Document jJ0Z3X6BvLpb6o7aqL8KppzO

Critical Reviews in Toxicology, 25(2): 133-163 (1995) Polychlorinated Biphenyls (PCBs) and Human Health: An Update* Renate D. Kimbrough, M.D. Institute for Evaluating Health Risks, 1101 Vermont Ave., N.W., Suite 608, Washington, D.C. 20005 * Funding was provided by the Institute for Evaluating Health Risks and the Aluminum Company of America. ABSTRACT: Polychlorinated biphenyls (PCBs) are a mixture of 209 different chlorinated biphenyl congeners (forms) of which 36 are environmentally relevant. PCBs are lipid (fat)-soluble, stable compounds. PCBs may be contaminated with more highly toxic polychlorinated dibenzofurans (PCDFs). Some PCDFs were primarily responsible for the two poisoning outbreaks -- Yusho and Yu-Cheng. Based on the reports on workers and the general population, no clear and convincing evidence that PCB exposures were casually associated with adverse health effects was advanced; this included cancer for a wide range of body burdens and exposures for serum PCB concentrations >1000 ppb (p.g/1) and adipose PCB levels >400 ppm (mg/kg). No meaningful reproductive problems have been identified in female capacitor workers. In the opinion of the review author, the available evidence for cancer and for reproductive effects is inconclusive. Adverse neurobehavioral effects in infants and young children have been reported in a study of women in the general population and a study of fish eaters and their offspring. The adverse effects observed in the two studies were not the same; the exposure assessments in both studies are not well defined and have many uncertainties. Subhuman primates appear to be more sensitive to reproductive and other adverse effects of PCBs than humans. Obvious external clinical signs are observed in the offspring of subhuman primates at dosage levels below those experienced by female capacitor workers and members of the general population prior to the control ofPCBs. KEY WORDS: polychlorinated biphenyls, carcinogenicity, humans, toxic equivalency factors. I. INTRODUCTION Polychlorinated biphenyls (PCBs) are a mix ture of 209 different chlorinated biphenyl conge ners (forms). According to McFarland and Clarice,1 only 36 of these congeners are environmentally relevant (Table 1). PCBs are lipid (fat)-soluble, stable compounds. Their stability depends on the number of chlorine atoms and the position of the chlorine atoms on the biphenyl molecule. If few chlorine atoms are attached to the biphenyl mol ecule or if at least one paraposition on the mol ecule is not occupied by a chlorine, then they can be more easily metabolized and eliminated. The lower chlorinated PCBs are more easily broken down in the environment. Thus, the composition of a given commercial mixture of PCBs changes 1040-8444/95/$.50 1995 by CRC Press, Inc. in the environment through selective degrada tions of congeners. PCBs were first produced commercially in the 1920s, although it was not until the 1950s that the industrial application of PCBs increased signifi cantly. They were used as capacitor, hydraulic, and transformer fluids in carbonless copying paper and as plasticizers in paint Through industrial applica tions, PCBs entered the environment and contami nated the food chain, most notably fish. In response to public concern over the envi ronmental build-up of PCBs and results from ani mal feeding studies, Monsanto Chemical Com pany, the sole U.S. producer of PCBs, voluntarily stopped open-ended uses of PCBs in 1971. Sub sequently, the lower chlorinated biphenyls, Aroclor 1016,1242, and some 1254 were produced.2 Prior DSW 148230 133 STLCOPCB4039715 TABLE 1 IUPAC Numbers for 36 PCB Congeners Considered Environmentally Relevant and Congeners Included in the Canadian Marine Analytical Standards Program (CLB-1) McFarland and Clark CLB-1 standard contains all congeners listed by McFarland and Clarke except Additional congeners that are included but according to McFarland and Clarke are irrelevant 18 37 15 37 70 31 44 74 40 49 81 54 52 119 60 70 123 86 74 126 103 77 157 121 81 158 129 87 167 137 99 168 141 101 169 143 105 - 177 154 114 159 118 171 119 173 123 182 126 ' 185 128 191 138 195 151 196 153 200 156 202 157 203 158 205 167 206 168 207 169 208 170 209 177 180 183 187 189 194 201 to 1971, other more highly chlorinated biphenyl mixtures such as Aroclor 1248, 1254, and 1260 were also marketed. In 1977 Monsanto Chemi cal Company ceased production of PCBs en tirely. Since the 1970s, levels of PCBs in the envi ronment have decreased overall;3 however, ca pacitors and transformers with PCBs as dielec tric fluid are still in operation and their containment is still of some concern. In addition to the presence of PCBs in capacitors and trans formers, PCBs are also formed through incinera tion, most likely by the dimerization of chlo robenzenes. The major PCBs formed through incineration are the coplanar congeners charac terized by 3,3',4,4' substitution. The formation of 2,6 substituted PCB congeners is less favored.4 It is unclear at this time whether formation 134 DSW 148231 STLCOPCB4039716 through combustion makes a major contribution to the input of PCBs into the environment. Since PCB production ceased in 1977, body burdens in the general population of the U.S. have decreased.3 5-6 The extent of this decrease is uncertain. Significant changes introduced over the years in PCB analysis and quantitation di rectly impact the validity of trend analyses and make time trend analysis difficult to interpret. PCBs may be contaminated with polychlori nated dibenzofurans (PCDFs), which are also a mixture of many homologues. If PCBs are heated to temperatures between 300C, PCDFs may be formed7 and may, in specific situations, be re sponsible for the toxic effects. Some PCDFs are much more toxic than PCBs in comparatively smaller amounts. In incomplete combustion situ ations, they may also be formed at temperatures between 600 and 800C.7a In the commercial mixtures of PCBs produced by Monsanto Chemi cal Company, PCDFs only occurred in trace amounts or were not detected at all. In Japan, in 1968, and in Taiwan, in 1978-1979, poisoning outbreaks of Yusho and Yu-Cheng occurred. The affected population had ingested rice-oil con taminated with PCDFs, PCBs, polychlorinated naphthalenes, dibenzodioxins, and quarterphenyls. In the poisoning outbreaks, the PCDFs, not the PCBs, caused the adverse human health effects.8 At similar PCB tissue levels in the gen eral population without corresponding levels of PCDFs, no adverse human health effects have been noted. Observations made in the Yusho and Yu-Cheng outbreaks should, therefore, not be construed as adverse effects of PCBs. In this review, the more recent PCB litera ture is summarized, data related to human health are examined in detail, and the relevance of re cent animal data to predict human health effects is discussed. The review begins with human ex posure pathways and a discussion of available data from air, water, and food monitoring. The literature on PCB metabolism in humans is re viewed as well as the literature regarding the human health effects of PCBs and the scientific evidence for reproductive, developmental, im munologic, and carcinogenic effects of PCBs. More recent animal data and the toxicity equiva lency factors and dose response relationship are discussed. II. HUMAN EXPOSURE A. Monitoring Data Monitoring data of PCB levels in humans provide some information about the extent of human exposure to PCBs. However, results from different laboratories are difficult to compare because chemical analytical methods and PCB quantitation have changed over the years and dif fer between laboratories.4 The analytical method of choice for many years was gas chromatogra phy using packed columns. With this method about 12 PCB peaks each containing many congeners can be separated in human milk,9 while with a more sophisticated method that uses glass capil lary columns, about 70 PCB congeners can be separated. The average levels in human milk in the U.S. are around 1 ppm (mg/kg) if more than six peaks of the gas chromatogram are used for quantitation and about 1.5 ppm (mg/kg) if 1 to 4 peaks are used.9 When similar quantitation meth ods are used, the average PCB levels in human milk do not differ much between industrialized countries. Differences in clean up and in the use of solvents for extraction also affect quantitation. PCBs are persistent and lipid soluble and accumulate in fluids and tissues. On a wet weight basis, PCB concentration in fatty tissue is much greater than in vital organs. On a lipid basis, the differences between blood, fatty tissue, and vital organs disappear. Overall, PCB levels in humans, in food, and in the environment seem to be declin ing in the U.S.6 Median PCB levels in whole serum in the general U.S. population now range between 2 and 7 ppb (jxg/l). Levels in workers and in some subpopulations of the general popu lation may still be higher because of past expo sures. Levels in adipose tissue and in human milk fat in industrialized countries average around 1 ppm (mg/kg)9 and fewer people overall have detectable levels of PCBs.3 1. PCBs in Air In air, PCB levels of 0.1 to 10 ng/m3 have been reported.4 Between 1 and 3 pg/m2 are deposited by air per year.4 Similar or higher aerial depositions have been reported for Lake Michigan OS* 148232 135 STLCOPCB4039717 and remote lakes surrounding the Great Lakes10 and in Sweden.11 For southern Lake Michigan, the atmospheric flux to the lake is around 7.5 pg/m2/year and the flux back to the atmo sphere is about 5.6 pg/m2/year, resulting in a net deposition of approximately 2 pg/m2/year. Based on mass balance analyses made prior to 1985,10 net atmospheric deposition and nonatmospheric sources each account for about half the total PCB input into Lake Michigan. 2. PCBs in Food Several years ago, the mean daily intake from food was calculated by Bennett to range from <0.1 to 1.9 |a.g per person per day.12 The Food and Drug Administration (FDA) estimated that for the years 1982 to 1984, the average human PCB in take was 0.042 pg/d. For a 70-kg person, the average PCB intake was 0.0006 pg/kg/d and for infants and toddlers it was 0.001 pg/kg/d.13 Ac cording to Fensterheim,3 PCB intake from food was 0.05 pg/d (0.0007 pg/kg bodyweight/d) for the years 1987 to 1989. B. Fish Consumption and PCB Serum Levels In the 1970s, consumption of Lake Michigan sport-caught fish produced higher PCB body bur dens in humans with a reported maximum of 360 ppb (pg/1) in whole serum of an angler who ate an average of 40 kg of fish per year. No adverse health effects associated with the PCB body burdens were identified.14 Many of these same anglers were studied again recently.16 The fish eaters now have a mean serum PCB level of 19 pg/1 with a range of 4.9 to 173.8 pg/1. How ever, the higher PCB levels had resulted from exposures prior to 1982. The mean PCB serum level in a group of non-fish-eating controls was 6.8 pg/1. Neither the controls nor the fish eaters showed a decline of PCB serum levels between 1982 and 1989,16 whereas DDT levels declined significantly. In a recent study of 801 Wisconsin anglers (Lake Michigan), PCB congener specific analy ses were performed on 192 serum samples.17 In 31 % of the samples, the PCB concentrations were below the limit of detection of 0.6 ppb (pg/1). The annual mean number of sport-caught fish meals ranged from 19.5 per year for individuals with no detectable PCB in serum to 32.3 per year for individuals with PCB levels >5 pg/1. A total of 13 congeners were identified in the 192 serum samples. The most frequently identified conge ners were 2,2',4,4',5,5' hexachlorobiphenyl (153), 2,2',3,4,4',5 hexachlorobiphenyl (138) and 2,2',3,4,4',5,5' hepatochlorobiphenyl (180). The numbers in parentheses are IUPAC (International Union for Pure and Applied Chemistry) num bers.4 The mean sum of these congeners was 2.2 ppb (pg/1). In another study during 1987 to 1989, repro ductive outcomes were examined in 1112 women from the Green Bay area of Wisconsin.18 Among the women, 100 participants were randomly se lected for serum PCB analysis. Detectable PCB levels were only found in 24 sera. The limit of detection for the PCB isomers was 0.6 pg/1. There were 13 PCB specific congeners determined. Only five of the congeners were present above the limit of detection, namely, 2,2',3,4,4',5' hexachloro biphenyl (138), 2,2\4,4',5,5' hexachlorobiphenyl (153), 2,2',3,4,4',5,5' (180) heptachlorobiphenyl (180), 2,2',3,4,5,5\6 heptachlorobiphenyl (187) and 2,3,4,4',5 pentachlorobiphenyl (118). Only isomer 153 was found in all samples. Isomer 187 was only identified in one sample. The arithmetic sum of these congeners ranged from 0.6 to 5.0 ppb (pg/1). No association between PCB exposure and adverse reproductive outcomes was found. Birth weight was found to be positively associated with fish consumption and PCB serum levels. It is known that greater fish consumption is associated with larger babies.19 Olsen et al.19 suggested that the association between fish consumption and birth size was related to marine n-3 fatty acids. These fatty acids improve placental blood flow by in creasing the ratio of prostacyclins to thromoxane. C. PCBs in Human Milk PCBs are present in human milk fat. A number of studies have examined the risk associated with ingesting such milk.20-21 Certain subpopulations such as the Inuit women from Arctic Quebec have re portedly higher PCB milk levels than other popu 136 DSW 148233 STLCOPCB4039718 lations. It has been suggested that their high con sumption of fish and marine mammals22 has re sulted in the higher PCB milk levels. In the Inuits, the mean concentration of PCBs in human milk fat was 3.6 ppm (mg/kg), while the comparison group of Caucasians from Quebec City and from Baie Comeau had mean PCB levels of0.77 ppm (mg/kg) in milk fat. However, high PCB milk levels were not present in the American Indians living on Cornwall Island in Upstate New York.23 They were exposed to contaminated fish because of the pres ence of a hazardous-waste site in their neighbor hood. Their arithmetic mean PCB concentration in milk fat of 0.554 ppm (mg/kg) was similar to a control group of 0.516 ppm (mg/kg). No association between PCB milk levels and fish consumption could be demonstrated in the American Indians liv ing on Cornwall Island. Concern has been raised that PCB levels in human milk are at times above the FDA guidelines of acceptable levels of PCBs in food 20 When such concerns are raised, the physiol ogy of the growing infant and the variations of the composition of human milk are not considered. Fur thermore, the benefits of breast feeding are also not included in the evaluation.21 Several nutritional com mittees have recommended a 6-month period of exclusive breast feeding. Beyond that point exclu sive breast feeding is nutritionally no longer desir able. The fat concentration in human milk averages 3.5 g/100 ml and the infant consumes 4.2 g/kg body wt/d of fat if it consumes 120 ml/kg body wt d of milk. Accepting 1 mg/kg of PCBs in hu man milk fat, the infant consumes 4.2 pg/kg/d or a total of 750 jig/kg of PCBs over a 6-month period. In 1987, a WHO work group estimated that the daily intake of PCBs for a 70-kg person was 0.2 pg/kg/d, or over a life time, a total of about 5.1 mg/kg PCBs. Thus, during an exclusive nursing period of6 months, the infant accumulates 6.8% of its lifetime PCB body burden in some industrialized countries. PCB levels in human milk fat in the U.S. have averaged about 0.5 ppm (mg/kg), resulting in a 3.4% uptake of the lifetime accumulation based on the estimates of WHO. D. Miscellaneous PCB Exposures of Specific Subpopulations In the past, segments of the general popula tion received additional exposures.14 Dairy cattle became contaminated in Ohio from PCB con taining paint used to paint the inside of silos between 1941 and 1970. The PCBs leached into the silage fed to dairy cattle and entered the human food chain through milk and cheese. PCBs were also present in breakfast cereals because PCBs leached out of the recycled paper used to make cereal boxes. Humans living in contaminated environments where food contamination is not an issue usually do not accumulate additional body burdens of PCBs. In 10 of 12 site-specific investigations, Stehr-Green et al.24 found that PCB serum levels of people considered to be exposed to PCB waste sites did not exceed those of a local or national reference population. In these waste sites, PCB concentrations reached 13% in soil and 0.002% in surface water leaching from the site. In two in stances an increased proportion of persons with higher PCB levels was found because occupa tional exposure and consumption of PCB con taminated fish occurred. In an earlier study of a community, PCB con taminated sludge was used as fertilizer in gar dens.25 The mean PCB levels for the controls were 24.4 and 17.4 ppb (]Xg/l) for the sludge us ers. This suggested that the PCB uptake from sludge through gardening was negligible. The mean levels in the sludge-treated soil was ap proximately 17 ppm (mg/kg) and ranged from 0.1 to 1700 ppm (mg/kg). Similar findings at low PCB concentrations in soil were made by Yafe and Reeder.26 In Greater New Bedford, MA, high concentrations of PCBs had been identified in sediment, wildlife, and fish. Seram specimens were obtained from 391 male and 449 female volunteers. These volunteers had resided in the area for at least 5 years. The preva lence of elevated PCB serum levels in this group was typical of an urban unexposed population. Only 1.3% of the study participants had PCB serum levels above 30 ppb (jig/1).27 Poor skin absorption of PCBs from soil and no direct contact may explain the lack of PCB uptake from PCB contaminated waste sites. Schmid et al.28 demonstrated that the maximum dermal uptake of PCBs was 6% of the absorption after oral intake of the same amount of PCBs. Absorption of PCBs in solution following inges tion is usually at least 90%.28 0SW 148234 137 STLCOPCB4039719 III. METABOLISM OF PCBs IN HUMANS Steele et al.29 estimated the half-life in hu mans for lower chlorinated biphenyls (Aroclor 1242) as 6 to 7 months and the corresponding half-life for the more highly chlorinated biphe nyls as 33 to 34 months. The reported half-lives of mixtures reflect the percent reduction of the individual congeners, resulting ultimately in a reduced number of more gradually excreted con geners in tissues. Buehler et al.30 studied the elimination of a I3C-labeled PCB mixture similar to Aroclor 1254. A single dose of 329 |ig/kg body weight was administered orally to a human volunteer and blood samples were taken over a period of 260 d. The half-lives for the isomers 2,2',3,4,4',5' hexachlorobiphenyl (138), 2,2',4,4',5,5' hexachlorobiphenyl (153), and 2,2',3,4,4',5,5' heptachlorobiphenyl (180) were 321, 338, and 124 d, respectively. The study illustrates that dif ferent persistent isomers30 are excreted at differ ent rates. Borlakoglu and Walker31 identified 12 PCB congeners that account for 80% of the total PCB adipose tissue burden in humans. Eight of these congeners appear to be persistent (IUPAC numbers 28, 118, 138, 153, 170, 180, 183, and 202). These compounds do not have meta-paravicinal protons and are mostly highly chlorinated. The two congeners, 2,4,4' trichlorobiphenyl and 2,3',4,4',5 pentachlorobiphenyl, represente the lowest chlorinated compounds in this group. The congeners used by Buehler et al.30 in their volun teer study were part of these 12 congeners. Wolff et al.32 determined PCB body burdens in male and female current and past capacitor workers in two factories with several years of exposure. Aroclor 1242, 1016, 1221, and some Aroclor 1254 were used. The lower chlorinated biphenyls in 61 fatty tissue samples ranged from 0.6 to 414 ppm (mg/kg), with a geometric mean of 28 ppm (mg/kg), and in 290 plasma samples from 6 to 2530 ppb (jJ.g/1). The more highly chlo rinated biphenyls ranged in fatty tissue from 1 to 165 ppm (mg/kg), with a geometric mean of 14 ppm (mg/kg) and in plasma from nondetectable levels to 546 ppb (JJ.g/1) with a geometric mean of 21 ppb (jj.g/1). It was reported, 25 workers in high exposure jobs had PCB plasma levels of 300 ppb (pg/1) or above. In workers with less than 5 years employment, the higher chlorinated biphenyl lev els were similar to the general population with a partition value between plasma and fatty tissue of about 160. In a follow-up study of 165 of these workers after cessation of exposure, the lower chlorinated biphenyl congeners similar in appear ance to Aroclor 1242 had decreased by 25 to 90%.33 The more highly chlorinated biphenyl congeners, similar in appearance to Aroclor 1254, did not decrease significantly as a whole, although six congener peaks declined by 15 to 25%.33 For workers with an initial geometric mean PCB se rum level of 100 ppb (p.g/1), a half-life of 4.8 years was estimated. For workers with a mean PCB serum concentration of 39 ppb (p.g/1), a half life of 17 years was calculated. These half-lives were longer than half-lives published by other authors. This may, in part, be explained by differ ent methods of quantitation and differences in the amount and congener distribution of the initial body burden. Phillips et al.34 studied the metabolism ofPCBs in 58 male workers with predominant exposure to Aroclor 1242 and 1016 until 1977. In 1977, the range of the lower chlorinated biphenyls in serum was 2 to 3300 ppb (|ig/l) with a median of 155 ppb (|Hg/l). The range of the more highly chlorinated biphenyls was 5 to 250 ppb (p.g/1) with a median of 39 ppb (p.g/1). In a follow-up study in 1985, the half-life was longer when the initial PCB level was lower. For the entire group of workers, the half-life for the lower chlorinated biphenyls (Aroclor 1242) was 2.6 and 4.6 years for the higher chlorinated biphenyls (Aroclor 1254). This illustrates that PCBs were eliminated. Elimination was faster for lower chlorinated mixtures and the half-life var ied for different congeners. Congener specific PCB determinations have been made in present and former capacitor work ers, transformer workers, workers without spe cific exposure to PCBs, and in human milk and serum of the general population.933-3536 The ca pacitor workers were exposed to Aroclor 1254, 1242, and 1016. The transformer workers were primarily exposed to Aroclor 1260. Based on the results by Fait et al.,36 the congener specific analy ses suggested that in serum, the total PCB con- OSW 148235 138 STLCOPCB4039720 centrations were underestimated while the sum of the congener specific PCB quantitation and the total PCB quantitation on packed columns were quite similar for adipose tissue. These differences between serum and adipose tissue may be a re flection of the limit of detection for PCB conge ners present in lower concentrations in serum. In currently exposed workers with higher levels, more congeners could be identified and quantitated. IV. CROSS-SECTIONAL AND CLINICAL STUDIES In this and in the section on mortality the published human studies are reviewed. These stud ies were performed because increased exposure to PCBs had occurred and because of adverse effects observed in animals studies. None of them were disease investigations or triggered by the occurrence of a specific disease. These studies should therefore be considered as hypothesis gen erating rather than hypothesis testing studies. In the laboratory animal studies the liver in some but not all species was a target organ for PCBs. Fur thermore, reproductive effects in female animals, immunotoxicity, hyperplasia of the mucosa of the stomach, chloracne-like lesions and hyperpigmen tation of the skin were reported in some studies in rodents or subhuman primates following expo sure to different PCB mixtures. In female rats an increase of hepatocellular carcinomas and a con comitant decrease in mammary gland tumors, tumors of the thyroid, and of the pituitary were noted following long-term exposure to Aroclor 1260 and Chlophen A-60. In reviewing the accumulated reports on stud ies in humans, it appears that investigators in the design of clinical and cross-sectional studies fo cused on the adverse effects observed in labora tory animals and also took other measures com monly performed in routine medical examinations such as cholesterol measurements, blood cell counts, and measurement ofblood pressure. Com parison groups matched on age, smoking, and socioeconomic status were usually not included in these studies. It is therefore unclear whether the observed findings made in these assorted groups were related to their PCB exposures. In examin ing whether the findings made in humans are related to PCBs consistency across studies at simi lar exposures should be determined. A dose-re sponse relationship should exist. Workers with higher and longer exposures as a group should be affected more, even though some people may be more sensitive to the toxic effects of a chemical than others. It must also be evaluated whether confounders and biases were examined and dis cussed and whether the measured perturbation represented a true adverse health effect. Finally, for biochemical measurements such as liver func tion tests slight increases or decreases outside the established normal range for a test should be con firmed by repeating the test before it is assumed that the result was abnormal. Such results may represent outliers of the normal range of the tests rather than abnormal clinical findings or may rep resent laboratory aberrations on a given day or artifacts because of high serum lipid concentra tions. Furthermore, some of these tests such as liver function tests are not very specific. To prop erly interpret such results, other tests would have to be conducted, particularly if only one or two tests of a battery of tests were abnormal. Without follow up and additional information gathering the observed findings should not simply be attrib uted to PCB exposure or to ill health. Follow up and an adequate diagnosis is even more important if these abnormalities occur relatively frequently in the general population. Specifically, they may be unrelated to the PCB exposure if only a few workers of a similarly exposed group show such manifestations. For instance, in Triana, AL, 458 persons were examined in a community-wide study.37 The study was initiated because this population ate fish with high levels of DDT (1 .l'-^^^-trichloroethylideneD&wI^-chlorobenzene] and its metabo lites. In addition to DDT, PCBs were also present in serum. The geometric mean PCB serum level in this population was 17.2 ppb (p.g/1), 80 to 90% of the study population had levels similar to those found in other communities at the time of the study. Older people had higher PCB serum levels and levels were higher in males than in females. PCB serum levels were also higher in members of the study population who reported higher alcohol consumption and in people with higher serum DSW 148236 139 STLCOPCB4039721 cholesterol levels. A liver function test, the gamma-glutamyl transpeptidase (GGT), was el evated in participants with higher PCB serum levels as an independent variable but when in cluded into multiple regression analyses, this posi tive association disappeared. In this particular population, high blood pressure occurred much more frequently than has been reported as na tional rates for a population of the same sex, age, and ethnic composition. Analysis of the data sug gested that a statistically significant positive cor relation existed between diastolic blood pressure and PCB serum levels. In this study, blood pres sure was only measured at one point in time. The authors stated that "in the absence of a control group the explanation for the excess prevalence of hypertension among Triana residents is debat able. However, the absence of a control group with blood pressure measurement does not weaken our observation that persons in the Triana study population with higher PCB serum levels tended to have higher blood pressure." For a diagnosis of hypertension several blood pressure measurements have to be taken over time. Although a statistical association was noted with diastolic blood pres sure, such an association could have occurred by chance. To adequately control for confounding by age, sex, obesity, and ethnicity might be im possible in a population with a high prevalence of high blood pressure particularly as PCB levels also increased with age. Lawton et al.38 studied capacitor workers that were exposed to the commercial mixtures of Aroclor 1016, 1242, and/or 1254. Air levels in 1975 around the capacitor filling and sealing area ranged from 200 to 2000 |ig/m3 with a geometric mean of 690 |ig/m3. Earlier air level measure ments were not available and later measurements made by the National Institute of Occupational Safety and Health (NIOSH) showed TWA values for personal air samples that averaged 168 pg/m3. Initially, the serum PCB levels ranged from 57 to 2207 ppb (pg/I). Lawton et al.38 found positive correlations with serum cholesterol, the liver func tion tests, GGT, and alanine aminotransferase. A positive association with high blood pressure was noted. However, when PCB serum levels were expressed on a lipid basis, the associations disap peared, suggesting that they were spurious and related to the concentration of the lipids in serum and indirectly to age. In this study PCBs also increased in serum with age and were higher in males than in females.37 The authors concluded that "these findings differ from those of most other recent studies of environmentally or occu pationally exposed populations in offering evi dence of microsomal enzyme induction which is consistent with the extensive LPCB (low poly chlorinated biphenyls) clearance also observed in these highly exposed workers. The evidence is statistical in nature and the associated parameters remained within their normal clinical ranges. The effects were therefore physiological rather than pathological and appeared to subside following the cessation of PCB use." Maroni et al.39 studied 40 male and 40 female workers from two plants with 6 or more years of exposure. A total of 67 workers were exposed to pyralene 3010 (PCB with 42% chlorination) in a plant manufacturing capacitors, and 13 additional workers were exposed to Apirolio, also a 42% chlorinated mixture ofPCBs. The 13 workers tested electrical capacitors by short-circuiting them. Oc casionally, during testing, the capacitors exploded and the PCBs dispersed. Air concentrations in the factory ranged from 48 to 275 jtg/m3, surface con tamination from 0.2 to 159 jig/cm2, and 2 to 27 |tg/cm2 were measured on the palms of the workers' hands. The PCB blood levels ranged from 41 to 1319 ppb (p.g/1) with mean total PCB levels of 337 and 292 ppb (|xg/l) for current and former workers, respectively. Four workers had definite chloracne. The mean PCB level of the chloracne patients was 450 ppb (jtg/1) with a range of 310 to 547 ppb (jJ.g/1). The frequency of abnormal liver function tests increased with increasing PCB se rum concentrations. Most workers had an enlarged liver on physical examination and the GGT was elevated. A case of chronic myelocytic leukemia and two cases of cavernous hemangioma were noted among the workers but were thought by the authors to be unrelated to the PCB exposure.39 According to the authors three of the workers with abnormal liver function tests had underlying dis ease that could have been the cause of the enlarged livers and only in a few additional cases was a well-defined mild liver failure present When the authors tested the frequency of hepatic involve- 140 DSW 148237 STLCOPCB4039722 ment against blood PCB concentrations a signifi cant trend was observed for trichlorobiphenyls but not for pentachlorobiphenyls. It was unclear from the report whether findings in females and males were the same. A closer examination of the results of the liver function tests listed in the paper showed most of them to be in the normal range of most laboratories to 10 to 45 U/l for AST (aspartate aminotransferase), and 10 to 35 U/l for ALT (ala nine aminotransferase). The normal ranges given in the paper were unusually low unless they did not represent international units. Only one worker had a definite elevated GGT; however, his exposure was less than the exposures of other workers with normal liver function tests. PCBs, other chlori nated compounds, medicinal drags, alcohol, and drags of abuse are known to induce GGT. Chloracne observed in four workers may have resulted from exposure to chlorinated dibenzofurans or other chemicals because electrical grade PCBs per se have not been reported to produce chloracne in most other plants where PCBs were used and exposure was similar or higher.38-4041 In earlier years, some workers in the U.S. exposed to PCBs also developed chloracne in unusual situations.41-43 However, exposure to chlo rinated dibenzofurans and perhaps chlorinated naphthalenes could not be ruled out in these ear lier chloracne reports. Colombi et al.44 studied 40 male and 27 fe male workers who made transformers and electri cal condensers with a mean age of 37 years and a mean exposure of 12 years. The PCBs used were Aroclor 1254 during 1949 and 1965, and Pyralene 3010 from Prodelec, France, with 42% chlorina tion since then. These workers smoked no more than one pack of cigarettes a day and drank no more than 80 g of alcohol/d. The controls con sisted of 30 male and 37 female volunteers with a mean age of 35 years. The mean PCB blood level of the exposed workers was 386 ppb (jJLg/1) with a range of 162 to 1319 ppb (p.g/1). Increased urinary uroporphyrin excretion in the PCB work ers was reported by the investigators with a nor mal uroporphyrin/coproporphyrin ratio. Careful examination of the data provided in the paper suggests that porphyrin levels were not reported for all participants, but no information is given on the number of workers with nondetectable levels. The ranges of the different porphyrins given for the exposed and the control group in the paper are not elevated above the normal ranges established by Hill et al.45 and personal communication ex cept for four coproporphyrin values. These four elevated coproporphyrin levels appear to be unre lated to the PCB exposure because no changes in the porphyrin pattern were observed. According to the authors the uroporphyrinrcoproporphyrin ratio was normal. The authors concluded that "at the present time it is not known whether PCBs can also be considered porphyrogenic to man; in other words, whether the observed alteration can progress to the subsequent stage of chemical porphyria. Although clinically evident porphyrias have never been reported in subjects with occupa tional exposure to PCBs or in subjects with acute or subacute PCB intoxication (Yusho patients), this possibility cannot be excluded." Smith et al.41 conducted a survey of three groups of workers occupationally exposed to PCBs in the U.S. One group made capacitors and was exposed to Aroclor 1242 and 1016. The other two groups were transformer maintenance workers for a public and a private utility company and were exposed to Aroclor 1254 or 1260. For capacitor workers, the air levels ranged from nondetectable levels to 264 |ig/m3, and the skin wipes ranged from 10 to 668 jig/cm2. The lower chlorinated PCB serum levels ranged from 2 to 3330 ppb (p.g/1) and the higher chlorinated PCBs from 1 to 250 ppb (pg/1). The hematology tests did not re veal any abnormalities related to PCB exposure. No consistent pattern of abnormalities was noted on physical examination. None of the participants had acneform lesions suggestive of chloracne. Although a correlation was noted by the authors between serum log HPCB (highly chlorinated PCBs) and diastolic blood pressure multiple lin ear regression analysis showed the apparent asso ciation to be attributable to an association prima rily with age and gender. Urinary creatinine, porphyrins, glucose, and steroids were normal. The log transformations oftwo liver function tests, AST and ALT, were positively and significantly correlated with log PCB serum levels. A positive correlation with triglycerides and an inverse cor relation with high-density lipoprotein cholesterol in fasting serum specimens was noted. Although DSW 148238 141 STLCOPCB4039723 statistically significantly different, it cannot be determined from the information given in the paper whether these findings have any clinical signifi cance because the ranges and means of the liver function tests were not provided. No clinical ab normalities related to PCB exposure were noted. According to the authors, "alterations in liver enzyme tests in association with increasing serum PCB concentration may not themselves be pre dictive of future chronic liver disease but may reflect liver microsomal enzyme induction." Chloracne or other occupationally related skin lesions were not observed. Air levels in the public utility company ranged from 37 to 215 Jig/m3. In the public util ity transformer workers, the lower chlorinated biphenyls in serum ranged from 1 to 59 ppb (jj.g/1) and the higher chlorinated biphenyls from l to 24 ppb (p.g/1). No hematologic changes that correlated with PCB exposure were noted after adjusting for confounders. Two liver function tests, AST and ALT, were positively and signifi cantly correlated with PCB levels. An inverse correlation with high-density lipoprotein choles terol and a positive correlation with triglycerides in fasting serum specimens were observed. Uri nary porphyrins, steroids, and glucose were nor mal. Results for the private utility company were similar. Air levels ranged from 0.4 to 82.3 pg/ m3, PCB concentrations on skin wipes ranged from 5 to 487 p.g/100 cm2. The serum PCB lev els for the lower chlorinated biphenyls ranged from 9 to 52 ppb (pg/1) and for the higher chlo rinated biphenyls from 3 to 250 ppb (pg/l). The results of hematological studies were within the normal range after adjusting for confounders. The age-related effects mentioned by the au thors have been well established. Patients with higher low-density lipoproteins and higher cho lesterol are more likely to be older, overweight, have arteriosclerosis, hypertension, and higher liver function tests.46-47 Such individuals also have higher PCB serum levels, as a proportionally larger amount of PCBs will remain in serum with higher lipid levels. Fischbein et al.40 conducted a clinical field survey of capacitor manufacturing workers, 168 males and 158 females. The workers had been exposed to some Aroclor 1254 and 1221, but mainly to Aroclors 1242 and 1016. They were between 30 and 70 years old and 69% had been exposed to PCBs for more than 10 years. Their general health was not affected and clinical labo ratory test results were within the normal range, except that there was a positive statistical correla tion between the AST liver function test and se rum levels of the more highly chlorinated biphe nyl congeners. It is difficult to interpret the abnormal liver function tests because the authors of the studies provided little detail and did not control for confounders. The authors noted that "the paucity of the physical findings in the PCB workers was striking and that this included a very low prevalence of abnormal liver findings although in a subgroup of the workers it was demonstrated that the cytochrome P-450 was induced in the liver." Additional studies of this type have been reviewed previously.48 Meigs et al.,43 Ouw et al.,49 and Elo et al.50 also noted slightly elevated liver function tests in one or more workers exposed to PCBs. To appro priately interpret these results the tests would have to be repeated in conjunction with other clinical laboratory work51 to ascertain the reason for the abnormal liver function tests. Furthermore, the serum PCB levels would have to be quantitated on a lipid basis. In the study by Elo et al.,50 worker exposure to polychlorinated dibenzofurans and perhaps other chemicals may have occurred. In the study by Meigs et al.,43 the PCBs were used as heat exchangers that may have resulted in the formation of chlorinated dibenzofurans. In conjunction with the Fischbein et al.40 study, Warshaw et al.52 surveyed the respiratory func tion of 309 male and female capacitor workers. An additional 66 workers were excluded because of exposure to asbestos, talc, or textile dust. The mean age of the males was 41.1 12.6 years and of the females it was 47.3 12.5 years. The average length of exposure for the male workers was 14.9 years (S.D. = 8.11) and 16.1 years (S.D. = 8.3) for the females. Results in a group of nonsmokers reported by Morris et al.53 were used for comparison. A total of 48% of the capacitor workers reported eye and upper respiratory irrita tion; 10% had experienced tightness of the chest; 35% of the males and 30% of the females were smokers or ex-smokers; 13% reportedly had 142 OSW 148239 STLCOPCB4039724 chronic bronchitis; an abnormal forced vital ca pacity was found in 14% of the males and 13% of the females; 13% of the males and 9% of the females had restrictive impairment; and intersti tial changes were seen on X-ray in 6 workers. The normal population of Morris et al.53 of 957 subjects had a 5.6% prevalence of abnormal vital capacity. Vital capacity measurement is af fected by age, height, and by the individual's effort to maximally inhale and exhale. In normal individuals the test is affected by smoking within the last hour before the test, use of bronchodilators, a heavy meal within the past 2 h, or an upper respiratory infection within the past 3 weeks be fore the test. NIOSH recommends that the test be performed in a sitting position unless the patient is obese. However, in the present study the test was performed in the standing position. Appar ently, 13% of the workers had "chronic bronchi tis". In chronic obstructive bronchitis prevalent in smokers forced vital capacity and forced expira tory volume are reduced. This observation raises questions about the appropriateness of the control group. Because only summary results were re ported, it is unclear whether the authors compared the results of the individual workers based on age and height to the results of Morris et al.53 or whether they compared means. In the original PCB study40 from which these workers were drawn 40% of the male and 53% of the female workers were older than 50 years. Furthermore, it was not clarified in the paper whether the observed chronic bronchitis was present in smokers or ex-smokers and whether they accounted for most of the work ers with abnormal pulmonary function tests. Kalina et al.54 studied chromosome aberra tions and sister chromatid exchanges (SCEs) in 32 PCB exposed workers in Czechoslovakia and two sets of 20 controls matched on smoking be havior. The authors noted a higher frequency of chromosome aberrations in the PCB workers ex posed for more than 10 years. SCE values were significantly increased in workers exposed for 15 years. According to the authors, the workers may also have been exposed to benzene and formalde hyde. The authors collected information on expo sure to medications, radiation, and other factors but did not report on these results. Increases in SCEs have been found in smokers, but more im portantly the workers were exposed to chemicals known to produce these types of aberrations. Because PCBs are relatively inert chemicals, they would not be expected to produce chromosome aberrations. Similar observations have not been reported in other worker populations. Kilbum et al.55 examined 14 firemen who had been exposed to byproducts generated in a trans former fire and found memory impairment, and impairment of cognitive function. No correlation was found with the firemen's serum or adipose tissue PCB levels, which were within the range of the general population. Following 2 to 3 weeks of supervised diet, daily exercise, sauna, and the administration of unsaturated fat and niacin, a significant improvement in behavioral perfor mance was noted. It is the opinion of this re viewer that the observations made in the firemen seem to be unrelated to their potential exposure to PCBs from the transformer fire. In another electric transformer fire in a State office building in Binghamton, NY, reported ex posure in fire fighters was positively related to PCB serum levels. However, the means and in dividual values were within the range reported in other studies unrelated to any unusual PCB exposure. Although significant correlations were observed between serum PCB levels and levels of liver enzymes and serum lipids, mean levels of these biochemical parameters were not asso ciated with reported exposure after an adjust ment for relevant covariables, such as age and alcohol consumption.56 These two studies of fire fighters exposed to transformer fires suggest that very little PCBs, if any, were taken up by these fire fighters. Alvares et al.57 measured the half-life of antipyrine in five workers occupationally exposed to Aroclor 1016 in acapacitor manufacturing plant. In the exposed workers, the antipyrine half-life was significantly lower (10.8 h) than in matched controls (15.6 h) of the general population. The authors found that Aroclor 1016 elicits the barbi turate type of inducing effects on drug metabo lism in humans. The authors also demonstrated this effect in rats.58 Recently, the suggestion has been made that PCBs may play a role in the development of breast cancer.59 Falck et al.59 found elevated lev- DSK 148240 143 STLCOPCB4039725 els of PCBs and DDT and its derivatives in mammary tissue of women with breast cancer when compared with women with benign breast cancer when compared with women with benign breast disease. The median age of women with benign breast disease was 58 and 64 years for women with malignant disease. Because chemi cals such as PCB and DDT (1,1'-(2,2,2trichloroethylidene)i>is[4-chlorobenzene]) in crease in humans with age, part of the observed increase may be explained by the differences in age distribution.37 A review of the data given in the paper indicates that several cancer patients did not have higher PCB levels than patients with benign breast disease. Slight but statisti cally significant differences in body burdens of persistent lipophilic compounds have also been reported for other cancers by other authors.60-65 On the other hand, Teufel et al.66 did not find elevated levels of PCBs in children with tumors or congenital malformations if PCB concentra tions in adipose tissue were compared with those of children without such lesions. No inferences can be drawn from these findings. The underly ing disease may affect body functions. Metabo lism, excretion, and distribution of these chemi cals may be altered in some situations affecting PCB levels. Furthermore, slightly higher con centrations of chlorinated aromatic compounds in different groups of people may be a reflection of differences in lifestyle. Industrial chemicals are usually slightly higher in urban populations, while the concentrations of pesticides are higher in people living in rural areas and in the South. It may also be fallacious to correlate tissue lev els with the clinical expression of a disease that may have existed in a dormant stage for some time in the past. The principal risk factors for cancer of the breast are marital status, parity, urban residence, older age at marriage and first pregnancy, lack of exercise, heredity and hor monal factors, early menarche, late menopause, and perhaps regular alcohol consumption.67-68 Low-level exposure to environmental chemicals such as PCBs and DDT and their derivatives would be inconsequential compared with these major risk factors. Based on these and other more recent studies. Key and Reeves concluded "that for PCBs there is no evidence for an asso ciation with breast cancer risk."681 In conclusion, based on the findings reported in the various studies reviewed in this section, no abnormal clinical findings that could be related to PCB exposure were made in the workers. The positive association with diastolic blood pressure when investigated further was found to be related to aging and to serum lipid concentrations that increase with aging in populations on a Western diet. It was also found that as lipids increase in serum so do PCBs. To eliminate this confounder PCBs should be expressed on a lipid basis in serum. By most authors, the slightly elevated liver function tests were thought to be an expression of the induction of mixed function oxidases (mi crosomal enzymes) rather than an expression of liver disease. Some study subjects had high-se rum lipids. High-serum lipids (turbidity) could have interfered in occasional study subjects with the liver function test measurements resulting in higher liver function tests as a laboratory error. This possibility would have to be considered only when occasional members of a group were af fected. The abnormal pulmonary function tests reported in one study may have been related to smoking rather than PCB exposure. The differ ences in body burdens between assorted groups with and without breast cancer or with benign or malignant disease were small, as were the number of study subjects. Although statistically signifi cantly different, they were within the range of levels found in the general population. It is un clear whether the observed differences in tissue levels could be the result of a selection bias unre lated to the disease or whether the differences in levels represent a proxy for other factors. Biologi cally, the small differences in amount of chemi cals that these different groups were conceivably exposed to make a causal association with the underlying disease implausible. V. REPRODUCTIVE AND DEVELOPMENTAL EFFECTS Taylor et al.69 studied the reproductive effects of PCB exposure in female capacitor workers by reviewing birth weight on birth certificates. The exposed group consisted of 51 infants of female capacitor workers. The mothers of these infants had been exposed to Aroclors 1254, 1242, and 144 DSW 148241 STLCOPCB4039726 1016 for 1 year or more prior to the birth of the infant. Their mean age was 32.4 6.3 years. Controls were 337 infants of other female (mean age of 26.5 4.9 years) workers in the same plant with indirect low exposure to PCBs. The gesta tional age at birth for the infants bom to mothers who had high exposure to PCBs was 6.6 d shorter than in the controls. After adjustment for gesta tional age, the exposed infants weighed an aver age 58 g less than the controls. According to the authors of this study, it was not possible to control for tobacco use, underlying medical conditions, maternal height, or a previous history of low birth weight. The small number of observations, lack of information on important influencing factors, and questionable biological significance of birth weight and gestational age differences ofthis small size made any conclusions about the effect of PCBs on birth weights tentative. In 1989, Taylor et al.70 reported an additional expanded reproductive study in the same capaci tor work force. The exposed group of this second study consisted of offspring of 200 female ca pacitor workers that had been exposed to Aroclors 1254,1242, or 1016 for 1 year or more prior to the birth of the infant The controls consisted of in fants of 205 other female workers in the same plant with indirect low exposure. Again, a signifi cant negative correlation among birth weight, gestational age, and estimated semm PCB levels was noted. An average decrease of 33 g in the birth weight of infants bom to mothers with direct exposure to PCBs was observed. According to the authors, the biological importance of such a small difference in birth weight has no clinical signifi cance for term infants. In North Carolina and Michigan, studies have been performed to elucidate the effect of PCBs on infants of the general population. In both studies, the authors concluded that postnatal exposure through human milk was not associated with any ill effects. They indicated, however, that a trans placental neurobehavioral effect could be demon strated. This effect was transient in one study (North Carolina) and persistent in the other. In the North Carolina cohort, a total of 856 infants of the general population were enrolled in the study as neonates and followed for several years.71 Selection of the study participants was not based on presumed higher exposures than the general population. PCBs and DDT and deriva tives were measured in human milk, semm, pla centa, and cord blood. Results of these measure ments were used to calculate prenatal exposure of the infants. In the final exposure assessment, PCB levels from cord blood and placenta were ex cluded because they were mostly below the limit of detection of the analytical method.72 In the assessment it was assumed that the mother weighed 60 kg, 25% of her weight was fat, the only means of excretion of PCBs for the mother was deposition ofPCBs into the fetus and through lactation, the daily dose was constant over a life time, and the mother was primiparous. However, these assumptions are erroneous as it was not considered that PCBs are metabolized and ex creted in humans with congener specific half lives of days, months, or years. In addition, Tilson et al.73 in a review article using the same data estimated the dose response relationship differ ently. These authors73 arrayed the crude neuro behavioral data (percent abnormal on the Brazelton neonatal behavioral assessment scale or means on the Bayley scales of infant development) of the North Carolina cohort by the PCB level in milk fat and estimated a no observed adverse effect level (NOAEL) by inspection. It was not stated in the paper how missing PCB values were treated. The NOAEL appeared to be about 3.4 ppm (mg/ kg) PCBs in human milk fat for the percent abnor mal on the Brazelton assessment or means on the Bayley scales. For visual recognition memory the level was 1.0 ppm (mg/kg) in human milk fat. In the Michigan study, a population eating fish from Lake Michigan was compared with a control group that did not consume such fish. In several papers describing these studies,74-78 neurobehavioral changes lower birth weight and a reduced gestation period were reported in as sociation with higher fish intake presumed to also represent higher PCB intake. According to Paneth,79 who reviewed these reports, exposure in this population was poorly defined, dose re sponse relationships were not well established, and other potential confounders such as expo sure to heavy metals, the mothers lifestyle and well being, and genetic make-up were not con sidered. The Michigan studies are based on a group of infants of which 242 were considered exposed because their mothers stated that they DSW 148242 14S STLCOPCB4039727 ate Lake Michigan fish and 71 were considered not be exposed because their mothers denied eating Lake Michigan fish. The PCBs were mea sured in samples of cord serum, serum from the mothers, and in the mothers milk. Results of PCB measurements were published in two papers.74 80 In the Schwartz et al.74 paper, it is unclear in how many of the 71 "controls" (non-fish eaters) and of the 242 "exposed" (fish eaters), PCBs were de tected or quantified because the controls and the exposed groups were combined in the report. Two standards were used for the quantitation of the PCBs: Aroclor 1016 and 1260. In a number of sera and milk specimens, PCBs could not be quan titated or were not detected.74 The mean maternal serum PCB concentrations were 5.5 ppb (jig/1) for Aroclor 1260 and 1.6 ppb (p.g/1) for Aroclor 1016..As expected, the cord serum levels were lower because lipid levels are lower in cord blood than in maternal blood. The non-fish-eating group had mean PCB serum levels of about 4 ppb (p.g/1) and those eating 52 to 183 meals of fish per year had mean PCB serum levels around 9 ppb (|ig/l). It is not clarified in the papers how the many nondetectable levels and the nonquantifiable lev els were accounted for in the correlation analyses. In contrast with the Michigan studies, no effect on birth weight was observed in the North Carolina studies,72 even though the cohort was much larger and the serum PCB levels of the mothers were about the same. The neurobehavioral effects in the Michigan and the North Carolina studies differed (Table 2), suggesting that in the two study groups the same region of the brain was not affected and that the two studies were inconsistent with each other.71 The children in both studies were followed for several years and a variety of tests were per formed (Table 2). At the time of delivery the in fants in the North Carolina study received a physi cal examination and the Brazelton neonatal behavioral assessment scales (BNBAS) were ad ministered. The BNBAS yields results of 27 be havioral scales and 20 reflexes. To simplify analy sis, the results were summarized into clusters81 using the clusters developed by Jacobson et al.76 For the BNBAS, the only cluster scores signifi cantly affected by PCBs or DDE were the tonicity and reflex scores. Higher PCB levels were associ ated with less muscle tone and activity. The au thors also noted that infants who were delivered by emergency Cesarean section or with low forceps or suction had a greater number of abnormal reflexes, but this did not change the significantly positive correlation with PCBs or DDE. Mothers with higher PCB serum levels were older and were more likely to consume one or more alcoholic beverages per week. It is not stated in the paper what the associa tion was between alcohol consumption and abnor mal reflexes.71,72 The BNBAS were performed during the first week in 51 % of the infants, in 20% during the second week, and in 16% during the third week of life. According to some investiga tors, the BNBAS should only be administered dur ing the first 3 d of life.71 Jacobson et al.76 also administered the BNBAS and reported a decrease in reflexes, but an increase in startle reflexes. The increase in startle reflexes was not noted by Rogan et al.71 and is usually not observed with hyporeflexia (decrease in reflexes as reported by Jacobson et al.)76 and hypotonia. In a subset of these children at the age of 4 years, Jacobson et al.83 noted a de crease in verbal and memory scales, while no such association was found among exposure and verbal and memory scales in the North Carolina cohort. In the North Carolina cohort, at 12 months of age, the psychomotor development was decreased as the calculated in utero PCB exposures increased. The observed neurobehavioral effects were no longer detectable at the ages of 3, 4, and 5 years in the North Carolina cohort,82 while this was not the case in Michigan.83-84 In addition to these epidemiology studies, Leoni et al.8S collected blood samples from 120 women who miscarried and 120 women who car ried their infant to term. PCBs reported as Fenclor 54 were significantly higher in women who mis carried. However, if PCBs were quantitated as decachlorobiphenyl following perchlorination, this statistically significant difference disappeared. Thus, the initial quantitation may have introduced an error that was removed when PCBs were quan titated as decachlorobiphenyl. These authors re ported that PCB levels correlated negatively with milk and fish consumption and positively with age and alcohol consumption. It is well known that age and alcohol consumption affect the preva lence of miscarriages. Because PCBs increase with age and have been shown to increase with alcohol 146 DSW 148243 STLCOPCB4039728 TABLE 2 Neurobehavioral Tests and Outcomes from Two Prospective Studies of Children Authors Tests Age Outcome Jacobson et. a|_ 75.76,78.80,63,84 Schwartz et al.74 Fein et al.77 Gladen72 Gladen and Rogan82 Rogan et al.71-81 Ballard maturity Brazelton neonatal behavioral assessment Infant fixation to novelty (illuminated pictures) McCarthy scales Mother Peabody picture vocabulary test Child behavioral record (adapted from Bayley infant behavior record) Brazelton test Bayley scales of infant development McCarthy scales Report cards 3d 7 months 4 years 3 d-weeks 6 months 12 months 3, 4, and 5 years 8 years Motoric immaturity; lability poor, startle T; reflexes i Less fixation with higher transplacental placental PCB exposure Verbal scales i; memory scales i; reduced activity based on composite rating between mother and examiner, test retest reliability moderate; r = 0.44 Outcomes unrelated to 4 year body burden; therefore effect trans placental? Hydroreflexia Verbal and memory scales normal Psychomotor development i with increased transplacental PCB exposure; unrelated to hyporeflexia noted at birth No association between mental and psychomotor scales and PCB exposure No relationship between poorer grades and PCB exposure consumption, they may in this case have served as a proxy for other parameters. In conclusion, the studies on reproductive effects show a slight but clinically insignificant effect on birth weight in highly exposed workers. Because the information in the worker studies was obtained from birth certificates, other factors that would affect birth weight could not be effec tively controlled and not all confounders may have been considered. In reproductive animal stud ies, effects on birth weight represent a sensitive indicator of fetal toxicity. It is remarkable that even with the very high exposures of female ca pacitor workers no clinically significant effect on birth weight of their offspring was noted. In the study of capacitor workers and in the North Carolina and Michigan studies, in utero exposures ofthe infants were estimated indirectly, which may have introduced a misclassification bias. In the Michigan studies exposure assess ment was partly based on fish consumption; how ever, the levels of PCBs and other chemicals in the consumed fish were unknown. In the North Carolina studies in indirectly assessing in utero exposure the erroneous assumption was made that humans do not excrete PCBs. In addition, analy- DSW 148244 147 STLCOPCB4039729 sis of a few milk samples does not capture the fluctuations of PCB levels in human milk over the months of nursing and during a given feeding period and are not a good estimate of maternal body burdens during pregnancy and transplacen tal exposure of the fetus. Thus, the estimated NOAELs calculated by Tilson et al.73 based on this information are also tenuous. In addition to the uncertainties in exposure assessment, the neurobehavioral tests used have not been standardized for large population stud ies. Because of their subjectivity, differences be tween examiners exist and customarily the inter pretation of these tests is combined with other clinical information. Merely by chance some sta tistically significant differences would be expected as these tests generate information on many end points. The fact that the two studies did not show the same results supports the possibility of a chance finding or may have been unrelated to PCBs be cause the exposures were not well defined. Some of the infants in the Michigan studies may have been premature (the lowest birth weight was 909.1 g), and the authors reported that the infants bom to mothers who consumed Lake Michigan fish were lighter75,77 than the controls. For premature infants the neurobehavioral test results would be different than for term infants. It is not stated in the published papers how immaturity might have affected the results. VI. IMMUNOLOGICAL EFFECTS Immunological responses to PCB exposure have been evaluated in a limited number of epide miology studies. None of these studies established a causal link between PCB exposure and immunotoxicity.8'86,87 Emmett86 evaluated delayed hypersensitivity of the dermal reaction to 0.1 cc intradermally injected mumps and trichophyton antigen in transformer maintenance and repair workers and a matched control group not exposed to PCBs. The delayed hypersensitivity reaction is a measure of a type IV delayed hypersensitivity reaction mediated by sensitized T-cells and mac rophages. The proportions of positive responses to mumps (92% for the exposed transformer main tenance workers, 89% control) and trichophyton (17% exposed, 8% control) did not differ signifi cantly.8687 The mean diameters of the skin reac tions to mumps antigen in the two groups were identical. The semm PCB concentrations in the 54 past and presently exposed workers tested with the antigens ranged from nondetectable levels to 300 ppb (p.g/1) and in the nonexposed comparison group they ranged from nondetectable levels to 15 ppb (fxg/l). Emmett86-87 also performed a bat tery of clinical laboratory tests in these PCBexposed transformer maintenance workers. Ac cording to the authors, except for a possible positive association between PCB serum levels and increases in mixed function oxidases, no other changes were positively associated with PCB se mm levels after the results were controlled for confounders. In conclusion, it is uncertain whether and at what semm PCB concentrations adverse effects would occur in humans, including immunotoxicity. Illnesses such as increased numbers of acute or chronic infections have not been observed, even in highly exposed individuals. However, such effects were reported in the Yusho and Yu-Cheng outbreaks. In these two poisoning episodes, expo sure to high concentrations of PCDFs also oc curred. The "PCB exposed" cohort of Emmett et al.86-87 had similar geometric mean semm PCB concentrations of 9.7 ppb (|ig/l) (highest level 300 p.g/1) as the steady-state semm level of 10 ppb (jxg/1) in a study of rhesus monkeys given 5 jig/ kg/d PCBs for 23 to 55 months.88 The rhesus monkeys were reported to have abnormal im mune function tests,88 suggesting that rhesus monkeys may not represent a good animal model for humans. VII. MORTALITY IN CAPACITOR AND TRANSFORMER WORKERS Several mortality studies ofworkers with very high and long-term exposure to PCBs have been conducted to determine cancer mortality. These studies (in the reviewer's opinion) have not pro vided any convincing evidence that PCBs are human carcinogens. To evaluate mortality studies in the aggregate and conclude whether a causal association exists is determined by several con 148 DS V*8245 STLCOPCB4039730 siderations. Results from similar studies should provide similar results. The same type and sitespecific cancers should be increased at similar exposures in comparable studies. Workers with higher and longer exposures and longer latency periods should be more highly affected. Expo sures should be well characterized. When com paring observed and expected cancers, the con fidence intervals should generally not include one. If the confidence intervals are wide (e.g., risk ratio = 3 and C.I. = 2-20) the precision of the risk ratio is low. For cancers known to have specific etiologies such as lung cancer and smok ing, control for confounders is important. How data are evaluated is also governed by the types of cancers observed. Some cancers such as the mesotheliomas are usually associated with ex posure to asbestos, while the prevalence of other cancers such as the carcinomas ofthe gastrointes tinal intestinal tract are thought to be affected by lifestyle and diet. A slight increase in liver (3 observed, 1.07 expected) and rectal (4 observed, 1.19 expected) cancers among 2567 male and female capacitor manufacturing workers in two facilities located in upstate New York and western Massachusetts were noted by Brown and Jones.89 One hepatocellular carcinoma occurred in a male worker at one plant. The other two carcinomas occurred in two fe males in the other plant. Similarly, one rectal cancer occurred in a male worker at one plant, while the other three rectal cancers occurred in females in another plant. In a follow-up study of the same cohort several years later, a continued excess of cancer of the liver was found (5 ob served, 1.9 expected), but no additional rectal cancers were observed and the previous excess in rectal cancer was no longer statistically signifi cant (4 observed, 1.9 expected).90 No additional tumors of the liver occurred in one of the plants. In the other plant one additional cancer of the liver was found in a female worker. Review of hospital records by Brown90 showed that not all of the liver cancers were primary liver cancers. Once Brown reclassified the tumors, the differences for liver tumors between the workers and the general population were no longer statistically signifi cant. In the statistical analysis the male and fe male workers were combined. Taylor91 followed the mortality of 6292 male and female capacitor workers in upstate New York. The overall cohort mortality was significantly less than the general population (510 observed vs. 614 expected; SMR 83, 95% C.I. 0.76--0.90) and no increase was noted for total or specific cancers. In Sweden, Gustavsson et al.92 performed an SMR analysis on 142 male capacitor workers. A total of 21 deaths was reported and 22.1 were expected. Neither total cancer deaths (7 observed, 5.39 expected) nor incident cases (7 observed, 7.58 expected) were elevated. The small sample size of this cohort precludes meaningful interpretation. Bertazzi et al.93 studied total and cancer mor tality in 2100 Italian workers employed in a capacitor manufacturing plant. Deaths for all causes were not affected in males (30 observed, 27.8 expected; SMR 108). The SMR analysis resulted in an excess of total cancers among males (14 observed, 5.5 expected; SMR 253, C.1.1.44-4.15), mainly in the digestive tract, the lung, and hematologic sites. Overall mortality in females exceeded expectation (34 observed, 25.8 expected, SMR 132 for the national comparison, 16.5 expected for the local comparison; SMR 206, 95% C.I. 1.45-2.85), as did total cancer mortality observed, 7.7 expected. The increase in all cause mortality in females was primarily the result of 9 accidental deaths of which 7 were non-work-related traffic accidents. In this co hort, work histories for employees were not avail able prior to 1954 and exposure was not well documented. Additionally, these workers were exposed to other chemicals in the plant. Accord ing to the authors "the cancer excess mortality was statistically significant only when compared with the local population. The local population was known to have a lower cancer mortality than the national population. Analysis by duration of exposure, latency, and year of first exposure did not reveal any definitive pattern or trend of mortality for any of the relevant causes." The authors in their report of the results also pointed out that for some tumors the latency period was very short, 3 months in one example, or workers had been hired at an advanced age. They state: "given this information no clearcut and definite conclusion regarding the association between cancer of the GI tract and exposure to PCBs can DSW 148246 149 STLCOPCB4039731 be drawn from the results of the study." The authors also pointed out in their discussion that "the limitations discussed did not permit a causal association to be either proved or dismissed." More recently, Sinks et al.94 conducted an SMR study of 3588 male and female workers at a capacitor plant and found a significant increase in melanocarcinomas (8 observed cancers vs. 2 expected). A nonsignificant increase in cancer of the brain was also noted. A dose-response relationship or increase with latency could not be established, nor were medical records avail able for all cases to confirm that all tumors were primary tumors. Furthermore, one of the melanocarcinomas originated in the gallbladder and should not have been coded as a tumor origi nating in skin. Another melanocarcinoma was diagnosed before the individual began employ ment at the plant. The observed number of deaths from all causes was 192 and the expected num ber was 283.3, with an SMR of 70 (95% C.I. 0.60-0.80). Similarly, the observed number of all cancers was lower than expected (54 ob served vs. 63.7 expected; SMR 80, 95% C.I. 0.60-1.10). Bahn et al.95 in a letter to the editor reported two melanocarcinomas from a group of 72 petro chemical workers exposed to PCBs in New Jer sey when only 0.04 were expected. These work ers were exposed to other chemicals and the potential confounding of these other exposures was not analyzed. In conclusion, the epidemiological studies in the aggregate have failed to provide consistent data regarding an increase in overall mortality or in specific cancer mortality of workers exposed to PCBs. The lack of a consistent increase in total or specific cancer mortality across studies does not suggest any particular trend between PCB expo sure and cancer mortality. The workers were pri marily exposed to the lower chlorinated mixtures such as Aroclor 1242 and 1016 with some expo sure to Aroclor 1254, while the rat studies show ing a carcinogenic effect of PCBs were conducted with a more highly chlorinated PCB mixture, Aroclor 1260 or Chlophen A-60. The majority of mortality studies were con ducted on relatively small study populations. Dif ferences in sample size, in the quality of PCB exposure assessment, in the length of exposure, time of exposure required to be eligible for inclu sion in the study, and differences in the cancer incidence of the control populations from differ ent countries preclude any comparisons across studies or combining the study populations for a "meta-analysis" as reported by the Industrial Dis ease Standards Panel, Report no. 2, Toronto, Ontario, December 1987 and published in the Ontario Gazette. The Panel stated that while only seven deaths from liver and biliary cancer oc curred, the association was strong and achieved statistical significance when these studies were combined in a meta analysis. However, the Panel did not consider that four cancers occurred in males, two in the Bertazzi93 study and one each in the Gustavsson92 and Brown90 studies. Among females, all three liver cancers and cancers of the biliary tracts occurred in one plant. Furthermore, a higher than expected number of liver cancers was not found in all mortality studies now avail able. The number of tumors in the different plants was very small and on follow-up in the Brown study,90 for instance, no further liver cancers or cancers of the biliary tract were found. It is also debatable whether tumors of the biliary tract and gallbladder should be combined with hepatocel lular carcinomas as their etiologies differ. The various statistically significantly increased sitespecific tumors may have occurred by chance. The findings are not consistent across studies, the actual number of specific tumors was small, the number of tumors did not increase with latency or were they primarily found in workers with the highest exposures. In the aggregate the results of these studies are inconclusive. Further follow-up of these and other workers may ultimately clarify the reported findings as the different cohorts age and more data and longer follow-up periods be come available. VIII. RECENT STUDIES IN LABORATORY ANIMALS A wide range of adverse effects of PCBs have been investigated in numerous animal stud ies over the past 20 years. At lower doses, repro ductive and developmental as well as immuno- 150 DSW 148247 STLCOPCB4039732 toxic effects have been reported primarily in subhuman primates and mink. It is beyond the scope of this manuscript to review these data in detail. They have been summarized recently.96-100 Only a few recent studies are mentioned here. The primary focus is on studies where effects were noted at comparatively low doses. Further more, the recent reevaluation of the liver tumors in rodents is summarized and the toxicity equiva lency approach is discussed. A. Reproductive Effects in Animals In a study of rhesus monkeys exposed to Aroclor 1248, Barsotti et al.101 reported decreased birth weights. However, toxicity was observed in the adult females at all dose levels and the de creased birth weights could have resulted from maternal toxicity. It has been suggested that a chemi cal should only be classified as a developmental toxicant, if it produces an effect on the conceptus at levels not or only slightly toxic to the mother.102 Barsotti and Van Miller103 also reported de creased birth weights in the offspring of rhesus monkeys receiving approximately 0.04 mg Aroclor 1016/kg body weight/d. The EPA104 has used the NOAEL of 0.01 mg/kg/d from this study to derive a guideline level for noncarcinogenic effects of PCBs. The EPA-IRIS,104-105 derived a guideline intake level of 1 x 10-4 mg/kg/d by dividing the NOAEL of 0.01 mg/kg/d from the Barsotti and Van Miller101 study by an uncertainty factor of 100. However, questions have been raised about the use of this study for regulatory purposes. Al though birth weights of animals in the high-dose group and the control group differed statistically, birth weights in both groups were within the range of historical measurements for rhesus monkeys reported by Van Wagenen and Catchpole.106 All adult subhuman primates were feral; however, the control animals were purchased 4 years before the experimental animals. Because the control animals were likely to be older than the experi mental animals and had been in captivity longer, the prepregnancy maternal weights may have been greater and may have influenced infant birth weights. Significant differences in genetic makeup may have existed between the two groups of monkeys. Furthermore, the monkey chow used in the study was found to be contaminated with polybrominated biphenyl (PBB) and other halogenated chemicals. This additional exposure may have affected the results. The results ofthe Barsotti and Van Miller103 study, therefore, are inconclu sive. Additional reproductive studies have been conducted in Macaca mulatta monkeys but have not been reported in detail.107 B. Immunotoxicity Immune function has been tested by several investigators in laboratory animal species exposed to PCBs.108-112 These data have been reviewed by Vos and Luster.100 PCBs have also been suspected of being a cofactor in phocine distemper virus (PDV) infections in harbor seals. When seals were infected with PDV under experimental conditions and some of them were also exposed to PCBs, no difference in mortality was noted between the two groups.113 It is unclear whether the dose given was sufficient. Additional on-going experiments in another laboratory may resolve this issue. In a series of three recent papers, Tryphonas et al.88114'115 reported results of immune function studies in subhuman primates. Five groups of rhesus monkeys (16 animals each) received di etary doses of Aroclor 1254 of 0, 0.005, 0.02, 0.04, or 0.08 mg/kg/d. After 23 months a statisti cally significant (p <0.05) dose-related decrease of less than half of the original values in antibody titers of the immunoglobulin IgM and IgG iso topes to SRBCs (sheep red blood cells) was noted. This difference has no clinical relevance. At least a fourfold change in titer count is necessary to be considered clinically meaningful.116 Lymphocyte levels were measured in the control and 80 pg/kg/d exposed group. A significant increase in T-suppressor (Ts cells [CD4 cells]) and a significant reduction in relative number of T-helper (TH) cells (CD8 cells) and in the TH/Ts ratio when compared with controls was reported. The num ber of B-lymphocytes (CD20 cells) and total lym phocytes or lymphocyte proliferation induced by pokeweed were not affected. Stimulation by stan dard mitogens produced no significant differences in lymphocyte transformation rates. Serum im- DSW 148248 151 STLCOPCB4039733 munoglobulin, serum protein, and hydrocortisone levels did not differ from controls. After 55 months of exposure, a statistically significant dose-re lated decrease in IgM antibodies but not in the IgG levels was observed in the immunized SRBCs monkeys. Antibody response to pneumococcal antigen and lymphocyte proliferation following stimulation by mitogens were similar in treated and control groups. A significant change in the relative number of CD2 cells when compared with controls was observed in all treated groups. Results for all other lymphocyte subpopulation analyses, including absolute number of total CD2 cells, were not affected. Results of a mixed lym phocyte culture assay and a monocyte activation assay were similar in control and treated groups.88 Hydrocortisone and interleukin-1 concentrations were not affected.88 In an additional publication, Tryphonas et al.115 reported that the serum complement (CH50) activity was significantly higher in all treated groups. However, a dose response relationship was not noted. A significant trend toward in creased natural killer cell activity was observed at the adjusted effector (E) to tumor (T) target cell ratio (E:T) at the 75:1 but not at the 50:1 or 25:1 ratios. No difference in activity was found when a pairwise comparison was made between the treatment and control group. Interferon levels were significantly higher than those of the controls in the 20 and 80 pg/kg/d PCB exposed groups but lower than controls in the 40 jig/kg/d group, and the same as the controls in the group given 5 |xg/ kg/d. Because a dose-response relationship was not found, these changes seem to be unrelated to PCB exposure. In conclusion, the immunological responses reported by Tryphonas et al.88'"4-115 in the aggre gate do not suggest immunotoxicity. The TH/Ts ratio reported to be lower in treated animals may be in error because cytotoxic T lymphocytes were excluded. CD8 cells and the CD4/CD8 ratio were not affected by PCB treatment."4 Furthermore, a positive control group as recommended by Luster et al."7 was not included in the studies.88 "4'"5 C. Carcinogenesis in Animals In several 2-year rat-feeding studies an in crease in tumors of the liver has been observed in Aroclor 1260 or Clophen 60 exposed animals. No such increase in liver tumors or any other tumors has been noted following the exposure to lower chlorinated mixtures. Recently, four 2-year rat feeding studies reporting hepatocellular carcino mas were reviewed."8 Since the results of the original rat bioassays were first reported, the cri teria for classification of hepatic proliferative le sions in the rat have been revised. In the original studies, liver lesions were diagnosed using the classification scheme of Squire and Levitt.119 Rat liver lesions were diagnosed as neoplastic nod ules (sometimes referred to as hyperplastic nod ules) or carcinomas. A revised classification for proliferative rat liver lesions was subsequently developed by the National Toxicology Program (NTP)120-121 and endorsed by the EPA.122 The cur rent NTP guidelines distinguish between hyperplasia, a nonneoplastic response and adenomas, a benign tumor.120122 The neoplastic nodule (hyperplastic nodule) category in the Squire and Levitt"9 system included both adenomas and hyperplasia. Based on the Institute for Evaluating Health Risks118 review, it was reaffirmed that chronic di etary exposure of rats, in three different studies, to 60% chlorinated PCB formulations (Aroclor 1260 or Clophen A 60) resulted in the development of benign and malignant liver tumors. Chronic expo sure of rats to a 54% chlorinated PCB formulation (Aroclor 1254) did not yield a statistically signifi cant increase of benign or malignant tumors.123 Rats chronically exposed to a PCB formulation that was 42% chlorinated (Clophen A 30) did not develop a statistically significant increase in be nign or malignant liver tumors. It was established by this review that not all PCB formulations are probable human carcino gens and that different PCB mixtures do not have the same quantitative potency to cause cancer. In its previous derivation of a cancer potency factor (CPF) for PCBs, the EPA104 considered the results of the three positive rat studies of PCB mixtures with 60% chlorination124-126 but used only the female rat data from Norback and Weltman126 to calculate the CPF of 7.7 (mg/kg/d). The CPF is also influenced by the way the dose given to the laboratory animals is extrapolated to humans. To extrapolate from rats to humans, body weight scal ing to the two thirds power was used by the EPA 152 DSW 148249 STLCOPCB4039734 but not by other federal agencies with a scaling factor of 5.85. The EPA based their scaling factor on the assumption that body surface is a better predictor for interspecies variations than body weight. Recently, the EPA, the FDA, and the Con sumer Products Safety Commission (CPSC) reached a consensus on a method for cross-species scaling and proposed it as a uniform compromise policy position. Using the approach proposed by these agencies, scaling would be performed on the basis of body weight raised to the three fourths power. This correlates to a scaling factor of 3.76 when using the rat and human body weights of 450 g and 70 kg, respectively. Use of the new scaling factor would reduce prior EPA estimates of CPFs by approximately 35%. Furthermore, the reassessed data underscore major differences in carcinogenic potential based on the degree of chlorination of the PCB mixture. Studies of PCB mixtures with 60% chlorination consistently result in a high incidence of rat liver tumors. Rats fed mixtures with 54 or 42% chlori nation did not develop a statistically significant increase in liver tumors. Comparison of the results of studies with Aroclor 1260124-126 showed a striking similarity in the nature of the tumor response even though three different strains of rats were used. The car cinogenic potential of 54% chlorinated PCB mix tures has also been tested in rats by Ito et al.127 and in two mouse studies.128-129 These studies were deemed inadequate to assess the carcinogenic potency of PCBs due to their short duration, lim ited number of animals, and the absence of infor mation on the animals' health.104 Other studies of 48 or 42% chlorinated PCB mixtures have not demonstrated a carcinogenic effect.127-130 Kimura and Baba130 conducted a study of the carcinogenicity of Kanechlor 400, a com mercial PCB mixture containing 48% chlorine. The authors observed changes in the liver, includ ing benign neoplasms, but they were not statisti cally significantly elevated. This study did not show that Kanechlor 400 was carcinogenic in the rat. In conclusion, based on the available data, 60% chlorinated PCB mixtures are carcinogenic in rats producing hepatocellular tumors. At similar dosage levels lower chlorinated PCBs do not ap pear to be carcinogenic in presently available stud ies. However, more comprehensive studies with lower chlorinated mixtures are needed to confirm these observations because the general population as well as the workers proportionally received more exposure to the lower chlorinated mixtures. IX. TOXICITY EQUIVALENCY FACTORS (TEFs) The concept of toxicity equivalency factors (TEFs) to estimate the relative toxicity of struc turally similar compounds evolved during the 1980s.131-132 The TEF approach provided a rela tively straightforward and simplistic method of evaluating complex mixtures. The TEF approach was first developed to assess the risks associated with air emissions of polychlorinated dibenzodioxins (PCDDs) and PCDFs formed during hightemperature incineration of industrial and mu nicipal waste.131-133 Subsequently, the EPA132 proposed interim guidelines for estimating risks associated with mixtures of PCDDs and PCDFs for other media as well. The TEF values are based on an estimate of relative toxicity of PCDDs and PCDFs using 2,3,7,8-TCDD as a reference. With the exception of 2,3,7,8-TCDD and hexachlorinated dibenzodioxins all TEFs for chronic toxicity and carcino genic potential of PCDDs and PCDFs are esti mated from results of short-term in vivo and in vitro tests. For most PCDDs and PCDFs, chronic toxicity studies have not been conducted. For each congener, one or more of the following end points were used to derive TEFs: enzyme induc tion, thymic atrophy, body weight gain, potential bioaccumulation, teratogenicity, immunotoxicity, and lethality.132 The TEF approach as implemented now has many uncertainties and has been questioned on scientific grounds. Neubert et al.134 outlined a number of prerequisites that would have to be fulfilled in order to improve the TEF approach: 1. The actions of the congeners must be strictly additive in the dose range to be evaluated. 2. The dose response curves for the various congeners must run parallel. 3. The organotropic manifestations of all congeners must be identical, also over the relevant dose ranges. DSW 148250 153 STLCOPCB4039735 4. Dose response curves for various toxico logical endpoints for a given congener must run parallel. 5. For extrapolations between species, the kinetics must be identical, or differences have to be taken into consideration. 6. With respect to risk assessment in humans, toxic or biological manifestations in the lower dose ranges are of special interest, and LDS0 values and effects induced in labo ratory animals by highly toxic doses are of minor importance. These prerequisites have either not been met for the PCDDs and PCDFs or the information is not available. Nonetheless, in the U.S. and in other countries some risk assessors have now pro posed to assess the combined potential toxicity using the TEF approach of 2,3,7,8-TCDD, PCDDs, PCDFs, and co-planar PCB congeners. In 1990, at an EPA Workshop on PCBs, it was decided not to adopt such a scheme for PCBs because of large data gaps.135 136 Research data collected since the 1990 Workshop indicate that calculated equiva lency values for some PCB congeners are inaccu rate by better than two orders of magnitude or incorrectly predict toxic endpoints.137*138 X. CONCLUSIONS: THE HUMAN RISK In a pilot study of four female cynomolgus monkeys (Macaca fascicularis) and four female rhesus monkeys (M. mulatto) receiving daily oral doses of 0.280 mg/kg Aroclor 1254, the monkeys developed chloracne-like lesions, edema of the eyelids, eye exudate, conjunctivitis, diy and ne crotic nail beds, and loss of nails.139 In addition, the rhesus monkeys had gingival hyperplasia and the cynomolgus monkeys had mammary gland exudate. Adult female rhesus monkeys on oral doses of 0.080 mg/kg body weight for 37 months developed chloracne-like lesions or a reduction of sebaceous glands107 (Dr. Douglas L. Arnold, per sonal communication). In the monkeys given oral doses of 0.080 mg/kg body weight and in mon keys on even lower doses, weight reduction was noted. These clinical signs would be easily recog nizable had they occurred in the hundreds of workers that have been studied and in the thou sands of workers that made capacitors. However, they have not been reported. In the offspring of monkeys receiving oral doses of 0.005 mg/kg body weight and at adipose tissue levels around 5 ppm (mg/kg) and serum levels around 10 ppb (p.g/1) (Tables 3a and 3b), facial pigmentation, enlargement of Meibomian glands, discolored and hyperkeratotic finger nails have been observed in offspring107 (Dr. D. L. Arnold, personal communication). These tissue levels are within the range of levels found in the general population, particularly before PCBs were phased out. Such changes would be easily recog nizable in human infants, but they have not been described in infants in the U.S. Many of the women working in capacitor plants with much higher tissue levels had children, but none of these clini cal signs were reported in the offspring by local hospitals or factory physicians. The studies of children exposed to low levels of environmental PCBs and other persistent chemicals also did not reveal such clinical signs. However, humans are capable of producing such lesions. They were described by Japanese and Taiwanese physicians initially unfamiliar with this syndrome in con junction with the Yusho and Yu-Cheng poisoning outbreaks caused mainly by PCDFs. In humans, for the most part, the doses re ceived can only be estimated, while more precise information is available for PCB-fed laboratory animals. However, PCB serum and adipose tissue levels have been determined in the general popu lation, in groups of workers, and some experi mental animals. Comparing tissue levels across species (Tables 3a and 3b) and correlating them to adverse effects and the cumulative dose may be helpful when extrapolating across species. Tissue levels could also be used to estimate daily PCB intake if the length of exposure and the amount excreted were known. For example, in rats dietaiy levels of 100 ppm (approximately 5 mg/kg body weight/d) of Aroclor 1242 or 1016 for 6 months resulted in PCB adipose tissue levels of about 200 ppm (mg/kg).140 Subhuman primates developed similar body burdens at five daily doses per week for 37 months of 0.08 mg/kg Aroclor 1254, a more persistent PCB mixture. Adipose tissue levels of about 1200 mg/kg fat were present in rats receiving 154 DSW 148251 STLCOPCB4039736 TABLE 3a Examples of Recent PCB Levels in Tissues (mg/kg) and in Blood (pg/l) in Humans and Adverse Effects Humans (N) Workers (221) Workers (39) Workers (196)38'140 (51)" (200)70 Workers (80) Humans general pop Aroclor Air levels iaiiy mean (pg/m3) Time exposed 1242 1016 ND-264 Years 1254 0.4-82 Years 1016 1242 1254 200-2000 17 Years 690 Pyralene 3010; Apirolio 48-275 (2-27 pg/cm2 Skin) 0.0006 ugAg diet Years Serum HPCB 1-250; LPCB 2-3330 LPCB 1-59; HPCB 1-24 57-2207 393a 41-1319; 337 current; 292 former ND-8 Adipose tissue Adverse effect No convincing effects reported (see text) No convincing effects (see text) 101" No convincing effects (see text); clinically insignificant decrease in birth weight Chloracne in 4 workers, most likely exposure to chlorinated dibenzofuran as well =1.0 No convincing effects reported (see text) Ref. 34, 41 34, 41 38, 142, 69, 70 39 9 Note: Analytical methods and quantitation differ. PCB levels may be higher or lower than reported.142 Partition value between serum and adipose tissue is about 160.33 The concentration in adipose tissue in workers would be between 6 and 235 ppm. a Composed of 363 ppb lower PCBs (eluting before DDE) and 30 ppb higher PCBs. b Serum lipid PCB, presumably equal to adipose tissue PCB on a fat basis. a total dose of 4500 mg/kg over a 6-month period, while monkeys receiving a total dose of 65 mg/kg body weight over a 37-month period (5 d/week) had tissue levels of about 109 mg/kg fat. The total dose given to the monkeys was 69 times less than the dose given to the rat, while the adipose tissue levels in the rat were only 10 times higher. Dosing in the rats occurred over a longer period ofthe rat's lifespan than in monkeys and they had also been held for an additional 4 months without being given PCB before the concentrations in adipose tissue were determined. During that time some depletion ofPCB tissue levels occurred. Betterdata are needed to determine whether retention of PCBs, particu larly the more biologically active congeners in subhuman primates, is more pronounced than in other species. Hayes,141 for instance, showed that for DDT, another lipophilic compound at dosage levels in the microgram range, rhesus monkeys stored proportionally more DDT in their adipose tissue than humans. At low doses of PCBs, the onset of toxicity occurs slowly in experimental animals, suggest ing that the total cumulative dose may be impor- 0SW 148252 155 STLCOPCB4039737 TABLE 3b Mean or Average PCB Levels in Tissues (mg/kg) and in Blood (ng/ml) in Different Species and Adverse Effects Species Rats Rats Rhesus monkeys infants Rhesus monkeys Mink Aroclor 1016 1242 1254 Clophen A-50 1248 1254 Clophen A-50; 1254 Daily dose (mg/kg) Months Adipose dosed Serum tissue Adverse effect Ref. 5 10 334 188 Enlarged liver cells 5 10 172 133 Enlarged liver cells 25 6 1192 Liver pathology (+4 recovery) 2 1.5 46 Liver necrosis 10 277 cholesterol T at 50 =2000 2 mg/kg. Serum 250 5114 bilirubin T, serum protein t, ALT T, and AST T at 50 mg/kg Transplacental Mothers Skin Facial edema, acne, and milk dosed 86-136 hyperpigmentation 140 140 146 147 148 0.005 37 10 5 Reduced number 149 0.020 30 24 term infants, 107 0.040 80 52 still-births, discoloration 0.80 130 109 of nail beds, facial pigmentation =2.0 3 180 Reproductive failure 149 =1.5 3 70 Note: Analytical methods and quantitation differ. Actual PCB levels may be higher or lower than reported.142 tant. However, at very low doses, such as 0.006 pg/kg/d, that the general population received dur ing the 1980s, a toxic dose would never be reached in a lifetime. In this case the total lifetime cumu lative dose would be 0.153 mg/kg body weight, while capacitor workers received a cumulative dose during a 200-d work period of 14.4 mg/kg body weight without showing ill effects. The relationship between the acute and the chronic toxic dose of a given chemical has been termed the chronicity factor or chronicity index. For some chemicals, the degree of accumulation, metabolism, and excretion determines these doses. For other chemicals, the type oftoxicity the chemi cals produce and the ease or inability of the ex posed to compensate or repair these perturbations affects this relationship. This concept is discussed in detail by Hayes.141 It appears that for PCBs the cumulative dose is more important than the daily dose unless very large doses (e.g., g/kg doses for rats) are given. The cumulative dose is associated with the accumulation of the PCBs in tissues. In subhuman primates adverse effects are still reported at adipose tissue concentrations of about 5 ppm (5 mg/kg), which were achieved at daily doses of 0.005 pg/kg body weight. Human work ers, at levels up to about 400 ppb (pg/1) serum and up to about 100 ppm (mg/kg) adipose tissue on a wet-weight basis, have not suffered from impaired health as a result of their exposure. Based on air levels and assumptions about dermal absorption, workers with these tissue levels could have re ceived daily doses between 70 and 140 pg/kg/d for months to years. The PCB tissue levels in these various studies were determined in different laboratories and, there fore, only rough comparisons can be made. To improve these data, all specimens would have to be analyzed in the same laboratory under strict quality control.142 Despite the shortcomings of the avail- 156 DSW 148253 STLCOPCB4039738 able data, humans appear to be less sensitive to PCBs than subhuman primates and no more sensi tive to the toxic effects of PCBs than other lesssensitive laboratory animals. However, the data are insufficient to make detailed comparisons. Workers with high PCB exposures have shown some elevation of liver function tests, particularly the GGT. The finding of an induction of hepatic microsomal monooxygenases in some workers from a capacitor plant suggested that drug me tabolism and metabolism of xenobiotics was in duced in humans at high PCB exposure. This induction would explain at least in part the shorter half-life of PCBs in highly exposed worker popu lations. However, GGT may also be elevated in alcoholic liver disease, following exposure to a number of therapeutic drugs and dmgs of abuse, and may represent an incidental finding in some workers and in the general population. In the U.S., in PCB exposed workers or in the adult general population, the combination of hyperpigmentation, chloracne, and discolored nails has never been observed. However, they were reported by the Japanese and the Taiwanese that developed Yusho or Yu-Cheng following expo sure to PCDFs.8 Similarly, they have been identi fied in adult subhuman primates following the administration of PCBs. These alterations of the skin and nails were very obvious even to the uninitiated and would not have been missed by health professionals or the workers themselves. Thus, there appeared to be a difference in re sponse to PCBs between subhuman primates and humans. On a quantitative basis, adult subhuman primates were either more sensitive to the toxic effects of PCBs or they preferentially retained the more toxic PCB congeners, while humans were able to excrete them more efficiently. Such differ ences in metabolism and excretion of specific congeners and differences in enzyme induction have been demonstrated with PCBs for several species.58,143,144 Toxicity in different species may be modu lated by species-specific differences in lipid me tabolism, quantitative differences in binding of PCBs to receptors in target organs, enzyme in duction, or other differences in the toxicokinetics of PCBs. Proportionally, humans may also be able to sequester more PCBs into adipose tissue away from vital organs and from skin. Instead of only extrapolating from laboratory animals the evaluation of the toxicity of PCBs should rely more on the available human data. REFERENCES 1. McFarland, V. A. and Clarke, J. U., Environmental occurrence, abundance, and potential toxicity of poly chlorinated biphenyl congeners: considerations for a congener-specific analysis. Environ. Health Perspect., 81, 225, 1989. 2. U.S. EPA, Polychlorinated biphenyls (PCBs): manu facturing, processing, distribution in commerce and use prohibitions: use in electrical equipment. Fed. Reg., 47, 37, 342, 1982. 3. Fensterheim, R. J-, Documenting temporal trends of polychlorinated biphenyls in the environment, Reg. Tox. Pharm., 1993, in press. . 4. Ballschmitter, K., Rappe, C, and Buser, H. R., Chemical properties, analytical methods and environ mental levels PCBs, PCTs, PCNs and PBBs, in Halo- genated Biphenyls, Terphenyls, Naphthalenes, Dibenzodioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier Biomedical Press, Amsterdam, 1989, 47. 5. Kimbrough, R. D-, Consumption of fish: benefits and perceived risk, J. Toxicol Environ. Health, 33,81,1991. 6. Robinson, P. E., Mack, G. A., Remmers, J., Levy, R., and Mohadjer, L., Trends of PCB, hexachloro- benzene, and ji-benzene hexachloride levels in the adipose tissue of the U.S. population. Environ. Res., 53, 175, 1990. 7. Morita, M., Nakagawa, JL, and Rappe, C., Poly chlorinated dibenzofuran (PCDF) formation from PCB mixture by heat and oxygen. Bull Environ. Contam. Toxicol, 19, 665, 1978. 7a. Erickson, M. D., Swanson, S. E., Flora, J. D., Jr., and Hinshaw, G. D., Polychlorinated dibenzofurans and other thermal combustion products from dielec tric fluids containing polychlorinated biphenyls. Environ. Sci. TechnoL, 23, 462, 1989. 8. Kuratsune, M., Yusho, with reference to Yu-Cheng, in Halogenated Biphenyls, Terphenyls, Naphthalenes, Dibenzodioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier, Amsterdam, 1989, 381. 9. Jensen, A. A., Background levels in humans, in Ha logenated Biphenyls, Terphenyls, Naphthalenes, Dibenzodioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier Biomedical Press, Amsterdam, 1989, 345. 10. Swackhamer, D. L. and Armstrong, D. E., Estima tion of the atmospheric and nonatmospheric contribu tions and losses of polychlorinated biphenyls for Lake Michigan on the basis of sediment records of remote lakes. Environ. Scl TechnoL, 20, 879, 1986. DSW 148254 157 STLCOPCB4039739 11. Larsson, P. and Okla, L., Atmospheric transport of chlorinated hydrocarbons to Sweden in 1985 com pared to 1973, Atmos. Environ., 23, 1699, 1989. 12. Bennett, B. G., Exposure of man to environmental PCBs -- an exposure commitment assessment, ScL Total Environ., 29, 101, 1983. 13. Gunderson, E., FDA total diet study, April 1982 -- April 1984, dietary intakes of pesticides, selected el ements and other chemicals, J. Assoc. Off. AnaL Chem., 71, 1200, 1988. 14. Anderson, H. A., General population exposure to environmental concentrations of halogenated biphe nyls, in Halogenated Biphenyls, Terphenyls, Naph thalenes, Dibenzodioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier Biomedical Press, Amsterdam, 1989, 325. 15. Hovinga, M. E., Sowers, M., and Humphrey, H. E. B., Environmental exposure and lifestyle predictors of lead, cadmium, PCB, and DDT levels in Great Lakes fish eaters.Arch. Environ. Health, 48,98,1993. 16. Hovinga, M. E., Sowers, M. F., and Humphrey, H. E. B., Historical changes in serum PCB and DDT levels in an environmentally exposed cohort, Arch. Environ. Contain. Toxicol, 22, 632, 1992. 17. Fiore, B. J., Anderson, H. A., Hanrahan, L. P., Olson, L. J., and Sonzogni, W. C., Sport fish con sumption and body burden levels of chlorinated hy drocarbons: a study of Wisconsin anglers. Arch. Environ. Health, 44, 82, 1989. 18. Dar, E., Kanarek, M. S-, Anderson, H. A., and Sonzogni, W. C-, Fish consumption and reproductive outcomes in Green Bay, Wisconsin, Environ. Res., 59, 189, 1992. 19. Olsen, S. F., Olsen, J., and Frische, G., Does fish consumption during pregnancy increase fetal growth? A study of the size of the newborn, placental weight and gestational age in relation to fish consumption during pregnancy, Int. J. Epid, 19, 971, 1990. 20. Calabrese, E. J., Human breast milk contamination in the United States and Canada by chlorinated hydro carbon insecticides and industrial pollutants: current status, J. Am. Coll Toxicol, 1, 91, 1982. 21. WHO, Assessment ofHealth Risks in Infants Associated with Exposure to PCBs, PCDDs and PCDFs in Breast MUk, Report 29, World Health Organization, Regional Office of Europe, Copenhagen, Denmark, 1988. 22. Dewailly, E^ Nantel, A., Weber, J.-P., and Meyer, F., High levels of PCBs in breast milk of Inuit women from arctic Quebec, Bull Environ. Contam. ToxicoL, 43, 641, 1989. 23. Fitzgerald, E. F-, Hwang, S.-A-, Brix, K. A., Bush, B., Quinn, J., and Cook, K., Chemical Contaminants in the Milk ofMohawk Womenfrom Akwesasne. New York State Department of Health, Albany, NY, 1992. 24. Stehr-Green, P. A., Welty, E., and Burse, V. W., Human exposure to polychlorinated biphenyls at toxic waste sites: investigations in the United States, Arch. Environ. Health, 43, 420, 1988. 158 25. Baker, E. L. J., Landrigan, P. J., Glueck, C. J., Zack, M. M. J., Liddie, J. A., Burse, V. W., Housworth, W. J., and Needham, L. L., Meta bolic consequences of exposure to polychlorinated biphenyls (PCB) in sewage sludge. Am. J. Epid., 112, 553, 1980. 26. Yaffe, B. A. and Reeder, B., An epidemiologic as sessment of exposure of children to polychlorinated biphenyls (PCBs) in a Toronto community. Can. J. Public Health, 80, 325, 1989. 27. Miller, D. T., Condon, S. K., Kutzner, S., Phillips, D. L., Krueger, E-, Timperi, R., Burse, V. W., Cutler, J., and Gute, D. M., Human exposure to polychlorinated biphenyls in Greater New Bedford, Massachusetts: a prevalence study. Arch. Environ. Contam. Toxicol., 20, 410, 1991. 28. Schmid, P., Buhler, F., and Schlatter, C., Dermal absorption of PCB in man, Chemosphere, 24, 1283, 1992. 29. Steele, G., Stehr-Green, P7 and Welty, E., Estimates of the biologic half-life of polychlorinated biphenyls in human serum, N. Engl J. Med, 314, 926, 1986. 30. Buehler, F,, Schmid, P,, and Schlatter, CL, Kinetics of PCB elimination in man, Chemosphere, 17,1717,1988. 31. Borlakoglu, J. T. and Walker, C. EL, Comparative aspects of congener specific PCB metabolism, Eur. J. Biochem., 14, 127, 1989. 32. Wolff, M. S., Fischbein, A., Thornton, J., Rice, C., Lilis, R., and Selikoff, I. J., Body burden of poly chlorinated biphenyls among persons employed in capacitor manufacturing, Int Arch. Occup. Environ. Health, 49, 199, 1982. 33. Wolff,M. S., Fischbein, A_, and SelikofT, L J., Changes in PCB serum concentrations among capacitor manu facturing workers. Environ. Res., 59, 202, 1992. 34. Phillips, D. I~, Smith, A. B., Burse, V. W., Steele, G. K., Needham, L. L., and Hannon, W. H-, Half-life of polychlorinated biphenyls in occupationally ex posed workers. Arch. Environ. Health, 44,351,1989. 35. Safe, S., Safe, L., and Mullin, Polychlorinated biphenyls: congener-specific analysis of a commer cial mixture and a human milk extract, J. Agric. Food Chem., 33, 24, 1985. 36. Fait,A,Grossman,EnSelf,S^Jeffiies,JnPenizzari, E. D., and Emmett, E. A*, Polychlorinated biphenyl congeners in adipose tissue lipid and serum of past and present transformer repair workers and a comparison group. Fund Appl ToxicoL, 12,42, 1989. 37. Kreiss, K., Zack, M. M., Kimbrough, R. D., Needham, L. L., Smrek, A. L., and Jones, B. T., Association of blood pressure and PCB level in cross sectional community study, JAMA, 245,2505, 1981. 38. Lawton, R. W., Ross, M. R., Feingold, J., and Brown, J. F., Jr., Effects of PCB exposure on bio chemical and hematological findings in capacitor workers. Environ. Health Perspect., 60, 165. 1985. 39. Maroni, M., Colombi, A., Arbosti, G-, Cantoni, S., and Foa, V., Occupational exposure to polychlori- DSW 148255 STLCOPCB4039740 nated biphenyls in electrical workers. II. Health ef fects, Br. J. Ind. Med., 38, 55, 1981. 40. Fischbein, A., Wolff, M. S-, Lilis, R., Thornton, J., and Selikoff, I. J., Clinical findings among PCB exposed capacitor manufacturing workers, N. Y. Acad. ScL, 320, 703, 1979. 41. Smith, A. B., Schloemer, J., Lowry, L. K., Smallwood, A. W., Ligo, R. N., Tanaka, S., Stringer, W., Jones, M., Hervin, R., and Glueck, C. J., Meta bolic and health consequences of occupational expo sure to polychlorinated biphenyls, Br. J. Ind. Med., 39, 361, 1982. 42. Jones, J. W. and Alden, H. S., An acneform dermatergosis, Arch. Dermatol Syphilology, 33, 1022, 1936. 43. Meigs, J. W., Albom, J. J., and Kartin, B. L., Chloracne from an unusual exposure to Aroclor, JAMA, 154, 1417, 1954. 44. Colombi, A., Maroni, M., Ferioii, A., Castoldi, M., Jun, L. K., Valla, C., and Foa, V., Increase in urinary porphyrin excretion in workers exposed to polychlorinated biphenyls, J. Appl. Toxicol., 2, 115, 1982. 45. Hill, R^ Bailey, S. L., and Needham, L. L., Devel opment and utilization of a procedure for measuring urinary porphyrins by high-performance liquid chro matography, J. Chromatogr., 232, 251, 1982. 46. Mathews, J. D-, Alcohol use, hypertension and coro nary heart disease, Clin. ScL Mol Med., 51 (Suppl.), 661, 1976. 47. Saxena, S., Korula, J., and Schulman, I. A^ A review of donor alanine aminotransferase testing. Arch. PcahoL Lab. Med., 113, 767, 1989. 48. Kimbrough, R. D. and Grandjean, P., Occupational exposure, in Halogenated Biphenyls, Terphenyls, Naphthalenes, Dibenzodioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier Biomedical Press, Amsterdam, 1989,485. 49. Ouw, H. EL, Simpson, G. R., and Siyali, D. S., Use and health effects of Aroclor 1242, a polychlorinated biphenyl, in an electrical industry. Arch. Environ. Health, 31, 189, 1976. 50. Elo, O., Vuojolahti, P., Janhunen, H., and Rantanen, J., Recent PCB accidents in Finland, Environ. Health Perspecu, 60, 315, 1985. 51. Wallach, J., Interpretation ofDiagnostic Tests, Little, Brown and Company, Boston, 1992, 3. 52. Warshaw, R., Fischbein, A., Thorton, J,, Miller, A., and Selikoff, I. J., Decrease in vital capacity in PCB-exposed workers in a capacitor manufacturing facility, N.Y. Acad. ScL, 320, 277, 1979. 53. Morris, J. F., Kolski, A., and Johnson, L. C., Spirometric results for healthy nonsmoking adults, Am. Rev. Resp. Disease, 103, 57, 1971. 54. Kalina, I., Sram, R. J., Konecna, H., and Ondrussekova, A., Cytogenic analysis of peripheral blood lymphocytes in workers occupationally exposed to polychlorinated biphenyls. Teratogen. Carcinog. Mutagen., 11,77,1991. 55. Kilburn, K. H., Warsaw, R. H., and Shields, M. G., Neurobehavioral dysfunction in firemen exposed to polychlorinated biphenyls (PCBs): possible improve ment after detoxification. Arch. Environ. Health, 44, 345, 1989. 56. Fitzgerald, E. F., Standfast, S. J., Youngblood, L. G-, Melius, J. M., and Janerich, D. T., Assessing the health effects of potential exposure to PCBs, dioxins, and furans from electrical transformer fires: the Binghamton State office building medical surveillance program. Arch. Environ. Health, 41, 368, 1986. 57. Alvares, A. P., Fischbein, A., Anderson, K. E., and Kappas, A., Alterations in drug metabolism in work ers exposed to polychlorinated biphenyls, Clin. Pharmacol Then, 22, 14, 1977. 58. Alvares, A. P-, Ueng, T.-H-, and Eiseman, J. L., Polychlorinated biphenyls (PCBs) inducible mono oxygenases in rabbits and mice: species and organ specificities. Life. ScL, 30, 747, 1982. 59. Falck, F., Jr., Ricci, A., Jr., Wolff, M. S., Godbold, J., and Deckers, P., Pesticides and polychlorinated biphenyl residues in human breast lipids and their relation to breast cancer. Arch. Environ. Health, 47, 146, 1992. 60. Casarett, L. J., Fryer, G. C., Yauger, W. L., Jr., and Klemmer, H-, Organochlorine pesticide residues in human tissue --Hawaii, Arch. Environ. Health, 17, 306, 1968. 61. Hoffman, W. S., Adler, H-, Fishbein, W. L, and Bauer, F. C., Relation of pesticide concentration in fat to pathological changes in tissues. Arch. Environ. Health, 15, 758, 1967. 62. Radomski, J. L~, Deichmann, W. B., and Clizer, E. E., Pesticide concentrations in the liver brain and adipose tissue of terminal hospital patients. Food Cosrnet. ToxicoL, 6, 209, 1968. 63. Unger, M. and Olsen, J, Organochlorine compounds in the adipose tissue of deceased people with and without cancer. Environ. Res., 23, 257, 1980. 64. Unger, M., Olsen, J., and Clausen, J., Organochlo rine compounds in tire adipose tissue ofdeceased people with and without cancer a statistical survey of some potential confounders, Environ. Res., 29, 371, 1982. 65. Unger, M., Kiaer, H., Blichert-Toft, M., Olsen, J., and Clausen, J., Organochlorine compounds in hu man breast fat from deceased with and without breast cancer and in a biopsy material from newly diagnosed patients undergoing breast surgery. Environ. Res., 34, 24, 1984. 66. Teufel, M., Bohn, 1., and Niessen, K. H., Belastung unserer Kinder mit Pestiziden, PCB und potentiell kritischen Anionen, Monatsschr. Kinderheilkd., 139, 442, 1991. 67. Graham, S., Zielezny, M., Marshall, J-, Priore, R., Freudenheim, J., Brasure, J., Haughey, B., Nasca, P., and Zdeb, M., Diet in the epidemiology of post menopausal breast cancer in the New York State co hort, Am. J. Epid., 136, 1327, 1992. DSW 148256 159 STLCOPCB4039741 68. Henderson, B. ., Ross, R. K., and Pike, M. C., Hormonal chemoprevention of cancer in women. Sci ence, 259, 633, 1993. 68a. Key, T. and Reeves, G., Organochlorines in the en vironment and breast cancer, Br. Med. J., 309, 23, 1994. 69. Taylor, P. R., Lawrence, C. E., Hwang, H.-L., and Paulson, A. S., Polychlorinated biphenyls: influence on birth weight and gestation. Am. J. Public Health, 74, 1153, 1984. 70. Taylor, P. R., Stelma, J. M., and Lawrence, C. E., The relation of polychlorinated biphenyls to birth weight and gestational age in the offspring of occu pationally exposed mothers. Am. J. Epid., 129,395, 1989. 71. Rogan, W. J., Gladen, B. C., McKinney, J. D., Carreras, N., Hardy, P., Thullen, J., Tingelstad, J., and Tully, M, Polychlorinated biphenyls (PCBs) and dichlorodiphenyl dichloroethene (DDE) in human milk: effects of maternal factors and previous lacta tion, Am. J. Public Health, 76, 172, 1986. 72. Gladen, B. C., Supplemental Premeeting Comments for Workshop on Developmental Neurotoxic Effects Associated with Exposure to PCBs. September 14--15, U.S. Environmental Protection Agency, Research Triangle Park, NC, 1992. 73. Tilson, H. A,, Jacobson, J. L., and Rogan, W. J., Polychlorinated biphenyls and. the developing ner vous system: cross-species comparisons, NeurotoxicoL TeratoL, 12, 239, 1990. 74. Schwartz, P. M., Jacobson, S. W., Fein, G., Jacobson, J. L., and Price, H. A., Lake Michigan fish consumption as a source of polychlorinated bi phenyls in human cord serum, maternal serum, and milk. Am. J. Public Health, 73, 293, 1983. 75. Jacobson, J. L., Jacobson, S. W., Schwartz, P. M., Fein, G. G., and Dowler, J. K., Perinatal exposure to an environmental toxin: a test of the multiple effects model, Dev. PsychoL, 20, 523, 1984. 76. Jacobson, J. L., Fein, G. G., Jacobson, S. W., and Schwartz, P. M., Factors and clusters for the Brazelton Scale. An investigation of the dimensions of neonatal behavior, Dev. PsychoL, 20, 339, 1984. 77. Fein, G. G., Jacobson, J. L., Jacobson, S. W., Schwartz, P. M., and Dowler, J. K, Prenatal expo sure to polychlorinated biphenyls: effects on birth size and gestational age, J. Pediatr., 105, 315, 1984. 78. Jacobson, S. W., Jacobson, J. L., Schwartz, P. M., and Fein, G. G., Intrauterine exposure of human newborns to PCBs: measures of exposure, in PCBs: Human and Environmental Hazards, D'ltri, F. M. and Kamrin, M., Eds., Butterworths, Boston, 1983, 319. 79. Paneth, N., Human reproduction after eating PCBcontaminated fish. Health Environ. Digest., 5,4,1991. 80. Jacobson, J. L., Fein, G. G., Jacobson, S. W., Schwartz, P. M., and Dowler, J. K., The transfer of polychlorinated biphenyls (PCBs) and polybrominated biphenyls (PBBs) across the human placenta into maternal milk. Am. J. Public Health, 74, 378, 1984. 81. Rogan, W. J., Gladen, B. C., McKinney, J. D., Carreras, N., Hardy, P., Thullen, J., Tinglestad, J., and Tully, M, Neonatal effects of transplacental expo sure to PCBs and DDE, J. Pediatr., 109, 335, 1986. 82. Gladen, B. C. and Rogan, W. J., Effects of perinatal polychlorinated biphenyls and dichlorodiphenyl dichloroethene on later development, J. Pediatr., 119, 58, 1991. 83. Jacobson, J. L., Jacobson, S. W., and Humphrey, H. E. B., Effects of in utero exposure to polychlori nated biphenyls and related contaminants on cogni tive functioning in young children, J. Pediatr., 116, 38, 1990. 84. Jacobson, J. L., Jacobson, S. W., and Humphrey, H. E. B., Effects of exposure to PCBs and related compounds on growth and activity in children, NeurotoxicoL TeratoL, 12, 319, 1990. 85. Leoni, V., Fabiani, L., Marinelli, G., Puccetti, G., Tarsitani, G. F., De Carolis, A., Vesda, N., Morine, A., Aleandri, V., Pozzi, V., Cappa, F., and Barbati, D_, PCB and other organochlorine compounds in blood of women with or without miscarriage: a hypothesis ofcorrelation, Ecotoxicol. Environ. Safety, 17,1,1989. 86. Emmett, E. A., Maroni, M-, Schmith, J. M., Levin, B. K., and Jefferys, J., Studies of transformer repair workers exposed to PCBs. I. Study design, PCB con centrations, questionnaire, and clinical examinations results. Am. J. Ind Med., 13, 415, 1988. 87. Emmett, E. A., Maroni, M., Jefferys, J., Schmith, JM Levin, B. K., and Alvares, A,, Studies of trans former repair workers exposed to PCBs. II. Results of clinical laboratory investigations. Am. J. Ind. Med., 14, 47, 1988. 88. Tryphonas, H., Luster, M. I., Schiffman, G., Dawson, L. L., Hodgen, M., Germolec, D., Hayward, S., Bryce, F., Loo, J. C. K., Mandy, F., and Arnold, D. L~, Effect of chronic exposure of PCB (Aroclor 1254) on specific and nonspecific im mune parameters in the rhesus monkey (Macaco mulatto). Fund. AppL ToxicoL, 16, 773, 1991. 89. Brown, D. P. and Jones, M., Mortality and industrial hygiene study of workers exposed to polychlorinated biphenyls. Arch. Environ. Health, 36, 120, 1981. 90. Brown, D. P., Mortality of workers exposed to poly chlorinated biphenyls -- an update. Arch. Environ. Health, 42, 333, 1987. 91. Taylor, P. R., The Health Effects of Polychlori nated Biphenyls, Harvard School of Public Health, Boston, 1988. 92. Gustavsson, P., Hogstedt, C., and Rappe, C., Short term mortality and cancer incidence in capacitor manu facturing workers exposed to polychlorinated biphe nyls, Am. J. Ind. Med., 10, 341, 1986. 93. Bertazzi, P. A., Riboldi, I~, Pesatori, A., Radke, L, and Zocchetti, G, Cancer mortality ofcapacitor manu facturing workers, Am. J. Ind Med, 11, 165, 1987. DSW 148257 160 STLCOPCB4039742 94. Sinks, T., Steele, G., Smith, A. B., Watkins, K., and Shults, R. A., Mortality among workers ex posed to polychlorinated biphenyls. Am. J. Epid., 136, 389, 1992. 95. Bahn, A. K., Rosenwaike, I., Herrmann, N., Grover, P., Stellman, J., and O'Leary, K., Mela noma after exposure to PCBs, N. Engl. J. Med., 295, 450, 1976. 96. McConnell, E. E., Acute and chronic toxicity and carcinogenesis in animals, in Halogenated Biphe nyls, Terphenyls, Naphthalenes, Dibenzodioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier Biomedical Press, Amsterdam, 1989, 161. 97. Morrissey, R. E. and Schwetz, B. A., Reproductive and developmental toxicity in animals, in Haloge nated Biphenyls, Terphenyls, Naphthalenes, Dibenzo dioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Elsevier Biomedical Press, Amsterdam, 1989, 195. 98. Zeiger, E., Genetic toxicity, in Halogenated Biphe nyls, Terphenyls, Naphthalenes, Dibenzodioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier Biomedical Press, Amsterdam, 1989, 227. 99. Goldstein, J. A. and Safe, S., Mechanism of action and structure-activity relationship for the chlorinated dibenzo-p-dioxins and related products, in Haloge nated Biphenyls, Terphenyls, Naphthalenes, Dibenzo dioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier Biomedical Press, Amsterdam, 1989, 239. 100. Vos, J. G. and Luster, M. L, Immune alterations, in Halogenated Biphenyls, Terphenyls, Naphthalenes, Dibenzodioxins and Related Products, Kimbrough, R. D. and Jensen, A. A., Eds., Elsevier Biomedical Press, Amsterdam, 1989, 295. 101. Barsotti, D. A., Marlar, R. J., and Allen, J. R., Reproductive dysfunction in rhesus monkeys exposed to low levels of polychlorinated biphenyls (Aroclor 1248), Food Cosmet. Toxicol., 14, 99, 1976. 102. Manson, J. M. and Kang, Y. J-, Test methods for assessing female reproductive and developmental toxi cology, in PrinciplesandMethods ofToxicology, Hayes, A. W,, Ed., Raven Press, New York, 1989, 311. 103. Barsotti, D. A. and Van Miller, J. P., Accumulation of a commercial polychlorinated biphenyl mixture (Aroclor 1016) in adult rhesus monkeys and their nursing infants. Toxicology, 30, 31, 1984. 104. U.S. EPA, Drinking Water Criteria Document for Polychlorinated Biphenyls (PCBs). Final Report, U.S. Environmental Protection Agency, PB89-192256. U.S. Department of Commerce, National Technical Infor mation Service (NTIS) Springfield, VA, 1988. 105. IRIS. Searchfor Polychlorinated Biphenyls. Integrated Risk Information Service. U.S. Environmental Pro tection Agency, Cincinnati, OH, and National Library of Medicine, Bethesda, MD, 1992. 106. Van Wagenen, G. and Catchpole, H. R., Physical growth of the rhesus monkey. Am. J. Phys. Anthropol., 23, 23, 1956. 107. Arnold, D. L., Bryce, F., Karpinski, K., Mes, J., Tryphonas, H., Truelove, J., and Zawidzka, Z. Z., Toxicity of polychlorinated biphenyls (Aroclor 1254) in adult monkeys as a consequence of continuous exposure and in infant monkeys exposed during preg nancy and nursing. Toxicologist (Abstr. 817), 11,220, 1991. 108. Smialowicz, R. J., Andrews, J. E., Riddle, M. M., Rogers, R. R., Luebke, R. W., and Copeland, C. B., Evaluation of the immunotoxicity of low-level PCB exposure in the rat. Toxicology, 56, 197, 1989. 109. Vos, J. G., Toxicology of PCBs for mammals and for birds. Environ. Health Perspect., 4, 105, 1972. 110. Loose, L. D., Silkworth, J. B-, Pittman, K. A., Benitz, K. F., and Mneller, W, Impaired host resistance to endotoxin and malaria in polychlorinated biphenyland hexachiorobenzene-treated mice. Infect. Immun., 20, 30, 1978. 111. Thomas, P. T. and Hindsill, R. D., Effect of poly chlorinated biphenyls on the immune responses of rhesus monkeys and mice, Toxicol. AppL Pharmacol., 44, 41, 1978. 112. Truelove, J-, Grant, D., Mes, J., Tryphonas, H., Tryphonas, L., and Zawidzka, Z-, Polychlorinated biphenyl toxicity in the pregnant cynomolgus mon key: a pilot study. Arch Environ. Contam. Toxicol., 11,583, 1982. 113. Harder, T. C., Willhaus, T., Leibold, W., and Liess, B., Investigations on course and outcome of phocine distemper virus infection in haibour seals (Phoca vitulina) exposed to polychlorinated biphenyls, J. Vet. Med. Sci., 39, 19, 1992. 114. Tryphonas, H., Hayward, S., O'Grady, L., Loo, J. C. K., Arnold, D. L., Bryce, F., and Zawidzka, Z. Z., Immunotoxicity studies of PCB (Aroclor 1254) in the adult rhesus (Macaco mulatto) monkey -- preliminary report, Int. J. ImmunopharmacoL, 11, 199, 1989. 115. Tryphonas, EL, Luster, M. I., White, K. L., Jr., Naylor, P. H., Erdos, M. R-, Burleson, G. R., Germolec, D., Hodgen, M., Hayward, S., and Arnold, D. L., Effects of PCB (Aroclor 1254) on non-specific immune parameters in Rhesus (Macaca mulatto) monkeys, lnt. J. ImmunopharmacoL, 13, 639, 1991. 116. Paul, J. R. and White, C., Serological Epidemiol ogy, Academic Press, New York, 1973. 117. Luster, M. I., Munson, A. E-, Thomas, P. T., Hosapple, M. P., Fenters, J. D., White, K. L., Lauer, L. D., Germolec, D. R., Rosenthal, G. J., and Dean, J. H, Methods evaluation: development of a testing battery to assess chemical-induced immunotoxicity: National Toxicology Program's guidelines for immunotoxicity evaluation in mice. Fund. Appl. ToxicoL, 10, 2, 1988. DSW 148258 161 STLCOPCB4039743 118. Moore, J. A., Hardisty, J. FI, Banas, D. A., and Smith, M. A., A comparison of liver tumor diagnoses from seven PCB studies in rats, Reg. Toxicol. Pharmacol., 20, 362, 1994. 119. Squire, R. A. and Levitt, M. H., Report of a work shop on classification of specific hepatocellular le sions in rats. Cancer. Res., 35, 3214, 1975. 120. Maronpot, R. R., Montgomery, C. A., Boorman, G. A., and McConnell, E. E., National Toxicology Program nomenclature for hepatoproliferative lesions of rats, Toxicol. Pathol., 14, 263, 1986. 121. McConnell, E. E., Solleveld, H. A., Swenberg, J. A., and Boorman, G. A., Guidelines for combining neo plasms for evaluation of rodent carcinogenesis stud ies, in Carcinogenicity: The Design, Analysis and Interpretation of Long-Term Animal Studies, Grice, H. C. and Cimineri, J. L., Eds., Springer-Verlag, New York, 1988, 183. 122. U.S. EPA, Proliferative Hepatocellular Lesions of the Rat: Review and Future Use in Risk Assessment, U.S. Environmental Protection Agency, EPA/625/386/011, 1986. 123. NCI, Bioassay ofAroclor 1254for Possible Carcino genicity, National Cancer Institute, Bethesda, MD, NCI-GC-TR-38, 1978. 124. Kimbrough, R. D., Squire, R. A., Linder, R. E., Strandberg, J. D., and Montali, R. J,, Induction of liver tumors in Sherman strain female rats by PCB Aroclor 1260,7. NatL Cancer Inst., 55, 1453, 1975. 125. Schaeffer, E., Gram, H., and Goessner, W., Pathol ogy of chronic polychlorinated biphenyl (PCB) feed ing in rats, Toxicol. AppL Pharmacol, 75,278,1984. 126. Norback, D. EL and Weltman, R. EL, Polychlori nated biphenyl induction of hepatocellular carcinoma in the Sprague-Dawley rat. Environ. Health Perspect., 60, 97, 1985. 127. Ito, N., Nagasaki, EL, Makiura, S., and Arai, M., Histopathoiogical studies on liver tumorigenesis in rats treated with polychlorinated biphenyls, GANN (Jpn. J. Cancer Res.), 65, 545, 1974. 128. Ito, N., Nagasaki, H-, Arai, M., Makiura, S., Sugihara, S., and Hirao, L, Histopathologic studies on liver tumorigenesis induced in mice by technical polychlorinated biphenyls and its promoting effect on liver tumors induced by benzene hexachloride, 7. NatL Cancer Inst., 51, 1637, 1973. 129. Linder, R. E. and Kimbrough, R. D., The induction of adenofibrosis and hepatomas of the liver in mice of the Balb/cj strain by polychlorinated biphenyls (Aroclor 1254), 7. NatL Ccmc. Inst., 53, 547, 1974. 130. Kimura, N. and Baba, T., Neoplastic changes in the rat liver induced by polychlorinated biphenyl, GANN, 64, 105,1973. 131. U.S. EPA, Interim Evaluation of Health Risks Asso ciated with Emissions of Tetrachlorinated Dioxins from Municipal Waste Resource Recovery Facilities, U.S. Environmental Protection Agency, Office of the Administrator, Washington, D.C., 1981, 132. U.S. EPA, Interim Proceduresfor Estimating Risks Asso ciatedwith Exposures to Mixtures ofChlorinatedDibenzo- p-Dioxins and Dibenzofurans (CDDs and CDFs) and 1989 Update. U.S. Environmental Protection Agency, Washington, D.C., EPA/625/3-89/016,1989. 133. U.S. EPA, PCB Disposal by Thermal Destruction. U.S. Environmental Protection Agency, National Tech nical Information Service, Springfield, VA, PB 82 241860, 1982. 134. Neubert, D., Golor, G., and Hagenmaier, H., Toxic TCDD -- equivalents, problems and applicability for risk assessment. Health Effects and Safety Assessment ofDioxins and Furans. Conference Proceedings (Janu ary 15-17 1990), The Toxicology Forum, Karlsruhe, Germany, 1990, 560. 135. Barnes, D. A., Alford-Stevens, A-, Birnbaum, L., Kutz, F. W, Wood, and Patton, D,, Toxicity equivalency factors for PCBs?, Quality Assurance: Good Practice Regulation, Law, 1, 70, 1991. 136. U.S. EPA. Workshop Report on Toxicity Equiva lency Factors for Polychlorinated Biphenyl Conge ners, U.S. Environmental Protection Agency, EPAI 625/3-91/020, 1991. 137. De Vito, M. J., Maier, W. E., Diliberto, J. J., and Birnbaum, L. S., Comparative ability of various PCBs, PCDFs, and TCDD to induce cytochrome P450 1A1 and 1A2 activity following four weeks of treatment. Fund. Appl. ToxicoL, 20, 125, 1993. 138. Walker, M. K. and Peterson, R. E., Potencies of polychlorinated dibenzo-p-dioxin, dibenzofuran, and biphenyl congeners, relative to 2,3,7,8-tetrachloro- dibenzo-p-dioxin, for producing early life stage mor tality in rainbow trout (Oncorhynchus mykiss), Aquat. ToxicoL, 21,219, 1991. 139. Arnold, D. L, Mes, J., Bryce, F., Karpinski, IC, Bickis, M. G., Zawidzka, Z. Z, and Stapley, R., A pilot study on the effects of Aroclor 1254 ingestion by rhesus and cynomolgus monkeys as a model for hu man ingestion of PCBs, Food Chem. ToxicoL, 28, 847, 1990. 140. Burse, V. W., Kimbrough, R. D., Villanueva, E. C-, Joinings, R. W., Linder, R. IL, and Sovocool, G. W., Polychlorinated biphenyls. Storage, distribution ex cretion and recovery: liver morphology after pro longed dietary ingestion. Arch. Environ. Health, 29, 301, 1974. 141. Hayes, W. J., Jr., Toxicology of Pesticides, Chap. 2, Williams & Wilkins, Baltimore, 1975, 37. 142. Lawton, R. W., Brown, J. F., Ross, M. R-, and Feingold, J., Comparability and precision of serum PCB measurements. Arch. Environ. Health, 40, 29, 1985. 143. Sipes, G. L, Slocumb, M. L., Chen, H.-S. G-, and Carter, D. E., 2,3,6,2',3',6,-HexachIorobiphenyl: dis tribution, metabolism, and excretion in the dog and the monkey, ToxicoL Appl. Pharmacol, 62, 317, 1982. 144. Sipes, G. L, Slocumb, M. L~, Perry, D. F., and Carter, D. E, 2,4,5,2',4',5'-Hexachlorobiphenyl: dis 162 DSW 148259 STLCOPCB4039744 tribution, metabolism, and excretion in the dog and the monkey, ToxicoL Appl. Pharmacol., 65,264,1982. 145. - Jacobson, S. W., Fein, G. G., Jacobson, J. L., Schwartz, P. M., and Dowler, J. K-, The effect of intrauterine PCB exposure on visual recognition memory. Child. Dev., 56, 853, 1985. 146. Kimbrough, R. D., Linder, R. E-, Burse, V. W., and Jennings, R. W., Adenofibrosis in the rat liver with persistence of polychlorinated biphenyls in adipose tissue. Arch. Environ. Health, 27, 390, 1973. 147. Baumann, M., Demi, ., Schaffer, ., and Greim, H., Effects of polychlorinated biphenyls at low dose levels in rats. Arch. Environ. Toxicol., 12, 509, 1983. 148. Allen, J. R. and Barsotti, D. A., Residual effects of polychlorinated biphenyls on adult nonhuman pri mates and their offspring, J. Toxicol. Environ. Health., 6, 55, 1980. 149. Mes, J., Arnold, D. L., Bryce, F., Davies, D. J., and Karpinski, K., The effect of long-term feeding of Aroclor 1254 to female rhesus monkeys on their poly chlorinated biphenyl tissue levels. Arch. Environ. Contam. Toxicol., 18, 858, 1989. 150. Kihlstrom, J. E., Olsson, M., Jensen, A. A., Johansson, A., Ahlbom, J., and Bergman, A., Ef fects of PCB on the reproduction of the mink (Mustela vison), Ambio, 21, 563, 1992. OSW 148260 163 STLCOPCB4039745