Document DG4vmrR0wM44GdzZ0kbQKJ53o

136 PCBs and Human Disease James et al Polychlorinated Biphenyl Exposure and Human Disease Robert C. James, PhD espite their discontinued manufac Harris Busch, MD, PhD Carlo H. Tamburro, MD, MPH Stephen M. Roberts, PhD Dture, the persistence of polychlori nated biphenyls (PCBs) in the envi ronment has caused them to remain a significant human health concern. Literature available to define the hu John D. Schell, PhD Raymond D. Harbison, PhD man toxicology of PCBs includes clin ical and scientific reports spanning over 50 years. This literature, though extensive, is somewhat fragmented. While there are many studies of po Polychlorinated biphenyls (PCBs) continue to be of great environmental and occupational health interest. This review summarizes the major clinical findings reported in individuals incurring the greatest PCB exposure--those persons working in the manufacture or repair of electrical capacitors or transformers. The potential target organs addressed in the studies reviewed include the liver, lungs, skin, cardiovascular system, nervous system, certain mdocrine systems, the blood/immune system, and the gastrointestinal and urinary tracts. After careful analysis, the weight ofevidence suggests the only adverse health effects attributable to high, occupational PCB exposures are dermal. This review confirms and extends the observations ofothers, ie, that the collective occupational experience with PCB fluids provides no evidence for adverse PCB effects on any other organ systems. tential PCB effects within discrete study populations, little is available in the way of comprehensive reviews of the human health effects of PCBs. Clinical studies of potential PCB health effects have been conducted in basically three types of study popula tions. The first represents ajather spe cial population--individuals poi soned with rice oil contaminated with PCBs. Two widespread outbreaks of contaminated rice oil poisoning have occurred, one in Japan in 1968 and the other in Taiwan in 1979. The symptoms of these rice oil poisonings, termed "Yusho" in Japan and "Yu- Cheng" in Taiwan, were similar and originally attributed to PCBs present in the oil. However, further examina tion of the poisoning incidents found several lines of evidence indicating that the symptoms were most prob ably caused by the presence of the more potently toxic polychlorinated dibenzofurans (PCDFs). This evi dence includes the observation that PCB levels in Yusho and Yu-Cheng From the Center for Environmental and Human Toxicology, University of Florida, Gaines victims with severe manifestations of ville, FL (Dr James, Dr Harbison, Dr Schell, Dr Roberts); Department ofPharmacology, Baylor disease were no greater than those of College of Medicine, Houston, TX (Dr Busch); and Liver Research Center and Division of Occupational Toxicology, Departments of Medicine and Pharmacology & Toxicology, Univer sity of Louisville, KY (Dr Tamburro). Address correspondence to: Dr Stephen M. Roberts, Center for Environmental and Human Toxicology, University of Florida, Progress Center, One Progress Boulevard, Box 17, Alachua, FL 32615. 0096-1736/93/3502-0136S03.00/0 healthy PCB-exposed workers, while their PCDF levels were much higher.1-2 Furthermore, studies in monkeys were able to duplicate Yusho-like symptoms with a mixture of PCBs and PCDFs resembling the toxic Copyright by American College of Occupational and Environmental Medicine rice oil, or with the PCDF component DSW 148216 STLCOPCB4039701 JOM Volume 35, Number 2, February 1993 137 alone, but not with a mixture contain ing the PCBs without PCDF.3 There is now general agreement among in vestigators most familiar with the rice oil poisoning incidents that they represent primarily PCDF intoxications2-4-12 and, therefore, have little relevance in defining the toxicity of PCBs in humans. They did, how ever, provide a stimulus for subse quent, more instructive studies of po tential PCB-related health effects. A second type of study population consists of individuals exposed to PCBs in the environment. Typically the source of environmental exposure in these studies was PCBs found in nearby soils or in the diet (usually in fish). While these studies conceivably have value in ascertaining the poten tial health effects of PCBs, they are often severely limited by low or equiv ocal PCB exposures and/or the pres ence of concurrent exposures to other chemicals that confound the interpre tation of results. The third type of study population is workers exposed to PCBs in an oc cupational setting. Exposures to PCBs among workers in some occupations (eg, capacitor or transformer manu facture) were very high, and some study populations contain workers with job-related exposures of 20 years of more. Both the magnitude and du ration of exposures in PCB worker studies provide the best opportunity to clearly observe which kinds of health effects might reasonably be at tributable to PCBs. As such, studies of PCB-exposed worker populations may be considered to be inherently the most valuable. The purpose of this review is to critically summarize available studies relevant to potential human health effects of PCBs. Clinical studies deal ing with both environmental and oc cupational PCB exposure have been considered, although, for the reasons listed above, the environmental stud ies have been given less weight in the final analyses. Where appropriate, in ferences from mortality studies have also been included in this analysis. Reports concerning Yusho and Yu- Cheng have been omitted since, as indicated above, it is doubtful that they can provide information relevant specifically to PCB toxicity. Potential health effects addressed in this review include effects on the skin, liver, lungs, serum lipids or lipoproteins, the cardiovascular system, the nervous system, the blood and immune sys tem, the gastrointestinal and urinary tracts, and some endocrine functions. Information concerning the potential human carcinogenicity or reproduc tive/developmental effects of PCBs will be the subjects of separate re views. Descriptions of the Occupational Cohorts The occupational cohort studies considered in this review are briefly described in tabular form (Tables 1 and 2). The largest cohorts of persons known to have experienced apprecia ble occupational PCB exposure in the workplace involved employees of companies manufacturing electrical equipment, particularly those persons who filled large capacitors and trans formers with PCB-containing fluids. Four large capacitor manufacturing plants located in the midwestem and northeastern United States have been studied in great detail (Table 1). These plants employed thousands of people during periods of moderate to heavy PCB usage, and several studies have been published that contain extensive clinical data derived from cross-sec tional surveys of workers from these plants. To a lesser extent, the health of employees from capacitor manu facturing plants from other countries have also been studied. These studies, although helpful, tend to be of lesser significance, either because of the small number of workers evaluated13' 17 or because the magnitude of PCB exposure to the cohort selected was probably not as great as that experi enced by US capacitor manufacture workers.18'20 Clinical observations have also been obtained from three cohorts of transformer repair workers, two cohorts of PCB manufacture workers, one cohort of silk thread manufacturers, one cohort of marine paint workers, and one group of em ployees exposed to heat transfer fluids (Table 2). As with the foreign capaci tor worker studies, these cohorts were, in general, considerably smaller than the US capacitor worker cohorts. Table 3 characterizes individual study qualities and deficiencies for the broad clinical surveys of PCB-exposed workers. All of these surveys per formed health and exposure history analyses and sought statistical corre lations between the clinical parame ters and PCB exposure (as determined primarily by serum PCB levels). A description of exposure was provided in most studies in that limited work place PCB measurements, hygiene de scriptions, or exposure duration data were at least summarily noted for the study population. The scope of the clinical surveys was often good, pro viding information on multiple or gans and systems, and a number of the surveys analyzed clinical parame ters pertaining to at least five of the nine general areas reviewed. However, few studies included values from a matched control group in their analy sis of the data. An important consideration in the evaluation of occupational cohort studies is the confounder of multiple chemical exposure. All of the studies reviewed involved exposure to other chemicals, principally those organic solvents, additives, or contaminants associated with industrial uses of PCB fluids or commercial mixtures con taining PCBs as a major component. Since the potently chloracnegenic PCDFs were generally present at part per million levels in commercial PCB fluids, low-grade PCDF exposure is common to all PCB-exposed cohorts. In addition, experience has shown that, with high temperatures and in the presence of oxygen, additional PCDFs may be formed from PCBs. As a consequence, a few of the studies14'16 are confounded by the fact that explosions of capacitors during stress testing may have created unusu ally high exposure to PCDFs for some of the workers in their study. Similar reservations are also held for the early dermatological survey of Meigs et al21 because this study involved worker exposure to vapors emanating from an open-air sump system where PCBs DSW 148217 STLCOPCB4039702 138 PCBs and Human Disease James et al TABLE 1 ^CB Capacitor Plant Cohort Study Information Cohort Size and Location PCB Usage Interval Study Authors and Year Indiana 3,588 total workers 1959-1977 Smith et al, 198238 Sinks etal, 199252 Study Type Number of Study Subjects Broad clinical survey Mortality analysis (192 deaths) 228 3,588 Massachusetts 1,559 workers in PCB-exposure jobs 1938-1977 500 current workers 1941-1977 Lawton et al, 198523 Brown and Jones, 198126 Brown, 198749 Acquavella et al, 198641 Broad clinical survey Mortality analysis (90 deaths) Mortality analysis (179 deaths) Broad clinical survey 194 1,599 1,607 205 New York 6,303 total workers at the two plant locations; 2,588 in PCBexposure jobs Plant 1: 1951-1977 Alvares et al, 197727 Liver function survey Plant 2:1946-1977 Fischbein et al, 197924 Broad clinical survey Warshaw et al, 197928 Lung function survey Brown and Jones, 198126 Mortality analysis (73 deaths) Fischbein et al, 1982,27a 198529a Dermal/ocular effect survey Fischbein, 198540 Liver function survey Lawton et al, 198653 Lung function survey ' Nicholson et al, 198750 Mortality analysis (188 deaths) Brown, 198749 Mortality analysis (116 deaths) Taylor, 1988s' . Mortality analysis (510 deaths) 5 326 326 . 968 ' 181-289 261 194 788 981 6,303 Australia 34 PCB-exposed workers 1951-1974 Ouw et al, 197613 Dermal effects/liver function 34 Italy 1,310-2,100 total workers 1946-1980 Maroni et al, 1981b15 Maroni et al, 198416 Bertazzi et al, 198720 Dermal effects/liver function Enzyme induction/porphyria Mortality analysis (64 deaths) 80 51 2,100 Japan 155 possibly exposed workers 1948-1972 Hara, 198518 Limited clinical survey 155 Sweden 145 exposed workers 1965-1978 Gustavsson et al, 198617 Mortality analysis (21 deaths) 142 were used as heat exchanger fluids. Last, the study of Jones and Alden22 is confounded by the fact that an im pure batch of benzene was used in the synthesis of the PCBs involved in that particular incident and the dermal le sions were ultimately attributed by the authors of this study to chlorinated contaminants rather than the PCBs. In addition to PCDFs, other chem ical exposures were associated with the use of PCB fluids and their pres ence should be considered a potential confounder of the results reported in the studies reviewed here. PCBs were commonly mixed with chlorinated benzenes (generally 30%-40% by vol ume) for use in large-capacity trans formers and capacitors. Upon com bustion, these chlorinated benzenes may form both PCDFs and polychlo rinated dibenzo-p-dioxins. Investiga tors of the major PCB capacitor worker cohorts23,24 noted the usage of chlorinated benzenes in the PCB fluids for capacitors, but other inves tigators of capacitor workers and transformer repair workers did not specify whether chlorinated benzenes were utilized. Unfortunately, in no case was exposure to chlorinated ben zenes measured and compared to that of the PCBs. Other likely but largely unspecified exposures to chemicals associated with PCB fluids include stabilizers like the alkylbenzenes or certain epoxides. Although these stabilizers were no doubt only a small component of the PCB fluids to which they were added, Lawrence25 has pointed out that their much higher vapor pressures may have resulted in a substantial work place exposure to these compounds. Another potentially confounding fac tor common to these studies was ex posure to chlorinated organic solvents like trichloroethylene and/or 1,1,1trichloroethane that were used to clean up PCB spills and to degrease equipment and final products. While some studies did address the potential confounder of exposure to other chemicals,26'28 there is little indication that other investigators considered the potential importance of the cohort's exposures to solvents or other work place chemicals. In addition to contaminants and stabilizers, PCB substitutes were in- DSw 148218 STLCOPCB4039703 JOM Volume 35, Number 2, February 1993 139 TABLE 2 Cohort Information for Transformer Repairmen and Other PCB-Exposed Workers Studied Cohort Size and Location PCB Usage Period Study Authors and Year Study Type Alabama PCB manufacturing; 24 production workers Delaware Locomotive repair; 120 total em ployees Illinois PCB manufacture; 89 exposed workers Midwestern United States Transformer repair; 93 exposed workers Washington, DC Switch gear repair; 120 total em-. ployees Other, United States PCBs used in heat exchange; 14 employees Japan Silk thread or marine paint factory workers; 112 workers Not specified Jones and Alden, 193622 Dermal survey 1939-1979? Chase et al, 198237 Broad clinical survey Not specified Zack and Musch, 1979s4 Mortality survey (30 deaths) Not specified Smith et at, 198238 Broad clinical survey Not specified Not specified Emmett, 198539 Emmett et al, 1988a, b29'30 Meigs et al, 195421 Liver function survey Broad clinical survey Dermal survey 1966-1972 Takamatsu et al, 198519 Limited clinical survey Number of Study Subjects 24 101 88 93 55 55 14 112 TABLE 3 Data-Base Characteristics for the Broad Clinical Surveys of PCB-Exposed Workers Study AuthorFischbein et al2< Lawton et al23 Smith et al38 Chase et al37 Emmett et al29,30 Acquavella et al41 Study features and confounders Study group size (exposed) Test values reported % abnormal reported Correlative studies Matched control group Health history Non-PCB exposures reported Organs/systems analyzed Skin Liver Hematopoietic Pulmonary Neurological symptoms Gastrointestinal symptoms Serum lipid levels Kidney Cardiovascular Endocrine PCB exposure characterization PCB air or surface levels Hygiene descriptions Serum PCB measurements Adipose PCB measurements Exposure duration reported Clinical evidence of exposure (chloracne) 326 No Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes No Yes Yes 194 321 101 Yes No Yes Yes No No Yes Yes Yes No No . . No Yes Yes Yes Yes Yes Yes 55 Yes No Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes No No No No No No No Yes No Yes Yes Yes Yes Yes Yes Yes Yes No No Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No No Yes Yes Yes No Yes Yes. No No Yes . No 205 No No Yes No Yes Yes Yes Yes Yes No No No Yes No No No Yes Yes Yes No No No corporated into the processes of all the major cohorts during 1977. The sub stitutes used included jS-chloranthroquinone, dioctylphthalates, dibutyl- sebacate, and mineral oil (New York, Massachusetts, northeastern US), and isopropylbiphenyl (Indiana). As many ofthe cohorts were examined (or reex amined) some time after 1977, the more recent daily exposure to these substitutes confounds any analysis of these studies. DSW 148219 STLCOPCB4039704 140 PCBs and Human Disease James et al It is unfortunate that only limited 'iscussions were provided by most aulors concerning the alternate chemi cal exposures that were present in these studies. It is apparently assumed by most investigators that only PCB exposures were to be associated with the clinical findings they reported. Since the actual impact of the con founding chemical exposure is un known, it is particularly important when reviewing the PCB literature that the observations of any study be confirmed by other investigations. Categorical Review of Clinical Studies Effects on the Skin Various skin conditions have been reported in humans occupationally exposed to PCBs. Dermal conditions observed in workers in these studies include dermatitis,15 temporary in flammation or edema of the skin and eyes,27a a thickening of the skin and fingernails,24 and chloracne.15,21,22,293 Complaints such as rashes or skin ir ritation have been typically reported in the absence of serious skin condi tions or documented high exposure to PCBs. These minor dermatological complaints may be more related to low-grade exposure to epoxide addi tives in the PCB fluids or caused by the workers' concomitant exposure to chlorinated benzenes, organic sol vents, or harsh detergents that were also present when viscous fluids like PCBs were used industrially.25,30 Other dermal effects such as thicken ing of the skin, temporary inflamma tion or edema of the skin and eyes, and chloracne appear to be related to the incidence and magnitude of the individual's dermal contact with PCB fluids or result from high exposure to heated PCB vapors, suggesting PCBs (or PCDFs) as the causative agent. A dermal effect associated with PCBs of particular interest is chlor acne. Chloracne is a dermatologic condition characterized by black heads, pustules and/or cysts, and gen erally appears on the malar regions of the face, the scalp, and the back of the neck. When induced by halogenated aromatic compounds, chloracne may sometimes be quite persistent. Only two early studies reported observing a relatively high incidence of chloracne associated with PCBs. In one of these studies, workers manufacturing PCB fluids were believed to have been ex posed to chlorinated byproducts gen erated by the use of impure benzene in the synthesis of PCBs.22 The second study consisted of workers exposed to leaking heat exchanger fluids,21 and, therefore, may represent a situation more like the PCDF exposures asso ciated with Yusho and. Yu-Cheng. Still, a low incidence of acne-like le sions has been reported in a number of the more recent studies of occupa tionally exposed PCB workers, and this condition would appear to be a potential effect of high, chronic ex posure to PCBs. The chloracne ob served in the occupational setting was generally associated with high body burdens of PCBs (eg, hundreds to sometimes thousands of parts per bil lion of PCBs in serum), though indi viduals with excessive PCB body bur dens (>1,000 ppb PCBs in serum) often do not develop chloracne.13,15 There are at least three possible expla nations for the inconsistent appear ance of chloracne in individuals with high body burdens. One potential ex planation is a difference in suscepti bility among workers to PCB-induced chloracne. A second explanation is a misclassification or misdiagnosis of the acne-like lesions due to the ab sence of histological confirmation. A third explanation for the inconsistent results is that the chloracne, and per haps also other oculodermal effects, are caused by a contaminant (ie, PCDFs) rather than PCBs.27a,29a PCDFs were not routinely measured in either the fluids or the individuals of any of these studies, and, based on the findings of Yusho and Yu-Cheng incidents, PCDFs are relatively potent chloracnegens. In contrast to the above reports, studies of environmental exposure to PCBs, primarily through contami nated fish consumption,31 contami nated sludge use,32 or residence near a PCB waste site33 have not demon strated the occurrence of chloracne or any other significant dermal effect or chronic skin disease. Studies examin ing low-level or short-term exposure to PCBs after a spill34 or through sew age treatment operations35 also failed to demonstrate significant acute or chronic dermal conditions attribut able to PCBs. Hepatic Effects Studies in animals indicate that PCBs induce hepatic drug metabolism and at higher dosages may cause liver toxicity.36 Evidence for both of these effects has been sought in clinical studies of PCB exposure. With respect to hepatic enzyme induction, three approaches have been used for its de tection. One approach has been to measure the rate of elimination of a test dose of antipyrine, a known sub strate for hepatic enzymes. Alvares et al27 found the average half-life of an tipyrine in five capacitor workers ex posed to PCBs to be significantly less than the average half-life in five con trol subjects (10.8 hours versus 15.6 hours), suggesting an hepatic induc tion effect associated with PCB expo sure. In contrast, a much larger series evaluated by Emmett et al,29,30 found the mean half-life of antipyrine in forty-seven PCB-exposed workers to be higher than, but not significantly different from, the average half-life in forty-four controls (12.4 hours versus 10.7 hours). The reason for this differ ence in observations is unclear, but could conceivably be due to differ ences in extent of exposure for the two study populations. ' A second approach has been through measurement of serum 7-glu- tamyl transpeptidase (GGT) activity, a parameter whose increase might, among other possibilities, suggest the occurrence of hepatic enzyme induc tion.37 None of the many studies that measured serum GGT activities found average values in PCB-exposed workers that were abnormal or signif icantly increased compared to control subjects. The evidence for serum GGT as an indicator of PCB exposure is equivocal: A number of investiga tors reported a correlation between serum PCB levels and serum GGT activities,23,29,32,37"40 while other inves tigators found no correlation between serum PCBs and serum GGT,24,41 or that such correlations are lost when OS W 148220 STLCOPCB4039705 JOM Volume 35, Number 2, February 1993 141 the results are adjusted for alcohol consumption in the subjects.28-36-37 The third approach has been to measure urinary D-glucaric acid lev els,16 a test that purportedly reflects changes in hepatic microsomal glucuronidation. While PCB-exposed workers had a significantly increased D-glucaric acid excretion in this study, no correlation between this measure ment and PCB blood levels was found, making it difficult to attribute these differences to PCBs. Currently, there is no clinical evi dence for toxic effects on the liver due to PCB exposure, even among electri cal workers with extensive PCB body burdens. In general, the results of a number of studies conducted from 1976 through 1988 have revealed that occupationally exposed groups were within the expected normal limits, or that the number ofpersons falling out side the normal range was not signif icantly greater than that anticipated.13-24-29,37'41 For example, studies by Ouw et al13 and Emmett29-39 either reported mean values for liver func tion tests that were shown to fall within the normal range for each test, or the mean values were not signifi cantly different from that of the con trol population. The study of Fischbein et al,24 one the largest study pop ulations examined to date, is also illustrative as they reported the fre quency of normal liver function test Values for their PCB-exposed popula tion to be 97.8% for serum aspartate aminotransferase (AST) activity, 92.8% for serum alanine aminotrans ferase (ALT) activity, 97.5% for serum lactate dehydrogenase (LHD) activity, 98.1 % for serum GGT activ ity, 98.8% for serum alkaline phos phatase activity, and 94.7% for serum bilirubin levels. These high percent ages are remarkably consistent with the fact that only about 95% of any normal population is expected to have values that fall within a "normal" range defined as two standard devia tions about the mean value. Given that 66% of the population examined by Fischbein et al24 was 40 years of age or older, and that 70% of this population had been exposed for more than 10 years, this study, per haps more than any other, indicates that the occupational exposures to PCBs did not produce liver toxicity. While correlations between- serum PCB levels and the numerical value for one or more different liver func tion tests have been reported, there is no liver function test for which such correlations have been consistently found, and in each case the correla tion was observed in a study where no excess number of abnormal test values was reported. For example, a correla tion between serum PCB levels and serum AST activity (or its log-trans formed value) has been reported in three studies,24-37-38 but was not found in four similar studies.23-29-39,40 The in termittent finding of these correla tions, correlations that are being re ported in populations with normal liver function tests, has no apparent toxicological or physiological mean ing and may instead merely arise from the large number of statistical com parisons being performed in these studies. An unusual observation of in creased incidence of hepatomegaly among PCB-exposed workers was re ported in one study.15 A total of 16 workers from a cohort of 80 were regarded as having "hepatic involve ment," either in the form of hepato megaly or one or more elevated liver enzyme tests. Among workers in plant B, where exposure included PCBs dis persed by explosions during stress test ing of capacitors, the incidence of hep atomegaly was especially high--more than 50% (7/13). While a relationship between serum PCB concentrations and the appearance of hepatomegaly was asserted in this study, there was no apparent relationship between serum PCB levels and severity of hep atomegaly. More importantly, the vir tual absence of hepatomegaly among the numerous other clinical studies of PCB-exposed workers13.23.24.29.30.32.37-4 makes it difficult to attribute this ob servation simply to PCB exposure. It is unlikely that the discrepancy be tween this study and others is a matter of degree of PCB exposure, since at least two other studies23,24 have con ducted physical examinations on study populations with serum PCB levels as high or higher (up to 3,850 ppb) than those in the Maroni et al14- 16 study group with no reported hep atomegaly. It is possible that the na ture of the exposure of this Italian population of PCB workers was unique, perhaps including an unusu ally high exposure to PCDFs formed during capacitor explosions. With re spect to liver function tests, abnormal results occurred randomly and were generally too small to be considered clinically significant. In no case were any of the small elevations in liver enzymes confirmed by retesting. Among the 67 workers in plant A, where exposures were not confounded by potential PCDF formation from capacitor explosions, the number of minimally elevated serum ALT, serum AST, and serum GGT values was that expected for a population of normal, healthy individuals (ie, <5%). Furthermore, no relationship was ap parent between the occasional abnor mal values and PCB exposure (Table 4); for individual liver tests, there were no significant differences in serum PCB levels between those with normal and abnormal test values. In conclusion, while the results of one study using antipyrine clearance have suggested hepatic enzyme induc tion in workers exposed occupation ally to PCBs, other studies using this and other approaches have not found induction attributable to PCB expo sure. The possibility of toxic effects in the liver associated with PCB expo sure has been examined in detail in several studies. Clinical measure ments reflective of liver toxicity from the various studies have been consist ently negative. In spite of evidence for hepatotoxic potential of PCBs in animal studies, the absence of adverse liver effects in PCB-exposed workers should not nec essarily be considered a surprising finding. The early concern for work place exposures was the prevention of liver and dermal toxicity,42-43 and this is clearly reflected in the initial thresh old limit values of 1.0 mg/m3 (42% chlorine mixtures) and 0.5 mg/m3 (54% chlorine mixtures) that were first established by the American Con ference of Governmental and Indus trial Hygienists in 1959. These guide lines were lower than those of 2.0 mg/ m3 (42% chlorine mixtures) and 1.0 DSW 148221 STLCOPCB4039706 142 PCBs and Human Disease James et al TABLE 4 Comparison of Mean Serum PCB Levels between Individuals with and without Abnormal Liver Function Indicators in the Cohort of Maroni et al (1981 b)15 Test (Group Results) Number of Workers Mean PCB Blood Level SEM (PPb) t-test (P Value)* Serum alanine aminotrans ferase Normal values Elevated values 10 627 123 6 352 76 NS (0.13) Serum aspartate aminotrans ferase , Normal values Elevated values . 13 3 542 104 411 142 NS (0.58) Serum ornithine-carbamoyl transferase Normal values Elevated values 10 440 110 6 663 136 NS (0.23) Serum 7-glutamyl transpepti dase Normal values Elevated values 8 8 648 + 157 400 60 NS (0.16) * Data derived from Maroni et al (1984)16 was subjected to the Student's f-test for unpaired values. NS (no significant difference at P < .05). mg/m3 (54% chlorine mixtures) pro posed earlier by Treon et al44 as ap parently protective of liver toxicity based on animal studies performed by these authors. The initial threshold limit values were also well below the 10.5 mg/m3 measured-air PCB con centrations in Massachusetts plants that Elkins4' reported as having pro duced no evidence of toxicity other than irritation. More recently, Lawton et al23 have pointed out that the doses and tissue levels of PCBs causing liver damage in rodents are some three to four orders of magnitude greater than those experienced occupationally. Effects on Serum Lipid or Lipoprotein Levels Many of the clinical studies of both occupational and environmental ex posure to PCBs have included meas urements of serum lipids, but no ad verse effect of PCB exposure on serum lipid levels has ever been demon strated. Clear evidence of an absence of significant clinical abnormalities in lipid metabolism has been provided in five studies of PCB-exposed work ers. Baker et al32 studied 148 persons exposed to PCBs both occupationally and environmentally and found no excess clinical abnormalities in serum triglycerides, total cholesterol, or highdensity lipoprotein (HDL)-cholesterol among exposed persons. Likewise, Smith et al38 studied 274 capacitor/ transformer workers who sustained high-level PCB exposures and found no excess clinical abnormalities in serum triglycerides, total cholesterol, or HDL-cholesterol among exposed persons. After comparing the mean values between PCB-exposed and matched control groups, no signifi cant differences were found by Chase et al37 in the mean serum triglycerides, total lipids, total cholesterol, or HDLcholesterol of 120 locomotive repair workers, or by Emmett29-39 in the mean serum triglycerides, total lipid, total cholesterol, HDL-cholesterol, low-density lipoprotein (LDL)-cholesterol, or very low-density lipoprotein (VLDL)-cholesterol of 55 transformer maintenance/repair workers. As with the serum liver enzyme correlations, a statistical correlation between serum PCB levels and the levels of some component of serum lipids have been reported in a number of studies. The two most common correlations were between serum PCBs and serum triglyceride lev els,23-29-37-38-45 and serum cholesterol levels.23-29-45 While some have postu lated that these correlations suggest a possible effect of PCBs on lipid me tabolism, a review of all available evi dence finds several facts that argue strongly against this supposition. First, no consistent correlation has been found in the clinical studies of either occupational or environmental exposures. Instead, the type of corre lation varies among the studies; some studies reported a correlation for PCBs with triglyceride levels and not cholesterol levels and others reported correlations for cholesterol levels and not triglyceride levels. Similar con flicting correlations are also found among the clinical studies of environ mental exposures to PCB.31"35-45 Sec ond, of five studies examining PCBs and serum lipids, none noted a signif icant excess of abnormal lipid values or a significant difference between control and PCB-exposed groups in total lipid, serum triglyceride, total cholesterol, HDL-cholesterol, LDLcholesterol, or VLDL-cholesterol lev els.29-32-37-39 Thus, as with the liver en zyme correlations, the correlations be tween serum PCB levels and serum lipid levels are found among persons with normal rather than elevated serum lipid values, and, consequently, these correlations have no toxicologi cal significance. Third, conflicting correlations have been observed both among environmentally exposed per sons (low exposure) and occupational (high exposure) studies, suggesting this phenomenon is unrelated to dose. As demonstrated by two recent stud ies,23-39 these correlations can be ex plained by the natural partitioning of PCBs to serum lipids. Emmett39 examined 55 transformer maintenance workers and initially found statistical correlations between log serum PCBs and log triglyceride, total cholesterol, and log VLDL con centrations. These observations of Emmett were consistent with the var ious correlations between serum PCBs and lipid levels that have been re ported by others. However, Emmett performed additional analyses to demonstrate that these correlations were of no medical consequence. DSW 148222 STLCOPCB4039707 JOM Volume 35, Number 2, February 1993 143 First, he demonstrated that the corre lations between serum lipid measure ments and serum PCB levels could be found even though the mean serum lipid levels of the PCB-exposed group were not significantly different from those of the control group. That is, these correlations existed even though the serum lipid levels of the PCBexposed population had not been ele vated by exposure. Second, no corre lation was found between adipose PCB levels, a better indicator of PCB body burden than serum measure ments, and any serum lipid compo nent. If tissue PCB levels do not cor relate with serum lipid levels, then one cannot propose some target organ tox icity as the cause of the elevation in serum lipid levels. Confirming this conclusion, Emmett found that all serum PCB and serum lipid correla tions became nonsignificant when he controlled for potentially confound ing variables such as alcohol intake, liver disease, history of diabetes, heart disease, etc. Based on these observa tions, Emmett proposed the serum PCB-lipid correlations could be ex plained chemically rather than toxi cologically. Due to their lipophilic na ture, the movement of PCBs into blood will be driven by the partition ing of PCBs between adipose tissue and lipids in the blood, principally triglycerides and cholesterol. There fore, the solubility of PCBs in the blood is a function of blood lipid con tent, and blood with higher triglycer ides or cholesterol will contain more PCBs. This finding indicates that the observed correlations exist because serum lipid levels are affecting serum PCB levels rather than PCB exposure affecting serum lipid levels. Lawton et al23 provided an exten sive analysis of clinical chemistry and hematological data from a group of 194 capacitor workers examined on two occasions. When correlation coef ficients were calculated for log serum PCB levels and serum lipids, the cor relations. were significant between PCBs and log serum triglycerides and cholesterol, similar to reports by oth ers. However, when the correlation coefficients were calculated using the log PCB concentration in serum lip ids, the associations disappeared. Thus, this finding provides further confirmation ihat the associations be tween gross serum PCB levels and serum lipid levels can be explained simply by the partitioning behavior of PCBs. The findings of these two studies23-39 provide insight concerning the nu merous inconsistent correlations be tween serum lipid levels and serum PCB concentrations reported by oth ers. First, the inconsistency in report ing may be explained, at least in part, by the failure of studies reporting cor relations to control for those variables of the individual (eg, age, alcohol con sumption, medical history) that may affect serum lipid levels. Second, these correlations are of no medical or tox icological significance because they exist in groups ofindividuals with nor mal lipid levels. Third, since there are no correlations between serum lipid levels and adipose tissue PCB levels, and as tissue levels are the better mea sure of body burdens and total dose received, there really is no correlation between PCB dose and serum lipid levels. Fourth, both studies demon strate that associations with gross serum PCB levels and serum lipid lev els are easily explained by the parti tioning behavior of PCBs. Effects on the Cardiovascular System Kreiss et al31 reported a significant correlation between log serum PCB levels and increases in diastolic blood pressure in a cohort of 458 Alabama residents exposed to higher-than-normal levels of PCBs and DDT via the ingestion of local contaminated fish. The magnitude of the changes in blood pressure was not reported, but the authors stated the rates of border line and definite hypertension were 30% higher than expected. Definitive conclusions from this study were pre vented by the lack of a matched con trol group, and later recognition that the co-linearity of PCB and DDT lev els in these individuals precluded as signment of any observation to either chemical alone.47 Several studies of environmentally exposed cohorts have subsequently attempted to repro duce this observation of elevated blood pressure, but have failed to do so. Stehr-Green et al33 found no excess of heart disease or hypertension in individuals living near PCB-contain- ing hazardous waste sites and, after controlling for confounding variables, found no correlation between log serum PCB levels and log diastolic blood pressure. A relationship be tween PCBs and hypertension was also not observed in another, much larger study of environmental expo sure to PCBs via consumption of con taminated fish.48 Mean heart rate and blood pressure were also normal in a cohort of sewage treatment workers potentially exposed to PCBs.35 There fore, the initial evidence suggesting that PCB exposure may affect blood pressure is not convincing and has not been supported by subsequent studies of environmentally exposed popula tions. If environmental exposure to PCBs does lead to hypertension or any other form of cardiovascular disease, an increase in cardiovascular disease and mortality should be easily ob served in studies of individuals occu pationally exposed to PCBs. Only two investigations have specifically re ported findings related to cardiovas cular status, but several studies have addressed potential cardiovascular ef fects through blood pressure measure ment, complete physical examina tions, chest radiographs, and/or med ical histories. None reported an excess occurrence of high blood pressure or any other cardiovascular abnormal ly. QfIS. 19,24,28,29,30,32,37 ^g |WQ studies reporting specific findings, Smith et al38 found that no correlation existed between log serum PCBs and diastolic blood pressure when adjusted for worker age and sex, and Lawton et al23 reported "no evidence for health impairments related to PCBs" after analyzing the incidence of hyperten sion and results of electrocardiograms and chest radiographs for each worker in their cohort. Lawton et al23 further noted that, "Despite the prevalence of cardiovascular risk factors (obesity, el evated cholesterol levels, smoking, etc) the mortality experience has been normal." Confirming this lack of ef fect of PCBs on the cardiovascular DSW 1A8223 STLCOPCB4039708 144 PCBs and Human Disease James et al system, none of the major mortality dies completed to date have found _uy association between PCBs and cardiovascular disease including hy pertension.26-49'52 In fact, a less-thanexpected mortality from heart diseases has been observed in two of these studies.51-52 Effects on the Lungs The relationship between pulmo nary function and PCB exposure has only been evaluated in some detail in three studies of two different popula tions occupationally exposed to PCBs. Pulmonary function tests and chest radiographs have been used, along with questionnaires dealing with res piratory-related symptoms. In an ini tial examination of pulmonary func tion in US (New York) capacitor workers, reported respiratory symp toms included upper respiratory tract irritation (50% of workers), wheezing (3%), tightness in the chest (10%), and "work-related" cough (14%).24-28 In `he absence of a comparison popula- on matched for age and smoking status, it is unclear which of these might be in excess. Spirometric tests found 14% of examined workers with diminished vital capacity and 11% with restrictive impairment. Chest ra diographs in all but one case of restric tive impairment were normal, an unusual finding if there was in fact pulmonary impairment due to occu pational exposure. In a follow-up study of the same population, no spirometric abnormal ities were found.53 In view of the ab sence of evidence of pulmonary im pairment on follow-up, and the failure of chest radiographs to confirm re strictive impairment in the first ex amination, the authors concluded that the original restrictive impairment finding was "artifactual due to test operator inexperience and inadequate expiratory efforts." No correlation of spirometric variables with past expo sure or serum PCB levels was found in the follow-up study, and, from this cohort, no evidence of an association between PCBs and respiratory disease has emerged. Emmett et al29,30 reported that 40% ofexposed transformer repair workers complained of wheezing compared to a 20% incidence among controls. However, the wheezing was not cor-' related with the extent of PCB expo sure (in terms of exposure history, log serum PCB levels, or log adipose PCB levels) or with objective measure ments of respiratory function. Among the respiratory parameters measured, only the forced expiratory volume in 1 second was significantly lower in the PCB-exposed workers compared with controls; but this difference did not correlate with PCB body levels and was eliminated when the results were corrected for smoking status. There fore, this study of transformer repair workers was similar to those of capac itor workers in that it provided no indication that chronic, occupational PCB exposures adversely affect pul monary function. Clinical investigations of respira tory function or complaints related to environmental PCB exposure are con sistent with the negative evidence pro vided by studies of occupationally ex posed persons.33,35 Furthermore, the available mortality analyses of PCB- exposed workers have not revealed a significant excess of deaths from non- malignant respiratory dis eases.17,50,51,52,54 Therefore, with the possible exception of the irritation that might be produced at some levels with any halogenated compound, the collective evidence demonstrates that chronic PCB exposure is not associ ated with pulmonary dysfunction or respiratory diseases. Effects on the Blood and Immune Systems Clinical studies examining blood cell counts and red blood cell param eters in PCB-exposed individuals have not reported an increased prevalence of any particular abnormality or gen eralized syndrome which suggests hematotoxicity. The extent of analyses in these studies varied considerably, from simple determinations of he moglobin content and hematocrit to complex statistical analyses of various hematologic parameters versus blood concentrations of higher chlorinated and lower chlorinated PCB homologs. In perhaps the most thorough study of its kind, Lawton et al23 compared serum PCB levels in capacitor workers with a variety of hematological pa rameters. Data were available from examinations of the workers con ducted in 1976 (n = 194) and 1979 (n = 174). No associations were found between PCB blood levels and param eters related to red blood cells. Ele vated lymphocyte counts were noted in the first examination (1976), but of these only an elevation in monocytes was confirmed in the second exami nation (1979). This increase, while statistically significant, was small (seen only in 11 of 194 workers) and had no apparent clinical importance, since mean values for all of the he matologic parameters analyzed were well within the standard range re ported. The authors concluded that serum PCB levels were related to "marginal" increases in monocyte counts that might have been related to the use of different analytical meth ods to assess this parameter in the two examination periods. Consistent with these negative findings, Taylor51 re ported that PCB capacitor worker mortality from diseases of the blood and blood-forming organs was not sig nificantly different from that expected based on national mortality rates. Only three clinical surveys involv ing persons environmentally exposed to PCBs have reported analyses of he matological parameters. None of these studies found an excess of he matological abnormalities.33'35 While studies examining immunocompetence in individuals exposed to PCBs are quite limited, what evidence is available does not suggest that oc cupational exposures to PCBs were immunotoxic. First, there is no evi dence from existing studies that PCB exposure adversely affects the levels of circulating immunocytes. In fact, leu kocyte count and differential blood cell counts from six clinical studies found no association between PCB exposure and these parame ters.15,24,32,34,37,38 Second, Emmett et al30 have recently compared 55 work ers exposed to PCBs and 56 nonexposed persons with respect to hyper sensitivity reactions as measured by dermal responses to mumps and tri chophyton antigens and found no sig OSW 148224 STLCOPCB4039709 JOM Volume 35, Number 2, February 1993 145 nificant differences in the response frequencies of the two groups. Third, the clinical studies of PCB-exposed workers in general provide indirect evidence that occupational contact with PCBs does not result in immune system dysfunction. Exposed PCB workers have been consistently de scribed as healthy overall, with no evi dence of increased frequency of infec tions or illnesses; and no association has been found between PCB expo sure and an excess mortality from in fectious diseases.51 Effects on the Kidneys Several clinical studies of PCB-ex posed populations have included tests for renal dysfunction or damage. The tests most commonly included were blood urea nitrogen (BUN) and cre atinine measurements, although other indices were occasionally analyzed. No study found evidence for an asso ciation between PCB exposure and renal toxicity or kidney disease. The infrequent abnormalities noted in some of the studies of larger worker populations lack control groups for comparison.23-24 These low frequen cies do not appear to be greater than would be anticipated in a healthy pop ulation. Lawton et al23 reported in creases in mean urinary specific grav ity and serum osmolarity, as well as a greater than expected prevalence of cells in the urine, but stated that meth odological deficiencies (failure to ob tain "clean catch" urine samples) and the historical prevalence of diabetes and urinary tract problems (kidney stones [4.1%] and urinary tract infec tions [7.2%]) among the workers may explain these findings independently of PCB exposure. Supporting evi dence for the lack of significant find ings in the clinical studies is the fact that mortality from genitourinary dis eases such as nephritis is not signifi cantly elevated among US capacitor workers.51 In a study of potential PCB expo sure via environmental contamina tion, Stehr-Green et al33 reported weak statistical correlations between log serum PCB levels and both uric acid (P = .08) and BUN (P = .05). However, mean values for these pa rameters fell within the expected ranges and the authors failed to asso ciate any increase in physician-diag nosed urinary tract problems with in creased serum PCB levels. Stark et al34 reported transient changes in serum electrolytes (slightly increased potas sium and decreased phosphorus) in people exposed to a transformer fluid spill, but BLJN, uric acid, and creati nine were unaffected and serum elec trolytes were normal at the 6-week follow-up examination. Nethercott and Holness35 similarly found no changes in BUN or serum creatinine in sewage treatment workers poten tially exposed to PCBs. Effects on the Gastrointestinal Tract Detailed evaluations of potential gastrointestinal effects of human ex posures to PCBs are currently unavail able. However, none of the clinical or mortality studies to date have noted an increased occurrence of gastric ul cer or other nonmalignant gastroin testinal diseases among PCB-exposed workers.51,54 In the clinical investiga tions of PCB-exposed workers, the in cidence of nonspecific complaints possibly relating to gastrointestinal function was reported,15-24-30-38 but no clear association with PCB exposure or a specific clinical effect was identi fied. Queries about gastrointestinal symptoms in other occupational cohorts23-37-39-41 or after environmental exposures to PCBs.31"34 apparently re vealed no excess complaints or asso ciations with PCB body levels. Effects on the Nervous System The current human evidence sug gests that PCB exposure is not associ ated with clinical impairment of the nervous system. Although some asso ciations between exposure and in creased rates of subjective symptoms have been made, the meaning of sub jective symptoms in the absence of objective clinical signs with regard to chemical exposure is always difficult to interpret. In the case of PCBs, sub jective symptoms have been generally inconsistent with medical histories, physical examination findings, and measures of PCB exposure. For ex ample, Fischbein et al24 reported an apparently high prevalence (158/326) of subjective neurological symptoms, primarily the common complaints of headache, nervousness, or fatigue among PCB capacitor workers. How ever, none of the subjective symptoms was related to duration of employ ment or serum PCB level, and the authors noted that, "routine neurolog ical examination did not reveal any remarkable prevalence of abnormali ties." Similarly, Smith et al38 reported the results of a questionnaire given to PCB-exposed workers in which com plaints of "loss of appetite" and "tin gling in the hands", were statistically correlated with log serum PCB con centrations "in the presence of confounders." These confounders appar ently could not be excluded from the analysis, and the meaning of these responses is impossible to determine without further investigation. Though physical signs of neurotoxicity were not specifically discussed by the au thors, Smith et al38 stated "No consist ent patterns of abnormalities were noted on physical examination at any study site." Finally, while Emmett et al30 reported higher rates of subjective symptoms (headache, loss of appetite, insomnia, and memory trouble) in transformer repair workers, these in vestigators found no corresponding excess of neurologic abnormalities in the medical histories or physical ex aminations of these workers. In summary, a number of occupa tional studies have failed to find any excess neurological deficits related to PCBs after performing complete phys ical examinations and/or medical his tory analyses.15-18-19-23-32-37-41 Consistent with the negative findings of these clinical studies, the mortality analyses have not revealed any significant ex cess of deaths from nervous system diseases among PCB capacitor workers.26-49-51 Effects on Endocrine Function The potential effects of PCB expo sure on endocrine function have only been, investigated to a limited extent in the clinical studies reported to date. No untoward effects on thyroid func tion were found in studies measuring serum T3 and/or T4 levels or on the prevalence of thyroid-related abnor malities.29,39 Likewise, no evidence of a PCB-induced effect on the pancreas OSW 148225 STLCOPCB4039710 146 PCBs and Human Disease James et al creas has been observed. Lawton et al23 reported that five of 40 elevated tod glucose measurements during ..wo clinical examinations of capacitor workers were probably related to dia betes while the remainder were appar ently due to testing inconsistencies and/or methodological problems. Stehr-Green et al33 reported that mean blood glucose and the relative risk of elevated blood glucose were not sig nificantly altered in people with >20 ppb serum PCBs, although a weak trend in increase was observed. Smith et al38 and Emmett et al29 observed no alterations in blood glucose measure ments among capacitor and trans former repair workers. Confirming the lack of endocrine effects reported in the clinical studies, no mortality study has revealed any excess of deaths related to endocrine organ ef fects, such as diabetes mellitus or pan creatitis. In fact, Taylor51 specifically reported a lower than expected mor tality from allergic, endocrine, meta bolic, and nutritional diseases, and from diabetes. While the present data 're limited, there is no evidence that CB exposure results in endocrine dysfunction or disease of any kind in occupationally or environmentally exposed persons. Conclusions Studies of PCB-exposed popula tions collectively suggest that the only adverse health effects attributable to PCBs in humans are dermal: chloracne, hyperpigmentation, and se quelae of chronic dermal and ocular irritation. These conditions occurred only in worker populations with rela tively high dermal and/or inhalation exposures. PCB-related dermal effects have not been clearly identified among worker populations receiving lesser exposure and were absent in populations environmentally exposed to PCBs through the consumption of contaminated fish, by living near PCB waste sites, or by potential contact vith PCB-contaminated soils. The as signment of PCBs as the causative agent for chloracne among workers with high PCB exposures is only ten tative. Reported incidences of chlor acne were typically greatest for occu pational environments in which PCBs were used as heat transfer fluids, rais ing the possibility, as was the case with Yusho and Yu-Cheng, of unusually high levels of contamination with po tently chloracnegenic PCDFs. The re lationship between PCB body burdens and chloracne is not necessarily con sistent with PCBs as causative agents--chloracne was only found in workers with high body burdens of PCBs, but many workers with very high body burdens did not develop chloracne. In the absence of measure ments of PCDF levels in these studies, a causative or contributing role of PCDFs in chloracne among PCB-ex posed workers cannot be ruled out. The collective occupational experi ence with PCB fluids provides no evi dence for adverse effects for other or gan systems, including the liver, the heart and circulatory system, the gas trointestinal and urinary tracts, the nervous system, the respiratory sys tem, the immune/hematopoietic sys tem, and some endocrine functions. The absence of adverse effects in hu mans despite a variety of toxic effects ofPCBs demonstrated in animal stud ies is consistent with the conclusions of other recent reviews of PCB toxi cology.55-58 This divergence in the findings reported for animal and hu man studies may result from many factors including species differences in susceptibility or sensitivity to PCB ef fects, as well as dosages tested in ani mal studies that were far greater than those found in even the highest of the occupational exposures. Confidence in the apparent absence of PCB-re lated diseases in humans is strength ened by the observation that this in formation is based primarily on stud ies of individuals with chronic, very high exposures to PCBs in industries where PCB use spanned over a 50year period. With the strictly regulated use and disposal of these chemicals, it is unlikely that human exposures to PCBs will ever again match those of workers in previous decades. There fore, it would appear that there is little basis for concern for organ system toxicity, at least among the broad cat egories covered in this review, result ing from present day exposures to these compounds. References 1. Kunita N, Kashimoto T, Miyata H, Fu kushima S, Hori S, Obana H. Causal agents of Yusho. Am J Ind Med. 1984;5:45-48. 2. Masuda Y, Kuroki H, Haraguchi K, Nagayama J. PCB and PCDF congeners in the blood and tissues of Yusho and YuCheng patients. Environ. Health Perspect. 1985;59:53-58. 3. Hori S, Obana H, Tanaka R, Kashimoto T. Comparative toxicity in rats of poly chlorinated biphenyls (PCBs), polychlo rinated quaterphenyls (PCQs) and poly chlorinated dibenzofurans (PCDFs) present in rice oil causing "Yusho." Eisei Kagaku. 1986;32:13-21. 4. Kashimoto T, Miyata H, Kunita S, et al. Role of polychlorinated dibenzofuran in Yusho (PCB poisoning). Arch Environ Health. 1981;36:321-326. 5. Masuda Y, Kuroki H, Yamaryo T, Har aguchi K, Kuratsune M, Hsu ST. Com parison of causal agents in Taiwan and Fukuoka PCB poisonings. Chemosphere. 1982;11:199-206. 6. Masuda Y, Yoshimura H. Polychlori nated biphenyls and dibenzofurans in patients with Yusho and their toxicolog ical significance: a review. Am J Ind Med. 1984;5:31-44. 7. Chen PH, Luo ML, Wong CK, Chen CJ. Polychlorinated biphenyls, dibenzofur- . ans, and quaterphenyls in the toxic ricebran oil and PCBs in the blood of pa tients with PCB poisoning in Taiwan. Am J Ind Med. 1984;5:133-145. 8. Kunita N, Hori S, Obana H, Otake T, Nishimura T, Ikegami N. Biological ef fects of PCBs, PCQs and PCDFs present in the oil causing Yusho and Yu-Cheng. Environ Health Perspect. 1985;59:79-84. 9. Miyata H, Fukushima S, Kashimoto T, Kunita N. PCBs, PCQs and PCDFs in tissues of Yusho and Yu-Cheng patients. Environ Health Perspect. 1985;59:59-66. 10. Kashimoto T, Miyata H, Fukushima S, Kunita N, Ohi G, Tung T. PCBs, PCQs and PCDFs in blood of Yusho and YuCheng patients. Environ Health Perspect. 1985;59:73-78. 11. Kashimoto T, Miyata H. Differences be tween Yusho and other kinds of poison ing involving only PCBs. In: Waid JS, ed. PCBs and the Environment. Vol III. Boca Raton, FL: CRC Press; 1986:1-26. 12. Kuratsune M. Yusho, with reference to Yu-Cheng. In: Kimbrough RD, Jensen AA, eds. Halogenated Biphenyls, Terphenyls, Naphthalenes, Debenzodioxons and Related Products. Amsterdam: El sevier; 1989;381 --400. 13. Ouw HK, Simpson GR, Siyali DS. Use and health effects of Aroclor 1242, a DSW 148226 STLCOPCB4039711 JOM Volume 35, Number 2, February 1993 147 polychlorinated biphenyl in an electrical industry. Arch Environ Health. 1976;31:189-194. 14. Maroni M, Colombi A, Cantoni S, Ferioli E, Foa V. Occupational exposure to polychlorinated biphenyls in electrical workers. I. Environmental and blood polychlorinated biphenyl concentra tions. BrJIndMed. 198 la;38:49. 15. Maroni M, Colombi A, Cantoni S, Ferioli E, Foa V. Occupational exposure to polychlorinated biphenyls in electrical workers. II. Health effects. BrJIndMed. 198 lb;38:55. 16. Maroni M, Colombi A, Ferioli A,. Foa V. Evaluation of porphyrinogenesis and enzyme induction in workers exposed to PCB. Med Lav. 1984;75:188-199. 17. Gustavsson P, Hogstedt C. Rappe C. Short-term mortality and cancer inci dence in capacitor manufacturing work ers exposed to polychlorinated biphenyls (PCBs). Am J Ind Med. 1986; 10:341-- 344. 18. Hara I. Health status and PCBs in blood of workers exposed to PCBs and of their children. . Environ Health Perspect. 1985;59:85-90. 19. Takamatsu M, Oki M, Maeda K, Inoue Y, Hirayama H, Yoshuzuka K. Surveys of workers occupationally exposed to PCBs and of Yusho patients. Environ Health Perspect. 1985;59:91-97. 20. Bertazzi PA, Riboldi L, Pesatori A, Radice L, Zocchetti C. Cancer mortality of capacitor manufacturing workers. Am J Ind Med. 1987;11:165-176. 21. Meigs JW, Albom JJ, Keith BL. Chloracne from an unusual exposure to Arodor. JAMA. 1954;154:1417-1418. 22. Jones JW, Alden HS. An acneform dermatergosis. Arch Dermatol Syphilol. 1936;33:1022-1034. 23. Lawton RW, Ross MR, Feingold J, Brown JF Jr. Effects of PCB exposure on biochemical and hematological findings in capacity workers. Environ Health Per spect. 1985;60:165. 24. Fischbein A, Wolff MS, Lilis R, Thorn ton J, SelikoffIJ. Clinical findings among PCB-exposed capacitor workers. Am NY AcadSci. 1979;320:703-715. 25. Lawrence C. PCB and melanoma. N Engl J Med. 1977;296:108. 26. Brown DP, Jones M. Mortality and in dustrial hygiene study of workers ex posed to polychlorinated biphenyls. Arch Environ Health. 1981 ;36:120-129. 27. Alvares AP, Fischbein A, Anderson KE, Kappas A. Alterations in drug metabo lism in workers exposed to polychlori nated biphenyls. Clin Pharmacol Ther. 1977;22:140-146. 27a.Fischbein A, Wolff MS, Bernstein J, Se likoff IJ, Thornton J. Dermatological findings in capacitor manufacturing workers exposed to dielectric fluids con taining polychlorinated biphenyls (PCBs). Arch Environ 1982;37:69-74. . Health. 28. Warshaw R, Fischbein A, Thornton J, Miller S, Selikoff IJ. Decrease in vital capacity in PCB-exposed workers in a capacitor manufacturing facility. Ann NY AcadSci. 1979;320:277-283. 29. Emmett EA, Maroni M, Schmith JM, Levin BK, Jefferys J. Studies of trans former repair workers exposed to PCBs. II. Results of clinical laboratory investi gations. Am J Ind Med. 1988b; 14:47-62. 29a.Fischbein A, Rizzo J, Solomon SJ, Wolff MS. Oculodermatological findings in workers with occupational exposure to polychlorinated biphenyls (PCBs). Br J ' Ind Med. 1985;42:426-430. 30. Emmett EA, Maroni M, Schmith JM, Levin BK, Jefferys J. Studies of trans former repair workers exposed to PCBs. I. Study design, PCB contaminations, questionnaire, and clinical examination results. Am J Ind Med. 1988a; 13:415-- All. 31. Kreiss K, Zack M, Kimbrough R, Need ham L, Smrek AL, Jones BT. Associa tion of blood pressure and polychlori nated biphenyls. JAMA. 1981 ;245:25052509. 32. Baker EL, Landrigan PJ, Glueck CJ, et al. Metabolic consequences of exposure to polychlorinated biphenyls (PCB) in sewage sludge. Am J Epidemiol. 1980;112:553-563. 33. Stehr-Green PA, Ross D, Liddle J, Welty E, Steele G. A pilot study of serum pol ychlorinated biphenyl levels in persons at high risk ofexposure in residential and occupational environments. Arch Envi ron Contam Toxicol. 1986;41:240-244. 34. Stark AD, Costas K, Chang HG, Vallet HL. Health effects of low-level exposure to polychlorinated biphenyls. Environ Res. 1986;41:174-183. 35. Nethercott J, Holness D. Health status of a group of sewage treatment workers in Toronto, Canada. Am Ind Hyg Assoc J. 1988;49:346-350. 36. Yoshimura H, Yoshihara S, Ozawa N, Miki M. Possible correlation between in duction modes of hepatic enzymes by PCBs and their toxicity in rats. Ann NY AcadSci. 1979;320:179-192. 37. Chase KH, Wong O, Thomas D, Bemey BW, Simon RK. Clinical and metabolic abnormalities associated with occupa tional exposure to polychlorinated bi phenyls (PCBs). J Occup Med. 1982;24: 109-114. 38. Smith AB, Schloemer J, Lowry LK, et al. Metabolic and health consequences of occupational exposure to polychlori nated biphenyls. Br J Ind Med. 1982; 39:361-369. 39. Emmett EA. Polychlorinated biphenyl exposure and effects in transformer re pair workers. Environ Health Perspect. 1985;60:185-192. 40. Fischbein A. Liver function tests in workers with occupational exposure to polychlorinated biphenyls (PCBs): com parison with Yusho and Yu-Cheng. En viron Health Perspect. 1985;60:145-150. 41. Acquavella JF, Harris NM, Nicolich MJ, Phillips SC. Assessment of clinical, met abolic, dietary, and occupational corre lations with serum polychlorinated bi phenyl levels among employees at an electrical capacitor manufacturing plant. J Occup Med. 1986;28:1177-1180. 42. Elkins HB. Industrial Toxicology. New York: John Wiley; 1958:152-153. 43. Irish DD. Chlorinated diphenyls. In: Patty FA, ed. Patty's Industrial Hygiene and Toxicology. Vol II 2nd ed. New York: Interscience Publishers; 1963: 1340-1343. 44. Treon JK, Cleveland FP, Cappel JW, Atchley RW. The toxicity of vapors of Aroclor 1242 and Aroclor 1254. Am Ind Hyg Assoc J. 1956;17:204-213. 45. Takamatsu M, Oki M, Maeda K, Inoue Y, Hirayama H, Yoshizuka K. PCBs in blood of workers exposed to PCBs and their health status. Am J Ind Med. 1984;5:59-68. 46. Steinberg KK, Freni-Titulaer LWJ, Rog ers TN, et al. Effects of polychlorinated biphenyls and lipemia on serum ana lytes. J Toxicol Environ Health. 1986;19:369-381. 47. Kreiss K. Studies on populations ex posed to polychlorinated biphenyls. En viron Health Perspect. 1985;60:193-199. 48. Massachusetts Department of Public Health. The Greater New Bedford PCB Health Effects Study 1984-1987. Boston: MDPH; 1987. 49. Brown DP, Mortality ofworkers exposed to polychlorinated biphenyls: an update. Arch Environ Health. 1987,42:333-339. 50. Nicholson WJ, Seidman H, Selikoff IJ. Mortality experience of workers exposed to polychlorinated biphenyls during manufacture ofelectrical capacitors. Pre liminary Report. Industrial Disease Standards Panel, Ontario Ministry ofLa bor, Toronto, Ontario, Canada; 1987. 51. Taylor PR. The Health Effects of Poly chlorinatedBiphenyls. Cambridge, Mass: Harvard University; 1988. Dissertation. 52. Sinks T, Steele G, Smith AB, Watkins K, Shultz R. Mortality among workers exposed to polychlorinated biphenyls. AmJ Epidemiol. 1992; 136:389-398. 53. Lawton RW, Ross MR, Feingold J. Spirometric findings in capacitor workers occupationally exposed to polychlori nated biphenyls. J Occup Med. 1986;28:453-456. 54. Zack JA, Musch DC. Mortality ofPCB Workers at the Monsanto Plant in Sauget, Illinois. St. Louis: Monsanto Co; 1979. 55. Kimbrough RD. Human health effects of polychlorinated biphenyls (PCBs) and OSW 148227 STLCOPCB4039712 148 polybrominated biphenyls (PBBs). Ann Rev Pharmacol Toxicol. 1987;27:87- 111. - j. Kimbrough RD. Polychlorinated bi phenyls: how do they affect human PCBs and Human Disease James et al health? Health Environ Digest. 1988;2:1-8. 57. Fischbein A, Rizzo JN. Polychlorinated . biphenyls: a brief review. Mt Sinai J Med. 1987;54:332-336. 58. Safe S. PCBs and human health. In: Safe S, ed. Polychlorinated Biphenyls (PCBs): Mammalian and Environmental Toxi cology. Berlin: Springer-Verlag; 1987: 133-145. DSw 148228 STLCOPCB4039713