equivalent dose, as listed in Table 5-3, for the general population (estimates covered 0.5- to 80-year-old
male and female individuals) to three subpopulations: 1-year-old children (males and females), 70-year-
old men, and 70-year-old women. For the general population and each subpopulation a Monte Carlo
simulation representing 10,000 individuals was conducted, and histograms of the resulting distribution
of human equivalent administered doses are shown in Figure 5-12, with corresponding statistics in Table
5-9. All groups used in these comparisons were limited to the GST-T1"".
The results shown above for differences in human equivalent dose values in different populations
are qualitatively what would be expected: a relatively broad distribution for the general population with
specific populations representing narrower components of that distribution. There are some differences
between men and women at 70 years of age, but neither of these would be greatly misrepresented by the
general population estimate. While 1-year-old children represent more of a distinct tail in the general
population, in this case the distribution of human equivalent concentrations in the general estimate is
lower than that seen in what would otherwise be considered a more sensitive population. This
difference most likely results from the higher specific respiration rate in children versus adults, which
allows them to eliminate more of orally ingested dichloromethane by exhalation, leading to lower
internal metabolized doses
259

-------
7581
7582
7583
7584
7585
7586
7587
7588
7589
7590
7591
7592
10
Human equivalent applied
dose distributions
9
8
	General
	70 yo Male
70 yo Female
	1 yo Child
7
6
5
4
3
2
£ 1
0
0.
0.3
0.5
Human equivalent applied dose (mg/kg-day)
0.7
0.9
Figure 5-12. Frequency density of human equivalent applied doses in specific populations
in comparison to a general population (0.5- to 80-year-old males and females) estimate for
an internal dose of 15.1 mg dichloromethane metabolized by CYP per liter liver per day; all
groups were restricted to the GST-T1 population).
Table 5-9. Statistical characteristics of human
equivalent applied doses in specific populations
of the GST-T1group

Human equivalent applied dose
(mg/kg-day)a
Population
Mean
5th percentile
1st percentile
All agesb
3.95 x 101
2.52 x 10-1
2.14 xlO1
1-year-old children
6.34 x 101
4.87 x 101
4.54 x 10
70-year-old men
3.18 xlO1
2.48 x 101
2.29 x 101
70-year-old women
2.64 x 101
2.03 x 10-1
1.85 x 101
aExposure levels predicted to result in 15.1 mg dichloromethane
metabolized via CYP pathway per liter liver per day (based on mean
BMDLio across acceptable models from Table 5-3).
b0.5- to 80-year-old males and females.
A similar comparison was made for inhalation HEC values, as shown in Figure 5-13 and Table
5-10. For HEC values, the distributions for 70-year-old men and women are both virtually
indistinguishable from the general population, and while 1-year-old children are clearly distinct, they are
260

-------
7593
7594
7595
7596
7597
7598
7599
7600
7601
7602
7603
7604
7605
7606
7607
7608
7609
7610
7611
7612
7613
7614
7615
7616
less different than in the human equivalent administered dose comparison and in this case are more
sensitive than the population in general. As described in detail in Appendix B, the allometric alveolar
ventilation constant, QAlvC, is about 28 L/hour-kg0'75 in a 1-year-old child but averages around 14
L/hour-kg0'75 in an adult. Combining this with the difference between a BW of 10 kg in that child and
70 kg in an "average" adult, the respiration rate per kg BW is about threefold higher in the child versus
adult. As noted above, for oral exposures this leads to faster elimination by respiration in children,
while for inhalation exposures it leads to higher uptake for a given air concentration.
The lack of difference in elderly adults versus the general population in HEC values is likely due
to the fact that the rate of exposure and rates of metabolism (the latter being the key dose metric) both
scale as BW0,75, with the scaling coefficients being either similar (respiration) or identical (metabolism)
among adults, who comprise the majority of the population, while for oral exposures the exposure rate is
normalized to total BW and scales as BW1, while elimination routes increase as BW0'75. Moreover, oral
exposures are simulated as occurring in a series of bolus exposures (drinking episodes) during the day,
and the higher body-fat content occurring in the elderly (see Appendix B) means that such a dose that
might saturate metabolism and therefore have a higher fraction exhaled in a leaner individual will tend
to be more sequestered in fat and slowly released, resulting in a higher fraction metabolized (less
saturation of metabolism) in a more obese individual. The difference among adults of different ages for
dosimetry from oral ingestion (bolus exposure) will be greater than the difference for inhalation
exposures. More careful examination of Figure 5-13 shows that the distribution for 70-year-old women,
for whom the fat fraction is estimated to be greatest, has a lower peak and higher upper tail than for the
general population. So the physiological differences do have some impact that is qualitatively consistent
with what is seen from oral exposure, given the mechanistic considerations described here. But the
impact of those differences is far less for inhalation exposure.
261

-------
7617
7618
7619
7620
7621
7622
7623
7624
7625
7626
7627
7628
7629
0.2
E" 0.18
Q.
¦|- 0.16
0
t5 0.14
re
£ 0.12
>
| 0.1
¦§ 0.08
>
£ 0.06
d>
§¦ 0.04
d>
1	0.02
0
Figure 5-13. Frequency density of HECs in specific populations in comparison to a
general population (0.5- to 80-year-old males and females) estimate for an internal
dose of 128.1 mg dichloromethane metabolized by CYP per liter liver per day ; all
groups restricted to the GST-T1population).
Table 5-10. Statistical characteristics of HECs in
specific populations of the GST-T1group
HEC (mg/m3)a
Population
Mean
5 th percentile 1st
percentile
All agesb
47.4
20.9
16.6
1-year-old children
24.7
144
12.1
70-year-old men
464
21.7
17.8
70-year-old women
50.0
22.2
18.0
aExposure levels predicted to result in 128.1 mg dichloromethane
metabolized via CYP pathway per liter liver per day (based on mean
BMDLio across acceptable models from Table 5-7).
b0.5- to 80-year-old males and females.
No data are available regarding toxicodynamic differences within a human population.
Therefore, a UF of 3 for possible differences in human toxicodynamic responses is intended to be
protective for sensitive individuals.
5.4. CANCER ASSESSMENT

Human equivalent
concentration distributions

	General population


	1 year old

: »
	70 yo male

:
	70 yo female



: t \ X

¦ f 1 \

B
\
¦ 1
-
10	20	30	40
Human equivalent concentration (ppm)
262

-------
7630
7631
7632
7633
7634
7635
7636
7637
7638
7639
7640
7641
7642
7643
7644
7645
7646
7647
7648
7649
7650
7651
7652
7653
7654
7655
7656
7657
7658
7659
7660
7661
7662
7663
7664
5.4.1. Cancer Oral Slope Factor
5.4.1.1.	Choice of Study/Data—with Rationale and Justification
No human data are available for the quantification of potential neoplastic effects from oral
exposures to dichloromethane. In the only chronic (2-year) oral exposure cancer bioassay, significant
increases in the incidence of liver adenomas and carcinomas was observed in male, but not female,
B6C3Fi mice exposed by drinking water, with incidence rates of 19, 26, 30, 31, and 28% in groups with
estimated mean intakes of 0, 61, 124, 177, and 234 mg/kg-day, respectively (trend p- value = 0.058) (see
Table 4-38 for group comparisons) (Serota et al., 1986b; Hazelton Laboratories, 1983). Evidence of a
trend for increased risk of liver tumors (described as neoplastic nodule or hepatocellular carcinoma) was
seen in female F344 rats, but not males, exposed via drinking water (p < 0.01) (Serota et al., 1986a).
However, the potential malignant characterization of the nodules was not described, and no trend was
seen in the data limited to hepatocellular carcinomas. The derivation of the cancer oral slope factor
(OSF) is based on the male mouse data (Serota et al., 1986b; Hazelton Laboratories, 1983) because of
their greater sensitivity compared to female mice and to male and female rats. The study authors
concluded that these increases were "within the normal fluctuation of this type of tumor incidence".
However, the trend Rvalue for these results is of borderline statistical significance and it may not be
reasonable to apply a correction for multiple comparisons given the lack of independence of the groups
and given a specific focus on the liver as a target organ. The development of liver tumors in B6C3Fi
mice is associated with metabolite production in this tissue via the GST metabolic pathway (section
4.7.3), a pathway that also exists in humans. Modeling intake, metabolism, and elimination of
dichloromethane in mice and humans is feasible. Thus, it is reasonable to apply the best available
PBTK models to estimate equivalent internal doses in mice and humans.
5.4.1.2.	Derivation of Oral Slope Factor
In a manner similar to the derivation of the noncancer toxicity values, PBTK models for
dichloromethane in mice and humans were used in the derivation of toxicity values (cancer OSF and
IUR) for cancer endpoints based on lung (for inhalation) and liver (for oral and inhalation) tumor data in
the mouse (Figure 5-14). A deterministic PBTK model for dichloromethane in mice was first used to
convert mouse drinking water or inhalation exposures to long-term daily average values of internal lung-
specific GST metabolism (GST metabolism in lung/lung volume) or liver-specific GST metabolism
(GST metabolism in liver/liver volume). The choice of this dose metric was made based on data
pertaining to the mechanism(s) involved in the carcinogenic response, specifically data supporting the
involvement of a GST metabolite(s). The evidence pertaining to the GST pathway is discussed in section
4.7, and includes the enhanced genotoxicity seen in bacterial and mammalian in vitro assays with the
introduction of GST metabolic capacity (Graves et al., 1994a) and the suppression by pretreatment with
263

-------
7665	a GSH depletory of the production of DNA SSBs seen in acute inhalation exposure to dichloromethane
7666	in mice (Graves et al., 1995).
7667
264

-------
Rodent Dose
Response Data
Rodent
PBTK Model
Estimates of Rodent
Internal Dose
BMD
Modeling
~u
£
o
£
~ <
Human Tumor Risk Factor
(internal dose)-1
Scaling
Factor
X
Benchmark Dose Analysis
0.5
0.4
0.3
0.2
0.1
0
10
20 30
Dose
40 50
60
Rodent Tumor Risk Factor
(internal dose)"1
(0.1/Rodent BMDL10)

Rodent Internal BMDL10

95% Lower Bound Estimate of Internal
Dose Associated with a 10% response
Multiply Human Tumor Risk Factor
By Distribution of Human Internal
Unit Doses
95th 99th
Distribution of Human Cancer
Oral Slope Factors or
Inhalation Unit Risks
Recommend value
+
Apply Age-Dependent Adjustment Factors
(ADAFs) for early life exposure
Probabilistic
Human PBTK
Model
Distribution of Human Internal
Doses from Unit Oral Doses
(1 mg/kg) or Inhalation
Concentrations (1 ug/m3)
. Monte Carlo
Sampling from
Distributions of
¦ Human PBTK
Model Parameters
7668
7669
7670
Figure 5-14. Process for deriving cancer OSFs and IURs by using rodent and human PBTK models.
Multistage	
	BMDL
	BMD
265

-------
7671
7672
7673
7674
7675
7676
7677
7678
7679
7680
7681
7682
7683
7684
7685
7686
7687
7688
7689
7690
7691
7692
7693
7694
7695
7696
7697
7698
7699
7700
7701
7702
7703
7704
7705
7706
7707
7708
The multistage cancer model (using BMDS version 2.0) was then fit to the tumor
incidence data and internal dose data for rodents, and BMDio and associated BMDLio values (for
a BMR of 10% extra risk) were calculated. A probabilistic PBTK model for dichloromethane in
humans, adapted from David et al. (2006) (see Appendix B), was used with Monte Carlo
sampling to calculate distributions of internal lung or liver doses associated with chronic unit
oral (1 mg/kg-day) or inhalation (1 (J,g/m3) exposures. The resulting distribution of human
internal doses was multiplied by a human internal dose tumor risk factor (in units of reciprocal
internal dose) to generate a distribution of OSFs or IURs associated with a chronic unit oral or
inhalation exposure, respectively.
The parameter statistics reported by David et al. (2006) include both the inter-individual
variability that would have been elucidated by the Bayesian analysis (variation between mean
values for each individual for which data were available) and uncertainty in those values. Since
EPA's objective is to account for both population variability and parameter uncertainty,
however, these statistics were primarily as-is (exceptions discussed in Appendix B) to define
population distributions. Assuming that these parameters are distributed independently, ignoring
the covariance that was likely represented in the actual posterior chains, will tend to over-
estimate the overall range of parameters and hence distribution of dose metrics in the population,
compared to what one would obtain if the covariance were explicitly included. Thus if the
covariance (i.e., the variance-covariance matrix) for the set of parameters had been reported by
David et al., it could have been used to narrow the predicted distribution of internal doses, or
equivalent applied doses. Lacking such information the approach used will not under-estimate
risk or over-estimate lower bounds on human equivalent exposure levels.
5.4.1.3.	Dose-Response Data
Data for liver tumors in male B6C3Fi mice following exposure to dichloromethane in
drinking water were used to develop oral cancer slope factors (Serota et al., 1986b; Hazelton
Laboratories, 1983). Significant increases in incidence of liver adenomas and carcinomas were
observed in male, but not female, B6C3Fi mice exposed for 2 years (Table 5-11). No significant
decreases in survival were observed in the treated groups of either sex compared with controls.
The at-risk study populations (represented by the denominators in the incidence data) were
determined by excluding all animals dying prior to 52 weeks.
5.4.1.4.	Dose Conversion and Extrapolation Methods: Cancer Oral Slope Factor
Dose conversion
The mouse PBTK model of Marino et al. (2006) was based on the PBTK model for
dichloromethane by Andersen et al. (1987), which was modified to include dichloromethane
metabolism in the lung compartment and kinetics of CO and COHb (Andersen et al., 1991). For
the mouse, physiological parameters and partition coefficients were adjusted to match those
266

-------
7709
7710
7711
7712
7713
7714
7715
7716
7717
7718
7719
7720
7721
7722
7723
7724
7725
7726
7727
7728
7729
reported in Andersen et al. (1991, 1987) and Clewell et al. (1993), respectively, while QCC,
VPR, and metabolic parameter distribution mean values were derived via MCMC model
calibration reported by Marino et al. (2006) (Appendix B). The model of Marino et al. (2006)
was used to simulate daily drinking water exposures comprising six discrete drinking water
episodes for specified times and percentage of total daily intake (Reitz et al., 1997) and to
calculate average lifetime daily internal doses for the male mouse data shown in Table 5-11. A
first-order oral uptake rate constant (ka) of 5 hours-1 was taken from Reitz et al. (1997) to
describe the uptake of dichloromethane from the gastrointestinal tract to the liver. Study-specific
BWs were not available, so reference BWs for male B6C3Fi mice in chronic studies (U.S. EPA,
1988a) were used. Based on evidence that metabolites of dichloromethane produced via the
GST pathway are primarily responsible for dichloromethane carcinogenicity in mouse liver
(summarized in section 4.7.3) and the assumption that these metabolites are sufficiently reactive
that they do not have substantial distribution outside the liver, the recommended selected internal
dose metric for liver tumors was daily mass of dichloromethane metabolized via the GST
pathway per unit volume of liver (Table 5-11). Figure 5-15 shows the comparison between
internal and external doses in the liver in mice and humans. The whole-body metabolism metric
was also examined. This metric would be more relevant under a scenario of slowly cleared
metabolites that undergo general circulation.
Table 5-11. Incidence data for liver tumors and internal liver doses,
based on GST metabolism dose metrics, in male B6C3Fi mice
exposed to dichloromethane in drinking water for 2 years

Nominal (actual)

Mouse internal


daily intake
Mouse liver
liver metabolism
Mouse whole body
Sex
(mg/kg-day)
tumor incidence3
doseb
metabolism dosec
Male
0(0)
24/125 (19%)
0
0
(BW =
60 (61)
51/199 (26%)
17.5
0.73
37.3 g)
125 (124)
30/99 (30%)d
63.3
2.65

185 (177)
31/98 (32%)d
112.0
4.68

250 (234)
35/123 (28%)d
169.5
7.1
""Hepatocellular carcinoma or adenoma, combined. Mice dying prior to 52 weeks were excluded
from the denominators. Cochran-Armitage trend p-valuc = 0.058.
bmg dichloromethane metabolized via GST pathway/L liver/day. Internal doses were estimated
from simulations of actual daily doses reported by the study authors.
0 Based on the sum of dichloromethane metabolized via the GST pathway in the lung plus the
liver, normalized to total BW (i.e., [lung GST metabolism (mg/day) + liver GST metabolism
(mg/day)]/kg B W. Units = mg dichloromethane metabolized via GST pathway in lung and
liver/kg-day.
dSignificantly (p < 0.05) different from control incidence by Fisher's exact test performed by
Syracuse Research Corporation.
Sources: Serotaetal., 1986b; Hazelton Laboratories, 1983
267

-------
7730
7731
7732
7733
7734
7735
7736
7737
7738
7739
7740
7741
7742
7743
7744
7745
7746
7747
7748
7749
7750
7751
7752
7753
1000
Liver GST dose
for oral exposure
_i
100
10
— - - Mouse
	Human mixed
	Human +/-
	Human +/+
1
10
100
Dose (mg/kg/day)
1000
Figure 5-15. PBTK model-derived internal doses (mg dichloromethane metabolized via
the GST pathway per liter liver per day) in mice and humans and their associated
external exposures (mg/kg-day) used for the derivation of cancer OSFs based on liver
tumors in mice. Six simulated drinking water episodes are described by Reitz et al. (1997).
The human metabolism rates were estimated using a computational sample of 1000 individuals
per dose, including random samples of the three GST-T1 polymorphisms (+/+, +/-, "Human
mixed" curve) or samples restricted to the GST +/+ or +/- populations, in the current U.S.
population based on data from Haber et al. (2002). Since a different set of samples was used
for each dose, some stochasticity is evident as the human points (values) do not fall on smooth
curves. Error bars indicate the range of 5th-95th percentile for the sub-populations sampled at
select concentrations.
Dose-response modeling and extrapolation
The multistage dose-response model was fit to the mouse liver tumor incidence and
PBTK model-derived internal dose data to derive mouse internal BMDio and BMDLio associated
with 10% extra risk (Table 5-12). Different polynomial models, and models dropping dose
groups starting with the highest dose group were compared based on adequacy of model fit as
assessed by overall % goodness of fit (/;-value > 0.10) and examination of residuals, particularly
in the region of the benchmark response (BMR). Appendix E-l provides details of the BMD
modeling results. The mouse liver tumor risk factor (extra risk per unit internal dose) was
calculated by dividing 0.1 by the mouse BMDLio for liver tumors.
268

-------
7754
7755
7756
7757
7758
7759
7760
7761
7762
7763
7764
7765
7766
7767
7768
7769
7770
7771
7772
7773
7774
7775
Table 5-12. BMD modeling results and tumor risk factors for internal dose metric
associated with 10% extra risk for liver tumors in male B6C3Fi mice exposed to
dichloromethane in drinking water for 2 years, based on liver-specific GST metabolism
and whole body GST metabolism dose metrics


x2


Allometric-
Tumor Risk Factor"
Internal
BMDS
goodness of
Mouse
Mouse
scaled human
Allometric-
dose metric
modelb
fit p- value
BMD10C
BMDL10C
BMDL10d
Scaling = 1.0 scaled
Liver-
MS (1,1)
0.56
73.0
39.6
5.66
2.53 x 10-3 1.77 x 10-2
specific






Whole-body
MS (1,1)
0.56
3.05
1.65
0.24
4.24 x 10"1
aLiver specific dose units = mg dichloromethane metabolized via GST pathway per liter tissue per day;
Whole-body dose units = mg dichloromethane metabolized via GST pathway in lung and liver/kg-
day)
bThe multistage (MS) model in EPA BMDS version 2.0 was fit to the mouse dose-response data
shown in Table 5-11 using internal dose metrics calculated with the mouse PBTK model. Numbers
in parentheses indicate (1) the number of dose groups dropped in order to obtain an adequate fit and
(2) the degree polynomial of the model.
°BMDio and BMDLio refer to the BMD-model-predicted mouse internal and its 95% lower confidence
limit, associated with a 10% extra risk for the incidence of tumors.
dMouse BMDL10 divided by (BWhuman/BWmouse)°25 = 7.
eDichloromethane tumor risk factor (extra risk per unit internal dose) derived by dividing the BMR
(0.1) by the mouse BMDLio and by the allometric-scaled human BMDL10„ for the scaling =1.0 and
allometric-scaled risk factors, respectively.
Linear extrapolation from the internal human BMDLio values (0.1/BMDLio) was used to
derive oral risk factors for liver tumors, based on tumor responses in male mice. Proposed key
events for dichloromethane carcinogenesis are discussed in sections 4.7 and 5.4.1.1. The linear
low-dose extrapolation approach for agents with a mutagenic mode of action was selected.
Application of allometric scaling factor
As discussed in section 4.7 and summarized in 5.4.1.2, several lines of evidence point to
the involvement of the GST metabolic pathway in the carcinogenic response seen in
dichloromethane. The role of specific metabolites has not been firmly established, however.
S-(chloromethyl)-glutathione is an intermediate to the production of formaldehyde through this
pathway (Hashmi et al., 1994). Formation of free hydrogen ion is also hypothesized, although
no direct evidence supporting this has been presented. The pattern of HPRT gene mutations seen
in CHO cells incubated with GST-complete mouse liver cytosol preparations suggest that
S-(chloromethyl)glutathione, rather than formaldehyde, is responsible for the mutagenic effects
associated with dichloromethane (Graves et al., 1996). DNA reaction products (e.g., DNA
adducts) produced by S-(chloromethyl)glutathione have not been quantified, possibly due to
potential instability of these compounds (Guengerich et al., 2003; Hashmi et al., 1994).
The question of the role of specific metabolites, and particularly how these metabolites
are transformed or removed is a key question affecting the choice of a scaling factor to be used in
conjunction with the internal dose metric based on rate of GST metabolism. If the key metabolite
is established and is known to be sufficiently reactive to not spread in systemic circulation, then
269

-------
7776
7777
7778
7779
7780
7781
7782
7783
7784
7785
7786
7787
7788
7789
7790
7791
7792
7793
7794
7795
7796
7797
7798
7799
7800
7801
7802
7803
7804
7805
7806
7807
7808
7809
7810
7811
7812
7813
it can be assumed that 1) the level of reactivity and rate of clearance (i.e., disappearance due to
local reactivity) for this metabolite, per volume tissue, is equal in rodents and humans; and (2)
risk is proportional to the long-term daily average concentration of the metabolite. Under these
assumptions, rodent internal BMDLio values based on tissue-specific dichloromethane
metabolism require no allometric scaling to account for toxicodynamic differences and predict
the corresponding level of human risk as a function of the metric (i.e., the scaling factor in Figure
5-14 was equal to 1.0). (A single metabolite is referenced, but the same argument holds in
general for more than one metabolite). Under this scenario and assumptions, humans and
rodents with the same long-term daily average metabolite formation per volume tissue (e.g.,
equal internal BMDLio) should both experience the same long-term average concentration of the
metabolite when the metabolite is highly reactive and hence experience the same extra risk.
Although the evidence points to a specific metabolic pathway and to site-specific actions
resulting from a reactive metabolite that does not escape the tissue in which it is formed, some
assumptions remain concerning this hypothesis. Specifically, the active metabolite(s) have not
been established, and data pertaining to the reactivity or clearance rate of these metabolite(s) are
lacking. Quantitative measurements of adducts of interest or of the half life of relevant
compounds in humans and in mice are not available. To address the uncertainties in the
available data it may be appropriate to use a scaling factor that addresses the possibility that the
rate of clearance for the metabolite is limited by processes that are known to scale allometrically,
such as blood perfusion or enzyme activity. This case would result in use of a mouse:human
dose-rate scaling factor of (BWhuman/BWmoUse)0'25 = 7 to adjust the mouse-based BMDLio values
downward. Using this internal dose metric (liver-specific metabolism with allometric scaling),
equivalent rodent and human internal BMDLio values result in a human liver tumor risk factor
(0.1/BMDLio) that is assumed equal to that for the mouse, given a 70-year lifetime exposure.
Another alternative that can be used is based on an allometrically-scaled whole-body
metabolism metric. In this case, less weight is given to the evidence of site-specificity, as this
metric allows for systemic circulation of the relevant metabolites.
The cancer toxicity values derived using each of these metrics and scaling factors (i.e.,
liver-specific metabolism with and without allometric-scaling and the whole-body metabolism
metric) are presented in the following tables. Considering the lack of data pertaining to
clearance rates or the actual AUC of the active carcinogenic metabolite(s) in mice and humans,
the OSF recommended by the EPA is based on the allometrically-scaled tissue-specific GST
metabolism rate dose metric.
Calculation of OSFs
The human PBTK model adapted from David et al. (2006) (see Appendix B), using
Monte Carlo sampling techniques, was used to calculate distributions of human internal dose
metrics of daily mass of dichloromethane metabolized via the liver-specific GST pathway per
270

-------
7814
7815
7816
7817
7818
7819
7820
7821
7822
7823
7824
7825
7826
7827
7828
7829
7830
7831
7832
7833
7834
7835
7836
7837
7838
7839
7840
7841
unit volume of liver resulting from a long-term average daily drinking water dose of 1 mg/kg
dichloromethane. In another analysis of whole body metabolism, a dose metric based on the
total metabolites formed in liver and lungs via GST metabolism per BW was used. The human
model used parameter values derived from Monte Carlo sampling of probability distributions for
each parameter, including MCMC-derived distributions for the metabolic parameters (David et
al., 2006). The drinking water exposures comprised six discrete drinking-water episodes for
specified times and percentage of total daily intake (Reitz et al., 1997) (Appendix B).
The distribution of cancer OSFs shown in Table 5-13 were derived by multiplying the
human oral liver tumor risk factors by the respective distributions of human average daily
internal doses resulting from chronic, unit oral exposures of 1 mg/kg-day dichloromethane.
Because adjustments for interindividual variability are not generally used or recommended in
cancer risk analysis, the mean slope factor was selected as the recommended value to be used in
deterministic risk assessments; other values at the upper end of the distribution are also
presented.
Consideration of Sensitive Human Subpopulations
An important issue in the derivation process used by EPA, pertaining to the use of the
human PBTK model, stems from the assumption regarding the population for which the
derivation should be applied. The inclusion of the GST-T1 null subpopulation in effect dilutes
the risk that would be experienced by those who carry a GST-T1 allele, by averaging in non-
responders (i.e., the GST-T1 7 genotype). Thus, the cancer OSF was derived specifically for
carriers of the GST-T1 homozygous positive (+/+) genotype, that is the population that would be
expected to be most sensitive to the carcinogenic effects of dichloromethane given the GST-
related dose metric under consideration. In addition, cancer values derived for a population
reflecting the estimated frequency of GST-T1 genotypes in the current U.S. population (20%
GST-T1 ^", 48% GST-T1 , and 32% GST-T1+/+, the "mixed" population) are also presented.
All simulations also included a distribution of CYP activity based on data from Lipscomb et al.
(2003).
271

-------
Table 5-13. Cancer OSFs for dichloromethane based on PBTK model-derived internal liver doses in B6C3F1 mice exposed
via drinking water for 2 years, based on liver-specific GST metabolism and whole body metabolism dose metrics, by
population genotype
Distribution of human internal
dichloromethane
doses from 1 mg/kg-day exposure"1
Resulting candidate human
OSFe (mg/kg-day) 1
Internal dose
metric and
Population
Human tumoi

95
th
99
th

95
th
99th
scaling factor"
genotypeb
risk factor0
Mean
percentile
percentile
Mean
percentile
percentile
Liver-specific,
GST-T1+/+
1.77
10~2
0.80 x 10"1
1.91 x
10 1
2.89 x
10 1
1.4 x 10 3
3.4 x
10 3
5.1 x 10~3
allometric-scaled
Mixed
1.77
10~2
0.45 x 1() 1
1.39 x
10 1
2.24 x
10 1
8.0 x 10^
2.5 x
10 3
4.0 x 10~3
Liver-specific,
GST-T1+/+
2.53
10~3
0.80 x 10"1
1.91 x
10 1
2.89 x
10 1
2.0 x 10"1
4.8 x
10 4
7.3 x 10~4
scaling =1.0
Mixed
2.53
10~3
0.45 x ]() 1
1.39 x
10 1
2.24 x
10 1
1.2 x 10 1
3.5 x
10 4
5.7 x 10~4
Whole-body,
GST-T1+/+
4.24
10"1
1.90 x 10~3
X
o
VO
10 3
7.20 x
10 3
8.1 x 10^
2.0 x
10 3
3.1 x 10~3
allometric-scaled
Mixed
4.24
10"1
1.08 x 1() '
3.40 x
10 3
5.49 x
10 3
4.6 x 10 1
1.4 x
10 3
2.3 x 10"3
aLiver specific dose units = mg dichloromethane metabolized via GST pathway per liter tissue per day; Whole-body dose units = mg
dichloromethane metabolized via GST pathway in lung and liver/kg-day.
bGST-Tl+/+ = homozygous, full enzyme activity; mixed = population reflecting estimated frequency of genotypes in current U.S. population:
20% GST-T . 48% GST-T1 . and 32% GST-T1+/+ (Haber et al., 2002).
Dichloromethane tumor risk factor (extra risk per unit internal dose per day) derived by dividing the BMR (0.1) by the allometric-scaled
human BMDLio and the mouse BMDL10 for the allometric-scaled and scaling =1.0 risk factors, respectively (from Table 5-12).
dMean, 95th, and 99th percentile of the human PBTK model-derived probability distribution of daily average internal dichloromethane dose
resulting from chronic oral exposure of 1 mg/kg-day.
eDerived by multiplying the dichloromethane tumor risk factor by the PBTK model-derived probabilistic internal doses from daily exposure to
1 mg/kg-day.
7842
7843
272

-------
7844
7845
7846
7847
7848
7849
7850
7851
7852
7853
7854
7855
7856
7857
7858
7859
7860
7861
7862
5.4.1.5.	Oral Cancer Slope Factor
The recommended cancer OSF for dichloromethane is 1 x 10 3 (mg/kg-day) 1 (rounded
from 1.4 x 10 3), and is based on liver tumor responses in male B6C3Fi mice exposed to
dichloromethane in drinking water for 2 years (Serota et al., 1986b; Hazelton Laboratories,
1983). The OSF was derived by using a tissue-specific GST metabolism dose metric, with
allometric scaling, for the population that is presumed to have the greatest sensitivity (the GST-
Tl+/+ genotype). The application of ADAFs to the cancer OSF is recommended and is described
in section 5.4.4.
5.4.1.6.	Alternative Derivation Based on Route-to-Route Extrapolation
For comparison, alternative cancer OSFs were derived via route-to-route extrapolations
from the data for liver tumors in male and female B6C3Fi mice exposed by inhalation for 2 years
(Mennear et al., 1988; NTP, 1986). This derivation, shown in Table 5-14, uses the cancer IUR
derived in section 5.4.2.4 (see Table 5-19 for these IUR values) and the distribution of human
internal dichloromethane exposures from 1 mg/kg-day exposure using the tissue-specific GST
metabolism dose metric (mg dichloromethane metabolized via the GST pathway per liter liver
per day). The oral cancer slope factor values based on the route-to-route extrapolations from
liver tumors in mice exposed by inhalation (Table 5-14) are about one order of magnitude lower
than those based on the liver tumor responses in mice exposed via drinking water.
273

-------
Table 5-14. Alternative route-to-route cancer OSFs for dichloromethane extrapolated from male B6C3Fi mouse
inhalation liver tumor incidence data using a tissue-specific GST metabolism dose metric, by population genotype
Internal dose metric
and scaling factor
Population
Genotype3
Human
tumor risk
factorb
Distribution of human internal
dichloromethane
doses from 1 mg/kg-day exposure0
Resulting candidate human
OSFd (mg/kg-day) 1
Mean
95th
percentile
99th
percentile
Mean
95th
percentile
99th
percentile
Liver-specific,
GST-T 1+/+
1.29
10~3
0.80 x 10"
1.91
10"1
2.89
10"1
1.0 x 10
"4 2.5
10"4
3.7
10"4
allometric-scaled
Mixed
1.29
10~3
0.45 x 1()
1.39
10"1
2.24
10"1
5.8 x 10
"5 1.8
10"4
2.9
10"4
Liver-specific,
GST-T 1+/+
1.84x
10"1
0.80 x 10"
1.91
10"1
2.89
10"1
1.5 x 10
"5 3.5
10 '
5.3
10 5
scaling =1.0
Mixed
1.84x
10"1
0.45 x 1()
1.39
10"1
2.24
10"1
8.3 x 10
~6 2.6
10 s
4.1
10 5
Whole-body
GST-T 1+/+
3.03
10~2
1.90 x 10"
4.60
10"3
7.20
10"3
5.8 x K)
5 1.4
10"4
2.2
10"4
metabolism
Mixed
3.03
10~2
1.08 x 1()
3.40
10"3
5.49
10"3
3.3 x 10
"5 1.0
10"4
1.7
10"4
aGST-Tl+/+ = homozygous, full enzyme activity; mixed = population reflecting estimated frequency of genotypes in current U.S. population:
20% GST-T . 48% GST-T 1 , and 32% GST-T1+/+ (Haber et al., 2002).
bDichloromethane tumor risk factor (extra risk per milligram dichloromethane metabolized via GST pathway per liter tissue per day) derived
by dividing the BMR (0.1) by the allometric-scaled human BMDLio and the mouse BMDL10 for the allometric-scaled and scaling =1.0 risk
factors, respectively (from inhalation unit risk data, Table 5-19).
°Mean, 95th, and 99 percentile of the human PBTK model-derived probability distribution of daily average internal dichloromethane dose
(milligrams dichloromethane metabolized via GST pathway per liter tissue per day) resulting from chronic oral exposure of 1 mg/kg-day.
dDerived by multiplying the dichloromethane tumor risk factor by the PBTK model-derived probabilistic internal doses from daily exposure
to 1 mg/kg-day.
7863
274

-------
7864
7865
7866
7867
7868
7869
7870
7871
7872
7873
7874
7875
7876
7877
7878
7879
7880
7881
7882
7883
7884
7885
7886
7887
7888
7889
7890
5.4.1.7. Alternative Based On Administered Dose
One comparison that can be made is with an alternative OSF based on liver tumors in
mice, using the external concentrations of dichloromethane in the mouse as converted to human
equivalent doses and then applying this by using BMD modeling to obtain the BMDLio and
resulting oral cancer risk. Mouse bioassay exposures were adjusted to human equivalent doses
as follows:
human equivalent dose =
(nominal daily intake BW scaling factor) x daily exposure adjustment factor
where BW scaling factor	= (BWhuman/BWm0Use) °'25 = 7
and
daily exposure adjustment factor = 5/7
The human equivalent administered doses for the 0, 60, 125, 185, and 250 mg/kg-day
dose groups used in the liver tumor analysis (Table 5-11) (from Serota et al. [1986b]) were 0,
6.12, 12.75, 18.87,and 25.51 mg/kg-day, respectively. The BMD modeling and OSF derived
from these values are shown in Table 5-15. The resulting OSF, based on the liver tumors in the
mouse, is approximately one order of magnitude higher than the current recommended value
obtained by using the mouse and human PBTK models.
Table 5-15. Cancer OSF based on a human BMDL10 using administered dose for
liver tumors in male B6C3F, mice exposed to dichloromethane in drinking water
for 2 years




x2

Cancer
Sex,
goodness of fit
Human Human
OSF d
tumor type BMDS model"
p- value
BMD10C BMDL10C
(mg/kg-day) 1
Male, liver MS (0,1)
0.55
19.4 10.4
1.0 x 10~2
aThe multistage (MS) model in EPA BMDS version 2.0 was fit to the mouse liver tumor data shown in
Table 5-11. The human equivalent doses for the 0, 60, 125, 185, and 250 mg/kg-day dose groups used in
the liver tumor analysis were 0, 6.12, 12.75, 18.87, and 25.51 mg/kg-day, respectively, based on
application of BW scaling factor = (BWhuman/BWmouse)°25 = 7 and adjusting for daily exposure by
multiplying by 5/7 days. Numbers in parentheses indicate (1) the number of dose groups dropped in order
to obtain an adequate fit, starting with the highest dose group, and (2) the degree polynomial of the model.
°BMDio and BMDLio refer to the BMD-model-predicted human equivalent administered dose (mg/kg-day)
and its 95% lower confidence limit, associated with a 10% extra risk for the incidence of tumors.
dCancer OSF (risk per mg/kg-day) = 0.1/human BMDLi0.
The administered dose methodology can be considered equivalent to using a single-
compartment, whole-body model of dichloromethane, where the internal dose metric is the AUC
of dichloromethane itself and clearance of dichloromethane scales from mice to humans as
Bw°75 xhe estimates based on the PBTK model, in contrast, use the rate of metabolism of
275

-------
7891
7892
7893
7894
7895
7896
7897
7898
7899
7900
7901
7902
7903
7904
7905
7906
7907
7908
7909
7910
7911
7912
7913
7914
7915
7916
7917
7918
7919
7920
dichloromethane (GST) as the metric. Another difference is that the administered dose
methodology does not account in any way for the GST polymorphism and so might be considered
as representing the general/mixed-GST-genotype population rather than the +/+ subpopulation.
5.4.1.8.	Previous IRIS Assessment: Cancer Oral Slope Factor
The previous IRIS assessment derived a cancer OSF of 7.5 x 10 3 (mg/kg-day )_1 by the
application of the multistage model to combined incidence of hepatocellular adenomas,
carcinomas, and neoplastic nodules from two studies. These were the 2-year drinking water
study of dichloromethane in B6C3Fi mice by the Hazelton Laboratories (1983) and the 2-year
inhalation study of dichloromethane in B6C3Fi mice by NTP (1986). The slope factor was the
arithmetic mean of two candidate slope factors, 1.2 x 10 2 (mg/kg-day) 1 (Hazelton Laboratories,
1983) and 2.6 x 10~3 (mg/kg-day )_1 (NTP, 1986). Since the NTP (1986) animal data were from
inhalation exposures, the estimated inhaled doses were calculated for mice and humans
(assuming near complete uptake into lung tissues and blood) and converted to administered
doses in units of mg/kg-day. Assumed inhalation rates of 0.0407 and 20 m3/day were used for
mice and humans, respectively. No adjustments were made for species differences in
metabolism or toxicokinetics.
5.4.1.9.	Comparison of Cancer Oral Slope Factors Using Different Methodologies
Cancer OSFs derived using different dose metrics and assumptions are summarized in
Table 5-16. The recommended OSF of 1 x 10 3 per mg/kg-day (rounded to one significant digit)
is based on a tissue-specific GST-internal dose metric, with allometric scaling (= 7), because of
some uncertainty regarding the rate of clearance of the relevant metabolite(s) formed via the
GST pathway. The value derived specifically for the GST-T1+/+ population is recommended to
provide protection for the population that is hypothesized to be most sensitive to the carcinogenic
effect. The values based on the GST-T1+/+ group are approximately two-fold higher than those
for the full population for the dose metrics used in this assessment (Table 5-16). Within a
genotype population, the values of the OSF among most of the various dose metrics vary by
about one to two orders of magnitude.
276

-------
7921
Table 5-16. Comparison of OSFs derived using various assumptions and metrics, based on tumors in male mice


Species,

Scaling
Mean OSF
Source
Population"
Dose metric
sex
Tumor
factor
(mg/kg-day) 1
(Table)
GST-T1+/+
Tissue-specific GST-metabolism rateb
Mouse, male
Liver
7.0
1.4 x 10"3
Table 5-13

Tissue-specific GST-metabolism rate
Mouse, male
Liver
1.0
2.0 >
< lO"
Table 5-13

Whole-body GST metabolism rate
Mouse, male
Liver
7.0
8.1 >
< 10"1
Table 5-13

Route-to-route extrapolation, tissue-specific metabolism
Mouse, male
Liver
7.0
1.0 >
< lO"
Table 5-14

Route-to-route extrapolation, tissue-specific metabolism
Mouse, male
Liver
1.0
1.5 >
< 10~5
Table 5-14

Route-to-route extrapolation, whole-body metabolism
Mouse, male
Liver
7.0
5.8 >
< 10~5
Table 5-14
Mixed
Tissue-specific GST-metabolism rateb
Mouse, male
Liver
7.0
8.0 >
< 10"1
Table 5-13

Tissue-specific GST-metabolism rate
Mouse, male
Liver
1.0
1.2 >
< 10"1
Table 5-13

Whole-body GST metabolism rate
Mouse, male
Liver
7.0
4.6 >
< 10"1
Table 5-13

Route-to-route extrapolation, tissue-specific metabolism
Mouse, male
Liver
7.0
5.8 >
< 10~5
Table 5-14

Route-to-route extrapolation, tissue-specific metabolism
Mouse, male
Liver
1.0
8.3 >
< 10^
Table 5-14

Route-to-route extrapolation, whole-body metabolism
Mouse, male
Liver
7.0
3.3 >
< 10~5
Table 5-14

Applied dose (human equivalent dose)
Mouse, male
Liver

1.0 >
( 10~2
Table 5-15

1995 IRIS assessment
Mouse, male
Liver

7.5 >
< 10~3

'GST-Tl = homozygous, full enzyme activity; Mixed = genotypes based on a population reflecting the estimated frequency of genotypes in the current
U.S. population: 20% GST-Tl^, 48% GST-Tl . and 32% GST-T1+/+ (Haber et al., 2002).
Bolded value is the basis for the recommended OSF of 1 x 10"3 per mg/kg-day.
7922
277

-------
7923
7924
7925
7926
7927
7928
7929
7930
7931
7932
7933
7934
7935
7936
7937
7938
7939
7940
7941
7942
7943
7944
7945
7946
7947
7948
7949
7950
7951
7952
7953
7954
7955
7956
7957
7958
7959
7960
5.4.2. Cancer Inhalation Unit Risk
5.4.2.1. Choice of Study/Data—with Rationale and Justification
As discussed in section 4.7, results from several cohort mortality studies of workers
repeatedly exposed to dichloromethane and several case-control studies provide some supporting
evidence of carcinogenicity in humans, specifically with respect to liver and brain cancer.
However, the epidemiologic studies do not provide adequate data to estimate exposure-response
relationships for dichloromethane exposure and these cancers.
Results from several bioassays provide sufficient evidence of the carcinogenicity of
dichloromethane in mice and rats exposed by inhalation, as well as adequate data to describe
dose-response relationships. As discussed in section 4.7.2, repeated inhalation exposure to
concentrations of 2,000 or 4,000 ppm dichloromethane produced increased incidences of lung
and liver tumors in male and female B6C3Fi mice (Maronpot et al., 1995; Foley et al., 1993;
Kari et al., 1993; Mennear et al., 1988; NTP, 1986). A weaker trend was seen with respect to
liver tumor incidence (described as neoplastic nodules or hepatic carcinomas) in female rats, but
this trend was not seen when limited to hepatic carcinomas (NTP, 1986). A statistically
significant increased incidence of brain tumors has not been observed in any of the animal cancer
bioassays, but a 2-year study using relatively low exposure levels (0, 50, 200, and 500 ppm) in
Sprague-Dawley rats observed a total of six astrocytoma or glioma (mixed glial cell) tumors
(combining males and females) in the exposed groups (Nitschke et al., 1988a). These tumors are
exceedingly rare in rats, and there are few examples of statistically significant trends in animal
bioassays (Sills et al., 1999). Male and female F344 rats exposed by inhalation to 2,000 or 4,000
ppm showed significantly increased incidences of benign mammary tumors (adenomas or
fibroadenomas) and the male rats also exhibited a low rate of sarcoma or fibrosarcoma in
mammary gland or subcutaneous tissue around the mammary gland (NTP, 1986).
The NTP inhalation study in B6C3Fi mice (NTP, 1986) was used to derive an IUR for
dichloromethane because of the completeness of the data, adequate sample size, and clear dose
response with respect to liver and lung tumors. The liver tumor incidence in male mice
increased from 44% in controls to 66% in the highest dose group; in females the incidence of this
tumor rose from 6 to 83%. For lung tumors, the incidence rose from 10 to 80% in males and
from 6 to 85% in females. Compelling evidence exists for the role of GST-mediated metabolism
of dichloromethane in carcinogenicity in mice (section 4.7.3), and both mice and humans possess
this metabolic pathway. Modeling intake, metabolism, and elimination of dichloromethane in
mice and humans is feasible. Thus, it is reasonable to apply the best available PBTK models to
estimate equivalent internal doses in mice and humans.
The mammary tumor data from the NTP (1986) study was also used to derive a
comparative IUR. However, the toxicokinetic or mechanistic events that might lead to
mammary gland tumor development in rats are unknown, and so a clear choice of the optimal
278

-------
7961
7962
7963
7964
7965
7966
7967
7968
7969
7970
7971
7972
7973
7974
7975
7976
7977
7978
7979
7980
7981
7982
7983
7984
7985
7986
7987
7988
7989
7990
7991
7992
7993
7994
internal dose metric could not be made. Thus, this derivation is based on the average daily AUC
for dichloromethane in blood. The role of CYP- or GST-mediated metabolism in the mammary
gland is uncertain, although both GST-T1 (Lehmann and Wagner, 2008) and CYP2E1 (El-Rayes
et al., 2003; Hellmold et al., 1998) expressions have been detected in human mammary tissue. It
is also possible that some metabolites enter systemic circulation from the liver and lung, where
they are formed.
The female rat liver cancer data from the NTP (1986) inhalation study was not used to
derive an IUR because the trend was weaker than that seen in the mouse (incidence increased
from 4% in controls to 10% in the highest dose group, trend p = 0.08), and because the effect
categorization included neoplastic nodule or hepatocellular carcinoma. The brain tumor data
seen in the Nitschke et al. (1988a) study in Sprague-Dawley rats were not used to develop an
IUR because of the low incidence of this rare tumor (a total of four astrocytoma or glioma
tumors in exposed males and two in exposed females). The mechanistic issues with respect to
mammary tumors and health effects issues with respect to brain tumors represent data gaps in the
understanding of the health effects of dichloromethane and relevance of the rat data to humans.
5.4.2.2.	Derivation of the Cancer Inhalation Unit Risk
The derivation of the IUR parallels the process described in section 5.4.1.2 for the cancer
OSF. Since modeling metabolism and elimination kinetics of dichloromethane in mice and
humans is feasible, it is reasonable to apply the best available PBTK models to determine
equivalent target organ doses in mice and humans.
5.4.2.3.	Dose-Response Data
Data for liver and lung tumors in male and female B6C3Fi mice following exposure to
airborne dichloromethane were used to develop IURs for dichloromethane (Mennear et al., 1988;
NTP, 1986). As discussed in section 5.4.1.8, the liver tumor dose-response data were also the
basis of an OSF, derived by route-to-route extrapolation using the PBTK models, to compare
with an OSF based on liver tumor data in mice exposed to dichloromethane in drinking water
(Serota et al., 1986b). In the NTP (1986) study, significant increases in incidence of liver and
lung adenomas and carcinomas were observed in both sexes of B6C3Fi mice exposed
6 hours/day, 5 days/week for 2 years (Table 5-17). Since significant decreases in survival were
observed in the treated groups of both sexes, the at-risk study populations (represented by the
denominators in the incidence data) were determined by excluding all animals dying prior 52
weeks.
279

-------
7995
7996
7997
7998
7999
8000
8001
8002
8003
8004
8005
8006
8007
8008
8009
8010
8011
5.4.2.4. Dose Conversion and Extrapolation Methods: Cancer Inhalation Unit Risk
Dose conversion
The PBTK model of Marino et al. (2006) for dichloromethane in the mouse was used to
simulate inhalation exposures of 6 hours/day, 5 days/week (Mennear et al., 1988; NTP, 1986)
and to calculate long-term daily average internal doses. Study-, group-, and sex-specific mean
BWs were used. Based on evidence that metabolites of dichloromethane produced via the GST
pathway are primarily responsible for dichloromethane carcinogenicity in mouse liver and lung
(summarized in section 4.7.3), and the assumption that these metabolites are sufficiently reactive
that they do not have substantial distribution outside these tissues, the recommended selected
internal dose metric for liver tumors and lung tumors were long-term average daily mass of
dichloromethane metabolized via the GST pathway per unit volume of liver and lung,
respectively (Table 5-17). Figure 5-16 show the comparison between inhalation external and
internal doses in the liver and lung, respectively, using this dose metric for the mouse and for the
human. A whole-body metabolism metric was also examined. This metric would be more
relevant under a scenario of slowly cleared metabolites that undergo general circulation
Table 5-17. Incidence data for liver and lung tumors and internal doses, based on GST
metabolism dose metrics, in male and female B6C3Fi mice exposed to dichloromethane via
inhalation for 2 years
Sex,
tumor type
BW (g)
External
dichloromethane
concentration
(ppm)
Mouse
tumor incidence
Mouse internal
tissue dose"
Mouse whole body
metabolism doseb
Male, liver0
--
0
22/50 (44%)e
0
0

34.0
2,000
24/47 (51%)
2,363.7
100.2

32.0
4,000
33/47 (70%)
4,972.2
210.7
Male, lungd
--
0
5/50 (10%)e
0
0

34.0
2,000
27/47 (55%)
475.0
100.2

32.0
4,000
40/47 (85%)
992.2
210.7
Female, liver0
--
0
3/47 (6%)e
0
0

30.0
2,000
16/46 (35%)
2,453.2
104.0

29.0
4,000
40/46 (87%)
5,120.0
217.0
280

-------
Table 5-17. Incidence data for liver and lung tumors and internal doses, based on GST
metabolism dose metrics, in male and female B6C3Fi mice exposed to dichloromethane via
inhalation for 2 years
Sex,
tumor type
BW (g)
External
dichloromethane
concentration
(ppm)
Mouse
tumor incidence
Mouse internal
tissue dosea
Mouse whole body
metabolism doseb
Female, lungd
--
0
3/45 (6%)e
0
0

30.0
2,000
30/46 (65%)
493.0
104.0

29.0
4,000
41/46 (89%)
1,021.8
217.0
aFor liver tumors: mg dichloromethane metabolized via GST pathway/L liver tissue/day from 6 hours per day, 5
days per week exposure; for lung tumors: mg dichloromethane metabolized via GST pathway/L lung tissue/day
from 6 hours per day, 5 days per week exposure.
''Based on the sum of dichloromethane metabolized via the GST pathway in the lung plus the liver, normalized to
total BW (i.e., [lung GST metabolism (mg/dav) + liver GST metabolism (mg/day)]/kg BW). Units = mg
dichloromethane metabolized via GST pathway in lung and liver/kg-day.
hepatocellular carcinoma or adenoma. Mice dying prior to 52 weeks were excluded from the denominators.
Bronchoalveolar carcinoma or adenoma. Mice dying prior to 52 weeks were excluded from the denominators.
Statistically significant increasing trend (by incidental and life-table tests; p < 0.01).
Sources: Memiearet al., 1988; NTP, 1986
8012
8013
8014
8015
8016
8017
10,000
1,000
Liver GST metabolism
©
w
o
"O
15
c
i_
4-1
c
100 L
Mouse
Human mixed GST
Human GST +/
Human GST +/+
i	i	i	rrrh
100	1,000
Inhalation concentration (ppm)
10,000
B.
281

-------
8018
8019
8020
8021
8022
8023
8024
8025
8026
8027
8028
8029
8030
8031
8032
8033
8034
8035
8036
8037
8038
8039
8040
8041
8042
8043
metabolism
Mouse
Human mixed GST
Human GST +/-
Human GST +/+
1,000
0.01
10	100	1,000	10,000
Inhalation concentration (ppm)
Figure 5-16. PBTK model-derived internal doses (mg dichloromethane metabolized via the
GST pathways per liter tissue per day) for liver (A) and lung (B) in mice and humans, and
their associated external exposures (ppm), used for the derivation of cancer inhalation unit
risks. Average daily doses were calculated from simulated mouse exposures of 6 hours/day, 5
days/week, while simulated human exposures were continuous. The GST metabolism rate in each
simulated human population was obtained by generating 1000 random samples from each population
(ages 0.5-80 years, males and females) for each exposure level, and calculating the average GST
metabolic rate for each sample.
Dose-response modeling and extrapolation
The multistage dose-response model was fit to the mouse tumor incidence and PBTK
model-derived internal dose data to derive mouse internal BMDio and BMDLio values associated
with 10% extra risk (Table 5-18). Different polynomial models and models dropping dose
groups starting with the highest dose group were compared based on adequacy of model fit as
assessed by overall y~ goodness of fit (p-value >0.10) and examination of residuals, particularly
in the region of the BMR (U.S. EPA, 2000c). Appendix E-2 provides details of the BMD
modeling results for the male. The mouse liver and lung tumor risk factors (extra risk per unit
internal dose) were calculated by dividing 0.1 by the mouse BMDLio for liver and lung tumors,
respectively.
Linear extrapolation from the internal BMDLio (0.1/BMDLio) was used to derive
inhalation risk factors for lung and liver tumors in male and female mice (Table 5-18). As
discussed in section 4.7, the linear low-dose extrapolation approach for agents with a mutagenic
mode of action was selected.
282

-------
8044
Table 5-18. BMD modeling results and tumor risk factors associated with 10% extra risk for liver and lung tumors in
male and female B6C3Fi mice exposed by inhalation to dichloromethane for 2 years, based on liver-specific GST
metabolism and whole body GST metabolism dose metrics
Tumor Risk Factor6
Internal dose

BMDS
goodness of fit
Mouse
Mouse
Allometric-
scaled human

Allometric-
metric3

modelb
/?-value
BMD10C
BMDL10C
BMDL10d
Scaling = 1.0
scaled
Liver-specific
Male, liver
MS (0,1)
0.40
913.9
544.4
77.8
1.84 x 10^
1.29 >
< 10~3

Male, lung
MS (0,1)
0.64
61.7
48.6
7.0
2.06 x 10-3
1.44 >
< 10~2

Female, liver
MS (0,2)
0.53
1224.1
659.7
94.2
1.52 x 10^
1.06 >
< 10~3

Female, lung
MS (0,1)
0.87
51.2
40.7
5.8
2.46 x 10-3
1.72 >
< 10~2
Whole body
Male, liver
MS (0,1)
0.40
38.7
23.1
3.3
--
3.03 >
< 10~2

Male, lung
MS (0,1)
0.66
13.1
10.3
1.5
--
6.80 >
< 10~2

Female, liver
MS (0,2)
0.53
51.9
28.0
4.0
--
2.50 >
< 10~2

Female, lung
MS (0,1)
0.88
10.8
8.6
1.2
--
8.14 >
< 10~2
aLiver specific dose units = mg dichloromethane metabolized via GST pathway per liter tissue per day; Whole-body dose units = mg
dichloromethane metabolized via GST pathway in lung and liver/kg-day)
bThe multistage (MS) model in EPA BMDS version 2.0 was fit to the mouse dose-response data shown in Table 5-17 using internal dose
metrics calculated with the mouse PBTK model. Numbers in parentheses indicate (1) the number of dose groups dropped in order to
obtain an adequate fit and (2) the degree polynomial of the model.
°BMDio and BMDLio refer to the BMD-model-predicted mouse internal dose and its 95% lower confidence limit, associated with a 10%
extra risk for the incidence of tumors.
dMouse BMDLio divided by (BWhllnKJBWnK,IKe)',2S= 7.
"Dichloromethane tumor risk factor (extra risk per unit internal dose) derived by dividing the BMR (0.1) by the mouse BMDLio and by the
allometric-scaled human BMDLi0„ for the scaling =1.0 and allometric-scaled risk factors, respectively.
8045
8046
8047
283

-------
8048
8049
8050
8051
8052
8053
8054
8055
8056
8057
8058
8059
8060
8061
8062
8063
8064
8065
8066
8067
8068
8069
8070
8071
8072
8073
8074
8075
8076
8077
8078
8079
8080
8081
8082
8083
8084
Application of allometric scaling factor
As discussed in section 5.4.1.4., the choice of a scaling factor is based on the question of
the role of specific metabolites, and particularly how these metabolites are transformed or
removed. If the key metabolite is established and is known to be sufficiently reactive to not
spread in systemic circulation, then it can be assumed that 1) the level of reactivity and rate of
clearance (i.e., disappearance due to local reactivity) for this metabolite, per volume tissue, is
equal in rodents and humans; and (2) risk is proportional to the long-term daily average
concentration of the metabolite. Under these assumptions, rodent internal BMDLio values based
on tissue-specific dichloromethane metabolism require no allometric scaling to account for
toxicodynamic differences and predict the corresponding level of human risk as a function of the
metric (i.e., the scaling factor in Figure 5-14 was equal to 1.0). (A single metabolite is
referenced, but the same argument holds in general for more than one metabolite). Under this
scenario and assumptions, humans and rodents with the same long-term daily average metabolite
formation per volume tissue (e.g., equal internal BMDLio) should both experience the same
long-term average concentration of the metabolite when the metabolite is highly reactive and
hence experience the same extra risk. However, some uncertainties remain concerning the
hypothesized role of a highly reactive metabolite in the carcinogenic effects seen with
dichloromethane. The active metabolite(s) have not been established, and data pertaining to the
reactivity or clearance rate of these metabolite(s) are lacking. For example, quantitative
measurements of adducts of interest or of the half life of relevant compounds in humans and in
mice are not available. To address these uncertainties, use of scaling factor that addresses the
possibility that the rate of clearance for the metabolite is limited by processes that are known to
scale allometrically, such as blood perfusion or enzyme activity, may be appropriate.. This case
would result in use of a mouse:human dose-rate scaling factor of (BWhuman/BWm0use)0'25 = 7 to
adjust the mouse-based BMDLio values downward. Using this internal dose metric (liver-
specific metabolism with allometric scaling), equivalent rodent and human internal BMDLio
values result in a human liver tumor risk factor (0.1/BMDLio) that is assumed equal to that for
the mouse, given a 70-year lifetime exposure. Another alternative that can be used is based on
an allometrically-scaled whole-body metabolism metric. In this case, less weight is given to the
evidence of site-specificity of the effects. As with the OSF derivations, the cancer toxicity
values derived using each of these metrics and scaling factors (i.e., liver-specific metabolism
with and without allometric-scaling and the whole-body metabolism metric) are presented in the
following tables. Considering the lack of data pertaining to clearance rates or the actual AUC of
the active carcinogenic metabolite(s) in mice and humans, the IUR recommended by the EPA are
based on the allometrically-scaled tissue-specific GST metabolism rate dose metric.
Calculation oflURs
284

-------
8085
8086
8087
8088
8089
8090
8091
8092
8093
8094
8095
8096
8097
8098
8099
8100
8101
8102
8103
8104
8105
8106
8107
8108
8109
8110
8111
8112
A probabilistic PBTK model for dichloromethane in humans adapted from David et al.
(2006) (see Appendix B) was used with Monte Carlo sampling to calculate distributions of
internal lung, liver, or blood doses associated with chronic unit inhalation (1 (J,g/m3) exposures.
The data on which the model is based indicate that relationship between exposure and internal
dose is linear at low doses. Parameters in the human PBTK model developed by David et al.
(2006) are distributions that incorporate information about dichloromethane toxicokinetic
variability and uncertainty among humans. Monte Carlo sampling was performed in which each
human model parameter was defined by a value randomly drawn from each respective parameter
distribution. The model was then executed by using the external unit exposure as input, and the
resulting human equivalent internal dose was recorded. This process was repeated for 10,000
iterations to generate a distribution of human internal doses.
The resulting distribution of IURs shown in Table 5-19 were derived by multiplying the
human internal dose tumor risk factor (in units of reciprocal internal dose) by the respective
distributions of human average daily internal dose resulting from a chronic unit inhalation
exposure of 1 |ig/m3 dichloromethane. Table 5-19 presents the analysis using the male data.
Analyses based on the female data produced very similar results, and are summarized in
Appendix F. Because adjustments for interindividual variability are not generally used or
recommended in cancer risk analysis, the mean slope factor was selected as the recommended
value to be used in deterministic risk assessments; other values at the upper end of the
distribution are also presented. As with the oral cancer slope factor derivation, the cancer IUR is
derived for a population composed entirely of carriers of the GST-T1 homozygous positive
genotype (the group that would be expected to be most sensitive to the carcinogenic effects of
dichloromethane), and a population reflecting the estimated frequency of GST-T1 genotypes in
the current U.S. population (20% GST-T1 7 , 48% GST-T1 , and 32% GST-T1+/+, the "mixed"
population). All simulations also included a distribution of CYP activity, based on data from
Lipscomb et al. (2003).
285

-------
8113
Table 5-19. IURs for dichloromethane based on PBTK model-derived internal liver and lung doses in B6C3Fi male mice exposed
via inhalation for 2 years, based on liver-specific GST metabolism and whole body metabolism dose metrics, by population
genotype
Distribution of human internal	„	... ^ .
, ,, , jt	, 3	Resulting candidate human
dichloromethane doses irom 1 ug/m	" t, , 3._i
d	IUR (jtg/m )
Internal dose metric
and scaling factor"
Population
genotypeb
Tumor
type
Human
tumor risk
factor0
Mea
95th
n percentile
99th
percentile
Mean
95th
percentile
99th
percentile
Tissue-specific,
GST-T1+/+
Liver
1.29
10~3
5.64 x 1
O^5 1.56
x 10~5
2.60
x 10~5
7.3 x 10~9
2.0 x
10"8
3.3 x
10"8
allometric-scaled
GST-T1+/+
Lung
1.44
10~2
3.31 x l
) ~ 8.55
x 10~7
1.34
x 10^
4.8 x K)
1.2 x
10"8
1.9 x
10"8

Mixed
Liver
1.29
10~3
2.62 x 1
0^ 8.65
x lO^5
1.45
x 10~5
3.4 x 1() 9
1.1 X
10^
1.9 x
10^

Mixed
Lung
1.44
10~2
1.81 x l
) " 5.67
x 10~7
9.84
x 10~7
2.6 x 10~9
8.2 x
10 9
1.4 x
10^
Tissue-specific,
GST-T1+/+
Liver
1.84
10"1
5.64 x 1
O^5 1.56
x 10~5
2.60
x 10~5
1.0 x 10"9
2.9 x
10 9
4.8 x
10~9
scaling =1.0
GST-T1+/+
Lung
2.06
10~3
3.31 x l
) ~ 8.55
x 10~7
1.34
X 10^
6.8 x 10~10
1.8 x
10 9
2.8 x
10~9

Mixed
Liver
1.84
10"1
2.62 x 1
O^5 8.65
X 10^
1.45
X 10"5
4.8 x 10
1.6 x
10 9
2.7 x
10~9

Mixed
Lung
2.06
10~3
1.81 x l
) ~ 5.67
X 10~7
9.84
X 10"7
3.7 x 10"10
1.2 x
10 9
2.0 x
10~9
Whole-body,
GST-T1+/+
Liver
3.03
10~2
1.53 x 1
) " 4.87
X 10~7
9.20
X 10~7
4.6 x l() 9
1.5 x
10^
2.8 x
10^
allometric-scaled
GST-T1+/+
Lung
6.80
10~2
1.53 x 1
) " 4.87
X 10~7
9.20
X 10"7
1.0 x 10~8
3.3 x
10^
6.3 x
10^

Mixed
Liver
3.03
10~2
8.76 x 1
0"8 3.20
X 10~7
6.76
X 10~7
2.7 x 10"9
9.7 x
10 9
2.1 x
10"8

Mixed
Lung
6.80
10~2
8.76 x 1
0"8 3.20
X 10~7
6.76
X 10"7
6.0 x 10~9
2.2 x
10"8
4.6 x
10"8
aTissue specific dose units = mg dichloromethane metabolized via GST pathway per liter tissue (liver or lung, respectively, for liver and lung
tumors) per day; Whole-body dose units = mg dichloromethane metabolized via GST pathway in lung and liver/kg-day.
bGST-Tl+/+ = homozygous, full enzyme activity;); mixed = population reflecting estimated frequency of genotypes in current U.S. population: 20% GST-T
48% GST-T 1 . and 32% GST-T1+/+ (Haber et al., 2002).
Dichloromethane tumor risk factor (extra risk per unit internal dose) derived by dividing the BMR (0.1) by the allometric-scaled human BMDL10 or by the
mouse BMDLio (from Table 5-18) for the allometric-scaled and scaling =1.0 risk factors, respectively.
dMean, 95th, and 99th percentile of the human PBTK model-derived probability distribution of daily average internal dichloromethane dose resulting from
chronic exposure to 1 |ig/m3 (0.00029 ppm).
"Derived by multiplying the dichloromethane tumor risk factor by the PBTK model-derived probabilistic internal doses from daily exposure to 1 (ig/m3
8114
286

-------
8115
8116
8117
8118
8119
8120
8121
8122
8123
8124
8125
8126
8127
8128
8129
8130
8131
8132
8133
8134
8135
8136
8137
8138
8139
8140
8141
8142
8143
8144
8145
8146
8147
8148
8149
8150
8151
8152
5.4.2.5. Cancer Inhalation Unit Risk
9	3 1	9	3 1
The recommended cancer IURs are 7 / 10 ((J,g/m ) and 5/10 ((J-g/m ) for the
development of liver and lung cancer, respectively, based on the mean for the GST-T1+/+
population (the group with the greatest presumed sensitivity). These values are based on male
B6C3Fi mice, using a tissue-specific GST metabolism dose metric, with allometric scaling
(Table 5-19). Risk estimates were slightly higher for liver tumors and essentially equivalent for
lung tumors in males compared to females (Appendix F), so the estimates for males were
selected for the candidate values.
Consideration of combined risk (summing risk across tumors)
With two significant tumor sites, focusing on the more sensitive response may
underestimate the overall cancer risk associated with exposure to this chemical. Following the
recommendations of the National Research Council (NRC, 1994) and the Guidelines for
Carcinogen Risk Assessment (U.S. EPA, 2005a), an upper bound on total risk was estimated in
order to gain some understanding of the total risk from multiple tumor sites in the selected data
set. Note that this estimate of overall risk describes the risk of developing either tumor type, not
just the risk of developing both simultaneously.
NRC (1994) stated that an approach based on counts of animals with one or more tumors
(or tumor-bearing animals) would tend to underestimate overall risk when tumor types occur
independently and that an approach based on combining the risk estimates from each separate
tumor type should be used. For dichloromethane, there is no reason to expect that the occurrence
of one tumor type depends on the incidence of the other, given the association of different dose
metrics with each tumor response. Therefore it appears reasonable to assume that the two tumor
types occur independently. However, simply summing upper limit risks may result in an
overestimation of overall of combined risk because of the statistical issues with respect to
summing variances of distributions. An additional challenge results from the use of different
internal dose metrics for different tumors, as is the case with the dose metrics based on tissue-
specific metabolism. Statistical methods based on a common metric can not be used with the
tissue-specific metabolism metric used in these derivations.
An alternative approach is to derive an upper bound on the combined risk estimates by
summing central tendency risks and calculating a pooled SD by using BMDio and BMDLio
values for liver and lung tumors. The SD associated with the IUR for each tumor site is
calculated as the difference between 95th percentiles of the distribution for upper bound and
maximum likelihood estimate IURs (based on either female or male mouse tumor risk factors),
divided by 1.645 (the relevant t statistic, assuming normal distributions of summed quantities).
Variances for each tumor site are the squares of the SDs. Pooled variance and SD are defined as
the sum of variances for lung and liver tumors and the square root of that sum, respectively.
Finally, the upper bound on the combined lung and liver cancer risk is determined by multiplying
287

-------
8153	the cumulative SD by 1.645 and adding it to the summed central tendency IURs. The
8154	calculations of these upper bound estimates for combined liver and lung tumor risks are shown in
8155	Table 5-20.
8156	Using this approach and the male mouse-derived risk factors, the combined human
8157	equivalent IUR values for both tumor types is 1 x 1CT8 (|ig/m3) 1 (rounded from l.lxl0 8) in
8158	the most sensitive (GST-T1+/+) population. This is the recommended inhalation cancer unit risk
8159	value to be used in deterministic risk assessments for chronic exposure to dichloromethane. The
8160	corresponding value for a population with the frequency distribution of GST-T1 genotypes
8161	currently found in the U.S. population is 6 x 10 9 (|ig/m3)
288

-------
8162
Table 5-20. Upper bound estimates of combined human IURs for liver and lung tumors resulting from lifetime
exposure to 1 jig/m3 dichloromethane, based on liver-specific GST metabolism and whole body metabolism dose
metrics, by population genotype
Upper bound on
Internal dose




Central
Variance of
Combined
combined tumor
metric and scaling
factor"
Population
genotypeb
Tumor site
Upper
bound IURC
tendency
IURd
tissue-specific
tumor risk6
tumor risk
SDf
risk8
(Ug/m3)1
Tissue-specific,
GST-T 1+/+
Liver
7.3 ;
< 10~9
4.3 ;
< 10~9
3.17 :
X 10~18




allometric-scaled

Lung
4.8 ;
< 10~9
3.8 ;
< 10~9
3.75 >
< 10~19






Liver or lung


8.1 ;
< 10~9


1.9
X 10~9
1.1 >
< 10~8

Mixed
Liver
3.4 :
< 10~9
2.0 :
< 10~9
6.84 >
< 10~19






Lung
2.6 :
< 10~9
2.1 ;
< 10~9
1.12 >
< 10~19






Liver or lung


4.1 :
< 10~9


8.9
X l(T10
5.5 :
< 10~9
Tissue-specific,
GST-T 1+/+
Liver
1.0 >
< 10~9
6.2 >
: 10-10
6.48 >
< io-20




scaling =1.0

Lung
6.8 x
: icr10
5.4 >
: 10-10
7.72 >
< 10~21

X lO"10




Liver or lung


1.2 :
< 10~9


2.7
1.6 :
< 10~9

Mixed
Liver
4.8 x
: icr10
2.9 >
: 10"10
1.40 >
< 10~20






Lung
3.7 x
: 1CT10
2.9 >
: 10"10
2.31 >
< 10~21






Liver or lung


5.8 >
: 10"10


1.3
X lO"10
7.9 x
: IO"10
Whole-body,
GST-T 1+/+
Liver
4.6 ;
< 10~9
2.8 :
< 10~9
1.29x
: l(T18




allometric-scaled

Lung
1.0 ;
< 10"8
8.2 :
< 10~9
1.84 >
< 10~18






Liver or lung


1.1 >
< 10"8


1.8
X 10~9
1.4 :
< 10~8

Mixed
Liver
2.7 :
< 10~9
1.6 :
< 10~9
4.23 >
< 10~19






Lung
6.0 :
< 10~9
4.7 :
< 10~9
6.04 >
< 10~19






Liver or lung


6.3 :
< 10~9


1.0
X 10~9
7.9 :
< 10~9
aTissue specific dose units = mg dichloromethane metabolized via GST pathway per liter tissue (liver or lung, respectively, for liver and
lung tumors) per day; Whole-body dose units = mg dichloromethane metabolized via GST pathway in lung and liver/kg-day.
bGST-Tl+/+ = homozygous, full enzyme activity;); mixed = population reflecting estimated frequency of genotypes in current U.S. population: 20%
GST-T . 48% GST-T1 . and 32% GST-T1+/+ (Haber et al., 2002).
Estimated at the human equivalent BMDLio (0. l/BMDLi0) (see Table 5-18).
Estimated at the human equivalent BMDi0 (0.1/BMD) (see Table 5-18).
"Calculated as the square of the difference of the upper bound and central tendency IURs divided by the t statistic, 1.645.
'Calculated as the square root of the sum of the variances for liver and lung tumors.
Calculated as the product of the cumulative tumor risk SD and the t statistic, 1.645, added to the sum of central tendency IURs.
8163
289

-------
8164
8165
8166
8167
8168
8169
8170
8171
8172
8173
8174
8175
8176
8177
8178
8179
8180
8181
8182
8183
8184
8185
8186
8187
8188
8189
8190
8191
8192
8193
8194
8195
8196
8197
8198
8199
5.4.2.6.	Comparative Derivation Based on Rat Mammary Tumor Data
Mammary gland tumor data from male and female F344 rats following an inhalation
exposure to dichloromethane were considered in development of a comparative IUR for
dichloromethane (Mennear et al., 1988; NTP, 1986). In both the male and female rats, there
were significant increases in the incidence of adenomas, fibroadenomas, or fibromas in or near
the mammary gland. These were characterized as benign tumors in the NTP report (NTP, 1986).
Increased numbers of benign mammary tumors per animal in exposed groups were also seen in
two studies of Sprague-Dawley rats (Nitschke et al., 1988a; Burek et al., 1984). An oral
(gavage) study in Sprague-Dawley rats reported an increased incidence of malignant mammary
tumors, mainly adenocarcinomas (8, 6, and 18% in the control, 100, and 500 mg/kg dose groups,
respectively), but the increase was not statistically significant. Data were not provided to allow
an analysis that accounts for differing mortality rates (Maltoni et al., 1988). There are
considerably more uncertainties regarding the interpretation of these data with respect to
carcinogenic risk compared with the data pertaining to liver and lung tumors. The trends were
driven in large part by benign tumors; adenocarcinomas and carcinomas were seen only in the
females, with incidences of 1, 2, 2, and 0 in the 0, 1000, 2000 and 4000 ppm exposure groups,
respectively. There are little data to guide the choice of relevant dose metric, and the
genotoxicity and mechanistic studies have not included mammary tissue. For these reasons, the
analysis and the calculation of the comparative IUR based on rat mammary tumor data are
presented in Appendix G. The IUR based on the female rat data was 1 x 10~7 (|ig/m3)
5.4.2.7.	Alternative Based on Administered Concentration
Another comparison that can be made is with an alternative IUR based on liver and lung
tumors in mice, using the external concentrations of dichloromethane in the mouse studies as
converted to HECs, and then applying this using BMD modeling to obtain the BMDLio and
resulting IUR. Mouse bioassay exposures were adjusted to HECs as follows:
•	Adjusting to continuous exposure: External concentrationadj = External
concentration x (6 hours/24 hours) x (5 days/7 days)
•	Concentrations in mg/m3 = concentrations in ppm x 84.93/24.45.
•	[Hb/gMHb/gjn = the ratio of blood:gas (air) partition coefficients in animals and
humans. Because the partition coefficient for mice (23.0) is higher than for humans
(9.7), a value of 1.0 was used, as per U.S. EPA (1994b) guidance.
Thus,
HECs = External concentration adj x [Hb/g]A/[Hb/g]H = External concentrationADj x 1
290

-------
8200
8201
8202
8203
8204
8205
8206
8207
8208
8209
8210
8211
8212
8213
8214
8215
8216
8217
8218
8219
8220
8221
8222
The HECs (mg/m3) for the 0, 2,000, and 4,000 ppm exposure groups were 0, 1,241, and
2,481 mg/m3, respectively. The BMD modeling and IURs derived from these values, in
conjunction with the liver and lung tumor data from Table 5-17 (NTP, 1986), are shown
in Table 5-21. The resulting IURs, based on the liver or lung tumors in the mouse, are
approximately one order of magnitude higher than the currently recommended value
obtained by using the mouse and human PBTK models.
Table 5-21. Inhalation units risks based on human BMDL10 values using
administered concentration for liver and lung tumors in B6C3Fi mice
exposed by inhalation to dichloromethane for 2 years
Sex,
tumor type
BMDS model3
1
goodness of fit
/?-value
BMD10b
BMDL10b
Inhalation
unit risk0
(jig/m3)"1
Male, liver
MS (0,1)
0.44
456.17
272.01
3.7 x 10~7
Male, lung
MS (0,1)
0.21
137.51
108.27
9.2 x 10~7
Female, liver
MS (0,2)
0.37
602.67
344.26
2.9 x 10~7
Female, lung
MS (0,1)
0.76
126.68
100.78
9.9 x 10~7
"The multistage (MS) model in EPA BMDS version 2.0 was fit to each of the four sets of mouse dose-response
data shown in Table 5-17. The HEC used in these models for the 0, 2,000,and 4,000 ppm exposure groups were
0, 1,241, and 2,481 mg/m3, respectively. Numbers in parentheses indicate (1) the number of dose groups
dropped in order to obtain an adequate fit, and (2) the lowest degree polynomial of the model showing an
adequate fit.
°BMDio and BMDLio refer to the BMD-model-predicted HECs (mg dichloromethane per cubic meter), and its
95% lower confidence limit, associated with a 10% extra risk for the incidence of tumors.
dIUR, (risk/mg-m3) = 0.1/human BMDLi0.
Sources: Mennear et al. (1988); NTP (1986).
The difference between the administered concentration methodology and PBTK-based
approaches depends on two key differences: the use of a dichloromethane-metabolite dose-
metric, rather than dichloromethane AUC, and the fact that the rate of dichloromethane
conversion to that metabolite is estimated in humans by using human data rather than default
allometric scaling (BW° 75). In addition, the administered concentration methodology does not
account in any way for the GST polymorphism and so might be considered as representing the
general/mixed-GST-genotype population rather than the +/+ subpopulation.
5.4.2.8. Previous IRIS Assessment: Cancer Inhalation Unit Risk
The IUR in the previous IRIS assessment was determined from the combined incidence
of liver and lung adenomas and carcinomas in B6C3Fi mice exposed to dichloromethane for
2 years by NTP (1986). A value of 4.7 x 10"7 ((j,g/m3)-1 was derived by the application of a
modified version of the PBTK model of Andersen et al. (1987), which incorporated the
291

-------
8223
8224
8225
8226
8227
8228
8229
8230
8231
8232
8233
8234
8235
8236
8237
8238
pharmacokinetics and metabolism of dichloromethane. Internal dose estimates, based on
dichloromethane metabolism via the GST pathway, were used and corrected for differences in
interspecies sensitivity by applying an interspecies scaling factor of 12.7, which was based on
dose equivalence adjusted to BW to the 2/3 power, to the human risks (Rhomberg, 1995; U.S.
EPA, 1987a).
5.4.2.8. Comparison of Cancer Inhalation Unit Risk Using Different Methodologies
In this assessment, cancer IURs derived by using different dose metrics and assumptions
were examined, as summarized in Table 5-22. The recommended IUR value of 1 x 10"8
(^ig/m3)-1 is based on a tissue-specific GST-internal dose metric, with allometric scaling, because
of the evidence for the involvement of highly reactive metabolites formed via the GST pathway.
The value derived specifically for the GST-T1+/+ population is recommended to provide
protection for the population that is hypothesized to be most sensitive to the carcinogenic effect.
The values based on the GST-T1+/+ group are approximately two to fivefold higher than those for
the full population, for all dose metrics used in this assessment. Within a genotype population,
the values of the IUR among the various dose metrics vary by about one to two orders of
magnitude.
292

-------
8239
Table 5-22. Comparison of IURs derived by using various assumptions and metrics




Scaling
IURb
Source
Population"
Dose metric
Species, sex,
Tumor type
factor
(jig/m3)"1
(Table)
GST-T1+/+
Tissue-specific GST-metabolism rate0
Mouse, male
Liver and lung
7.0
1.1 x 10"8
Table 5-20
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Liver
7.0
7.3
x 10~9
Table 5-19
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Lung
7.0
4.8
x 10~9
Table 5-19
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Liver and lung
1.0
1.6
X 10~9
Table 5-20
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Liver
1.0
1.0
X 10~9
Table 5-19
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Lung
1.0
6.8
X 10~10
Table 5-19
GST-T1+/+
Whole-body GST metabolism rate
Mouse, male
Liver and lung
7.0
1.4
X 10"8
Table 5-20
GST-T1+/+
Whole-body GST metabolism rate
Mouse, male
Liver
7.0
4.6
X 10~9
Table 5-19
GST-T1+/+
Whole-body GST metabolism rate
Mouse, male
Lung
7.0
1.0
X 10"8
Table 5-19
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Liver and lung
7.0
5.5
X 10~9
Table 5-20
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Liver
7.0
3.4
X 10~9
Table 5-19
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Lung
7.0
2.6
X 10~9
Table 5-19
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Liver and lung
1.0
7.9
X 10~10
Table 5-20
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Liver
1.0
4.8
X 10~10
Table 5-19
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Lung
1.0
3.7
X 10~1CI
Table 5-19
Mixed
Whole-body GST metabolism rate
Mouse, male
Liver and lung
7.0
7.9
X 10~9
Table 5-20
Mixed
Whole-body GST metabolism rate
Mouse, male
Liver
7.0
2.7
X 10~9
Table 5-19
Mixed
Whole-body GST metabolism rate
Mouse, male
Lung
7.0
6.0
X 10~9
Table 5-19

Administered concentration (HEC)
Mouse, male
Liver

3.7
X 10~7
Table 5-21

Administered concentration (HEC)
Mouse, male
Lung

9.2
X 10~7
Table 5-21

1995 IRIS assessment0
Mouse, male
Liver and lung
12.7
4.7
X 10"7

aGST-Tl+/+ = homozygous, full enzyme activity; Mixed = genotypes based on a population reflecting the estimated frequency of genotypes in the current
U.S. population: 20% GST-T1 . 48% GST-T1 . and 32% GST-T1+/+ (Haber et al., 2002).
' 'Based on mean value of the derived distributions
Bolded value is the basis for the recommended IUR of lx 10~8jig/m3 per mg/kg-day.
293

-------
8240
8241
8242
8243
8244
8245
8246
8247
8248
8249
8250
8251
8252
8253
8254
8255
8256
8257
8258
8259
8260
8261
8262
8263
8264
8265
8266
8267
8268
8269
8270
8271
8272
8273
8274
8275
8276
8277
5.4.3. Differences Between Current Assessment and Previous IRIS PBTK-based
Assessment
To better understand the changes in assessment risk predictions between previous EPA
evaluations and the current assessment, the differences in PBTK model parameters for the
B6C3Fi mouse were evaluated. Values that differed significantly between the model version
used previously and that of Marino et al. (2006), along with derived group parameters that lend
further insight, are shown in Table 5-23.
The tissue:air partition coefficients in Table 5-23 show that, while several of the blood:air
partition coefficients appear to differ significantly between the two models, the corresponding
blood:air partition coefficient does not, and it is the latter that can be more indicative of long-
term equilibration between the tissue (tissue group) and air. Thus, the partition coefficients, the
ones that most significantly differ are the blood:air and liver:air partition coefficients that are,
respectively, 2.8- and 2.6-fold higher in the current version. The increased blood:air partition
coefficient results in a tendency for simulated blood concentrations to rise more quickly and
reach higher values, other parameters being equal. The significantly increased QCC and VPR
contribute even more to this difference, resulting in an even faster rise to steady state during
inhalation exposure simulations, though also more rapid delivery to the liver (decreasing blood-
flow limitation of hepatic metabolism) and more rapid exhalation. The increased liver:air
partition coefficient leads to higher predicted liver concentrations (again, other parameters being
equal) and hence higher rates of metabolism.
For metabolism, a much reduced oxidative metabolism is seen, which at low doses
depends on Vmaxc/Km. The revised hepatic metabolism is over 40% lower and the total of lung +
liver metabolism is 50% lower than previously used. This lower rate of metabolism means that
far less of parent dichloromethane will be removed through metabolism, and hence predicted
blood concentrations will be higher still, relative to the impact of changes in partition coefficient,
QCC, and VPR, as noted above.
The result of having higher predicted blood and liver dichloromethane concentrations is
that, while the GSH-pathway metabolic constant, kfC, is virtually the same for the mouse liver in
both cases, the much higher concentration of dichloromethane available will lead to a much
higher predicted rate of metabolism via this pathway. For the lung, since the lung:liver ratio
(A2) is 43% higher in the model of Marino et al. (2006), the relative increase will be even
greater.
Because the revised rate of GST metabolism in mice was estimated by using data with
CYP2E1 inhibited by a suicide inhibitor, there is considerable confidence in the relative rate of
metabolism by these two pathways, and the GST pathway in particular. The partition
coefficients used by Marino et al. (2006) are as measured by Clewell et al. (1993) and expected
to be at least as reliable as those used in the 1995 assessment. Considering that the revised
PBTK model does an excellent job of reproducing closed-chamber gas uptake data that were not
294

-------
8278	available for calibration of the 1987 model, as well blood concentrations after intravenous
8279	injection, we can have fairly high confidence in its predictions.
8280	The net result of these model changes is that, under mouse bioassay conditions, the
8281	predicted dose metrics for liver and lung cancer, i.e., GST-mediated metabolism, are higher than
8282	those obtained with the previous model, resulting in a lower risk estimated per unit of dose.
8283
Table 5-23. Comparison of key B6C3Fi mouse parameters differing
between prior and current PBTK model application
Marino et al. (2006); mean	U.S. EPA
Parameter"	values as applied (posterior) (1988b, 1987a, b)
Partition coefficients
PB blood/air
23
8.29
PF fat/blood
5.1
14.5
PF • PB (fat/blood) • (blood/air) = fat/air
117.3
120.2
PL liver/blood
1.6
1.71
PL PB (liver/blood) (blood/air) = liver/air
36.8
14.2
PLu lung (tissue)/blood
0.46
1.71
PLu PB (lung/blood) (blood/air) = lung/air
10.6
14.2
PR rapidly perfused/blood
0.52
1.71
PRPB rapidly perfused/air
12.0
14.2
PS slowly perfused/blood
0.44
0.96
PS PB slowly perfused/air
10.1
7.96
Flow rates


QCC cardiac output (L/hour/kg°74)
24.2
14.3
VPR ventilation:perfusion ratio
1.45
1.0
Metabolism parameters


Vmaxc maximum CYP metabolic rate (mg/hour/kg°7)
9.27
11.1
Km CYP affinity (mg/L)
0.574
0.396
VmaxC/Km (L/hour/kg07)
16.1
28
Al ratio of lung Vmaxc to liver Vmaxc
0.207
0.416
Total lung + liver Vmaxc/Km
19.5
39.7
kfC first-order GST metabolic rate constant
1.41
1.46
(kg03/hr)
0.196
0.137
A2 ratio of lung kfC to liver kfC
1.69
1.66
Total lung + liver kfC
Parameters not listed differed by less than 10% between versions. See Table 3 -5 and associated text for details.
8284
8285
8286	The other piece of the PBTK-based risk estimation is the human model. In updating the
8287	parameter estimates for the human model (see Appendix B for details), the oxidative metabolism
8288	Vmaxc/Km approximately doubled, which leads to lower predicted blood concentrations of
8289	dichloromethane available for metabolism by GST. In addition, the liver GST was reduced by
8290	almost 60%, and the lung:liver GST ratio decreased by almost fivefold, for a net change in lung
295

-------
8291
8292
8293
8294
8295
8296
8297
8298
8299
8300
8301
8302
8303
8304
8305
8306
8307
8308
8309
8310
8311
8312
8313
8314
8315
8316
8317
8318
8319
8320
8321
8322
8323
8324
8325
8326
8327
8328
GST of over 90%. Given the larger human data set available to David et al. (2006) and the
sophisticated Bayesian analysis used to recalibrate the model, the expectation is that these values
are more reliable than the values used in the 1995 IRIS assessment.
Since actual rates of metabolism at a given exposure level also depend on respiration rate
and blood flows, these changes in metabolic parameters do not completely determine the relative
(predicted) dosimetry. But the difference in cancer risk predictions between the current and
previous assessments is primarily explained by the overall prediction of higher GST-mediated
dosimetry in the mouse (at bioassay conditions) and lower human GST metabolism (due in part
to greater predicted clearance of dichloromethane by oxidative metabolism). In addition to these
changes in PBTK parameters, the reduction of scaling factor from 12.7 to 7 is a significant factor
in the overall change from the previous assessments.
5.4.4. Application of Age-Dependent Adjustment Factors (ADAFs)
The available dichloromethane studies do not include an evaluation of early-life
susceptibility to dichloromethane cancer risk. In the absence of this type of data and if a
chemical follows a mutagenic mode of action for carcinogenicity, like dichloromethane, the
Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to Carcinogens
(U.S. EPA, 2005b) recommends that ADAFs be applied to the cancer values. Since the OSF of 1
x 10 3 (mg/kg-day) 1 and the IUR of 1 x 10 8 (|ig/m3) 1 were calculated from adult
dichloromethane exposures, early-life cancer susceptibility has not been accounted for in these
values and ADAFs need to be applied in combination with exposure informations when
estimating cancer risks that include early-life exposures. Sample calculations that incorporate
ADAFs into the cancer risks are presented in subsequent sections. Additional examples of
evaluations of cancer risks incorporating early-life exposure are provided in section 6 of the
Supplemental Guidance (U.S. EPA, 2005b).
In the Supplemental Guidance (U.S. EPA, 2005b), ADAFs are established for three age
groups. An ADAF of 10 is applied for age groups less than 2 years, 3 is applied for ages 2 to
<16 years, and 1 is applied for 16 years and above (U.S. EPA, 2005b). The 10- and 3-fold
adjustments in cancer values are combined with age-specific exposure estimates, when early-life
exposure considerations need to be included in cancer risk estimates. The most current
information on usage of ADAFs can be found at www.epa.eov/cancerguidelines. For estimation
of risk, the Supplemental Guidance (U.S. EPA, 2005b) recommends obtaining and using age-
specific values for exposure and cancer potency. In the absence of age-specific values cancer
potency values, as is true for dichloromethane, age-specific cancer values are estimated by using
the appropriate ADAFs. Using this process, a cancer risk is derived for each age group. The
risks are summed across the age groups to get the total cancer risk for the age-exposure period of
interest.
296

-------
8329
8330
8331
8332
8333
8334
8335
8336
8337
8338
8339
8340
8341
8342
8343
8344
8345
8346
8347
8348
8349
8350
8351
8352
8353
8354
8355
8356
5.4.4.1. Application of ADAFs in Oral Exposure Scenarios
Sample calculations incorporating the use of ADAFs are presented for three exposure
duration scenarios. These scenarios include full lifetime exposure (assuming a 70-year lifespan),
and two 30-year exposures from ages 0-30 and ages 20-50. A constant dichloromethane
exposure of 1 mg/kg-day was assumed for each scenario.
Table 5-24 lists the four factors (ADAFs, OSF, assumed exposure, and duration
adjustment) that are needed to calculate the partial cancer risk based on the early age-specific
group. The partial cancer risk for each age group is the product of the four factors in columns 2-
5. Therefore, the partial cancer risk following daily dichloromethane oral exposure in the age
group 0 to <2 years is the product of the values in columns 2-5 or 10 x (1 x 10~3) x l x 2/70 =
2.9 x 10 4 The partial risks that are listed in the last column of Table 5-24 are added together to
get the total risk. Thus, a 70-year (lifetime) risk estimate for continuous exposure to 1 mg/kg-
day dichloromethane is 1.7 x 10 3 per mg/kg-day, which is adjusted for early-life susceptibility
and assumes a 70-year lifetime and constant exposure across age groups.
Table 5-24. Application of ADAFs to dichloromethane cancer risk
following a lifetime (70-year) oral exposure
Age group
(years)
ADAF
Unit risk
(per mg/kg-
day)
Exposure
concentration
(mg/kg-day)
Duration
adjustment
Partial risk
0 - <2
10
1 x 1(T3
1
2 years/
70 years
2.9 x 10~4
2-<16
3
1 x 1(T3
1
14 years/
70 years
6.0 x 10~4
>16
1
1 x 1(T3
1
54 years/
70 years
7.7 x 10~4




Total risk
1.7 x 10"3
In calculating the cancer risk for a 30-year constant exposure to dichloromethane at an
exposure level of 1 mg/kg-day from ages 0-30, the duration adjustments would be 2/70, 14/70,
and 14/70 and the partial risks for the three age groups would be 2.9 x 10 4, 6.0 x 10 4, and 2.0 x
10 4, which would result in a total risk estimate of 1.1 x 10 3.
In calculating the cancer risk for a 30-year constant exposure to dichloromethane at an
exposure level of 1 mg/kg-day from ages 20-50, the duration adjustments would be 0/70, 0/70,
and 30/70. The partial risks for the three groups are 0, 0, and 4.3 x 10 4, which would result in a
total risk estimate of 4.3 x 10 4,
5.4.4.2. Application of ADAFs in Inhalation Exposure Scenarios
297

-------
8357
8358
8359
8360
8361
8362
8363
8364
8365
8366
8367
8368
8369
8370
8371
8372
8373
8374
8375
8376
8377
8378
8379
8380
8381
8382
8383
8384
8385
8386
Sample calculations incorporating the use of ADAFs are presented for three exposure
duration scenarios involving inhalation exposure. These scenarios include full lifetime exposure
(assuming a 70-year lifespan) and two 30-year exposures from ages 0-30 and ages 20-50. A
constant dichloromethane inhalation exposure of 1 (J,g/m3 was assumed for each scenario.
Similar to the oral exposure scenarios presented in section 5.4.4.1, Table 5-25 lists the
four factors (ADAFs, unit risk, assumed exposure, and duration adjustment) that are needed to
calculate the partial cancer risk based on the early age-specific group. The partial cancer risk for
each age group is the product of the four factors in columns 2-5. Therefore, the partial cancer
risk following daily dichloromethane inhalation exposure in the age group 0 to <2 years is the
product of the values in columns 2-5 or 10 x (1 x 10~8) x l x 2/70 = 2.9 x 10~9. The partial
risks that are listed in the last column of Table 5-25 are added together to get the total risk. Thus,
a 70-year (lifetime) risk estimate for continuous exposure to 1 (J,g/m3 dichloromethane is 1.8 x
10 8 per (J,g/m3, which is adjusted for early-life susceptibility and assumes a 70-year lifetime and
constant exposure across age groups.
Table 5-25. Application of ADAFs to dichloromethane cancer risk
following a lifetime (70-year) inhalation exposure


Unit risk
Exposure


Age group

(per mg/kg-
concentration
Duration

(years)
ADAF
day)
(mg/kg-day)
adjustment
Partial risk
0 - <2
10
1 x KT8
1
2 years/
70 years
2.9 x 10~9
2-<16
3
1 x KT8
1
14 years/
70 years
6.0 x 10~9
>16
1
1 x KT8
1
54 years/
70 years
Total risk
7.7 x 10~9
1.7 x 10"8
In calculating the cancer risk for a 30-year constant exposure to dichloromethane at a
level of 1 |ig/m3 from ages 0-30, the duration adjustments would be 2/70, 14/70, and 14/70, and
the partial risks for the three age groups are 2.9 x 10 9, 6.0 x 10 9, and 2.0 x 10 9, These partial
risks result in a total risk estimate of 1.1 x 10 8,
In calculating the cancer risk for a 30-year constant exposure to dichloromethane at a
level of 1 |ig/m3 from ages 20-50, the duration adjustments would be 0/70, 0/70, and 30/70, and
the partial risks for the three age groups are 0, 0, and 4.3 x 10 9, resulting in a total risk estimate
of 4.3 x 10~9.
5.4.5. Uncertainties in Cancer Risk Values
The derivation of cancer risk values often involves a number of uncertainties in the
extrapolation of dose-response data in animals to cancer risks in human populations. Several
types of uncertainty have been quantitatively integrated into the derivation of the recommended
298

-------
8387
8388
8389
8390
OSFs and IURs for dichloromethane, while others are qualitatively considered. Table 5-26 and
the ensuing discussion summarize the principal uncertainties identified, their possible effects on
the cancer risk values, decisions made in the derivations, and justifications for the decisions.
Table 5-26. Summary of uncertainty in the derivation of cancer risk
values for dichloromethane
Consideration/
approach
Impact on cancer
risk value
Decision
Justification
Selection of
data set
Selection of
target organ
Selection of
extrapolation
approach
Selection of
dose metric
Selection of an
alternative data set or
target organ could
change the
recommended cancer
risk values.
Selection of a target
organ could change
the recommended
cancer risk values.
Selection of
extrapolation
approach could
change the
recommended cancer
risk values.
Selection of dose
metric could change
the recommended
cancer risk values.
Select Serota et al.
(1986b) and NTP
(1986) as principal
studies to derive
recommended liver
and lung cancer risks
for humans from
responses in mice.
Examine cancer risk
values based on
alternative tumor
responses (mammary
gland tumors in rats);
identification of
potential brain cancer
risk as a data gap.
Examine cancer risk
values based on
alternative approaches.
Evidence of GST
involvement supports
focus on this pathway.
Cancer risk estimates
based on alternative
(tissue-specific versus
whole-body) metrics
examined.
The NTP (1986) inhalation bioassay with
mice provides the strongest cancer
responses (liver and lung tumors) and the
best dose-response data in the animal
database. The Serota et al. (1986b) mouse
drinking water study provides the best oral
dose-response data for liver tumors.
Dichloromethane carcinogenicity appears to
be mediated by a metabolic pathway that is
also present in humans (i.e., the GST
pathway). In combination with the animal
results, epidemiological studies provide
evidence of increased risks for liver and
biliary duct tract cancer but are limited by a
number of factors discussed in sections
4.1.3.6 and 4.1.3.7.
Inhalation cancer risk values based on
mammary tumors in rats are about one
order of magnitude higher than risk values
based on liver or lung tumors in mice, but
the evidence for mammary gland tumors
from dichloromethane exposure is less
consistent than evidence for liver and lung
tumors.
Oral cancer risk values based on route-to-
route extrapolation from the NTP (1986)
inhalation mouse bioassay were about one
order of magnitude lower than values based
on liver tumors in orally exposed mice
(Serota et al., 1986b) (see Table 5-16) but
are inherently less certain than the values
based on oral exposure due to the influence
of route of exposure on toxicokinetics.
Inhalation and oral liver cancer risk values
derived using a tissue-specific GST
metabolism dose metric were slightly
higher than values derived using a whole-
body GST metabolism dose metric; for lung
tumors, the reverse pattern is seen. The
values based on liver or lung tumors using
the tissue-specific GST metabolism are
recommended based on the evidence of site
locality of effects
299

-------
8391
8392
8393
8394
8395
8396
8397
8398
8399
8400
8401
8402
8403
Table 5-26. Summary of uncertainty in the derivation of cancer risk
values for dichloromethane
Consideration/
approach
Impact on cancer
risk value
Decision
Justification
Dose-response
modeling
Low-dose
extrapolation
Interspecies
extrapolation of
dosimetry and
risk
Sensitive
subpopulations
Human risk values
could increase or
decrease, depending
on fits of alternative
models
Human risk values
would be expected to
decrease with the
application of
nonlinear tumor
responses in low-
dose regions of dose-
response curves.
Alternative values for
PBTK model
parameters and cross-
species scaling factor
could increase or
decrease human
cancer risk values.
Differences in CYP
and GST metabolic
rates could change
cancer risk values.
Use multistage dose-
response model to
derive a BMD.
Use linear
extrapolation of risk in
low-dose region.
Use PBTK model and
allometric scaling for
the primary dose
metric.
CYP variability
incorporated in the
PBTK model; separate
risk estimates
generated for the
presumed most
sensitive (GST-T1+/+)
genotype
The multistage model has biological
support and is the model most consistently
used in EPA cancer assessments.
Linear extrapolation from the human tumor
risk factors was used to derive cancer risk
values for oral and inhalation exposures.
The linear low-dose extrapolation approach
for agents with a mutagenic mode of action
was selected.
Application of rodent and human PBTK
models reduced uncertainty on cancer risk
values due to interspecies differences in
toxicokinetics. Examination of impact of
different values for key parameters in
human model, and sensitivity analysis of
rodent PBTK model parameters identified
influential metabolic parameters for which
little or no experimental data exist (see
Interspecies Extrapolation of Dosimetry of
Risk section, below).
No data are available to determine the range
of human toxicodynamic
variability/sensitivity, including whether
children are more sensitive than adults. The
toxicokinetic effect of the GST-T1
polymorphism is included in the human
PBTK model, as are other sources of
variability in GST and CYP metabolic
parameters.
Data selections for derivation of IUR and OSF
The database of animal bioassays identifies the liver and lung as the most sensitive target
organs for dichloromethane-induced tumor development. These effects demonstrate a dose-
response relationship in mice exposed orally (liver only) or by inhalation (liver and lung).
Statistically significant increases in benign mammary gland tumors were observed in one study
of F344 rats exposed by inhalation to 2,000 or 4,000 ppm (Mennear et al., 1988; NTP, 1986),
and evidence for a tumorigenic mammary gland response in Sprague-Dawley rats was limited to
increased numbers of benign mammary tumors per animal at levels of 50-500 ppm (Nitschke et
al., 1988a) or 500-3,500 ppm (Burek et al., 1984). An oral (gavage) study in female Sprague-
Dawley rats reported an increased incidence of malignant mammary tumors, mainly
adenocarcinomas (8, 6, and 18% in the control, 100, and 500 mg/kg dose groups, respectively),
300

-------
8404
8405
8406
8407
8408
8409
8410
8411
8412
8413
8414
8415
8416
8417
8418
8419
8420
8421
8422
8423
8424
8425
8426
8427
8428
8429
8430
8431
8432
8433
8434
8435
8436
8437
8438
8439
8440
8441
but the increase was not statistically significant. Data were not provided to allow an analysis that
accounts for differing mortality rates (Maltoni et al., 1988). The toxicokinetic or mechanistic
events that might lead to mammary gland tumor development in rats are unknown, although
CYP2E1 (El-Rayes et al., 2003; Hellmold et al., 1998) and GST-T1 expression has been detected
in human mammary tissue (Lehmann and Wagner, 2008). Rare CNS tumors were observed in
one study in rats, a study spanning a relatively low range of exposures (0-500 ppm). These
cancers were not seen in two other studies (NTP, 1986; Burek et al., 1984) in rats, both involving
higher doses (1,000-4,000 ppm), or in a similar high-dose study (NTP, 1986) in mice. The
relative rarity of the tumors precludes the use of the low-dose exposure study in a quantitative
dose-response assessment. The in vivo genotoxicity and mechanistic data in rodents provide a
detailed sequence of steps from generation of reactive metabolites to mutagenic effects, such as
DNA-protein cross-links and DNA strand breaks. Further, the toxicokinetic pathways implicated
in production of the putative carcinogenic metabolites in animals also exist in humans. Thus,
there is high confidence that the dose-response data for liver and lung cancer in mice represents
the best data currently available for derivation of human cancer risks. A more complete
understanding of the carcinogenic potential of dichloromethane would be achieved by addressing
data gaps identified with respect to issues regarding potential risk and mechanisms relating to
brain cancer and mammary tumors. The available epidemiologic studies provide some evidence
of an association between dichloromethane and brain cancer (see Section 4.1.3.7.1). The
available epidemiologic studies do not provide an adequate basis for the evaluation of the role of
dichloromethane in breast cancer because there are currently no cohort studies with adequate
statistical power and no case-control studies with adequate exposure methodology to examine
this relationship (see section 4.1.3.7.6)
Target organ
The liver and lung tumor incidence from chronic exposure biassays provide clear
evidence of the carcinogenic potential of dichloromethane exposure. The biassays are supported
by a substantial literature of genotoxicity and mechanistic studies (summarized in section 4.5).
The evidence for mammary gland tumors from dichloromethane exposure is based primarily on
observations of benign tumors in rats with inhalation exposure (NTP, 1986). Derivation of
cancer potenciy values based on these data are presented in Appendix G. The potential brain
cancer risk, suggested by a limited number of these relatively rare tumors in both animal and
human studies, is identified as a data gap which would benefit from additional research.
Extrapolation approach
A route-to-route extrapolation from the NTP (1986) inhalation mouse bioassay was used
to develop an oral cancer slope value. This value is inherently less certain than the values based
on oral exposure due to the influence of route of exposure on toxicokinetics.
301

-------
8442
8443
8444
8445
8446
8447
8448
8449
8450
8451
8452
8453
8454
8455
8456
8457
8458
8459
8460
8461
8462
8463
8464
8465
8466
8467
8468
8469
8470
8471
8472
8473
8474
8475
8476
8477
8478
Dose Metric
There is considerable data supporting the role of GST-related metabolism of
dichloromethane in carcinogenicity, as described in sections 4.5.1 and 4.7. Pretreatment of mice
with buthionine sulphoximine, a GSH depletor, caused a decrease, to levels seen in controls, in
the amount of DNA damage detected immediately after in vivo exposure in liver and lung tissue
(Graves et al., 1995). Although the results of Landi et al. (2003) indicate that GST activity is not
needed for the observation of DNA damage by the comet assay from some trihalomethanes (e.g.,
bromodichloromethane), the results for dichloromethane were much weaker and of uncertain
significance.
Dose-response modeling
Because of the adequacy of the fit of the multistage model to the data, little modeling
uncertainty would be expected to be introduced by the choice of this model. Application of the
multistage model allowed for estimation of a point of departure in the lower region of exposure
for observable cancer effects.
Low-dose extrapolation
The mode of action is a key consideration in determining how risks should be estimated
for low-dose exposure. The in vitro and in vivo genotoxicity data suggest that mutagenicity is
the most plausible mode of action, although key mutagenic events in the development of liver or
lung tumors have not been identified. No data are available that provide an adequate rationale
for choosing a nonlinear dose response in the low-dose region. Because a mutagenic mode of
action is most plausible, a linear-low-dose extrapolation approach was used to estimate OSFs and
IURs.
Interspecies extrapolation of dosimetry and risk
Target organ dosimetry for neoplastic mouse responses and estimation of equivalent
internal human doses were accomplished using PBTK models for dichloromethane in mice and
humans. Uncertainty in the ability of the PBTK models to estimate animal and human internal
doses from lifetime bioassay low-level environmental exposures may affect the confidence in the
cancer risk extrapolated from animal data. Uncertainties in the mouse and human model
parameter values were integrated quantitatively into parameter estimation by utilizing
hierarchical Bayesian methods to calibrate the models at the population level (David et al., 2006;
Marino et al., 2006). The use of Monte Carlo sampling to define human model parameter
distributions allowed for derivation of human distributions of dosimetry and cancer risk,
providing for bounds on the recommended risk values.
302

-------
8479
8480
8481
8482
8483
8484
8485
8486
8487
8488
8489
8490
8491
8492
8493
8494
8495
8496
8497
8498
8499
8500
8501
8502
8503
8504
8505
8506
8507
8508
8509
8510
8511
8512
8513
8514
8515
8516
A detailed discussion of PBTK model structure (CYP rate equation) and parameter
uncertainties is provided in Section 5.3. While the structure and equations used in the existing
model have been described in multiple peer-review publications over the past two decades, there
are discrepancies between dichloromethane kinetics observed in vitro and the model parameters
obtained from in vivo data. However, an alternative (dual-binding-site) CYP metabolic equation
appears to resolve these discrepancies. Integration of the alternate rate equation into the PBTK
modeling, and then quantitative risk assessment, will likely require several years of further
research, and hence is beyond the scope of the current assessment. Since the GST activity in the
current model is within a factor of three of that measured in vitro (when both are evaluated on a
per gram of liver basis), the impact of that model uncertainty is also expected to be no more than
a factor of three. Sensitivity to the human PBTK parameter distributions was evaluated by
rescaling the parameters to the mean values obtained by David et al. (2006) for a specific data set
(DiVincenzo and Kaplan, 1981) for which the GST activity was close to a numerical average of
those obtained across individual data sets. When this was done, the upper bound estimates on
GST dosimetry (for low, fixed inhalation or oral exposures) in the GST-T1 +/+ subpopulation
increased by over an order of magnitude, as did the estimate of the mean activity for an
inhalation exposure, although the estimated mean GST activity for an oral exposure only
increased about two-fold. So while correspondence of the in vivo GST activity with that
measured in vitro suggests a lower degree of quantitative uncertainty, it is possible that revision
of the PBTK model could have a larger impact. The ultimate impact will depend on how
revisions effect model predictions for both the animal and the human. If the predicted GST
metabolism per unit exposure increases in both mice and humans by a similar factor, there will
be little impact on the risk estimate. But if the GST activity predicted in the mouse is decreased
by a factor of 3, while that in the human is increased by a factor of 3, for example, then the net
impact would be an increase of 9-fold in human risk estimates.
The PBTK animal models were utilized deterministically; i.e., the single-value parameter
estimates for the rat PBTK model were used for rat dosimetry simulations and the mean
parameter estimates from the Bayesian analysis of Marino et al. (2006) were used for the mouse
dosimetry simulations. To assess the effect of using point estimates of parameter values for
calculation of rodent dosimetry, a sensitivity analysis was performed to identify model
parameters most influential on the predictions of dose metrics used to estimate oral and
inhalation cancer risks. As was described in the RfD and RfC sensitivity analysis calculation,
this procedure used a univariate analysis in which the value of an individual model parameter
was perturbed by an amount (A), in the forward and reverse direction (i.e., an increase and
decrease from the nominal value), and the change in the output variable was determined. Results
are for the effects of a perturbation of ±1% from the nominal value of each parameter on the
output values at the end of a minimum of 10,000 simulated hours. This time was chosen to
achieve a stable daily value of the dose metric, given that the simulated bioassay exposures did
303

-------
8517
8518
8519
8520
8521
8522
8523
8524
8525
8526
8527
8528
8529
8530
8531
8532
8533
8534
8535
8536
8537
8538
8539
8540
8541
8542
8543
8544
not include weekend exposures. The exposure conditions represented the lowest bioassay
exposure, resulting in significant increases in the critical effect. For inhalation exposures in
mice, the blood:air partition coefficient, followed closely by the first-order GST-mediated
metabolism rate (kfC), had the greatest impact on the dose metric for liver cancer (mg
dichloromethane metabolized via GST pathway per liter liver per day) (Figure 5-17). For
drinking water exposures in mice, the kfC, followed by the CYP-mediated maximum reaction
velocity (Vmaxc), affected the liver cancer dose metric to the greatest extent (Figure 5-18). For
mice inhaling dichloromethane, the lung cancer dose metric (mg dichloromethane metabolized
via GST pathways per liter lung per day), like the liver cancer metric, was highly affected by the
kfC and the blood:air partition coefficient (Figure 5-19). However, since GST-mediated lung
metabolism is calculated as a constant fraction of the liver metabolism rate (A2 x kfC), the lung
cancer dose metric was most sensitive to the proportional yield of liver GST-mediated metabolic
activity attributed to the lung. The blood:air partition coefficient was experimentally determined,
lending high confidence to its value. Values for the three metabolic parameters were determined
by computational optimization against data sets not directly measuring dichloromethane or its
metabolites in the target/metabolizing tissues. It is uncertain how alternative values for these
three parameters would affect the estimation of animal BMDL10 values and, ultimately, the OSFs
and IURs.
KFC
A2
VMAXC
| PB
£ VSC
(0
VLC
VPR
QCC
-0.75 -0.5 -0.25 0 0.25 0.5 0.75 1
Normalized sensitivity coefficient
Inhalation exposure: liver GST
Figure 5-17. Sensitivity coefficients for long-term mass GST-mediated
metabolites per liver volume from a long-term average daily inhalation
concentration of 2000 ppm in mice. KFC = GST-mediated metabolism rate; A2 =
proportion of liver GST metabolism attributed to the lung; VMAXC = CYP-mediated
maximum rate of metabolism; PB = blood:air partition coefficient; VSC = slowly
perfused tissue volume; VLC = liver volume; VPR = Ventilation perfusion ratio ;
QCC = cardiac output constant.
304

-------
8545
8546
8547
8548
8549
8550
8551
8552
8553
8554
8555
8556
8557
8558
8559
8560
8561
8562
8563
8564
a)
a>
E
(0
i_
(0
a.
KFC
A2
KA
VMA
PB
VSC
VLC
VPR
QCC
Oral exposure: liver GST
¦
C
-0.75 -0.5 -0.25 0 0.25 0.5 0.75
Normalized sensitivity coefficient
Figure 5-18. Sensitivity coefficients for long-term mass GST-mediated metabolites
per liver volume from a long-term average daily drinking water concentration of
500 mg/L in mice. KFC = GST-mediated metabolism rate; A2 = proportion of liver
GST metabolism attributed to the lung; KA = oral absorption rate from gut; VMAXC =
CYP-mediated maximum rate of metabolism; PB = blood:air partition coefficient; VSC
= slowly perfused tissue volume; VLC = liver volume; VPR = Ventilation perfusion
ratio ; QCC = cardiac output constant.
a)
a>
E
(0
i_
(0
a.
KFC
A2
VMAXC
PB
VSC
VLC
VPR
QCC
Inhalation exposure: lung GST
]
-0.75 -0.5 -0.25 0 0.25 0.5 0.75
Normalized sensitivity coefficient
Figure 5-19. Sensitivity coefficients for long-term mass GST-mediated metabolites
per lung volume from a long-term average daily inhalation concentration of 500
ppm in mice. KFC = GST-mediated metabolism rate; A2 = proportion of liver GST
metabolism attributed to the lung; VMAXC = CYP-mediated maximum rate of
metabolism; PB = blood:air partition coefficient; VSC = slowly perfused tissue volume;
VLC = liver volume; VPR = Ventilation perfusion ratio ; QCC = cardiac output
constant.
305

-------
8565
8566
8567
8568
8569
8570
8571
8572
8573
8574
8575
8576
8577
8578
8579
8580
8581
8582
8583
8584
8585
8586
8587
8588
8589
8590
8591
8592
8593
8594
8595
8596
8597
8598
8599
8600
8601
8602
The comparison of the OSF derived from the oral exposure data and from the route-to-
route extrapolation from the inhalation data provides a crude measure of the uncertainty in
recommending a human OSF based on the available rodent bioassay data. The oral cancer slope
factor based on route-to-route extrapolations from liver tumors in mice exposed by inhalation are
about an order of magnitude lower than those based on the liver tumor responses in mice
exposed via drinking water. This difference may be explained, at least partially, by the
heterogeneity of hepatic cell types within the sinusoid, resulting in regio-specific toxicity. Oral
exposure may result in a higher internal exposure of hepatocytes in the periportal region
(particularly those lining the portal vein, through which all gastrointestinal-absorbed
dichloromethane passes) than in the centrilobular region (Syracuse Research Corporation [SRC],
1989). Further, the metabolic capacity of hepatic cells is also regio-specific, with higher CYP
activity found in the centrilobular region compared to the periportal region. Thus, liver perfusion
via the systemic arterial circulation (through which inhaled dichloromethane would be
introduced) or portal drainage of the gastrointestinal tract may influence regio-specific
hepatotoxicity, resulting in the route-of-exposure effects on toxicity. The available PBTK
models do not have the capability to predict regio-specific disposition of dichloromethane in the
liver.
There is uncertainty as to whether the reactivity of the toxic dichloromethane metabolites
is sufficiently high enough to preclude systemic distribution. Therefore, alternative derivations
of cancer risk values were performed under the assumption that high reactivity leads to complete
clearance from the tissue in which the active metabolite is formed (scaling factor = 1.0). The
difference in scaling factor (7.0 for allometric scaling versus 1.0) results in a 7-fold decrease in
estimated cancer toxicity values. Using a whole-body GST metabolism dose metric, the
resulting OSF and IUR for liver and lung cancer were approximately five-fold higher than when
tissue-specific dose metrics were used (Table 5-16 and Table 5-22). The mechanistic data
support the notion that reactive metabolites produced in the target tissues are not well distributed
and produce deleterious effects in the metabolizing tissues soon after generation. Thus, there is
less uncertainty in the cancer risk values derived by using a tissue-specific GST metabolism dose
metric compared with those derived using a whole-body GST metabolism dose metric.
Sensitive human populations
Possible sensitive populations include persons with altered CYP (e.g., obese individuals,
alcoholics, diabetics, and the very young) and GST (e.g., GST-T1 homozygous conjugators)
metabolic capacity. The PBTK model includes an estimate of the variability of CYP metabolism
(sixfold variation), within the general population but does not specifically address what could be
greater variation in these other groups. However, the known polymorphisms for GST-T1
expression were integrated into the human model. The distributions of human IUR values (from
which the recommended [i.e., mean] values were taken) show that the 99th percentiles are
306

-------
8603
8604
8605
8606
8607
8608
8609
8610
8611
8612
8613
8614
8615
8616
8617
8618
8619
8620
8621
8622
8623
8624
8625
8626
8627
8628
8629
8630
8631
8632
8633
8634
8635
8636
8637
approximately 4-fold and 6-fold higher than means for liver and lung cancer. For the distribution
of OSFs, the 99th percentile is approximately threefold higher than mean for liver cancer.
To further characterize the potential sensitivity of specific subpopulations, internal dose
distributions for oral exposure to 1 mg/kg-day or inhalation exposure to 1 mg/m3 were estimated
for 1-year-old children and 70-year-old men and women to compare with the broader population
results used to estimate cancer risks above. Since the recommended cancer risk estimate is based
on the GST-T1+/+ subpopulation, this analysis was also restricted to that subpopulation, so that
only the factors of age and gender would differ. The impact of considering other GST-T1 groups
can be seen where risk estimates for the GST-T1 and entire population mix are given above.
Specification of age- and gender-specific parameters are as described in Appendix B. This
sensitivity analysis is qualitatively similar to that described previously for the noncancer
assessments of dichloromethane, where the variability in human equivalent administered dose
and HEC values was estimated.
For this analysis, however, consideration of exclusively GST-T1+/+ individuals will
clearly narrow any estimate of variability. This analysis will also differ from that for noncancer
effects in that the inverse of the former relationship is being considered (i.e., the variation in a
specific internal dose for a fixed exposure is being computed, whereas for the human equivalent
administered dose and HEC the variability in exposure levels corresponding to a fixed internal
dose are estimated. The results of this analysis for oral exposures are shown in Figure 5-20 and
Table 5-27 and for inhalation exposures in Figure 5-21 and Table 5-28.
For the oral exposure analysis, the distribution of internal doses shows little variation
among the different age/gender groups (Figure 5-20, Table 5-27). The cancer analysis is based
on a very low internal dose, where little enzymatic saturation is expected to occur, allowing for
efficient first-pass metabolism, which is independent of differences in respiration; differences
will be more significant at the higher doses analyzed for the noncancer human equivalent applied
dose. Thus, the consideration of only GST metabolism and the narrower range of metabolic rate
for that pathway in the +/+ population at low oral exposure rates results in minimal age/gender
sensitivity differences (the 7-year-old female is only 5% more sensitive from pharmacokinetic
factors than the general population).
For inhalation, an internal liver GST dose (mean value) about 2.5 times higher in the
child than the general population is predicted due to the higher inhalation rate. The results for
the liver GST dose for inhalation, Figure 5-21 and Table 5-28 indicate that the 70-year-old male
and female populations are only slightly shifted from the general population, while the
population for the 1-year-old child is a distinct upper tail of the general distribution.
307

-------
8638
8639
8640
8641
8642
8643
8644
8645

	70 yo Male
— - - 70 yo Female
Internal dose distribution,
GST-T1 ++ population,
oral ingestion
General
1 yo child
0 0.05 0.1 0.15 0.2 0.25 0.3
Internal dose (mg GST metabolites/L liver/day)
Figure 5-20. Histograms for a liver-specific dose of GST metabolism (mg
GST metabolites per liter liver per day) for the general population (0.5- to
80-year-old males and females) and specific age/gender groups, within the
population of GST-T1+/+ genotypes, given a daily oral dose-rate of 1 mg/kg-
day dichloromethane.
Table 5-27. Statistical characteristics of human internal
doses for 1 mg/kg-day oral exposures in specific
populations
8646
Internal dose (mg/L liver per day)3
Population
Mean
95th percentile 99th percentile
All agesb
7.96 :
*10"2
1
.91 :
xlO-1
2
.89 :
<10-1
1-year-old children
6.60 :
xio-2
1
.47 :
xlO-1
2
.05 :
xio-1
70-year-old men
8.22 :
xio-2
1
.97 :
xlO-1
2
.98 :
xio-1
70-year-old women
8.66 :
<10"2
2
.18 :
xlO"1
3
.37 :
X10"1
aLiver-specific GST-T1 metabolism in GST-T1
orally to 1 mg/kg-day dichloromethane.
b0.5- to 80-year-old males and females.
individuals exposed
308

-------
8647
8648
8649
8650
8651
8652
8653
8654
8655
8656
8657
8658
0.25
Internal dose distribution,
GST-T1 ++ population,
1 mg/m3 inhalation
	All ages, M & F
— - - 1 yo Child
	70 yo Male
	70 yo Female
10
20
30
40
ng GST metabolites/L liver/day
Figure 5-21. Histograms for liver-specific dose of GST metabolism (mg GST
metabolites per liter liver per day) for the general population (0.5- to 80-year-
old males and females) and specific age/gender groups, within the population
of GST-T1+/+ genotypes, given a continuous inhalation exposure to 1 mg/m3
dichloromethane.
Table 5-28. Statistical characteristics of human internal
doses for 1 mg/m3 inhalation exposures in specific
subpopulations

Internal dose (mg/L liver per day)3
Population
Mean
95th percentile
99th percentile
All agesb
5.63 x 10"6
1.56 x 10"5
2.60s x 10"5
1-year-old children
1.41 x 10"5
3.30 x 10"5
4.70 x 10"5
70-year-old men
4.36 x 10"6
1.11 x 10"5
1.62 x 10"5
70-year-old women
3.55 x 10"6
9.41 x 10-6
1.45 x 10"5
aLiver-specific GST-T1 metabolism in GST-T1+/+ individuals
exposed continuously by inhalation to 1 mg/m3 diehloromethane.
b0.5- to 80-year-old males and females.
309

-------
8659
8660
8661
8662
8663
8664
8665
8666
8667
8668
8669
8670
8671
8672
8673
8674
8675
8676
8677
8678
8679
8680
8681
8682
8683
8684
8685
8686
8687
8688
8689
8690
8691
8692
8693
8694
8695
8696
6. MAJOR CONCLUSIONS IN Till CHARACTERIZATION OF HAZARD AND
DOSE RESPONSE
6.1. HUMAN HAZARD POTENTIAL
Dichloromethane (CASRN 75-09-2), also known as methylene chloride, is a colorless
liquid with a penetrating ether-like odor. It is produced by the direct reaction of methane with
chlorine at either high temperatures or low temperatures under catalytic or photolytic conditions.
The principal uses for dichloromethane have been in paint strippers and removers, as a propellant
in aerosols, in the manufacture of drugs, pharmaceuticals, film coatings, electronics, and
polyurethane foam, and as a metal-cleaning solvent.
Dichloromethane is rapidly absorbed through both oral administration and inhalation
exposure with a steady-state saturation occurring with inhalation. Results from studies of
animals show that, following absorption, dichloromethane is rapidly distributed throughout the
body and has been detected in all tissues that have been evaluated. Metabolism of
dichloromethane involves two primary pathways. Dichloromethane is metabolized to CO and to
a lesser extent CO2 in a CYP-dependent oxidative pathway (CYP2E1) that is predominant at low
exposure levels. The other major pathway for dichloromethane metabolism involves the
conjugation of dichloromethane to GSH, catalyzed by GST (GST-T1). This results in the
formation of a GSH conjugate that is eventually metabolized to CO2. The conjugation of
dichloromethane to GSH results in the formation of two reactive intermediates that have been
hypothesized to be involved in dichloromethane carcinogenicity, S-(chloromethyl)glutathione
and formaldehyde. Formation of formaldehyde leads to several covalent modifications of
cellular macromolecules, including DNA-protein cross-links (Casanova et al., 1996) and RNA-
formaldehyde adducts (Casanova et al., 1997). Evidence is also available that
S-(chloromethyl)glutathione can result in both DNA SSBs and DNA mutations, presumably
through DNA alkylation (Green, 1997; Graves and Green, 1996; Graves et al., 1996, 1994a;
Hashmi et al., 1994). However, DNA reaction products (e.g., DNA adducts) produced by S-
(chloromethyl)glutathione have not been found, possibly due to potential instability of these
compounds (Watanabe et al., 2004; Hashmi et al., 1994).
Information on noncancer effects in humans exposed orally to dichloromethane are
restricted to case reports of neurological impairment (general CNS depression), liver and kidney
effects (as severe as organ failure), and gastrointestinal irritation in individuals who ingested
amounts ranging from about 25 to 300 mL (Chang et al., 1999; Hughes and Tracey, 1993). The
animal toxicity database identifies hepatic effects (hepatic vacuolation, nonneoplastic liver foci)
as the critical dose-dependent noncancer endpoint associated with oral exposure to
dichloromethane. The most frequently observed liver effect was hepatocyte vacuolation, seen
with drinking water exposure (90 days) in F344 rats at >166 mg/kg-day and B6C3Fi mice at
310

-------
8697
8698
8699
8700
8701
8702
8703
8704
8705
8706
8707
8708
8709
8710
8711
8712
8713
8714
8715
8716
8717
8718
8719
8720
8721
8722
8723
8724
8725
8726
8727
8728
8729
8730
8731
8732
8733
8734
586 mg/kg-day (Kirschman et al., 1986) and with gavage exposure (14 days) in CD-I mice at
333 mg/kg-day (Condie et al., 1983). Hepatocyte degeneration or necrosis was observed in
female F344 rats exposed in drinking water for 90 days to 1,469 mg/kg-day (Kirschman et al.,
1986) and in female F344 rats exposed by gavage for 14 days to 337 mg/kg-day (Berman et al.,
1995). In the chronic-duration (104-week) study, liver effects were described as nonneoplastic
foci in F344 rats exposed to drinking water doses between 50 and 250 mg/kg-day (Serota et al.,
1986a). In the reproductive oral administration studies, no significant effect on reproductive
function or parameter was observed in rats up to 225 mg/kg-day (General Electric Co., 1976) or
in mice up to 500 mg/kg-day (Raje et al., 1988). The NOAEL and LOAEL for altered
neurological functions in female F344 rats were 101 and 337 mg/kg-day (as reported by Moser et
al., 1995).
Acute inhalation exposure of humans to dichloromethane has been associated with
cardiovascular impairments due to decreased oxygen availability from COHb formation and
neurological impairment from interaction of dichloromethane with nervous system membranes
(Bos et al., 2006; ACGIH, 2001; AT SDR, 2000; Cherry et al., 1983; Putz et al., 1979; Gamberale
et al., 1975; Winneke, 1974). Relatively little is known about the long-term neurological effects
of chronic exposures, although there are studies that provide some evidence of an increased
prevalence of neurological symptoms among workers with average exposures of 75-100 ppm
(Cherry et al., 1981) and long-term effects on some neurological measures (i.e., possible
detriments in attention and reaction time in complex tasks) in workers whose past exposures
were in the 100-200 ppm range (Lash et al., 1991). These studies are limited by the relatively
small sample sizes and low power for detecting statistically significant results for these
endpoints.
Following repeated inhalation to dichloromethane, the liver is the most sensitive target
for noncancer toxicity in rats and mice. Lifetime exposure was associated with hepatocyte
vacuolation and necrosis in F344 rats exposed to 1,000 ppm 6 hours/day (Mennear et al., 1988;
NTP, 1986), hepatocyte vacuolation in Sprague-Dawley rats exposed to 500 ppm 6 hours/day
(Nitschke et al., 1988a; Burek et al., 1984), and hepatocyte degeneration in B6C3Fi mice
exposed to 2,000 ppm 6 hours/day (lower concentrations were not tested in mice) (Mennear et
al., 1988; NTP, 1986). Other effects observed include renal tubular degenerations in F344 rats
and B6C3Fi mice at 2,000 ppm, testicular atrophy in B6C3Fi mice at 4,000 ppm, and ovarian
atrophy in B6C3Fi mice at 2,000 ppm.
Other studies with inhalation exposure to dichloromethane revealed no significant effects
on reproductive performance in rats (up to 1,500 ppm) (Nitschke et al., 1988b), although some
evidence of a decrease in fertility index was seen in male mice exposed to 150 and 200 ppm
(Raje et al., 1988), and no adverse effects on fetal development of mice or rats exposed up to
1,250 ppm were seen by Schwetz et al. (1975). Decreases in fetal BW and changes in behavioral
habituation were observed in Long-Evans rats exposed to 4,500 ppm during the gestational
311

-------
8735
8736
8737
8738
8739
8740
8741
8742
8743
8744
8745
8746
8747
8748
8749
8750
8751
8752
8753
8754
8755
8756
8757
8758
8759
8760
8761
8762
8763
8764
8765
8766
8767
8768
8769
8770
8771
8772
period (Bornschein et al., 1980; Hardin and Manson, 1980). Exposure-related noncancer effects
on the lungs consisted of foreign-body pneumonia in rats exposed to 8,400 ppm 6 hours/day for
13 weeks (NTP, 1986), Clara cell vacuolation in mice exposed to 4,000 ppm 6 hours/day for
13 weeks (Foster et al., 1992), and pulmonary congestion in guinea pigs exposed to 5,000 ppm
7 hours/day for 6 months (Heppel et al., 1944). Several neurological mediated parameters
including decreased activity (Kjellstrand et al., 1985; Weinstein et al., 1972; Heppel and Neal,
1944), impairment of learning and memory (Alexeef and Kilgore, 1983), and changes in
responses to sensory stimuli (Rebert et al., 1989) are reported from acute and short-term
dichloromethane exposure. Evidence of a localized immunosuppressive effect in the lung,
resulting from inhalation dichloromethane exposure, was seen in an acute exposure (3 hours,
100 ppm) study in CD-I mice (Aranyi et al., 1986).
Numerous in vitro studies have demonstrated mutagenic and genotoxic effects
associated with dichloromethane exposure. For example, bacterial assays, yeast, and fungi
provide evidence that the mutagenic action of dichloromethane in bacterial systems is enhanced
by metabolic activation (e.g., Dillon et al., 1992; Jongen et al., 1982; Gocke et al., 1981).
Positive results from assays of DNA damage with in vitro mammalian systems provide support
that dichloromethane genotoxicity is linked to metabolism by GST enzymes (Graves et al., 1996,
1995, 1994b). Consistent evidence for several genotoxic endpoints in target tissues (liver and
lung) in mice following in vivo exposure to dichloromethane provides supporting evidence that
GST-pathway metabolites are key actors in the mutagenic and carcinogenic mode of action for
dichloromethane. Pretreatment of mice with buthionine sulphoximine, a GSH depletor, caused a
decrease to levels seen in controls in the amount of DNA damage detected immediately after in
vivo exposure in liver and lung tissue, indicating GSH involvement in the genotoxic process
(Graves et al., 1995). DNA damage (detected by the comet assay) was also reported in liver and
lung tissues from male CD-I mice sacrificed 4 hours after administration of a single oral dose of
1,720 mg/kg of dichloromethane (Sasaki et al., 1998). In this study, DNA damage in lung and
liver was not detected 3 hours after dose administration, and no DNA damage occurred at either
time point in several other tissues in which a carcinogenic response was not seen in chronic
animal cancer bioassays (e.g., stomach, kidney, bone marrow). The weight of evidence from
these studies suggests that dichloromethane is carcinogenic by a mutagenic mode of action.
Dichloromethane is "likely to be carcinogenic in humans" by the inhalation and oral
routes of exposure under the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a).
Results from several 2-year bioassays provide adequate evidence of the carcinogenicity of
dichloromethane in mice and rats exposed by inhalation, as well as adequate data to describe
dose-response relationships. Oral exposure to dichloromethane produced statistically significant
increases in hepatocellular adenomas and carcinomas in male B6C3Fi mice (Serota et al., 1986b;
Hazelton Laboratories, 1983). Inhalation exposure to concentrations of 2,000 or 4,000 ppm
dichloromethane produced increased incidences of lung and liver tumors in B6C3Fi mice
312

-------
8773
8774
8775
8776
8777
8778
8779
8780
8781
8782
8783
8784
8785
8786
8787
8788
8789
8790
8791
8792
8793
8794
8795
8796
8797
8798
8799
8800
8801
8802
8803
8804
8805
8806
8807
8808
8809
8810
(Maronpot et al., 1995; Foley et al., 1993; Kari et al., 1993; Mennear et al., 1988; NTP, 1986).
Significantly increased incidences of benign mammary tumors (adenomas or fibroadenomas)
were observed in male and female F344/N rats exposed by inhalation to 2,000 or 4,000 ppm
(Mennear et al., 1988; NTP, 1986). A statistically significant increased incidence of brain or
CNS tumors has not been observed in any of the animal cancer bioassays, but a 2-year study
using relatively low exposure levels (0, 50, 200, and 500 ppm) in Sprague-Dawley rats observed
a total of six astrocytoma or glioma (mixed glial cell) tumors in the exposed groups (Nitschke et
al., 1988a). These tumors are exceedingly rare in rats, and there are few examples of statistically
significant trends in animal bioassays (Sills et al., 1999).
6.2. DOSE-RESPONSE
6.2.1. Oral RfD
The available oral toxicity data for animals identify hepatic effects (hepatic vacuolation,
nonneoplastic liver foci) as the most sensitive noncancer endpoint associated with chronic oral
exposure to dichloromethane. The 104-week drinking-water study in F344 rats (Serota et al.,
1986a) was selected as the principal study for RfD derivation because the study provided a
sensitive endpoint (nonneoplastic liver foci) and used lower doses in comparison to other chronic
oral administration studies. In this study, four doses (6, 52, 125, and 235 mg/kg-day in males; 6,
58, 136, and 263 mg/kg-day in females) were used. A NOAEL of 6 mg/kg-day in males and
females and a LOAEL of 52 (male) and 58 (female) mg/kg-day for nonneoplastic alterations of
liver foci was identified.
An RfD of 7 x 10 3 mg/kg-day is recommended for use in humans. The RfD derivation
process involved first fitting all available dichotomous models in BMDS version 2.0 to the
incidence data for male rats. The male data were used because a greater sensitivity was seen in
males compared with females in this study. A dose metric of average daily mass of
dichloromethane metabolized via the CYP pathway per unit volume of liver was derived from an
EPA-modified rat PBTK model (see Appendix C). This metric was chosen because there are no
data to support the role of a specific metabolite in the development of the noncancer liver lesions
seen in oral and inhalation exposure studies and the CYP-metabolism dose metric was
determined to be most consistent with the data. Then, the lower 95% confidence limit on the
dose associated with a 10% risk for liver lesions (BMDLio) was derived, based on the best fitting
model (in terms of the value of the AIC and examination of model fit and residuals). Because
the metric is a rate of metabolism, rather than the concentration of putative toxic metabolites, and
the clearance of these metabolites may be slower per volume tissue in the human compared with
the rat, this rodent internal dose metric for noncancer effects was adjusted by dividing by a
pharmacokinetic scaling factor to obtain a human equivalent internal BMDLio. This BMDLio
was then converted to the human equivalent dose by using a human PBTK model (adapted from
David et al., 2006; see Appendix B) that utilizes Monte Carlo sampling techniques to provide a
313

-------
8811
8812
8813
8814
8815
8816
8817
8818
8819
8820
8821
8822
8823
8824
8825
8826
8827
8828
8829
8830
8831
8832
8833
8834
8835
8836
8837
8838
8839
8840
8841
8842
8843
8844
8845
8846
8847
8848
distribution of human equivalent doses. The first percentile of the distribution of human
equivalent doses was chosen to include the most sensitive population, while staying within
bounds of what is considered computationally stable. The first percentile human equivalent
administered dose was used as a point of departure and was divided by a composite UF of 30 (3
[10°5] to account for uncertainty about interspecies toxicodynamic equivalence, 3 [10°5] to
account for uncertainty about toxicodynamic variability in humans, and 3 [10°5] for database
deficiencies) to arrive at an RfD of 7 x 10 3 mg/kg-day.
Use of the mean value (3.6 x 10 1 mg/kg-day) of the human equivalent administered dose
distribution instead of the 1st percentile, with an additional UF of 3 (10°5) to account for human
toxicokinetic variability, would yield a candidate RfD of 4 x 10 3, which is relatively similar to
the recommended RfD of 1 x 10 3.
6.2.2. Inhalation RfC
The liver is the most sensitive target for noncancer toxicity in rats and mice, following
repeated inhalation exposure to dichloromethane. Nonneoplastic liver lesions (specifically,
hepatic vacuolation) in rats are the critical noncancer effect from chronic dichloromethane
inhalation in animals. Inhalation bioassays with Sprague-Dawley rats identified the lowest
inhalation LOAEL for nonneoplastic liver lesions in the database: 500 ppm (6 hours/day,
5 days/week for 2 years) (Nitschke et al., 1988a; Burek et al., 1984). Nitscke et al. (1988a)
identified a NOAEL of 200 ppm for hepatocyte vacuolation in female rats. Because the Nitschke
et al. (1988a) study more adequately covers the range spanning the BMR compared with the
study by Burek et al. (1984), the former study was selected as the principal study for derivation
of a chronic inhalation RfC.
An RfC of 0.2 mg/m3 is derived based on the observed critical effect in the principal
study. As was described above for the RfD, the RfC derivation process was based on a dose
metric of average daily mass of dichloromethane metabolized via the CYP pathway per unit
volume of liver. This metric was derived from an EPA-modified rat PBTK model (see Appendix
C). Then, the lower 95% confidence limit on the dose associated with a 10% risk for liver
lesions (BMDLio) was derived, based on the best fitting model in terms of the value of the AIC
and examination of model fit and residuals. Because the metric is a rate of metabolism, rather
than the concentration of putative toxic metabolites, and the clearance of these metabolites may
be slower per volume tissue in the human compared with the rat, this rodent internal dose metric
for noncancer effects was adjusted by dividing by a pharmacokinetic scaling factor to obtain a
human-equivalent internal BMDLio. This BMDLio was then converted to the HEC by using a
human PBTK model (adapted from David et al., 2006; see Appendix B) that utilizes Monte
Carlo sampling techniques to provide a distribution of HECs.
The first percentile HEC was used as a point of departure. This percentile was chosen
because it included the most sensitive population while staying within bounds of what is
314

-------
8849
8850
8851
8852
8853
8854
8855
8856
8857
8858
8859
8860
8861
8862
8863
8864
8865
8866
8867
8868
8869
8870
8871
8872
8873
8874
8875
8876
8877
8878
8879
8880
8881
8882
8883
8884
8885
8886
considered computationally stable. This point of departure was divided by a composite UF of
100 (3 [10°5] to account for uncertainty about interspecies toxicodynamic equivalence, 3 [10°5]
to account for uncertainty about toxicodynamic variability in humans, and 10 for database
deficiencies) to arrive at an RfC of 0.2 mg/m3.
Use of the mean value (47.36 mg/m3)of the HEC distribution instead of the 1st percentile,
with an additional UF of 3 (10°5) to account for human toxicokinetic variability would yield a
candidate RfC identical to the recommended value of 0.2 mg/m3. In addition, two comparison
values derived from occupational studies produced values of 3.5 mg/m3 (Cherry et al., 1983) and
0.55 mg/m3 (Lash et al., 1991). The animal-derived candidate RfC is preferable to the human-
derived candidate RfC because of the uncertainties about the exposure durations for the workers
in the Cherry et al. (1983) study and uncertainties regarding the exposures and effect sizes in
Lash et al. (1991) and because the RfC based on the rat data is more health protective.
6.2.3. Uncertainties in Reference Dose and Reference Concentration Values
One data uncertainty identified is the potential for neurodevelopmental effects. Animal
bioassays have not identified gross or microscopic effects on neural tissues from long-term
exposures or single (Schwetz et al., 1975) or multigenerational (Nitschke et al., 1988b)
developmental toxicity studies. However, behavioral changes were observed in pups born to rats
exposed to high levels (4,500 ppm) of dichloromethane (Bornschein et al., 1980; Hardin and
Manson, 1980); 4,500 ppm was the only dose used in this study. Thus uncertainty exists as to
the development of neurological effects from lower gestational exposures in animals, or in
humans. In addition, immunotoxicity data revealed an additional area of data uncertainty
specifically with respect to inhalation exposure. Data from Aranyi et al. (1986) demonstrated
evidence of immunosuppression, following a single 100 ppm dichloromethane exposure for 3
hours in CD-I mice. The weight of evidence for nonneoplastic effects in humans and animals
suggests that the development of liver lesions is the most sensitive effect, with a UF applied
because of the lack of neurodevelopmental studies and, for the RfC, the uncertainty regardingthe
lack of a low dose developmental study.
The extrapolation of internal dichloromethane dosimetry from nonneoplastic rat
responses to human risk was accomplished by using PBTK models for dichloromethane in rats
and humans. Uncertainties in rat and human dosimetry used for RfD and RfC derivation can
arise from uncertainties in the PBTK models to accurately simulate the toxicokinetics of
dichloromethane for animals under bioassay conditions and humans experiencing relatively low,
chronic environmental exposures. Further, the dose metric used in the models is the rate of
metabolism to a putative toxic metabolite, rather than the concentration (average or area under
the concentration curve of the metabolite), so the model specifically fails to account for rodent-
human differences in clearance or removal of the toxic metabolite. A scaling factor, based on
BW ratios, was used to account for this difference.
315

-------
8887
8888
8889
8890
8891
8892
8893
8894
8895
8896
8897
8898
8899
8900
8901
8902
8903
8904
8905
8906
8907
8908
8909
8910
8911
8912
8913
8914
8915
8916
8917
8918
8919
8920
8921
8922
8923
8924
Uncertainties in the human population model were quantitatively accounted for by
utilizing hierarchical Bayesian calibration methods during model development (David et al.,
2006; Marino et al., 2006). The rat model was modified and utilized in a deterministic manner.
Data were not available to perform a hierarchical Bayesian calibration in the rat, but
uncertainties in the rat model predictions were assessed qualitatively. For both oral and
inhalation exposures, the liver volume, followed closely by the volume of slowly perfused
tissues, had the greatest impact on the internal dose of mg dichloromethane metabolized via CYP
pathway per liter tissue per day. This was due to the fact that the dose metric is a tissue-specific
concentration, the majority of CYP metabolism is attributed to the liver, and changes in liver
volume have a greater impact on the total CYP metabolism than either of the individual Vmax
values. There is high confidence in the values used for volume of liver and slowly perfused
tissues in the rat, as these are well studied (Brown et al., 1997). Therefore, the uncertainties
associated with use of the rat PBTK model should not markedly affect the values of the RfD and
RfC.
An additional uncertainty inherent in this process, however, is the lack of knowledge
concerning the most relevant dose metric (e.g., a specific metabolite) within the context of the
development of the noncancer liver effects. This basic research question represents a data gap,
and this uncertainty is not addressed quantitatively or qualitatively in the assessment.
The effect of dichloromethane on human populations that are sensitive due to
pharmacokinetic differences was addressed quantitatively by using a human probabilistic PBTK
model to generate distributions of human exposures likely to occur given a specified internal
BMDLio. The model and resulting distributions take into account the known differences in
human physiology and metabolic capability with regard to dichloromethane dosimetry. The first
percentile values of the distributions of human equivalent doses (Table 5-3) and HECs (Table 5-
7) served as points of departure for candidate RfDs and RfCs, respectively, to protect
toxicokinetically sensitive individuals. No data are available regarding toxicodynamic
differences within a human population. Therefore, a UF of 3 for possible differences in human
toxicodynamic responses is intended to be protective for sensitive individuals.
6.2.4. Oral Cancer Slope Factor
The recommended oral cancer slope factor for dichloromethane is 1 x 10 3 (mg/kg-day) ',
which is based on liver tumor responses in male B6C3Fi mice exposed to dichloromethane in
drinking water for 2 years (Serota et al., 1986b; Hazelton Laboratories, 1983). This value was
derived by using a tissue-specific GST metabolism dose metric with allometric scaling to
account for uncertainty regarding the reactivity and clearance of the metabolite(s) involved in the
carcinogenic response.
There was only one adequate oral exposure cancer bioassay (Serota et al., 1986a, b;
Hazelton Laboratories, 1983) evaluating the carcinogenic potential of orally administered
316

-------
8925
8926
8927
8928
8929
8930
8931
8932
8933
8934
8935
8936
8937
8938
8939
8940
8941
8942
8943
8944
8945
8946
8947
8948
8949
8950
8951
8952
8953
8954
8955
8956
8957
8958
8959
8960
8961
8962
dichloromethane in F344 rats and B6C3Fi mice. Significant increases in incidence of liver
adenomas and carcinomas were observed in male (trendp-walue = 0.058) but not female B6C3Fi
mice (Serota et al., 1986b; Hazelton Laboratories, 1983). In F344 rats (Serota et al., 1986a), no
increased incidence of liver tumors was seen in male rats, and the pattern in female rats was
characterized by a jagged stepped pattern of increasing incidence of hepatocellular carcinoma or
neoplastic nodules; a similar pattern, but based on more sparse data, was seen when limited to
hepatocellular carcinomas. Statistically significant increases in tumor incidences were observed
in the 50 and 250 mg/kg-day groups (incidence rates of 10 and 14%, respectively) but not in the
125 mg/kg-day group (incidence rate of 3%). Incidence was also increased (10%) in a group
exposed for 78 weeks followed by a 26-week period of no exposure. The derivation of the oral
cancer slope factor is based on the male mice data because of their greater sensitivity to liver
cancer compared with female rats.
A modified mouse PBTK model of Marino et al. (2006) was used to approximate the
internal dose of daily dichloromethane (mg) metabolized via the GST pathway per unit volume
of liver from the daily oral administered doses. This approach was taken based on evidence that
GST-pathway metabolites produced from dichloromethane are primarily responsible for
dichloromethane carcinogenicity in mouse liver. The multistage dose-response model (BMDS
version 2.0) was used to fit the mouse liver tumor incidence and PBTK model-derived internal
dose data and derive a mouse internal BMD and BMDL associated with 10% extra risk
(BMDLio). The human BMDLio was derived by multiplying the mouse BMDLio by allometric
scaling factor (BWhUman/BWm0use)0'25 ~ 7). Linear extrapolation from the internal human
BMDLio (0.1/BMDLio) was used to derive oral risk factors for liver tumors based on tumor
responses in male mice. The linear low-dose extrapolation approach for agents with a mutagenic
mode of action was selected because GST-metabolism of dichloromethane is expected to occur
at and below exposures producing the mouse BMDLio, even though CYP2E1 metabolism is
expected to be unsaturated and to represent the predominant metabolic pathway in the liver.
Currently, there are no data from chronic oral cancer bioassays in mice providing support for a
nonlinear dose-response relationship.
Probability distributions of human oral cancer slope factors were derived by using a
human PBTK model (adapted from David et al. [2006]; see Appendix B). The cancer reference
values (OSF and IUR) were derived for a sensitive population: a population composed entirely
of carriers of the GST-T1+/+ homozygous genotype (that is, the group that would be expected to
be most sensitive to the carcinogenic effects of dichloromethane). In addition, cancer values
derived for a population reflecting the estimated frequency of GST-T1 genotypes in the current
U.S. population (20% GST-T1 , 48% GST-T1 , and 32% GST-T1+/+) were presented. All
simulations also included a distribution of CYP activity based on data from Lipscomb et al.
(2003). The mean OSF based on liver tumors in mice exposed to dichloromethane in drinking
water, 1 x 10 3 (mg/kg-day) ', based on what is assumed to be the most sensitive of the
317

-------
8963
8964
8965
8966
8967
8968
8969
8970
8971
8972
8973
8974
8975
8976
mi
populations, the GST-T1+/+ group, is the recommended OSF to be used in deterministic risk
assessments for chronic oral exposures to dichloromethane.
An OSF derived from the liver tumor data in the Serota et al. (1986b) study, using
administered dose dosimetry rather than PBTK modeling, is approximately one order of
magnitude higher than the current recommended value of 1 x 10 3 (per mg/kg-day). There is
approximately one to two orders of magnitude difference among the values based on different
dose metrics, scaling factors, and populations (Table 6-1).
The recommended OSF of 1 x 10 3 (per mg/kg-day) is based on a tissue-specific GST
internal dose metric with allometric scaling. Although the involvement of the GST pathway in
carcinogenic response has been established, some uncertainty remains as to the metabolite(s)
involved and the rate at which those metabolites are cleared. The value derived specifically for
the GST-T1+/+ population is recommended to provide protection for the population that is
hypothesized to be most sensitive to the carcinogenic effect. Application of ADAFs to the
cancer OSF is recommended in combination with appropriate exposure data when assessing
risks associated with early-life exposure (see section 5.4.4 for more details).
318

-------
8979
Table 6-1. Comparison of OSFs derived by using various assumptions and metrics, based on liver tumors
in male mice


Species,

Scaling
Mean OSF
Source
Population"
Dose metric
sex
Tumor
factor
(mg/kg-day) 1
(Table)
GST-T1+/+
Tissue-specific GST-metabolism rateb
Mouse, male
Liver
7.0
1.4 x 10"3
Table 5-13

Tissue-specific GST-metabolism rate
Mouse, male
Liver
1.0
2.0 >
< 10"1
Table 5-13

Whole-body GST metabolism rate
Mouse, male
Liver
7.0
8.1 >
< 10"1
Table 5-13

Route-to-route extrapolation, tissue-specific metabolism
Mouse, male
Liver
7.0
1.0 >
< 10"1
Table 5-14

Route-to-route extrapolation, tissue-specific metabolism
Mouse, male
Liver
1.0
1.5 >
< 10~5
Table 5-14

Route-to-route extrapolation, whole-body metabolism
Mouse, male
Liver
7.0
5.8 >
< 10~5
Table 5-14
Mixed
Tissue-specific GST-metabolism rateb
Mouse, male
Liver
7.0
8.0 >
< 10"1
Table 5-13

Tissue-specific GST-metabolism rate
Mouse, male
Liver
1.0
1.2 >
< 10"1
Table 5-13

Whole-body GST metabolism rate
Mouse, male
Liver
7.0
4.6 >
< 10"1
Table 5-13

Route-to-route extrapolation, tissue-specific metabolism
Mouse, male
Liver
7.0
5.8 >
< 10~5
Table 5-14

Route-to-route extrapolation, tissue-specific metabolism
Mouse, male
Liver
1.0
8.3 >
< lO^5
Table 5-14

Route-to-route extrapolation, whole-body metabolism
Mouse, male
Liver
7.0
3.3 >
< 10~5
Table 5-14

Applied dose (human equivalent dose)
Mouse, male
Liver

1.0 >
( 10~2
Table 5-15

1995 IRIS assessment
Mouse, male
Liver

7.5 >
< 10~3

aGST-Tl+/+ = homozygous, full enzyme activity; Mixed = genotypes based on a population reflecting the estimated frequency of genotypes in the current
U.S. population: 20% GST-T1 . 48% GST-T1 . and 32% GST-T1+/+ (Haber et al., 2002).
Bolded value is the basis for the recommended OSF of 1 x 10"3 per mg/kg-day.
319

-------
8980
8981
8982
8983
8984
8985
8986
8987
8988
8989
8990
8991
8992
8993
8994
8995
8996
8997
8998
8999
9000
9001
9002
9003
9004
9005
9006
9007
9008
9009
9010
9011
9012
9013
9014
9015
9016
9017
6.2.5. Cancer Inhalation Unit Risk
The recommended cancer IUR is 1 x 10 8 (|ig/m3) 1 for the development of liver and lung
cancers, based on data from male B6C3Fi mice, using a tissue-specific GST metabolism dose
metric. Data for liver and lung tumors in male and female B6C3Fi mice, following exposure to
airborne dichloromethane, were used to develop IURs for dichloromethane (Mennear et al.,
1988; NTP, 1986). This study was selected as the principal study to derive an IUR for
dichloromethane because of the completeness of the data, adequate sample size, and clear dose
response. In the NTP (1986) study, significant increases in incidence of liver and lung adenomas
and carcinomas were observed in both sexes of B6C3Fi mice exposed 6 hours/day, 5 days/week
for 2 years.
The PBTK model of Marino et al. (2006) for dichloromethane in the mouse was used to
calculate long-term daily average internal liver doses. The selected internal dose metrics for
liver tumors and lung tumors were long-term average daily mass of dichloromethane
metabolized via the GST pathway per unit volume of liver and lung, respectively. This approach
was taken based on evidence that GST-pathway metabolites produced from dichloromethane are
primarily responsible for dichloromethane carcinogenicity in mouse liver. The multistage dose-
response model (BMDS version 2.0) was used to fit the mouse liver tumor incidence and PBTK
model-derived internal dose data and derive a mouse internal BMD and BMDL associated with
10% extra risk (BMDLio). The human BMDLio was derived by multiplying the mouse BMDLio
by allometric scaling factor (BWhuman/BWmoUse)0'25 ~ 7). A linear extrapolation approach using
the internal human BMDLio for liver and lung tumors was used to calculate human tumor risk
factors by dividing the BMR of 0.1 by the human BMDL for each tumor type. Currently, there
are no data from chronic inhalation cancer bioassays in mice or rats providing support for a
nonlinear dose-response relationship.
The human PBTK model (adapted from David et al. [2006]; see Appendix B) provided
distributions of human internal dose metrics of daily mass of dichloromethane metabolized via
the GST pathway per unit volume of liver and lung resulting from chronic inhalation exposure to
a unit concentration of 1 (J,g/m3 dichloromethane (0.00029 ppm). As with the OSF, the cancer
IUR was derived for a sensitive population: a population composed entirely of carriers of the
GST-T1 homozygous positive genotype (that is, the group that would be expected to be most
sensitive to the carcinogenic effects of dichloromethane). In addition, cancer values derived for
a population reflecting the estimated frequency of GST-T1 genotypes in the current U.S.
population (20% GST-T1 7, 48% GST-T1 , and 32% GST-T1+/+) were also presented. The
distributions of IURs for liver or lung tumors were generated by multiplying the human tumor
risk factor for each tumor type and sex by the distribution of internal doses from chronic
exposure to 1 (J,g/m3 dichloromethane. A procedure to combine risks for liver and lung tumors
using different dose metrics for the different tumors (i.e., liver-specific and lung-specific
320

-------
9018
9019
9020
9021
9022
9023
9024
9025
9026
9027
9028
9029
9030
9031
9032
9033
9034
9035
9036
9037
9038
9039
9040
9041
9042
9043
9044
9045
9046
9047
9048
9049
9050
metabolism for liver and lung tumors, respectively), was used to derive the recommended IUR of
1 x 10~8 (^g/m3)-1 based on what is assumed to be the most sensitive of the populations, the
GST-T1 + + group.
The current recommended IUR value of 1 x 10~8 ((j,g/m3)-1 is approximately 1.5 orders
of magnitude lower than the previous IRIS value of 4.7 x 10~7 (^g/m3)-1 and similar to the
occupational exposure-based risk value of 4.17 x io~8 ((j,g/m3)-1 promulgated by OSHA (1997)
and derived from an estimated risk of 3.62 x 10"3 for a lifetime occupational inhalation exposure
of 25 ppm. The current use of the updated mouse PBTK model, with blood and tissue
equilibrium partition coefficients and metabolic parameters updated with MCMC calibration,
resulted in approximately three- and four-fold increases in the values of internal liver and lung
dose metrics, respectively, associated with the dichloromethane exposure concentrations,
compared with estimates from the model used in the previous IRIS assessment. For a unit
inhalation exposure, the mean internal lung GST dose predicted for the entire population
predicted by the MCMC updated human PBTK model is approximately thirteen-fold lower
compared with the human PBTK model used in the U.S. EPA (1995) assessment. The mean
internal liver GST dose, however, is approximately the same as (only 16% higher than) that
obtained with the previous PBTK parameters. For unit oral exposures, the mean internal liver
GST dose predicted by the MCM updated model is about 80% of that predicted using the
previous parameters, while the mean whole-body GST dose is predicted to be about 50% of that
predicted using the previous parameters.
An IUR derived from the liver tumor data of the NTP (1986) study using applied
concentration dosimetry rather than PBTK modeling, 3.7 x 10 7 (|ig/m3) ', is approximately one
order of magnitude higher than the currently recommended value of 1 x 10 8 (|ig/m3) 1 (Table
6-2). There is approximately one- to two- orders of magnitude difference among the values
based on different dose metrics, scaling factors, and populations.
The recommended IUR value of 1 x io~8 (p,g/m3)-1 is based on a tissue-specific GST-
internal dose metric with allometric scaling. Although the involvement of the GST pathway in
carcinogenic response has been established, some uncertainty remains as to the metabolite(s)
involved and the rate at which those metabolites are cleared. The value derived specifically for
the GST-T1+/+ population is recommended to provide protection for the population that is
hypothesized to be most sensitive to the carcinogenic effect. Application of ADAFs to the
cancer IUR is recommended when assessing risks associated with early-life exposure (see
section 5.4.4 for more details).
321

-------
851
Table 6-2. Comparison of IURs derived by using various assumptions and metrics




Scaling
IURb
Source
Population"
Dose metric
Species, sex,
Tumor type
factor
(jig/m3)"1
(Table)
GST-T1+/+
Tissue-specific GST-metabolism rate0
Mouse, male
Liver and lung
7.0
1.1 x 10"8
Table 5-20
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Liver
7.0
7.3
< 10~9
Table 5-19
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Lung
7.0
4.8
< 10~9
Table 5-19
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Liver and lung
1.0
1.6
< 10~9
Table 5-20
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Liver
1.0
1.0
< 10~9
Table 5-19
GST-T1+/+
Tissue-specific GST-metabolism rate
Mouse, male
Lung
1.0
6.8 x
10~10
Table 5-19
GST-T1+/+
Whole-body GST metabolism rate
Mouse, male
Liver and lung
7.0
1.4
< 10~8
Table 5-20
GST-T1+/+
Whole-body GST metabolism rate
Mouse, male
Liver
7.0
4.6
< 10~9
Table 5-19
GST-T1+/+
Whole-body GST metabolism rate
Mouse, male
Lung
7.0
1.0
< 10~8
Table 5-19
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Liver and lung
7.0
5.5
< 10~9
Table 5-20
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Liver
7.0
3.4
< 10~9
Table 5-19
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Lung
7.0
2.6
< 10~9
Table 5-19
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Liver and lung
1.0
7.9 x
10~10
Table 5-20
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Liver
1.0
4.8 x
10~10
Table 5-19
Mixed
Tissue-specific GST-metabolism rate
Mouse, male
Lung
1.0
3.7 x
10~10
Table 5-19
Mixed
Whole-body GST metabolism rate
Mouse, male
Liver and lung
7.0
7.9
< 10~9
Table 5-20
Mixed
Whole-body GST metabolism rate
Mouse, male
Liver
7.0
2.7
< 10~9
Table 5-19
Mixed
Whole-body GST metabolism rate
Mouse, male
Lung
7.0
6.0
< 10~9
Table 5-19

Administered concentration (HEC)
Mouse, male
Liver

3.7
< 10~7
Table 5-21

Administered concentration (HEC)
Mouse, male
Lung

9.2
< 10~7
Table 5-21

1995 IRIS assessment0
Mouse, male
Liver and lung
12.7
4.7
< 10"7

aGST-Tl+/+ = homozygous, full enzyme activity; Mixed = genotypes based on a population reflecting the estimated frequency of genotypes in the current U.S.
population: 20% GST-T1 . 48% GST-T1 . and 32% GST-T1+/+ (Haber et al., 2002).
' 'Based on mean value of the derived distributions
Bolded value is the basis for the recommended IUR of 1 x 10"8 ng/m3 per mg/kg-day.
9053
322

-------
9054
9055
9056
9057
9058
9059
9060
9061
9062
9063
9064
9065
9066
9067
9068
9069
9070
9071
9072
9073
9074
9075
9076
9077
9078
9079
9080
9081
9082
9083
9084
9085
9086
9087
9088
9089
9090
9091
6.2.6. Uncertainties in Cancer Risk Values
The database of animal bioassays identifies the liver and lung as the most sensitive target
organs for dichloromethane-induced tumor development, and there is high confidence that the
dose-response data for liver and lung cancer in mice represent the best available data for
derivation of human cancer risks. A dose-response relationship was seen with respect to liver
cancer in mice exposed orally and with respect to liver and lung cancer in mice exposed by
inhalation. Statistically significant increases in benign mammary gland tumors were observed in
one study of F344 rats exposed by inhalation to 2,000 or 4,000 ppm (Mennear et al., 1988; NTP,
1986); evidence for a tumorigenic mammary gland response in Sprague-Dawley rats was limited
to increased numbers of benign mammary tumors per animal at levels of 50-500 ppm (Nitschke
et al., 1988a) or 500-3,500 ppm (Burek et al., 1984). An oral (gavage) study in female Sprague-
Dawley rats reported an increased incidence of malignant mammary tumors, mainly
adenocarcinomas (8, 6, and 18% in the control, 100, and 500 mg/kg dose groups, respectively),
but the increase was not statistically significant. Data were not provided to allow an analysis that
accounts for differing mortality rates (Maltoni et al., 1988). The toxicokinetic or mechanistic
events that might lead to mammary gland tumor development in rats are unknown, although
CYP2E1 (El-Rayes et al., 2003; Hellmold et al., 1998) and GST-T1 expression has been detected
in human mammary tissue (Lehmann and Wagner, 2008). Rare CNS tumors were observed in
one study in rats, a study spanning a relatively low range of exposures (0-500 ppm) (Nitschke et
al., 1988a). These cancers were not seen in two other studies in rats, both involving higher doses
(1,000-4,000 ppm) (NTP, 1986; Burek et al., 1984), or in a similar high-dose study in mice
(NTP, 1986). The relative rarity of the tumors precludes the use of the low-dose exposure study
(Nitschke et al., 1988a) in a quantitative dose-response assessment. The available epidemiologic
studies provide some evidence of an association between dichloromethane and brain cancer. The
available epidemiologic studies do not provide an adequate basis for the evaluation of the role of
dichloromethane in breast cancer because there are currently no cohort studies with adequate
statistical power and no case-control studies with adequate exposure methodology to examine
this relationship.
There is uncertainty as to whether the reactivity of the toxic dichloromethane metabolites
is sufficiently high enough to preclude systemic distribution. Therefore, alternative derivations
of cancer risk values were performed under the assumption that high reactivity leads to complete
clearance from the tissue in which the active metabolite is formed (scaling factor = 1.0). The
difference in scaling factor (7.0 for allometric scaling versus 1.0) results in a 7-fold decrease in
estimated cancer toxicity values. Using a whole-body GST metabolism dose metric, the
resulting OSF and IUR for liver and lung cancer were approximately five-fold higher than when
tissue-specific dose metrics were used (Table 5-16 and Table 5-22). The mechanistic data
support the notion that reactive metabolites produced in the target tissues are not well distributed
and produce deleterious effects in the metabolizing tissues soon after generation. Thus, there is
323

-------
9092
9093
9094
9095
9096
9097
9098
9099
9100
9101
9102
9103
9104
9105
9106
9107
9108
9109
9110
9111
9112
9113
9114
9115
9116
9117
9118
9119
9120
9121
9122
9123
9124
9125
less uncertainty in the cancer risk values derived by using a tissue-specific GST metabolism dose
metric compared with those derived using a whole-body GST metabolism dose metric.
Uncertainty in the ability of the PBTK models to estimate animal and human internal
doses from lifetime bioassay low-level environmental exposures may affect the confidence in the
cancer risk extrapolated from animal data. Uncertainties in the mouse and human model
parameter values were integrated quantitatively into parameter estimation by utilizing
hierarchical Bayesian methods to calibrate the models at the population level (David et al., 2006;
Marino et al., 2006). The use of Monte Carlo sampling to define human model parameter
distributions allowed for derivation of human distributions of dosimetry and cancer risk,
providing for bounds on the recommended risk values. However, the PBTK animal models were
utilized deterministically, and a sensitivity analysis was performed to identify model parameters
most influential on the predictions of dose metrics used to estimate oral and inhalation cancer
risks. For inhalation exposures in mice, the blood:air partition coefficient, followed closely by
the first-order GST-mediated metabolism rate, had the greatest impact on the dose metric for
liver cancer (mg dichloromethane metabolized via GST pathway per liter liver per day). For
drinking water exposures in mice, the first-order GST-mediated metabolism rate, followed by the
CYP-mediated maximum reaction velocity (Vmaxc) affected the liver cancer dose metric to the
greatest extent. For mice inhaling dichloromethane, the lung cancer dose metric (mg
dichloromethane metabolized via GST pathways per liter lung per day), like the liver cancer
metric, was highly affected by the first-order GST-mediated metabolism rate and the blood:air
partition coefficient. However, the lung cancer dose metric was most sensitive to the
proportional yield of liver GST-mediated metabolic activity attributed to the lung. The blood:air
partition coefficient was experimentally determined, lending high confidence to its value. In
contrast, values for the three metabolic parameters were determined by computational
optimization against data sets not directly measuring dichloromethane or its metabolites in the
target/metabolizing tissues. It is uncertain how alternative values for these three parameters
would affect the estimation of animal BMDL10 values and, ultimately, IURs and OSFs. In
addition, specific uncertainty remains concerning the human PBTK parameter distributions. In
addition, while the structure and equations used in the existing model have been described
extensively in peer-reviewed publications, uncertainty remains concerning the model structure,
and specifically the potential of an alternative (dual-binding-site) CYP metabolic rate equation
for dichloromethane Integration of the alternate rate equation into the PBTK modeling, and then
quantitative risk assessment, will likely require several years of further research, and hence is
beyond the scope of the current assessment.
324

-------
7. REFERENCES
ACGIH (American Conference of Governmental Industrial Hygienists). (2001) Dichloromethane. In:
Documentation of the threshold limit values and biological exposure indices. 7th edition. Cincinnati, OH:
American Conference of Governmental Industrial Hygienists.
Alexeef, GV; Kilgore, WW. (1983) Learning impairment in mice following acute exposure to
dichloromethane and carbon tetrachloride. J Toxicol Environ Health 11:569-581.
Allen, J; Kligerman, A; Campbell, J; et al. (1990) Cytogenetic analyses of mice exposed to
dichloromethane. EnvironMol Mutagen 15:221-228.
Anders, MW; Sunram, JM. (1982) Transplacental passage of dichloromethane and carbon monoxide.
Toxicol Lett 12:231-234.
Andersen, ME; Clewell, HJ, III; Gargas, ML; et al. (1987) Physiologically based pharmacokinetics and the
risk assessment process for methylene chloride. Toxicol Appl Pharmacol 87:185-205.
Andersen, ME; Clewell, HJ, III; Gargas, ML; et al. (1991) Physiologically based pharmacokinetic
modeling with dichloromethane, its metabolite, carbon monoxide, and blood carboxyhemoglobin in rats
and humans. Toxicol Appl Pharmacol 108:14-27.
Andrae, U; Wolff, T. (1983) Dichloromethane is not genotoxic in isolated rat hepatocytes. Arch Toxicol
52:287-290.
Angelo, MJ; Pritchard, AB; Hawkins, DR; et al. (1986a) The pharmacokinetics of dichloromethane. I.
Disposition in B6C3Fi mice following intravenous and oral administration. Food Chem Toxicol
24(9):965-974.
Angelo, MJ; Pritchard, AB; Hawkins, DR; et al. (1986b) The pharmacokinetics of dichloromethane. II.
Disposition in Fischer 344 rats following intravenous and oral administration. Food Chem Toxicol
24(9):975-980.
Anthony, T. (1979) Methylene chloride. In: Kirk, RE; Othmer, DF; eds. Kirk-Othmer encyclopedia of
chemical technology. 3rd edition. New York, NY: John Wiley & Sons; pp. 686-693.
Anundi, H; Lind, ML; Friis, L; et al. (1993) High exposures to organic solvents among graffiti removers.
Int Arch Occup Environ Health 65:247-251.
Aranyi, C; O'Shea, WJ; Graham, JA; et al. (1986) The effects of inhalation of organic chemical air
contaminants on murine lung host defenses. Fundam Appl Toxicol 6:713-720.
Arcus-Arth, A; Blaisdell, RJ. (2007) Statistical distributions of daily breathing rates for narrow age groups
of infants and children. Risk Anal 27:97-110.
Astrand, I; Ovrum, P; Carlsson, A. (1975) Exposure to methylene chloride. I. Its concentration in alveolar
air and blood during rest and exercise and its metabolism. Scand J Work Environ Health 1:78-84.
Atkinson, R. (1989) Kinetics and mechanisms of the gas-phase reactions of the hydroxyl radical with
organic compounds. JPhys Chem Ref Data 1:63.
ATSDR (Agency for Toxic Substances and Disease Registry). (2000) Toxicological profile for methylene
chloride. Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA. Available
online at http://www.atsdr.cdc.gov/toxpro2.html.
325

-------
Bakinson, MA; Jones, RD. (1985) Gassings due to methylene chloride, xylene, toluene, and styrene
reported to Her Majesty's factory inspectorate 1961-80. Br J Ind Med 42:184-190.
Bell, BP; Franks, P; Hildreth, N; et al. (1991) Methylene chloride exposure and birth weight in Monroe
County, New York. Environ Res 55:31-39.
Berman, E; Schlicht, M; Moser, VC; et al. (1995) A multidisciplinary approach to toxicological screening:
I. Systemic toxicity. J Toxicol Environ Health 45:27-143.
Bernauer, U; Vieth, B; Ellrich, R; et al. (2000) CYP2E1 expression in bone marrow and its intra- and
interspecies variability: approaches for a more reliable extrapolation from one species to another in the risk
assessment of chemicals. Arch Toxicol 73:618-624.
Blair, A; Hartge, P; Stewart, P; et al. (1998) Mortality and cancer incidence of aircraft maintenance
workers exposed to trichloroethylene and other organic solvents and chemicals: extended follow up. Occup
Environ Med 55:161-171.
Blocki, FA; Logan, MSP; Baoli, C; et al. (1994) Reaction of rat liver glutathione s-transferase and bacterial
dichloromethane dehalogenase with dihalomethanes. J Biol Chem 269(12):8826-8830.
Bogaards, JJP; vanOmmen, B; vanBladeren, PJ. (1993) Interindividual differences in the in vitro
conjugation of methylene chloride with glutathione by cytosolic glutathione S-transferase in 22 human liver
samples. Biochem Pharmacol 45(10):2166-2169.
Bornschein, RL; Hastings, L; Manson, JM. (1980) Behavioral toxicity in the offspring of rats following
maternal exposure to dichloromethane. Toxicol Appl Pharmacol 52:29-37.
Bos, PMJ; Zeilmaker, MJ; van Eijkeren, JCH. (2006) Application of physiologically based
pharmacokinetic modeling in setting acute exposure guideline levels for methylene chloride. Toxicol Sci
91(2):576-585.
Boublik, T; Fried, V; Hala, E. (1984) The vapor pressures of pure substances: selected values of the
temperature dependence of the vapour pressures of some pure substances in the normal and low pressure
region. Vol. 17. 2nd revised edition. Amsterdam, Netherlands: Elsevier; p. 42.
Bowen, SE; Batis, JC; Paez-Martinez, N; Cruz, SL. (2006) The last decade of solvent research in animal
models of abuse: mechanistic and behavioral studies. Neurotoxicol Teratol 28(6): 636-647..
Breslow, NE; Day, NE. (1987) Statistical methods in cancer research. Volume II. The design and analysis
of cohort studies. IARC Sci Publ. 82:1-406.
Briving, C; Hambergerm, A; Kjellstrand, P; et al. (1986) Chronic effects of dichloromethane on amino
acids, glutathione and phosphoethanolamine in gerbil brain. Scand J Work Environ Health 12:216-220.
Brown, RP; Delp, MD; Lindstedt, SL; et al. (1997) Physiological parameter values for physiologically-
based pharmacokinetic models. Toxicol Ind Health 13(4):407-484.
Brown-Woodman, PDC; Hayes, LC; Huq, F; et al. (1998) In vitro assessment of the effect of halogenated
hydrocarbons: chloroform, dichloromethane, and dibromoethane on embryonic development of the rat.
Teratology 57:321-333.
Bruhn, C; Brockmoller, J; Kerb, R; et al. (1998) Concordance between enzyme activity and genotype of
glutathione S-transferase theta (GST-T1). Biochem Pharmacol 56:1189-1193.
326

-------
Brzezinski, MR; Boutelet-Bochan, H; Person, RE; et al. (1999) Catalytic activity and quantitation of
cytochrome P-450 2E1 in prenatal human brain. J Pharmacol Exp Ther 289:1648-1653.
Burek, JD; Nitschke, KD; Bell, TJ; et al. (1984) Methylene chloride: a two-year inhalation toxicity and
oncogenicity study in rats and hamsters. Fundam Appl Toxicol 4:30-47.
Cagiano, R; Ancona, D; Cassano, T; et al. (1998) Effects of prenatal exposure to low concentrations of
carbon monoxide on sexual behaviour and mesolimbic dopaminergic function in rat offspring. Br J
Pharmacol 125(4):909-915.
Callen, DF; Wolf, CR; Philpot, RM. (1980) Cytochrome P-450 mediated genetic activity and cytotoxicity
of seven halogenated aliphatic hydrocarbons in Saccharomyces cerevisiae. Mutat Res 77:55-63.
Candrian, U; You, M; Goodrow, T; Maronpot, RR; Reynolds, SH; Anderson, MW (1991). Activation of
protooncogenes in spontaneously occurring non-liver tumors from C57BL/6 x C3H F1 mice. Cancer Res
51:1148-1153.
Cantor, KP; Stewart, PA; Brinton, LA; et al. (1995) Occupational exposures and female breast cancer
mortality in the United States. J Occup Environ Med 37:336-348.
Carlsson, A; Hultengren, M. (1975) Exposure to methylene chloride. III. Metabolism of 14C-labelled
methylene chloride in rat. Scand J Work Environ Health 1:104-108.
Casanova, M; Deyo, DF; Heck, Hd'A. (1992) Dichloromethane (methylene chloride): metabolism to
formaldehyde and formation of DNA-protein cross links in B6C3Fi mice and Syrian golden hamsters.
Toxicol Appl Pharmacol 114:162-165.
Casanova, M; Conolly, RB; Heck, Hd'A. (1996) DNA-protein cross-links (DPX) and cell proliferation in
B6C3Fi mice but not Syrian golden hamsters exposed to dichloromethane: pharmacokinetics and risk
assessment with DPX as dosimeter. Fundam Appl Toxicol 31:103-106.
Casanova, M; Bell, DA; Heck, Hd'A. (1997) Dichloromethane metabolism to formaldehyde and reaction of
formaldehyde with nucleic acids in hepatocytes of rodents and humans with and without glutathione
S-transferase T1 and Ml genes. Fundam Appl Toxicol 37:168-180.
Chang, Y-L; Yang, CC; Deng, JF; et al. (1999) Diverse manifestations of oral methylene dichloride
poisoning: report of 6 cases. Clin Toxicol 37(4):497-504.
Cherry, N; Venables, H; Waldron, HA; et al. (1981) Some observations on workers exposed to methylene
chloride. Br J Ind Med 38:351-355.
Cherry, N; Venables, H; Waldron, HA. (1983) The acute behavioural effects of solvent exposure. J Soc
Occup Med 33:13-18.
Clewell, HJ, III; Gearhart, JM; Andersen, ME. (1993) Analysis of the metabolism of methylene chloride in
the B6C3Fi mouse and its implications for human carcinogenic risk. Submitted to Occupational Safety and
Health Administration, U.S. Department of Labor, Washington, DC. Docket # H-071, Exhibit #96.
Clewell, HJ; Gentry, PR; Covington, TR; et al. (2004) Evaluation of the potential impact of age- and
gender-specific pharmacokinetic differences on tissue dosimetry. Toxicol Sci 79:381-393.
CMR (Chemical Marketing Reporter). (1973) Chemical profile on methylene chloride. Chemical
Marketing Reporter, October 22, 1973.
327

-------
CMR (Chemical Marketing Reporter). (1979) Chemical profile on methylene chloride. Chemical
Marketing Reporter, August 6, 1979.
CMR (Chemical Marketing Reporter). (1982) Chemical profile on methylene chloride. Chemical
Marketing Reporter, July 12, 1982.
CMR (Chemical Marketing Reporter). (1997) Chemical profile on methylene chloride. Chemical
Marketing Reporter, November 24, 1997.
CMR (Chemical Marketing Reporter). (2000) Chemical profile on methylene chloride. Chemical
Marketing Reporter, October 9, 2000.
Cocco, P; Heineman, EF; Dosemeci, M. (1999) Occupational risk factors for cancer of the central nervous
system (CNS) among US women. Am J Ind Med 36:70-74.
Cohen J. (1987) Statistical power analysis for the behavioral sciences. Hillsdale NJ: Lawrence Erlbaum
Associates; p. 36-37.
Condie, LW; Smallwood, CL; Laurie, RD. (1983) Comparative renal and hepatotoxicity of halomethanes:
bromodichloromethane, bromoform, chloroform, dibromochloromethane, and methylene chloride. Drug
Chem Toxicol 6(6):563-578.
Crebelli, R; Benigni, R; Franekic, J; et al. (1988) Induction of chromosome malsegregation by halogenated
organic solvents in Aspergillus nidulans: unspecific or specific mechanism? Mutat Res 201:401-411.
Crebelli, R; Carere, A; Leopardi, P; et al. (1999) Evaluation of 10 aliphatic halogenated hydrocarbons in
the mouse bone marrow micronucleus test. Mutagenesis 14(2):207-215.
Dankovic, DA; Bailer, AJ. (1994) The impact of exercise and intersubject variability on dose estimates for
dichloromethane derived from a physiologically based pharmacokinetic model. Fundam Appl Toxicol
22:20-25.
David, RM; Clewell, HJ; Gentry, PR; et al. (2006) Revised assessment of cancer risk to dichloromethane II.
Application of probabilistic methods to cancer risk determinations. Regul Toxicol Pharmacol 45:55-65.
De Salvia, MA; Cagiano, R; Carratu, MR; et al. (1995) Irreversible impairment of active avoidance
behavior in rats prenatally exposed to mild concentrations of carbon monoxide. Psychopharmacology
(Berl) 122(1):66-71.
Dearfield, KL; Moore, MM. (2005) Use of genetic toxicology information for risk assessment. Environ
Mol Mutagen 46(4):236-245.
DeMarini, DM; Shelton, ML; Warren, SH; et al. (1997) Glutathione S-transferase-mediated induction of
GC—>AT transitions by halomethanes in Salmonella. Environ Mol Mutagen 30:440-447.
Devereux, TR; Foley, JF; Maronpot, RR; et al. (1993) Ras proto-oncogene activation in liver and lung
tumors fromB6C3Fi mice exposed chronically to methylene chloride. Carcinogenesis 14(5):795-801.
Dillon, D; Edwards, I; Combes, R; et al. (1992) The role of glutathione in the bacterial mutagenicity of
vapour phase dichloromethane. Environ Mol Mutagen 20:211-217.
Dinse GE; Lagakos SW. (1982) Nonparametric estimation of lifetime and disease onset distributions from
incomplete observations. Biometrics. 38(4):921-932.
328

-------
DiVincenzo, GD; Kaplan, CJ. (1981) Uptake, metabolism, and elimination of methylene chloride vapor by
humans. Toxicol Appl Pharmacol 59:130-140.
DiVincenzo, GD; Yanno, FJ; Astill, BD. (1971) The gas chromatographic analysis of methylene chloride in
breath, blood, and urine. Am Ind Hyg Assoc J 32:387-391.
DiVincenzo, GD; Yanno, FJ; Astill, BD. (1972) Human and canine exposures to methylene chloride vapor.
Am Ind Hyg Assoc J 33:125-134.
Doherty, AT; Ellard, S; Parry, EM; et al. (1996) An investigation into the activation and deactivation of
chlorinated hydrocarbons to genotoxins in metabolically competent human cells. Mutagenesis 11(3):247-
274.
Dosemeci, M; Cocco, P; Chow, WH. (1999) Gender differences in risk of renal cell carcinoma and
occupational exposures to chlorinated aliphatic hydrocarbons. Am J Ind Med 36:54-59.
Dumas, S; Parent, ME; Siemiatychi, J. (2000) Rectal cancer and occupational risk factors: a hypothesis-
generating, exposure-based case-control study. Int J Cancer 87:874-879.
Eastmond, DA; Hartwig, A; Anderson, D. et al. (2009). Mutagenicity testing for chemical risk assessment:
update of WHO/IPCS harmonized scheme. Mutagenesis 24:341-349.
Ehrlich, R. (1980) Interaction between environmental pollutants and respiratory infections. Environ Health
Perspect 35:89-100
El-Masri, HA; Bell, DA; Portier, CJ. (1999) Effects of glutathione transferase theta polymorphism on the
risk estimates of dichloromethane to humans. Toxicol Appl Pharmacol 158:221-230.
El-Rayes, BF; Ali, S; Heilbrun, LK; et al. (2003). Cytochrome P450 and glutathione transferase expression
in human breast cancer. Clin Cancer Res 9(5): 1705-1709.
Engstrom, J; Bjurstrom, R. (1977) Exposure to methylene chloride. Content in subcutaneous adipose tissue.
Scand J Work Environ Health 3:215-224.
Estill, CF; Spencer, AB. (1996) Case study: control of methylene chloride exposures during furniture
stripping. Am Ind Hyg Assoc J 57:43-49.
Fechner, G; Ortmann, C; Du Chesne, A; et al. (2001) Fatal intoxication due to excessive dichloromethane
inhalation. Forensic Scilnt 122:69-72.
Fechter, LD (1987) Neurotoxicity of prenatal carbon monoxide exposure. Res Rep Health Eff Inst. 12: 3-
22.
Fodor, GG; Prajsnar, D; Schlipkoter, H-W. (1973) Endogenous CO formation by incorporated halogenated
hydrocarbons of the methane series. Staub Reinhalt Luft 3 3:260-261.
Foley, JF; Tuck, PD; Ton, T-VT; et al. (1993) Inhalation exposure to a hepatocarcinogenic concentration of
methylene chloride does not induce sustained replicative DNA synthesis in hepatocytes of female B6C3Fi
mice. Carcinogenesis 14(5)811-817.
Forster, HV; Graff, S; Hake, CL; et al. (1974) Pulmonary-hematologic studies on humans during exposure
to methylene chloride. Prepared by the Department of Environmental Medicine, Medical College of
Wisconsin, Milwaukee, WI, for the National Institute for Occupational Safety and Health, Centers for
Disease Control and Prevention, U.S. Department of Health and Human Services, Cincinnati, OH.
Available from the National Technical Information Service, Springfield, VA; PB82-151697.
329

-------
Foster, JR; Green, T; Smith, LL; et al. (1992) Methylene chloride-an inhalation study to investigate
pathological and biochemical events occurring in the lungs of mice over an exposure period of 90 days.
Fundam Appl Toxicol 18:376-388.
Foster, JR; Green, T; Smith, LL; et al. (1994) Methylene chloride: an inhalation study to investigate
toxicity in the mouse lung using morphological, biochemical, and Clara cell culture techniques.
Toxicology 91:221-234.
Friedlander, BR; Hearne, T; Hall, S. (1978) Epidemiologic investigation of employees chronically exposed
to methylene chloride. J Occup Med 20:657-666.
Fujimoto, K; Arakawa, S; Watanabe, T; et al. (2007) Generation and functional characterization of mice
with a disrupted glutathione S-transferase, theta 1 gene. Drug Metab Dispos. 35:2196-2202.
Fuxe, K; Andersson, K; Hansson, T; et al. (1984) Central catecholamine neurons and exposure to
dichloromethane. Selective changes in amine levels and turnover in tel- and diencephalic and Na nerve
terminal systems and in the secretion of anterior pituitary hormone in the male rat. Toxicol 29:293-305.
Gamberale, F; Annwall, G; Hultengren, M (1975) Exposure to methylene chloride. II. Psychological
function. Scand J Work Environ Health 1:95-103.
Gargas, ML; Clewell, HJ; Andersen, ME. (1986) Metabolism of inhaled dihalomethanes in vivo:
differentiation of kinetic constants for two independent pathways. Toxicol Appl Pharmacol 82:211-223.
Garrett, NE; Lewtas, J. (1983) Cellular toxicity in Chinese hamster ovary cells culture. I. Analysis of
cytotoxicity endpoints for twenty-nine priority pollutants. Environ Res 32:455-465.
Garte, S; Gaspari, L; Alexandrie, AK; et al. (2001) Metabolic gene polymorphism frequencies in control
populations. Cancer Epidemiol Biomarkers Prev 10:1239-1248.
General Electric Co. (1976) Dichloromethane and ninety day oral toxicity study in rats. Prepared by the
International Research and Development Corporation, Mattawan, MI for the Plastics Tech Department,
General Electric Co., Pittsfield, MA. Submitted under TSCA Section 8D; EPA Document No. 86-
878210707; NTIS No. OTS0205887.
Gibbs, GW. (1992) The mortality of workers employed at a cellulose acetate and triacetate fibers plant in
Cumberland, Maryland: a "1970" cohort followed 1970-1989 [final report]. Prepared by Safety Health
Environment International Consultants Corporation, Winterburn, Alberta, Canada, for the Hoechst
Celanese Corporation, Somerville, NJ.
Gibbs, GW; Amsel, J; Soden, K. (1996) A cohort mortality study of cellulose triacetate-fiber workers
exposed to methylene chloride. J Occup Environ Med 38(7):693-697.
Giustino, A; Cagiano, R; Carratu, MR; et al. (1999) Prenatal exposure to low concentrations of carbon
monoxide alters habituation and non-spatial working memory in rat offspring. Brain Res 844(1-2):201-
205.
Glatzel, W; Tietze, K; Gutewort, R; et al. (1987) Interaction of dichloromethane and ethanol in rats:
toxicokinetics and nerve conduction velocity. Alcoholism: Clin Exp Res 11:450-455.
Gocke, E; King, MT; Eckhardt, K; et al. (1981) Mutagenicity of cosmetic ingredients licensed by the
European communities. MutatRes 90:91-109.
Gomez, MR. (1996) Exposure determinants needed to improve the assessment of exposure. Am J Ind Med
29:569-570.
330

-------
Gomez, MR; Cocco, P; Dosemeci, M; et al. (1994) Occupational exposure to chlorinated aliphatic
hydrocarbons: job exposure matrix. Am J Ind Med 26:171-183.
Goodman, DG; Maronpot RR; Newbene PM, Popp JA; Squire RA. (1994) Proliferative and selected other
lesions in the liver of rats. Gl-5. In: Guides for Toxologic Pathology. STP/ARP/AFIP, Washington DC: 1-
24.
Goulle, JP; Lacroix, C; Vaz, E; et al. (1999) Fatal case of dichloromethane poisoning. J Anal Toxicol
23:380-383.
Graves, RJ; Green, T. (1996) Mouse liver glutathione S-transferase mediated metabolism of methylene
chloride to a mutagen in the CHO/HPRT assay. Mutat Res 367:143-150.
Graves, RJ; Callander, RD; Green, T. (1994a) The role of formaldehyde and S-chloromethylglutathione in
the bacterial mutagenicity of methylene chloride. Mutat Res 320:235-243.
Graves, RJ; Coutts, C; Eyton-Jones, H; et al. (1994b) Relationship between hepatic DNA damage and
methylene chloride-induced hepatocarcinogenicity inB6C3Fi mice. Carcinogenesis 15(5):991-996.
Graves, RJ; Coutts, C; Green, T. (1995) Methylene chloride-induced DNA damage: an interspecies
comparison. Carcinogenesis 16(8): 1919-1926.
Graves, RJ; Trueman, P; Jones, S; et al. (1996) DNA sequence analysis of methylene chloride-induced
HPRT mutations in Chinese hamster ovary cells: comparison with the mutation spectrum obtained for
1,2-dibromoethane and formaldehyde. Mutagenesis ll(3):229-233.
Green, T. (1983) The metabolic activation of dichloromethane and chlorofluoromethane in a bacterial
mutation assay using Salmonella typhimurium. Mutat Res 118:277-288.
Green, T. (1989) A biological database for methylene chloride risk assessment. In: Travis, CC; ed.
Biologically based methods for cancer risk assessment. New York, NY: Plenum Press, p. 289-300.
Green, T. (1997) Methylene chloride induced mouse liver and lung tumours: an overview of the role of
mechanistic studies in human safety assessment. Hum Exp Toxicol 16:3-13.
Guengerich, FP. (1997) Mechanisms of mutagenicity of DNA adducts derived from alkyl and vinyl halides.
Jpn J Toxicol Environ Health 43:69-82.
Haber, LT; Maier, A; Gentry, PR; et al. (2002) Genetic polymorphisms in assessing interindividual
variability in delivered dose. Reg Toxicol Pharmacol 35:177-197.
Hall, AH; Rumack, BH. (1990) Methylene chloride exposure in furniture-stripping shops: ventilation and
respirator use practices. J Occup Med 32:33-37.
Hallier, E; Schroder, KR; Asmuth, K; et al. (1994) Metabolism of dichloromethane (methylene chloride) to
formaldehyde in human erythrocytes: influence of polymorphism of glutathione transferase theta (GST Tl-
1). Arch Toxicol 68:423-427.
Hansch, C; Leo, A; Hoekman, D. (1995) Exploring QSAR. Hydrophobic, electronic, and steric constants.
ACS professional reference book. Washington, DC: American Chemical Society; p. 3.
Hardie, DWF. (1964) Methylene chloride. In: Kirk-Othmer encyclopedia of chemical technology. 2nd
edition. New York, NY: John Wiley & Sons; pp. 111-119.
331

-------
Hardin, BD; Manson, J. (1980) Absence of dichloromethane teratogenicity with inhalation exposure in rats.
Toxicol Appl Pharmacol 52:22-28.
Hashmi, M; Dechert, S; Dekant, W; et al. (1994) Bioactivation of [13C]dichloromethane in mouse, rat, and
human liver cytosol: 13C nuclear magnetic resonance spectroscopic studies. Chem Res Toxicol 7:291-296.
Haufroid, V; Ligocka, D; Buysschaert, M; et al. (2003) Cytochrome P4502E1 (CYP2E1) expression in
peripheral blood lymphocytes: evaluation in hepatitis C and diabetes. Eur J Clin Pharmacol 59:29-33.
Haun, CC; Harris, ES; Darmer, KI, Jr. (1971) Continuous animal exposure to methylene chloride. In:
Proceedings of the 2nd conference of environmental toxicology; August 31-September 2; Fairborn, OH;
Paper No. 10; AMRL-TR-71-120. Aerospace Medical Research Laboratory, Wright-Patterson Air Force
Base, OH; pp. 125-135. Available from the National Technical Information Service, Springfield, VA;
AD751432 (individual paper); AD746660 (entire proceedings).
Haun, CC; Vernot, EH; Darmer, KI, Jr; et al. (1972) Continuous animal exposure to low levels of
dichloromethane. Proceedings of the 3rd conference of environmental toxicology; October 25-27;
Fairborn, OH; Paper No. 12; AMRL-TR-72-120. Aerospace Medical Research Laboratory, Wright-
Patterson Air Force Base, OH; pp. 199-208. Available from the National Technical Information Service,
Springfield, VA; AD773766.
Hazelton Laboratories. (1983) 24-Month oncogenicity study of methylene chloride in mice [final report].
Performed by Hazelton Laboratories America, Inc., Vienna, VA for the National Coffee Association, New
York, NY; Submitted under TSCA Section 4; EPA Document No. 45-8303005; NTIS No. OTS0206132.
Hearne, FT; Pifer, JW. (1999) Mortality study of two overlapping cohorts of photographic film base
manufacturing employees exposed to methylene chloride. J Occup Environ Med 41(12): 1154-1169.
Hearne, FT; Grose, F; Pifer, JW; et al. (1987) Methylene chloride mortality study: dose-response
characterization and animal model comparison. J Occup Med 29(3):217-228.
Hearne, FT; Pifer, JW; Grose, F. (1990) Absence of adverse mortality effects in workers exposed to
methylene chloride: an update. J Occup Med 32(3):234-240.
Hegi, ME; Soderkvist, P; Foley, JF; et al. (1993) Characterization of p53 mutations in methylene chloride-
induced lung tumors from B6C3Fi mice. Carcinogenesis 14(5):803-810.
Heineman, EF; Cocco, P; Gomez, MR; et al. (1994) Occupational exposure to chlorinated aliphatic
hydrocarbons and risk of astrocytic brain cancer. Am J Ind Med 26:155-169.
Heineman, EF; Gomez, MR; Dosemeci, M; et al. (1996) Methylene chloride and brain cancer: interpreting
a new study in light of existing literature. Am J Ind Med 30:506-507.
Hellmold, H; Rylander, T; Magnusson, M; et al. (1998) Characterization of cytochrome P450 enzymes in
human breast tissue from reduction mammaplasties. J Clin Endocrinol Metab 83(3):886-895.
Heppel, LA; Neal, PA. (1944) Toxicology of dichloromethane (methylene chloride). II. Its effect upon
running activity in the male rat. J Ind Hyg Toxicol 26(1): 17-21.
Heppel, LA; Neal, PA; Perrin, TL; et al. (1944) Toxicology of dichloromethane (methylene chloride). J
Ind Hyg Toxicol 26:8-16.
Herr, DW; Boyes, WK. (1997) A comparison of the acute neuroactive effects of dichloromethane, 1,3-
dichloropropane, and 1,2-dichlorobenzene on rat flash evokes potentials (FEPs). Fundam Appl Toxicol
35:31-48.
332

-------
Hines, RN. (2007) Ontogeny of human hepatic cytochromes P450. J Biochem Mol Toxicol 21:169-75.
Holbrook, MT. (2003) Methylene chloride. In: Kirk-Othmer encyclopedia of chemical technology. New
York, NY: John Wiley & Sons. Available online at
http://www.mrw.interscience.wiley.com/kirk/kirk_search_fs.html. (subscription required).
Horvath, AL. (1982) Halogenated hydrocarbons: solubility-miscibility with water. New York, NY: Marcel
Dekker, Inc., p. 543.
Hu, Y; Kabbler, SL; Tennant, AH et al. (2006) Induction of DNA-protein crosslinks by dichloromethane in
a V79 cell line transfected with the murine glutathione-S-transferase theta 1 gene. MutationRes 607:231-
239.
Hughes, NJ; Tracey, JA. (1993) A case of methylene chloride (nitromors) poisoning, effects on
carboxyhaemoglobin levels. Hum Exp Toxicol 12:159-160.
IARC (International Agency for Research on Cancer). (1999) Dichloromethane. IARC monographs on the
evaluation of carcinogenic risk of chemicals to humans. Vol. 71. Re-evaluation of some organic chemicals,
hydrazine and hydrogen peroxide. Lyon, France: International Agency for Research on Cancer, pp. 251-
315. Available online at http://inchem.org/documents/iarc/vol71/004-dichloromethane.html.
Infante-Rivard, C; Siemiatycki, J; Lakhani, R; et al. (2005) Maternal exposure to occupational solvents and
childhood leukemia. Environ Health Perspect 113:787-792.
Ingelman-Sundberg, M. (2004) Human drug metabolising cytochrome P450 enzymes: properties and
polymorphisms. Naunyn-Schmiedeberg's Arch Pharmacol 369:89-104.
Jacubovich, RM; Landau, D; Dayan, YB; et al. (2005) Facial nerve palsy after acute exposure to
dichloromethane. Am J Ind Med 48:389-392.
Johnsrud, EK; Koukouritaki, SB; Divakaran, K; et al. (2003) Human hepatic CYP2E1 expression during
development. J Pharmacol Exp Ther 307:402-407.
Jongen, WMF; Alink, GM; Koeman, JH. (1978) Mutagenic effect of dichloromethane on Salmonella
typhimurium. MutatRes 56:245-248.
Jongen, WMF; Lohman, PHM; Kottenhagen, MJ; et al. (1981) Mutagenicity testing of dichloromethane in
short-term mammalian test systems. MutatRes 81:203-213.
Jongen, WMF; Harmsen, EGM; Alink, GM; et al. (1982) The effect of glutathione conjugation and
microsomal oxidation on the mutagenicity of dichloromethane inX typhimurium. MutatRes 95:183-189.
Jonsson, F; Johanson, G. (2001) A Bayesian analysis of the influence of GST-T1 polymorphism on the
cancer risk estimate for dichloromethane. Toxicol Appl Pharmacol 174:99-112.
Jonsson, F; Johanson, G. (2003) The Bayesian population approach to physiological toxicokinetic-
toxicodynamic models—an example using the MCSim software. Toxicol Lett 138:143-150.
Jonsson, F; Bois, F; Johanson, G. (2001) Physiologically based pharmacokinetic modeling of inhalation
exposure of humans to dichloromethane during moderate to heavy exercise. Toxicol Sci 59:209-218.
Juronen. E: Tasa. G: Uuskiila, M;, Pooga. M: Mikelsaar. AY (1996). Purification, characterization and
tissue distribution of human class theta glutathione S-transfera.se Tl-1. Biochem Mol Biol Int. 39:21-9
333

-------
Kanada, M; Miyagawa, M; Sato, M; et al. (1994) Neurochemical profile of effects of 28 neurotoxic
chemicals on the central nervous system in rats (1). Effects of oral administration on brain contents of
biogenic amines and metabolites. Ind Health 32:145-164.
Kanno, J; Foley, JF; Kari, F; et al. (1993) Effect of methylene chloride inhalation of replicative DNA
synthesis in the lungs of female B6C3Fi mice. Environ Health Perspect 101(Suppl. 5):271-276.
Kari, FW; Foley, JF; Seilkop, SK; et al. (1993) Effect of varying exposure regimens on methylene chloride-
induced lung and liver tumors in female B6C3Fi mice. Carcinogenesis 14(5):819-826.
Karlsson, JE; Rosengren, LE; Kjellstrand, P; et al. (1987) Effects of low-dose inhalation of three
chlorinated aliphatic organic solvents on deoxyribonucleic acid in gerbil brain. Scand J Work Environ
Health 13:453-458.
Kayser, MF; Vuilleumier, S. (2001) Dehalogenation of Dichloromethane by Dichloromethane
Dehalogenase/Glutathione S-Transferase Leads to Formation of DNA Adducts. J Bacteriology 183:5209-
5212.
Kelly, M. (1988) Case reports of individuals with oligospermia and methylene chloride exposures. Reprod
Toxicol 2:13-17.
Kernan, GJ; Ji, BT; Dosemeci, M; et al. (1999) Occupational risk factors for pancreatic cancer: a case-
control study based on death certificates from 24 US states. Am J Ind Med 36:260-270.
Kim, SK; Kim YC. (1996) Effect of a single administration of benzene, toluene or m-xylene on
carboxyhaemoglobin elevation and metabolism of dichloromethane in rats. J Appl Toxicol 16:437-444.
Kim, RB; O'Shea, D; Wilkinson, GR. (1995) Interindividual variability of chlorzoxazone 6-hydroxylation
in men and women and its relationship to CYP2E1 genetic polymorphisms. Clin Pharmacol Ther
57(6):645-655.
Kim, NY; Park, SW; Suh, JK. (1996) Two fatal cases of dichloromethane or chloroform poisoning. J
Forensic Sci 41:527-529.
Kirschman, JC; Brown, NM; Coots, RH; et al. (1986) Review of investigations of dichloromethane
metabolism and subchronic oral toxicity as the basis for the design of chronic oral studies in rats and mice.
Food Chem Toxicol 24(9):943-949.
Kitchin, KT; Brown, JL. (1989) Biochemical effects of three carcinogenic chlorinated methanes in rat liver.
Teratog Carcinog Mutagen 9:61-69.
Kiellstrand, P; Holmquist, B; Jonsson, I; et al. (1985) Effects of organic solvents on motor activity in mice.
Toxicol 35:35-46.
Kolodner, K; Cameron, L; Gittlesohn, A; et al. (1990) Morbidity study of occupational exposure to
methylene chloride using a computerized surveillance system. Center of Occupational Health and
Environmental Health, Johns Hopkins School of Hygiene and Public Health, Baltimore, MD. Submitted to
TSCA under Section 8D; EPA Document No. 86-900000421; NTIS No. OTS0522984.
Korzekwa, KR; Krishnamachary, N; Shou, M; et al. (1998) Evaluation of atypical cytochrome P450
kinetics with two-substrate models: evidence that multiple substrates can simultaneously bind to
cytochrome P450 active sites Biochemistry 37:4137-4147.
334

-------
Kramers, PGN; Mout, HCA; Bissumbhar, B; et al. (1991) Inhalation exposure inDrosophila mutagenesis
assays: experiments with aliphatic halogenated hydrocarbons, with emphasis on the genetic activity profile
of 1,2-dichloroethane. Mutat Res 252:17-33.
Landi, S; Naccarati, A; Ross, MK; et al. (2003) Induction of DNA strand breaks by trihalomethanes in
primary human lung epithelial cells. Mutat Res 538:41-50.
Landry, TD; Ramsey, JC; McKenna, MJ. (1983) Pulmonary physiology and inhalation dosimetry in rats:
development of a method and two examples. Toxicol Appl Pharmacol 71:72-83.
Lanes, SF; Cohen, A; Rothman, KJ; et al. (1990) Mortality of cellulose fiber production workers. Scand J
Work Environ Health 16:247-251.
Lanes, SF; Rothman, KJ; Dreyer, NA; et al. (1993) Mortality update of cellulose fiber production workers.
Scand J Work Environ Health 19:426-428.
Lash, AA; Becker, CE; So, Y; et al. (1991) Neurotoxic effects of methylene chloride: are they long lasting
in humans? Br J Ind Med 48:418-426.
Lefevre, PA; Ashby, J. (1989) Evaluation of dichloromethane as an inducer of DNA synthesis in the
B6C3Fi mouse liver. Carcinogenesis 10(6): 1067-1072.
Lehmann, L; Wagner, J. (2008) Gene expression of 17beta-estradiol-metabolizing isozymes: comparison of
normal human mammary gland to normal human liver and to cultured human breast adenocarcinoma cells.
Adv Exp Med Biol 617:617-624.
Lehnebach, A; Kuhn, C; Pankow, D. (1995) Dichloromethane as an indicator of cytochrome c oxidase in
different tissues of rats. Arch Toxicol 69:180-184.
Leighton, DT, Jr; Calo, JM. (1981) Distribution coefficients of chlorinated hydrocarbons in dilute air-water
systems for groundwater contamination applications. J Chem Eng 26:382-385.
Leikin, JB; Kaufman, D; Lipscomb, JW; et al. (1990) Methylene chloride: report of five exposures and two
deaths. Am J Emerg Med 8:534-537.
Leitao, MM, Jr; White, P; Cracchiolo, B. (2008). Cervical cancer in patients infected with the human
immunodeficiency virus. Cancer 112:2683-2689.
Leuschner, F; Neumann, B-W; Htibscher, F. (1984) Report on subacute toxicological studies with
dichloromethane in rats and dogs by inhalation. Arzneim Forsch/Drug Res 34(12): 1772-1774.
Lewis, RJ, Sr; ed. (1997) Hawley's condensed chemical dictionary. 13th edition. New York, NY: John
Wiley & Sons, Inc., p. 736.
Lide, DR; ed. (2000) CRC handbook of chemistry and physics. 81st edition. Boca Raton, FL: CRC Press;
pp. 3-206.
Lipscomb, JC; Garrett, CM; Snawder, JE. (1997) Cytochrome P450-dependent metabolism of
trichloroethylene: interindividual differences in humans. Toxicol Appl Pharmacol 142:311-318.
Lipscomb, JC; Teuschler, LK; Swartout, J; et al. (2003) The impact of cytochrome P450 2El-dependent
metabolic variance on a risk-relevant pharmacokinetic outcome in humans. Risk Anal 23(6):1221-1238.
Lowenfels, AB; Maisonneuve, P. (2005) Risk factors for pancreatic cancer. J Cell Biochem 95:649-656.
335

-------
Lucas, D; Ferrara, R; Gonzalez, E; et al. (1999) Chlorzoxazone, a selective probe for phenotyping CYP2E1
in humans. Pharmacogenetics 9:377-388.
Lucas, D; Ferrara, R; Gozales, E; et al. (2001) Cytochrome CYP2E1 phenotyping and genotyping in the
evaluation of health risks from exposure to polluted environments. Toxicol Lett 124:71-81.
Madle, S; Dean, SW; Andrae, U; et al. (1994) Recommendations forthe performance of UDS tests in vitro
and in vivo. Mutation Res 312:263-285.
Mahmud, M; Kales, SN. (1999) Methylene chloride poisoning in a cabinet worker. Environ Health
Perspect 107:769-772.
Mainwaring, GW; Williams, SM; Foster, JR; et al. (1996) The distribution of Theta-class glutathione S-
transferase in the liver and lung of mouse, rat and human. Biochem J 318:297-303.
Mainwaring, GW; Foster, JR; Green, T. (1998) Nuclear and cellular immunolocalization of theta-class
glutathione S-transferase GST-T1-1 in the liver and lung of the mouse. Biochem J 329:431-432.
Maltoni, C; Cotti, G; Perino, G. (1988) Long-term carcinogenicity bioassays on methylene chloride
administered by ingestion to Sprague-Dawley rats and Swiss mice and by inhalation to Sprague-Dawley
rats. Ann NY Acad Sci 534:352-366.
Manno, M; Rugge, M; Cocheo, V. (1992) Double fatal inhalation of dichloromethane. Hum Exp Toxicol
11:540-545.
Marino, DJ; Clewell, HJ; Gentry, PR; et al. (2006) Revised assessment of cancer risk to dichloromethane:
part I Bayesian PBPK and dose-response modeling in mice. Regul Toxicol Pharmacol 45:44-54.
Maronpot, RR; Devereux, TR; Hegi, M; et al. (1995) Hepatic and pulmonary carcinogenicity of methylene
chloride in mice: a search for mechanisms. Toxicol 102:73-81.
Marriott, HM; Dockrell, DH. (2007) The role of the macrophage in lung disease mediated by bacteria. Exp
Lung Res 33:493-505.
Marsch, GA; Mundkowski, R; Morris, BKJ; et al. (2001) Characterization of Nucleoside and DNA
Adducts Formed by S-(l-Acetoxymethyl)glutathione and Implications for Dihalomethane-Glutathione
Conjugates ChemRes Toxicol 14:600-608.
Marsch, GA; Botta, S; Martin, MV; et al. (2004) Formation and mass spectrometric analysis of DNA and
nucleoside adducts by S-(l-acetoxymethyl)glutathione and by glutathione S-transferase-mediated
activation of dihalomethanes. Chem Res Toxicol 17:45-54.
Mathews, JM; Raymer, JH; Etheridge, AS; Velez, GR; Bucher, JR (1997) Do endogenous volatile organic
chemicals measured in breath reflect and maintain CYP2E1 levels in vivo? Toxicol Appl Pharmacol
146:255-260.
Mattsson, JL; Albee, RR; Eisenbrandt, DL. (1990) Neurotoxicologic evaluation of rats after 13 weeks of
inhalation exposure to dichloromethane or carbon monoxide. Pharmacol Biochem Behav 36:671-681.
McKenna, MJ; Zempel, JA. (1981) The dose-dependent metabolism of [14C]methylene chloride following
oral administration to rats. Food Cosmet Toxicol 19:73-78.
336

-------
McKenna, MJ; Zempel, JA; Braun, WH. (1982) The pharmacokinetics of inhaled methylene chloride in
rats. Toxicol Appl Pharmacol 65:1-10.
Mendoza-Cantu, A; Castorena-Torres, F; et al. (2004) Genotype and allele frequencies of polymorphic
cytochromes P450 CYP1A2 and CYP2E1 in Mexicans. Cell Biochem Funct 22:29-34.
Mennear, JH; McConnell, EE; Huff, JE; et al. (1988) Inhalation and carcinogenesis studies of methylene
chloride (dichloromethane) inF344/nrats andB6C3Fi mice. Ann NY Acad Sci 534:343-351.
Meyer, DJ; Coles, B; Pemble, SE; et al. (1991) Theta, a new class of glutathione transferases purified from
rat and man. Biochem J 274:409-414.
Miksys, S; Tyndale, RF. (2004) The unique regulation of brain cytochrome P450 2 (CYP2) family enzymes
by drugs and genetics. Drug Metab Rev 36(2):313-333.
Mirsalis, JC; Tyson, CK; Steinmetz, KL; et al. (1989) Measurement of unscheduled DNA synthesis and s-
phase synthesis in rodent hepatocytes following in vivo treatment: testing of 24 compounds. Environ Mol
Mutagen 14:155-164.
Moser, VC; Cheek, BM; MacPhail, RC. (1995) A multidisciplinary approach to toxicological screening:
III. Neurobehavioral toxicity. J Toxicol Environ Health 45:173-210.
Narotsky, MG; Kavlock, RJ. (1995) A multidisciplinary approach to toxicological screening: II.
Developmental toxicity. J Toxicol Environ Health 45:145-171.
Nelson, HH; Wiencke, JK; Christiani, DC; et al. (1995) Ethnic differences in the prevalence of the
homozygous deleted genotype of glutathione S-transferase theta. Carcinogenesis 16(5): 1243-1245.
NIOSH (National Institute of Occupational Safety and Health). (1986) Methylene chloride. Current
intelligence bulletin 46. National Institute of Occupational Safety and Health, Centers for Disease Control
and Prevention, U.S. Department of Health and Human Services, Cincinnati, OH; DHHS (NIOSH)
Publication No. 86-114. Available online at: http://www.cdc.gov/niosh/86114_46.html (accessed June 29,
2006).
Nishimura, M; Yaguti, H; Yoshitsugu, H. (2003) Tissue distribution of mRNA expression of human
cytochrome P450 isoforms assessed by high-sensitivity real-time reverse transcription PCR. Yakugaku
Zasshi 123(5):369-375.
Nitschke, KD; Burek, JD; Bell, TJ; et al. (1988a) Methylene chloride: a 2-year inhalation toxicity and
oncogenicity study in rats. Fundam Appl Toxicol 11:48-59.
Nitschke, KD; Eisenbrandt, DL; Lomax, LG; et al. (1988b) Methylene chloride: two-generation inhalation
reproductive study in rats. Fundam Appl Toxicol 11:60-67.
NLM (National Library of Medicine). (2003) Dichloromethane. HSDB (Hazardous Substances Data
Bank). National Library of Medicine, National Institutes of Health, U.S. Department of Health and Human
Services, Bethesda, Maryland. Available online at http://toxnet.nlm.nih.gov/.
Norman, WC, III; Boggs, P. (1996) Flawed estimates of methylene chloride exposures. Am J Ind Med
30:504-509.
NRC (National Research Council). (1983) Risk assessment in the federal government: managing the
process. Washington, DC: National Academy Press.
337

-------
NRC (National Research Council). (1994) Science and judgment in risk assessment. Washington, DC:
National Academy Press.
NTP (National Toxicology Program). (1986) Toxicology and carcinogenesis studies of dichloromethane
(methylene chloride) (CAS No. 75-09-2) in F344/N rats and B6C3F1 mice (inhalation studies). Public
Health Service, U.S. Department of Health and Human Services; NTP TR 306. Available from the National
Institute of Environmental Health Sciences, Research Triangle Park, NC. Available online at
http://ntp.niehs.nih.gov/ntp/htdocs/LT_rpts/tr306.pdf
Oda, Y; Yamazaki, H; Thier, R; et al. (1996) A new Salmonella typhimurium NM5004 strain expressing rat
glutathione S-transferase 5-5: use in detection of genotoxicity of dihaloalkanes using an SOSlumu test
system. Carcinogenesis 17:297-302.
Ognenovski, VM; Marder, W; Somers EC; et al. (2004) Increased incidence of cervical intraepithelial
neoplasia in women with systemic lupus erythematosus treated with intravenous cyclophosphamide. J
Rheumatol 31:1763-1767.
Oh, SJ; Kim, SK; Kim, YC. (2002) Role of glutathione in metabolic degradation of dichloromethane in
rats. Toxicol Lett 129:107-114.
Ojajarvi, A; Partanen, T; Ahlbom, A; et al. (2001) Risk of pancreatic cancer in workers exposed to
chlorinated hydrocarbon solvents and related compounds: a meta-analysis. Am J Epidemiol 153:841-850.
O'Neil, MJ; Smith, A; Heckelman, PE; et al. (2001) Methylene chloride. The Merck index: an
encyclopedia of chemicals, drugs, and biologicals. Whitehouse Station, NJ: Merck & Co., Inc.; p. 1082.
Osterman-Golkar, S; Hussain, S; Walles, S; et al. (1983) Chemical reactivity and mutagenicity of some
dihalomethanes. ChemBiolInteract46:121-130.
OSHA (Occupational Safety and Health Administration). (1997) Occupational exposure to methylene
chloride. Federal Register 63(7): 1383-1658.
Ott, MG; Skory, LK; Holder, BB; et al. (1983a) Health evaluation of employees occupationally exposed to
7methylene chloride: general study design and environmental considerations. Scand J Work Environ
Health 9(Suppl. 1): 1—7.
Ott, MG; Skory, LK; Holder, BB; et al. (1983b) Health evaluation of employees occupationally exposed to
methylene chloride: mortality. Scand J Work Environ Health 9(Suppl. 1):8—16.
Ott, MG; Skory, LK; Holder, BB; et al. (1983c) Health evaluation of employees occupationally exposed to
methylene chloride: clinical laboratory evaluation. Scand J Work Environ Health 9(Suppl. 1): 17-25.
Ott, MG; Skory, LK; Holder, BB; et al. (1983d) Health evaluation of employees occupationally exposed to
methylene chloride: twenty-four hour electrocardiographic monitoring. Scand J Work Environ Health
9(Suppl. l):26-30.
Ott, MG; Skory, LK; Holder, BB; et al. (1983e) Health evaluation of employees occupationally exposed to
methylene chloride: metabolism data and oxygen half-saturation pressures. Scand J Work Environ Health
9(Suppl. 1):31—38.
Ott, MG; Carlo, GL; Steinberg, S; et al. (1985) Mortality among employees engaged in chemical
manufacturing and related activities. Am J Epidemiol 122:311-322.
Pankow, D. (1988) Enhancement of dichloromethane-induced carboxyhemoglobinemia by isoniazid
pretreatment. Biomed Biochem Act 3:293-295.
338

-------
Pankow, D; Hoffmann, P. (1989) Dichloromethane metabolism to carbon monoxide can be induced by
isoniazid, acetone and fasting. Arch Toxicol Suppl 13:302-303.
Pankow, D; Jagielki, S. (1993) Effect of methanol or modifications of the hepatic glutathione concentration
on the metabolism of dichloromethane to carbon monoxide in rats. Hum Exp Toxicol 12:227-231.
Pankow, D; Kretschmer, S; Weise, M. (1991a) Effect of pyrazole on dichloromethane metabolism to
carbon monoxide. Arch Toxicol Suppl 14:246-248.
Pankow, D; Matschiner, F; Weigmann, H-J. (1991b) Influence of aromatic hydrocarbons on the
metabolism of dichloromethane to carbon monoxide in rats. Toxicol 68:89-100.
Pegram, RA; Andersen, ME; Warren, SH; et al. (1997) Glutathione S-transferase-mediated mutagenicity of
trihalomethanes in Salmonella typhimurium: contrasting results with bromodichloromethane and
chloroform. Toxicol Appl Pharmacol 144:183-188.
Perocco, P; Prodi, G. (1981) DNA damage by haloalkanes in human lymphocytes cultured in vitro. Cancer
Lett 13:213-218.
Putz, VR; Johnson BL; Setzer, JV. (1979) A comparative study of the effects of carbon monoxide and
methylene chloride on human performance. J Environ Pathol Toxicol 2:97-112.
Quondamatteo, F; Schulz, TG; Bunzel, N; et al. (1998) Immunohistochemical localization of glutathione
S-transferase-Tl in murine kidney, liver, and lung. Histochem Cell Biol 110:417-423.
Raimondi, S; Paracchini, V; Autrup, H; et al. (2006) Human genome epidemiology (HuGE) review. Meta-
and pooled analysis of GST-T1 and lung cancer: a HuGE-GSEC review. Am J Epidemiol 164(11): 1027-
1042.
Raje, R; Basso, M; Tolen, T; et al. (1988) Evaluation of in vivo mutagenicity of low-dose methylene
chloride in mice. J Am Coll Toxicol 7(5):699-703.
Ramsey, JR; Andersen, ME. (1984) A physiologically-based description of the inhalation pharmacokinetics
of styrene in rats and humans. Toxicol Appl Pharmacol 73:159-175.
Raphael, M; Nadiras, P; Flacke-Vordos, N. (2002) Acute methylene chloride intoxication—a case report on
domestic poisoning. Eur J Emerg Med 9:57-59.
Rebert, CS; Matteucci, MJ; Pryor, GT. (1989) Acute effects of inhaled dichloromethane on the EEG and
sensory-evoked potentials of Fischer-344 rats. Pharmacol Biochem Behav 34:619-629.
Reitz, RH. (1991) Estimating the risk of human cancer associated with exposure to methylene chloride.
Ann 1st Super Sanita 27(4):609-614.
Reitz, RH; McDougal, JN; Himmelstein, MW; et al. (1988a) In vitro metabolism of methylene chloride in
human and animal tissues: use in physiologically based pharmacokinetic models. Toxicol Appl Pharmacol
97:230-246.
Reitz, RH; Mendrala, AL; Park, CN; et al. (1988b) Incorporation of in vitro enzyme data into the
physiologically-based pharmacokinetic (PB-PK) model for methylene chloride: implications for risk
assessment. Toxicol Lett 43:97-l 16.
Reitz, RH; Mendrala, AL; Guengerich, FP. (1989) In vitro metabolism of methylene chloride in human and
animal tissues: use in physiologically based pharmacokinetic models. Toxicol Appl Pharmacol 97:230-
246.
339

-------
Reitz, RH; Hays, SM; Gargas, ML. (1997) Addressing priority data needs for methylene chloride with
physiologically based pharmacokinetic modeling. Prepared for the Agency for Toxic Substances and
Disease Registry, Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA, on
behalf of the Halogenated Solvents Industry Alliance (HSIA), Arlington, VA.
Rhomberg, L. (1995) Use of quantitative modeling in methylene chloride risk assessment. Toxicology
102:95-114.
Riley, EC; Fassett, DW; Sutton, WL. (1966) Methylene chloride vapor in expired air of human subjects.
Am Ind Hyg Assoc J 27:341-348.
Rioux, JP; Myers, RA. (1988) Methylene chloride poisoning: a paradigmatic review. J Emerg Med 6:227-
238.
Rodkey, FL; Collison, HA. (1977) Effect of dihalogenated methanes on the in vivo production of carbon
monoxide and methane by rats. Toxicol Appl Pharmacol 40:39-47.
Rodriguez-Arnaiz, R. (1998) Biotransformation of several structurally related 2B compounds to reactive
metabolites in the somatic w/w+ assay of Drosophila melanogaster. Environ Mol Mutagen 31(4):390-401.
Roldan-Aijona, T; Pueyo, C. (1993) Mutagenic and lethal effects of halogenated methanes in the Aratest
of Salmonella typhimurium: quantitative relationship with chemical reactivity. Mutagenesis 8(2): 127—131.
Rosengren, LE; Kjellstrand, P; Aurell, A; et al. (1986) Irreversible effects of dichloromethane on the brain
after long term exposure: a quantitative study of DNA and the glial cell marker proteins S-100 and GFA.
Br J Ind Med 43:291-299.
Roth, RP; Drew, RT; Lo, RJ; et al. (1975) Dichloromethane inhalation, carboxyhemoglobin concentrations,
and drug metabolizing enzymes in rabbits. Toxicol Appl Pharmacol 33:427-437.
Rothman, KJ; Greenland, S. (1998) Precision and validity in epidemiologic studies. In: Modern
epidemiology. 2nd edition. Philadelphia, PA: Lippincott-Raven Publishers; pp. 115-134.
Sakai, T; Morita, Y; Wakui, C. (2002) Biological monitoring of workers exposed to dichloromethane using
head-space gas chromatography. J Chromat B 778:245-250.
Sasaki, YF; Saga, A; Akasaka, M; et al. (1998) Detection of in vivo genotoxicity of haloalkanes and
haloalkenes carcinogenic to rodents by the alkaline single cell gel electrophoresis (comet) assay in multiple
mouse organs. Mutat Res 419:13-20.
Savolainen, H; Pfaffli, P; Tengen M; et al. (1977) Biochemical and behavioural effects of inhalation
exposure to tetrachloroethylene and dichloromethane. J Neuropathol Exp Neurol 36:941-949.
Savolainen, H; Kurppa, K; Pfaffli, P; et al. (1981) Dose-related effects of dichloromethane in rat brain in
short-term inhalations exposure. Chem-Biol Inter 34:315-322.
Schwetz, BA; Leong, BKJ; Gehring, PJ. (1975) The effect of maternally inhaled trichloroethylene,
perchloroethylene, methyl chloroform, and methylene chloride on embryonal and fetal development in
mice and rats. Toxicol Appl Pharmacol 32:84-96.
Searles, J; McPhail, HA. (1949) Methylene chloride, CH2C12. In: Kirk, RE; Othmer, DF; eds.
Encyclopedia of chemical technology. New York, NY: Interscience Encyclopedia, Inc.; pp. 747-751.
Selgrade, MK; Gilmour, MI. (2006) Immunotoxicology of inhaled compounds-assessing risks of local
immune suppression and hypersensitivity. J Toxicol Environ Health A 69:827-844.
340

-------
Serota, DG; Thakur, AK; Ulland, BM; et al. (1986a) A two-year drinking water study of dichloromethane
in rodents. I. Rats. Food Chem Toxicol 24(9):951-958.
Serota, DG; Thakur, AK; Ulland, BM; et al. (1986b) A two-year drinking water study of dichloromethane
in rodents. II. Mice. Food Chem Toxicol 24(9):959-963.
Sheldon, T; Richardson, CR; Elliott, BM. (1987) Inactivity of methylene chloride in the mouse bone
marrow micronucleus assay. Mutagenesis 2(l):57-59.
Sherratt, PJ; Pulford, DJ; Harrison, DJ; et al. (1997) Evidence that human class Theta glutathione S-
transferase Tl-1 can catalyse the activation of dichloromethane, a liver and lung carcinogen in the mouse.
Biochem J 326:837-846.
Sherratt, PJ; Williams, S; Foster, J; et al. (2002) Direct comparison of the nature of mouse and human GST
Tl-1 and the implications on dichloromethane carcinogenicity. Toxicol Appl Pharmacol 179:89-97.
Shimada, T; Yamazaki, H; Mimura, M; et al. (1994) Interindividual variations in human liver cytochrome
P-450 enzymes involved in the oxidation of drugs, carcinogens and toxic chemicals: studies with liver
microsomes of 30 Japanese and 30 Caucasians. J Pharmacol Exp Ther 270(l):414-423.
Shusterman, D; Quinlan, P; Lowengart, R; et al. (1990) Methylene chloride intoxication in a furniture
refinisher. A comparison of exposure estimates utilizing workplace air sampling and blood
carboxyhemoglobin measurements. J Occup Med 32:451-454.
Sills, RC; Hailey, JR; Neal, J; et al. (1999) Examination of low-incidence brain tumor responses in F344
rats following chemical exposures in National Toxicology Program carcinogenicity studies. Toxicol Pathol
27(5):589-599.
Simula, TP; Glancey, MJ; Wolf, CR. (1993) Human glutathione S -1 ra n s fc ra se -ex p re s s i ng Sa Im one Ha
typhimurium tester strains to study the activation/detoxification of mutagenic compounds: studies with
halogenated compounds, aromatic amines and aflatoxin Bl. Carcinogenesis 14(7): 1371—1376.
Slikker, W, Jr; Andersen, ME; Bogdanffy, MS; et al. (2004) Dose-dependent transition in mechanisms of
toxicity: case studies. Toxicol Appl Pharmacol 201:226-294.
Soden, KJ. (1993) An evaluation of chronic methylene chloride exposure. J Occup Med 35(3):282-286.
Soden, KJ; Marras, G; Amsel, J. (1996) Carboxyhemoglobin levels in methylene chloride-exposed
employees. J Occup Environ Med 38(4):367-371.
Spirtas, R; Stewart, PA; Lee, JS; et al. (1991) Retrospective cohort mortality study of workers at an aircraft
maintenance facility. I. Epidemiological results. Br J Ind Med 48:515-530.
SRC (Syracuse Research Corporation). (1989) A review of in vitro test methodology for assessment of
hepatotoxicity with a view to application to chemical mixtures. Prepared for U.S. EPA, Environmental
Criteria and Assessment Office, U.S. Environmental Protection Agency; SRC TR-89-205.
Stephens, EA; Taylor, JA; Kaplan, N; et al. (1994) Ethnic variation in the CYP2E1 gene: polymorphism
analysis of 695 African-Americans, European-Americans and Taiwanese. Pharmacogenetics 4:185-192.
Stewart, RD; Fisher, TN; Hosko, MJ; et al. (1972a) Experimental human exposure to methylene chloride.
Arch Environ Health 25:342-348.
Stewart, RD; Fisher, TN; Hosko, MJ; et al. (1972b) Carboxyhemoglobin elevation after exposure to
dichloromethane. Science 176:295-296.
341

-------
Stewart, PA; Lee, JS; Marano, DE; et al. (1991) Retrospective cohort mortality study of workers at an
aircraft maintenance facility. II Exposures and their assessment. Br J Ind Med 48:531-537.
Stott, WT; McKenna, MJ. (1984) The comparative absorption and excretion of chemical vapors by the
upper, lower, and intact respiratory tract of rats. Fundam Appl Toxicol 4:594-602.
Sweeney, LM; Gargas, ML; Strother, DE; Kedderis, GL. (2003) Physiologically based pharmacokinetic
model parameter estimation and sensitivity and variability analyses for acrylonitrile disposition in humans.
Toxicol Sci. 71:27-40.
Sweeney, LM; Kirman, CR; Morgott, DA; et al. (2004) Estimation of interindividual variation in oxidative
metabolism of dichloromethane in human volunteers. Toxicol Lett 154:201-216.
Takeshita, H; Mogi, K; Yasuda, T; et al. (2000) Postmortem absorption of dichloromethane: a case study
and animal experiments. Int J Legal Med 114:96-100.
Taskinen, H; Lindbohm, ML; Hemminki, K. (1986) Spontaneous abortions among women working in the
pharmaceutical industry. Br J Ind Med 43:199-205.
Tay, P; Tan, KT; Sam, CT. (1995) Fatal gassing due to methylene chloride—a case report. Singapore Med
J 36:444-445.
Teschke, K; Olshan, AF; Daniels, JL; et al. (2002) Occupational exposure assessment in case-control
studies: opportunities for improvement. Occup Environ Med 59:575-593.
Thier, R; Foest, U; Deutschmann, S; et al. (1991) Distribution of methylene chloride in human blood. Arch
Toxicol Suppl 14:254-258.
Thier, R; Taylor, JB; Pemble, SE; et al. (1993) Expression of mammalian glutathione S-transferase 5-5 in
Salmonella typhimurium TA1535 leads to base-pair mutations upon exposure to dihalomethanes. Proc Natl
Acad Sci USA 90:8576-8580.
Thier, R; Wiebel, FA; Hinkel, A; et al. (1998) Species differences in the glutathione transferase GST-T1-1
activity towards the model substrates methyl chloride and dichloromethane in the liver and kidney. Arch
Toxicol 72:622-629.
Thilagar, AK; Kumaroo, V. (1983) Induction of chromosome damage by methylene chloride in CHO cells.
MutatRes 116:361-367.
Thilagar, AK; Back, AM; Kirby, PE; et al. (1984) Evaluation of dichloromethane in short term in vitro
genetic toxicity assays. Environ Mutagen 6:418-419.
Thomas, AA; Pinkerton, MK; Warden, JA. (1972) Effects of low level dichloromethane exposure on the
spontaneous activity of mice. Proceedings of the 3rd conference of environmental toxicology; October 25-
27; Fairborn, OH; Paper No. 14; AMRL-TR-72-130. Aerospace Medical Research Laboratory, Wright-
Patterson Air Force Base, OH; pp. 223-238. Available from the National Technical Information Service,
Springfield, VA; AD773766.
Tomenson, JA; Bonner, SM; Heijne, CG; et al. (1997) Mortality of workers exposed to methylene chloride
employed at a plant producing cellulose triacetate film base. Occup Environ Med 54:470-476.
Treluyer, JM; Cheron, G; Sonnier M; et al. (1996) Cytochrome P-450 expression in sudden infant death
syndrome. Biochem Pharmacol 52:497-504.
342

-------
Trueman, RW; Ashby, J. (1987) Lack of UDS activity in the livers of mice and rats exposed to
dichloromethane. Environ Mol Mutagen 10:189-195.
U.S. Coast Guard. (1999) Dichloromethane. CHRIS: hazardous chemical data. U.S. Coast Guard,
Department of Transportation, Washington, DC. Available online at
http://www.chrismanual.com/findform.htm.
U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk assessment of
chemical mixtures. Federal Register 51(185):34014-34025. Available online at
http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for mutagenicity risk assessment.
Federal Register 51(185):34006-34012. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (1987a) Update to the health assessment document and
addendum for dichloromethane (methylene chloride): pharmacokinetics, mechanism of action and
epidemiology [review draft]. Office of Health and Environmental Assessment, Office of Research and
Development, Washington, DC; EPA/600/8-87/030A. Available from the National Technical Information
Service, Springfield, VA, PB87228565, and online at
http ://nepis .epa.gov/Exe/ZyPURL. cgi?Dockey=30001GFH .txt.
U.S. EPA (Environmental Protection Agency). (1987b) Technical analysis of new methods and data
regarding dichloromethane hazard assessments [review draft]. Office of Health and Environmental
Assessment, Office of Research and Development, Washington, DC; EPA/600/8-87/029A. Available from
the National Technical Information Service, Springfield, VA, PB87-228557, and online at
http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=30001GAN.txt.
U.S. EPA (Environmental Protection Agency). (1988a) Recommendations for and documentation of
biological values for use in risk assessment. Environmental Criteria and Assessment Office, Office of
Health and Environmental Assessment, Cincinnati, OH; EPA/600/6-87/008. Available from the National
Technical Information Service, Springfield, VA, PB88-179874/AS, and online at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=34855.
U.S. EPA (Environmental Protection Agency). (1988b) The impact of pharmacokinetics on the risk
assessment of dichloromethane. Office of Health and Environmental Assessment, Office of Research and
Development, Washington, DC; EPA/600/D-88/219. Available from National Technical Information
Service, Springfield, VA; PB89-173249.
U.S. EPA (Environmental Protection Agency). (1991) Guidelines for developmental toxicity risk
assessment. Federal Register 56(234):63798-63826. Available online at
http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (1994a) Interim policy for particle size and limit
concentration issues in inhalation toxicity: notice of availability. Federal Register 59(206):53799.
Available online at http://www.epa.gov/EPA-PEST/1994/October/Day-26/pr-l 1.html.
U.S. EPA (Environmental Protection Agency). (1994b) Methods for derivation of inhalation reference
concentrations and application of inhalation dosimetry. Environmental Criteria and Assessment Office,
Office of Health and Environmental Assessment, Cincinnati, OH; EPA/600/8-90/066F. Available from the
National Technical Information Service, Springfield, VA, PB2000-500023, and online at
http://cfpub.epa.gov/ncea/raf/recordisplay.cfm?deid=71993.
U.S. EPA (Environmental Protection Agency). (1995) Use of the benchmark dose approach in health risk
assessment. Risk Assessment Forum, Washington, DC; EPA/630/R-94/007. Available from the National
343

-------
Technical Information Service, Springfield, VA, PB95-213765, and online at
http://cfpub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA (Environmental Protection Agency). (1996) Guidelines for reproductive toxicity risk assessment.
Federal Register 61(212):56274-56322. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (1998) Guidelines for neurotoxicity risk assessment.
Federal Register 63(93):26926-26954. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.
U.S. EPA (Environmental Protection Agency). (2000a) Science policy council handbook: risk
characterization. Office of Science Policy, Office of Research and Development, Washington, DC;
EPA/100-B-00-002. Available online at http://www.epa.gov/OSA/spc/pdfs/prhandbk.pdf.
U.S. EPA (Environmental Protection Agency). (2000b) Benchmark dose technical guidance document
[external review draft]. Risk Assessment Forum, Washington, DC; EPA/630/R-00/001. Available online at
http ://cfpub. epa. gov/ncea/cfm/
nceapublication.cfm?ActType=PublicationTopics&detype=DOCUMENT&subject=BENCHMARK+DOS
E&subjtype=TITLE&excCol=Archive.
U.S. EPA (Environmental Protection Agency). (2000c) Supplementary guidance for conducting health risk
assessment of chemical mixtures. Risk Assessment Forum, Washington, DC; EPA/630/R-00/002.
Available online at http://cfpub.epa.gov/ncea/raf/chem_mix.cfm.
U.S. EPA (Environmental Protection Agency). (2000d) Toxicological review of vinyl chloride. Integrated
Risk Information System (IRIS), National Center for Environmental Assessment, Washington, DC;
EPA/635/R-00/004. Available online at http://www.epa.gov/iris.
U.S. EPA (Environmental Protection Agency). (2002) A review of the reference dose concentration and
reference concentration processes. Risk Assessment Forum, Washington, DC; EPA/630/P-02/002F.
Available online at http://cfpub.epa.gov/ncea/raf/raf_pubtitles.cfm?detype=document&excCol=archive.
U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk assessment. Federal
Register 70(66): 17765-18717. Available online at http://www.epa.gov/cancerguidelines.
U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility
from early-life exposure to carcinogens. Risk Assessment Forum, Washington, DC; EPA/630/R-03/003F.
Available online at http://www.epa.gov/cancerguidelines.
U.S. EPA (Environmental Protection Agency). (2006a) Science policy council handbook: peer review. 3rd
edition. Office of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-
06/002. Available online at http://www.epa.gov/OSA/spc/2peerrev.htm.
U.S. EPA (Environmental Protection Agency). (2006b) A framework for assessing health risk of
environmental exposures to children. National Center for Environmental Assessment, Washington, DC;
EPA/600/R-05/093F. Available online at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=158363.
Vieira, I; Sonnier, M; Cresteil, T. (1996) Developmental expression of CYP2E1 in human liver.
Hypermethylation control of gene expression during the neonatal period. Eur J Biochem 238:476-483.
Warbrick, EV; Kilgour, JD; Dearman, RJ; et al. (2003) Inhalation exposure to methylene chloride does not
induce systemic immunotoxicity in rats. J Toxicol Environ Health A 66:1207-1219.
Warholm, M; Alexandrie, AK; Hogberg, J; et al. (1994) Polymorphic distribution of glutathione transferase
activity with methyl chloride in human blood. Pharmacogenetics 4:307-311.
344

-------
Watanabe K, LibermanRG, Skipper PL, Tannenbaum SR, Guengerich FP. (2007). Analysis of DNA
adducts formed in vivo in rats and mice from 1,2-dibromoethane, 1,2-dichloroethane, dibromomethane, and
dichloromethane using HPLC/accelerator mass spectrometry and relevance to risk estimates. Chem Res
Toxicol. 20:1594-600.
Weinstein, RS; Boyd, DD; Back, KC. (1972) Effects of continuous inhalation of dichloromethane in the
mouse: morphologic and functional observations. Toxicol Appl Pharmacol 23:660-679.
Wells, VE; Schrader, SM; McCamon, CS; et al. (1989) Letter to the editor. Reprod Toxicol 3:281-282.
Westbrook-Collins, B; Allen, JW; Sharief, Y; et al. (1990) Further evidence that dichloromethane does not
induce chromosome damage. J Appl Toxicol 10(2):79-81.
Winneke, G. (1974) Behavioral effects of methylene chloride and carbon monoxide as assessed by sensory
and psychomotor performance. In: Xintaras, C; Johnson, BL; de Groot, I; eds. Behavioral toxicology: early
detection of occupational hazards. Washington, DC: National Institute for Occupational Safety and Health,
Center for Disease Control, Public Health Service, U.S. Department of Health, Education and Welfare; pp.
130-144.
Wirkner, K; Damme, B; Peolchen, W; et al. (1997) Effect of long-term ethanol pretreatment on the
metabolism of dichloromethane to carbon monoxide in rats. Toxicol Appl Pharmacol 143:83-88.
Withey, JR; Karpinski, K. (1985) The fetal distribution of some aliphatic chlorinated hydrocarbons in the
rat after vapor phase exposure. Biol Res Pregnancy Perinatol 6(2):79-88.
Zarrabeitia, MT; Ortega, C; Altuzarra, E; et al. (2001) Accidental dichloromethane fatality: a case report. J
Forensic Sci 46:726-727
Zeiger; E. (1990) Mutagenicity of 42 chemicals in Salmonella. Environ Mol Mutagen 16:32-54.
Zielenska, M; Ahmed, A; Pienkowska, M; et al. (1993) Mutational specificities of environmental
carcinogens in the lacl gene of Escherichia coli. VI. Analysis of methylene chloride-induced mutational
distribution in Uvr+ and UvrB- strains. Carcinogenesis 14(5):789-794.
345

-------