Porphyria cutanea tarda

UNTIL THE 1970s, hexachlorobenzene (HCB), an organochlorine compound used widely in the environment, was used as a seed dressing to prevent fungal disease on grain. The main environmental sources of HCB encompass the use of some chlorinated pesticides, incomplete combustion, old dump sites, and the manufacture of chlorinated solvents and chlorinated pesticides.[1] Hexachlorobenzene is highly lipophilic, inert, chemically stable, and accumulates in living organisms. In serum, concentrations of HCB in the general population are usually below 1.5 ng/ml.[2,3] Hexachlorobenzene produces experimental porphyria in laboratory animals[4,5] and is a potent carcinogen in rodents, in which it induces hepatic and thyroid neoplasms.[6,7] Adverse reproductive effects (e.g., ovarian toxicity, alterations in ovarian esteroidogenesis,[8,9] depletion of [T.sub.3] and [T.sub.4],[10] goiter,[11] hypothyroidism[12]) have also been reported in laboratory animals. Information on the effects of chronic exposure to HCB in humans is scarce. Most data are related to the accidental exposure in Turkey during the time period between 1955 and 1959 when people consumed HCB-treated wheat grain intended to be used for seed grain. The outbreak affected more than 3 000 people--mostly children--among whom disturbances in porphyrin metabolism and skin lesions compatible with porphyria cutanea tarda (PCT) were observed.[13,14] Enlargement of the thyroid gland, Parkinsonism, and other neurological effects were also described among the affected people.[15] The internal dose of HCB was never measured.

Unusually high levels of HCB were detected in the air (35 ng/[m.sup.3], 24-h average) and in sera (26 ng/ml) of volunteer inhabitants of Flix, a rural village of 5 000 inhabitants in the south of Catalonia (Spain), located near an electrochemical factory, the only existing industry in this municipality. The factory was built in 1898 and has produced volatile chlorinated solvents during the past several decades. Grimalt et al.[16] conducted a preliminary study from 1980 to 1989 on mortality and cancer incidence in this population, and they found an excess incidence for soft-tissue sarcoma, thyroid cancer, and brain cancer in males. In light of these results, we conducted a cross-sectional study to assess the health status and effects potentially associated with HCB (i.e., PCT, thyroid disorders, Parkinson's disease, cancer, and reproductive outcomes). Additional follow-up of the population is also underway. In the current study, we described the health status of the aforementioned population and compared the results with the Catalan National Health Survey. We also assessed the cross-sectional association between exposure variables, including HCB internal dose, and health effects, which included serum levels of thyroid-stimulating hormone (TSH) and creatinine and total urine porphyrin levels.

Method and Materials

This study was conducted in June 1994. Given the minimal frequency of potential effects under investigation and the small size of the population, the cross-sectional study encompassed all residents of the population aged 14 y or older (N = 4 178). Trained interviewers administered a questionnaire that solicited demographic characteristics, lifestyle, perceived health, and chronic health conditions excerpted from the National Health Survey of Catalonia. We also included questions about residence, occupation, and adverse health effects potentially related to HCB exposure. We asked the females to provide a brief reproductive history of spontaneous abortion, low birth weight, and congenital malformation. We also requested 40-mi blood samples and 24-h urine and fecal specimens from a random sample, stratified by age, of the same population (n = 777), as well as from some volunteers of the nonrandomized population. Informed consent for inclusion in the study was obtained from all subjects.

The participation scheme is shown in Figure 1. Interviewers initially asked the random-sample population to answer the questionnaire and, on the basis of information provided, we also asked their relatives and neighbors from the nonrandomized population to answer the questionnaire. Briefly, a total of 1 800 persons responded to the questionnaire; this number represented 43% of the total population older than 14 y of age. A higher response rate was obtained for the random sample. Biological samples were obtained from 615 subjects, but 7 cases were not included because they were not residents of Flix. From the remaining 608 subjects who provided biological samples, 328 were from the random sample (42.2% of the random sample) and 280 from the nonrandomized population.

Nonresponses primarily resulted from our inability to contact the person after three visits at home at different times (44%); 36.6% of the individuals refused to participate in the study. The response rate was higher for females than men, but we noted no gender differences with respect to causes of nonresponse. Nonrespondent females were typically older than respondent women, but no differences by age were found among men. We observed no differences in sociodemographic or health variables between participants who provided biological samples and all participants. In addition, we did not observe differences between the randomized and nonrandomized samples apart from age; this resulted from our intentional age stratification of the random sample (Table 1).

[Figure 1 ILLUSTRATION OMITTED]

Table 1.--Characteristics of the Population Studied

(*) Values within parentheses are standard deviations.

We requested information via questionnaire on (a) perceived health (i.e., excellent, very good, good, fair, and poor); (b) 16 chronic conditions (i.e., hypertension, heart diseases, varicose veins, arthritis, allergies, asthma, chronic bronchitis, diabetes, intestinal ulcer, urinary problems, high cholesterol, cataracts, cutaneous problems, constipation, depression/anxiety, and embolism); and (c) and all diseases for which an association with HCB exposure had been previously suggested (i.e., PCT, thyroid disorders, Parkinson's disease, cancer, and reproductive outcomes). We also evaluated levels of TSH and creatinine in serum, as well as total urinary porphyrin, by the spectrofluorometric method. The standards for normal values were 0.6-1.4 ng/ml for creatinine and 0.25-5 mlU/ml for TSH, and the higher normal limit for urinary total porphyrin was 300 nmol/l.[17]

We used the medical records of the Primary Health Center of Flix and the Cancer Register of Tarragona (i.e., a population-based cancer registry that covered the region [since 1980] in which Flix is located) to perform a validity study of the self-reported diagnoses. Given the incompleteness of clinical records, validation of self-reported diagnoses was limited. However, we used the confirmed diagnoses (i.e., 50% of hypothyroidism, 41% of goiter, 50% of Parkinson's, 28% of renal diseases, 18% of the total number of cancers or tumors reported as diagnosed since 1980 in the questionnaire, 63% of low-birth-weight infants, and 33% of congenital malformations) for a sensitive analysis of the association between HCB exposure and health.

We used data from a reference population with the same socioeconomic conditions as Flix, excerpted from the Health Survey of Catalonia (HSC), to compare perceived health and prevalence of self-reported chronic conditions. The HSC was a cross-sectional survey that included 15 000 subjects and was based on a representative sample of the noninstitutionalized population of Catalonia, Spain. For comparative purposes we selected individuals from the HSC who were older than 14 y of age and who lived in municipalities with less than 20 000 inhabitants in rural areas.

We analyzed organochlorine compounds in sera by gas chromatography (GC) coupled to electron capture detection and GC coupled to chemical ionization negative-ion mass spectrometry. A Varian Star 3400 coupled to a Finnigan Mat INCOS XL was used for the analyses. Detectable levels of HCB (ng/ml) were found in all cases, and quantifiable concentrations of [Beta]-hexachlorocyclohexane ([Beta]-HCH), 2,2-bis[p-chlorophenyl]-1,1-dichloroethylene (p,p'-DDE) and total polychlorinated biphenyls (summation of 28, 52, 101, 118, 138, 153, and 1 80 congeners) were found in more than 85% of cases. Details of the methodology have been reported elsewhere.[18]

Organochlorine compounds distributions in serum were skewed; therefore, we used natural logarithmic transformations in the analyses. We measured the association between exposure variables and health outcomes by the odds ratio, using logistic regression models. Exposure variables were selected from the multivariate process, according to their relation to HCB. We used EGRET to perform the analyses. Statistical significance was defined as p [is less than] .05 (two-sided), and 95% confidence intervals (Cls) were determined.

Results

Serum levels of the organochlorine compounds are shown in Table 2. HCB concentrations were the highest (median = 16.5 ng/ml) of any chemicals studied. In women, the average serum level of HCB increased by age, duration of residence in Flix, and by consumption of local fish (Table 3). In men, levels were highest in the middle age group, and levels increased with (a) duration of residence, (b) proximity to the electrochemical factory, (c) local fish consumption, and (d) factory occupation. Place of birth and pesticide use were not associated with HCB levels. The association between HCB level and duration of residence disappeared after we adjusted for age, and the association between HCB level and proximity to the factory disappeared after we adjusted for occupation (given the tendency of workers to live close to the factory). Duration of employment was not associated with HCB level. The adjusted association between occupation and HCB levels was very strong (adjusted change in past and current workers = 9.3 ng/ml and 35.7 ng/ml, respectively; p [is less than] .01 for both). With the exception of HCB, we did not observe any statistically significant differences between occupation and serum levels of organochlorine compounds. The same pattern was observed in both randomized and nonrandomized populations.

Table 2.--5th, 25th, Median, 75th, and 95th Percentiles of Organochlorine Levels (ng/ml) in Serum of the Population Studied in Flix, Spain (1994; n = 608)

Notes: HCB = hexachlorobenzene, [Beta]-HCH = [Beta]-hexachlorocyclohexane, p,p'-DDE = 2,2-bis[p-chlorophenyl]-1,1-dichloroethylene, and PCBs = polychlorinated biphenyls.

Table 3.--Levels of HCB (ng/ml), by Exposure Variables and Sex, in Randomized and Nonrandomized Participants in Flix, Spain (1994; n = 608)

Note: GM = geometric mean, and CI = confidence interval.

Prevalence of fair/poor perceived health and number of self-reported chronic conditions were higher among females younger than 45 y in the study population than in the reference population (Table 4). Young males also presented a higher prevalence of chronic conditions than older males. In contrast, the average number of chronic conditions in the older groups was lower in the study population than in the reference population. No subjects reported a history of PCT, although 6 persons (1.6% and 0.6% of males and women, respectively) had total urinary porphyrin levels that exceeded reference values. Hypothyroidism and goiter occurred more frequently among females (1.8% and 2.6%, respectively) than males (0.1% and 0.3%, respectively). The frequencies of Parkinson's disease were 0.4% and 0.5% among females and men, respectively. Creatinine levels that exceeded the upper reference limit occurred more frequently in males than females (2.8% versus 0.6%, respectively). A benign or malignant neoplasm was reported by 13% of the females and 4.5% of the men. Declared spontaneous abortion, low birth weight, and congenital malformations were reported by 1 5%, 2.9%, and 1.6%, respectively, of females who were younger than 65 y of age.

Table 4.--Perceived Health and Mean Numbers of Self-Reported Chronic Conditions in the Study Population of Flix (n = 1 800) and in the Reference Population of Catalonia (n = 4 743)

Note: CI = confidence interval.

(*) p < .05, contrast for the difference between Flix and the reference populations.

Men who had ever worked in the electrochemical factory had a higher prevalence of each of the potentially HCB-related diseases than nonworkers (Table 5). Total urinary porphyrin and creatinine above the upper reference limit were also higher among workers, although the risk was not statistically significant. No increased risks were observed among women. Given the small numbers, we performed a multivariate analysis-including all exposure variables related to HCB exposure, tobacco and alcohol consumption, gender, and age--only for perceived health (OR associated with occupation = 1.1; 95% CI = 0.8, 1.5), mean number of reported chronic health conditions (OR = 1.1; 95% CI = 0.7, 1.4), self-reported cancer prevalence (OR = 1.4; 95% CI = 0.8, 2.2), and confirmed cancer (OR = 1.9; 95% CI = 0.5, 7.6). To control for possible problems resulting from "no response," we conducted a sensitivity analysis in a random population sample. The adjusted OR associated with occupation in the electrochemical factory for validated neoplasms was 2.5 (95% CI = 0.2, 12.9), but perceived health did not change.

Table 5.--Association between Occupation (Ever Versus Never Worked in the Electrochemical Factory) and Perceived Health, Self-Reported Chronic Conditions, and Hexachlorobenzene (HCB)-Related Diseases and Biochemical Markers, by Sex, in Flix, Spain (1994; n = 1 800)

Note: TSH = thyroid-stimulating hormone, and CI = confidence interval.

(*) Number of ever/never workers in the electrochemical factory.

([dagger]) ORa -- odds ratio adjusted by age.

([double dagger]) Numbers of ever/never workers in the electrochemical factory were 60/936.

([sections]) Numbers of ever/never workers in the electrochemical factory were 46/719.

([parallel]) Numbers of ever/never workers in the electrochemical factory were 169 and 84 males and 18 and 333 females, respectively.

(#) Numbers of ever/never workers in the electrochemical factory were 167 and 83 males and 22 and 336 females, respectively.

We found no statistically significant differences in HCB serum levels between Females who reported any of the potentially HCB-related diseases and those who did not. females who reported neoplasms had an age-adjusted geometric mean (GM) of HCB of 14.8 ng/ml (95% Cl = 12.2, 1 6.8), compared with females without any of the potentially HCB-related diseases (GM = 14.1 ng/ml; 95% Cl = 13.7, 14.5). In men, only subjects who had high levels of total porphyrin in a 24-h period presented higher levels of HCB (GM = 63.3 rig/mi; 95% Cl: 35.5, 131.3) than males without any of the poten- HCB-related diseases (GM = 19.8 ng/ml; 95% CI = 18.1, 28.6).

Discussion

The serum HCB levels in our general population sample are the highest ever reported--being between 50 and 100 times higher than levels reported in other populations.[2,3,19] In contrast, concentrations of [Beta]-HCH, p,p'-DDE, and PCBs were in the range reported in other studies.[20-22] Perceived health and prevalence of chronic health conditions in this population did not differ from other rural populations in Catalonia. However, males who had ever worked in the electrochemical factory showed a significant increase of a priori pathologies that were related to HCB, compared with inhabitants who never had worked at the factory. Male workers had the highest levels of HCB, but not of the other organochlorine compounds studied.

The lack of specific effects in the general population suggests that the level of exposure to HCB--although reportedly being the highest levels resulting from environmental pollution--was insufficient to cause an observable increase of the endpoints in this cross-sectional study. The prevalence of reported goiter, hypothyroidism, Parkinson's disease, and reproductive outcomes, as well as serum levels of THS, creatinine, and total urinary porphyrines, were in the range observed in other populations who were not characterized by any specific high-exposure situation.[23-26] Perhaps the levels attained by the general population were lower than a toxic level. Workers seemed to have reached such a level.

Chronic exposure to HCB among workers may have been associated with specific health outcomes, although results were not statistically significant. Several observations support this statement. The prevalence of each of the potentially HCB-related diseases, as well as levels of biomarkers (e.g., creatinine, uroporphyrines), were higher in workers than in nonworkers--a finding that confirms specificity to the results. The strength of the associations for cancer increased when we restricted the analysis to the randomized population and when a more valid outcome (e.g., the confirmed diagnostics) was used. The association was observed in males only, and this finding is consistent with the fact that males were the most highly exposed inhabitants because they were employed at the factory. The temporal pattern was probably in the appropriate causative order inasmuch as the production of HCB as a byproduct was initiated decades ago.

The main limitation of the present study was the cross-sectional observation of the outcome and the exposure, the small size of the population under study, and the low response rate. First, investigators find it difficult to evaluate the effects of health-related migration, which may be associated with adaptation and differential nonresponse. Second, the biomarker of current exposure might not adequately represent levels in the past. Migration out of the village (between the census of 1991 and 1996 the population size decreased by 1 3%) or temporal migration at the time of the study (i.e., some of the inhabitants spent summers in a second residence off the coast) was the main source of nonresponse. If migration was related to health, it is likely that healthy people stayed in the town inasmuch as most of the migration resulted from early retirement. This bias would likely produce an attenuation of any association. No differences were observed between random and nonrandom participants with respect to their perceived health or prevalence of self-reported diseases, and no pattern suggesting overreporting of chronic health conditions was observed among the study population, compared with the reference population from the Catalan National Health Survey. Moreover, participants were unaware of which health outcomes were under study, and they were unaware of their organochlorine compound levels. No differences in HCB levels were observed between the random and nonrandom population samples. In addition, the response rate among current workers (41.8% of total current workers) was the same as that of the total study population; again no differences were observed between workers and the rest of participants with respect to prevalence of self-reported chronic conditions. Finally, the odds ratio for fair/poor perceived health among workers was lower than the specific odds ratio for most of the HCB-related diseases. Overall, these arguments suggested that both the lack of effects in the general population and the higher frequency of potentially HCB-related diseases among workers were not the result of a selection bias.

The association at the individual level between health endpoints and occupation was more consistent and stronger than with the biomarker of exposure (i.e., serum levels of HCB). Occupational history is likely to better document the historical exposure to HCB than current internal dose levels. Chlorinated solvents that generate HCB as a byproduct (e.g., trichloroethylene) have been produced in large quantities since 1960.[16] The estimated half-life for most of the organochlorine compounds ranges from 2 to 15 y.[27] The use of "ever employment in the factory" as an exposure variable, moreover, allowed us to include the entire study population in the analysis and to minimize the biases of disease-related changes in exposure. Our use of a biomarker of exposure, however, has been crucial to our understanding the average group levels according to different exposure characteristics, as well as to improve our knowledge of the kinetics of HCB.[28]

Confounding by other compounds released from the same factory in the past may be another limitation of this study. This factory produced chloroform, carbon tetrachloride, dichloro-dyphenyl-trichloroethane (DDT) PCBs, and other chlorinated solvents, but no information was available for levels of exposure. Some investigators have reported that these chlorinated compounds are related--albeit weakly--with health effects (e.g., soft-tissue sarcoma,[29,30] thyroid cancer,[31,32] Parkinson's disease[33,34]) similar to those associated with HCB exposure. The fact that the common byproduct of the production of all pesticides and chlorinated solvents was HCB, and current serum levels of DDE, PCBs, and [Beta]-HCH were similar to levels in other geographic areas, suggested that HCB had always been the major pollutant in this population. The total dioxin equivalents in the area under study fulfill the ambient air-quality standard for "no significant health effect."[16] A common limitation, however, in environmental observational studies is the inability of investigators to separate the individual effects of each pollutant from a given mixture of pollutants.

A short latency period for some of the pathologies under study--mainly cancer--could be an other limitation of the study for that part of the population that migrated to Flix in relatively late periods. Our results on cancer risk among workers, however, are in agreement with findings of our preliminary study of cancer incidence in this population.[16] The follow up of this population will help us to evaluate diseases with very long latency.

This is the first epidemiological study of the health effects of HCB since the outbreak in Turkey in the 1950s, at which time HCB dose was not measured. The population of Flix has the highest levels of HCB ever recorded. The results suggest that this high exposure to HCB does not affect the health status of the general population, although it could have some specific effects on the health of workers in the factory who are the most highly exposed subjects. The study did not allow us to control for other possible confounding exposures. The follow up of the studied population will contribute new information for us to evaluate the effects of HCB. These results point to the necessity of undertaking a multicenter study in workers exposed to HCB.

This study was partially financed by a grant from the Fondo de Investigacion Sanitaria, FIS 93/0006-01, and by a grant from the Departament de Sanitat i Seguretat Social, Generalitat de Catalunya.

Submitted for publication December 16, 1997; revised; accepted for publication September 24, 1998.

Requests for reprints should be sent to Dr. Maria Sala, Institut Municipal d'Investigacio Medica, c/Doctor Aiguader, 80, 08003 Barcelona, Spain.

References

[1.] Jacoff F, Scarberry R, Rosa D. Source assessment of hexachlorobenzene from the organic chemical manufacturing industry. In: Morris CR, Cabral JRP (Eds). Hexachlorobenzene: Procedings of an International Symposium. Lyon, France: International Agency for Research on Cancer, IARC Scientific Publications no 77, 1996; pp 31-37.

[2.] Needham LL, Burse VW, Head SL, et al. Adipose tissue/serm partioning of chlorinated hydrocarbon pesticides in humans. Chemosphere 1990; 20:975-80.

[3.] Krauthacker B. Levels of organochlorine pesticides and polychlorinated biphenyls (PCBs) in human milk and serum collected from lacting mothers in the northern Adriatic area of Yugoslavia. Bull Environ Contam Toxicol 1991; 46:797-802.

[4.] Rogan WJ, Gladen BC. PCBs, DDE, and child development at 18 and 24 months. Ann Epidemiol 1991; 1:407-13.

[5.] Furst P, Furst C, Wilmers K. Human milk as a bioindicator for body burden of PCDDs, PCDFs, organochlorine pesticides, and PCBs. Environ Health Perspect 1994; 102(Suppl 1):187-93.

[6.] Cabral JRP, Shubik P, Mollner T, et al. Carcinogenic activity of hexaclorobenzene in hamsters. Nature 1977; 269:510-11.

[7.] Cabral JRP, Mollner T, Raitano F, et al. Carcinogenesis of hexachlorobenzene in mice. Int J Cancer 1979; 23:47-51.

[8.] Foster WG, Pentick JA, McMahon A, et al. Body distribution and endocrine Toxicity of hexachlorobenzene (HCB) in the female rat. J Appl Toxicol 1993; 13:79-83.

[9.] Jarrell J, McMahon A, Villeneuve D, et al. Hexachlorobenzene toxicity in the monkey primordial germ cell without induced porphyria. Reprod Toxicol 1993; 7:41-47.

[10.] van Raaij JAGM, Frijters CMG, van den Berg JK. Hexachlorobenzene-induced hypothyroidism. Involvement of different mechanisms by parent compound and metabolite. Biochem Pharmacol 1993; 46:1385-91.

[11.] Smith A, Dinsdale D, Cabral J, et al. Goiter and wasting induced in hamsters by hexachlorobenzene. Toxicology 1987; 60: 343-49.

[12.] van Raaij JA, Frijters CM, van den Berg KJ. Hexachlorobenzene-induced hypothyroidism. Involvement of different mechanisms by parent compound and metabolite. Biochem Pharmacol 1993; 46:1385-91.

[13.] Schmid R. Cutaneous porphyria in Turkey. New Engl J Med 1960; 263:397-98.

[14.] Cam C, Nigogosyan G. Acquired toxic porphyria cutanea tarda due to hexachlorobenzene. JAMA 1963; 183:88-91.

[15.] Peters H, Gocmen A, Cripps D, et al. Epidemiology of Hexachlorobenzene-Induced porphyria in Turkety. Arch Neurol 1982; 39:744-49.

[16.] Grimalt JO, Sunyer J, Moreno V, et al. Risk excess of soft-tissue sarcoma and thyroid cancer in a community exposed to airborne organochlorinated compound mixtures with a high hexachlorobenzene content. Int J Cancer 1994; 56:200-03.

[17.] Blake D, Poulos V, Rossi R. Diagnosis of porphyria. Recommended methods for peripheral laboratories. Clin Biochem Revs 1992; 13(suppl):2-13.

[18.] Otero R, Santiago-Silva M, Grimalt JO. Hexachlorobenzene in human blood serum. J Chromatog 1997; 778:87-94.

[19.] Jarrell JF, Villeneuve D, Franklin C, et al. Contamination of human ovarian follicular fluid and serum by chlorinated organic compounds in three Canadian cities. Can Med Assoc J 1993; 148:1321-27.

[20.] Wolff MS, Schecter A. Use of PCB blood levels to assess potential exposure following an electrical transformer explosion. J Occup Med 1992; 34:1079-83.

[21.] Schantz SL, Jacobson JL, Humphrey HE, et al. Determinants of polychlorinated biphenyls (PCBs) in the sera of mothers and children from Michigan farms with PCB-contaminated silos. Arch Environ Health 1994; 49:452-58.

[22.] Kocan A, Petrik J, Drobna B, et al. Levels of PCBs and some organochlorine pesticides in the human population of selected areas of the Slovak Republic. I. Blood. Chemosphere 1994; 29: 2315-25.

[23.] Okamura K, Nakashima T, Ueda K, et al. Thyroid disorders in the general population of Hisayama Japan, with special references to prevalence and sex differences. Int J Epidemiol 1987; 16:545-49.

[24.] Tison F, Dartigues JF, Dubes L, et al. Prevalence of Parkinson's disease in the elderly: a population study in Gironde, France. Acta Neurol Scand 1994; 90:111-15.

[25.] Nordentoft M, Lou HC, Hansen D, et al. Intrauterine growth retardation and premature delevery: the influence of maternal smoking and psychosocial factors. Am J Public Health 1996; 86: 347-54.

[26.] Correa A, Gray RH, Cohen R, et al. Ethylene glycol ethers and risk of spontaneous abortion and subfertility. Am J Epidemiol 1996; 143:707-17.

[27.] Woodruff T, Wolff MS, Davis DL, et al. Organochlorine exposure estimation in the study of cancer etiology. Environ Res 1994; 65:132-44.

[28.] To-Figueras J, Sala M, Otero R, et al. Metabolism of hexachlorobenzene in humans: association between serum levels and urinary metabolites in a highly exposed population. Environ Health Perspect 1997; 105:78-83.

[29.] Wingren G, Fredrikson M, Brage HN, et al. Soft tissue sarcoma and occupational exposures. Cancer 1990; 66:806-11.

[30.] Serraino D, Franceschi S, La Vecchia C, et al. Occupation and soft-tissue sarcoma in northeastern Italy. Cancer Causes Control 1992; 3:25-30.

[31.] Saracci R, Kogevinas M, Bertazzi PA, et al. Cancer mortality in workers exposed to chlorophenoxy herbicides and chlorophenols. lancet 199t; 338:1027-32.

[32.] Wingren G, Hatschek T, Axelson O. Determinants of papillary cancer of the thyroid. Am J Epidemiol 1993; 138:482-91.

[33.] Hubble JP, Cao T, Hassanein RES, et al. Risk factors for Parkinson's disease. Neurology 1993; 43:1693-97.

[34.] Gorrell JM, DiMonte D, Graham D. The role of the environment in Parkinson's Disease. Environ Health Perspect 1996; 104: 652-54.

COPYRIGHT 1999 Heldref Publications
COPYRIGHT 2000 Gale Group