Influence of Metabolic Genotypes on Biomarkers of Exposure to 1,3-Butadiene in Humans¹

BD³ is a chemical widely used for rubber, resin, and latex production, i.e., butadiene/styrene synthetic rubber. It belongs to the family of diolefin compounds, aliphatic hydrocarbons with conjugated double bonds that, because of their reactivity, are used extensively in the chemical industry. BD is a gas at room temperature. Exposure may occur by inhalation in chemical plants where the monomer is produced and/or polymerized. For prevention of risks arising from exposure to BD, a concentration of 2 ppm (4.4 mg/m³) as a maximal time-weighted average (TWA) during an 8-h shift is recommended by the American Conference of Governmental Hygienists (1). Concentrations of 15 ppm (34 mg/m³) and 5 ppm (11 mg/m³) are established as technical exposure limits (TRK) by the German Federal Ministry for Employment and Social affairs (AGS Committee for hazardous substances) for processing after polymerization or loading and other applications, respectively (2). In Sweden, the current permissible occupational exposure limit (8-h average) is 0.5 ppm (1 mg/m³).

Besides exposures in occupational settings, BD is a ubiquitous pollutant; major sources of exposure in the general population are cigarette smoke and automobile exhausts. Main stream and side-stream smoke of one cigarette have been found to contain 14–75 and 205–361 μg of BD, respectively, depending on the brand (3). A study performed in Sweden measured airborne BD levels in outdoor urban air in the range 0.5–5 μg/m³ attributable to traffic exhausts (4), whereas a recent study conducted in the United Kingdom showed BD concentrations in indoor and outdoor environments ranging from not detected to 10.8 μg/m³ (5).

In 1999, the IARC classified BD as a probable carcinogen to humans (group 2A; Ref. 6). In 2000, the ninth report on carcinogens, edited by the Department of Health and Human Services in the United States, classified BD as “known to be a human carcinogen” (7).

BD, once inhaled and absorbed through the respiratory tract, is oxidized by CYP enzymes (CYP2E1 and CYP2A6) to EB (8, 9). CYP enzymes (CYP2E1 and CYP3A4) also mediate the further oxidation of EB to 1,2:3,4-diepoxybutane (10). EB may be hydrolyzed to 1,2-dihydroxy-3-butene and further oxidized to 1,2-dihydroxy-3,4-epoxybutane. Epoxides form covalent bonds with proteins and DNA (11, 12, 13, 14, 15, 16) and are genotoxic, both in vitro and in vivo, causing point mutations, CAs, and SCEs (reviewed in Ref. 17). Detoxification of epoxides may occur via hydrolysis and/or conjugation with glutathione. These reactions may be spontaneous or enzyme-mediated by mEH and GST (8, 18). The glutathione conjugates, after hydrolysis to cysteine derivatives, are acetylated and excreted in urine as mercapturic acids (19, 20, 21). Data in mice suggest that ADH may also be involved in the mercapturic acid formation pathway (22).

Among experimental animals, different species show large differences in susceptibility to BD-induced cancer. Mice are much more responsive than rats to lower doses and acquire larger numbers and more types of tumors (reviewed in Ref. 23). The higher susceptibility of mice has been related to increased formation of genotoxic species, i.e., the epoxides (EB, 1,2:3,4-diepoxybutane, and 1,2-dihydroxy-3,4-epoxybutane), after experimental exposure to BD (24, 25, 26). The activities of enzymes involved in BD metabolism are claimed to be responsible for such differences; in fact, mice seem to be more prone to formation of epoxides and less efficient in their detoxification.

In humans, the same metabolic pathways as those described in rodents have been partially confirmed. In workers exposed to BD during monomer production and polymerization, the presence of the urinary mercapturic acids MI and MII have been detected (21, 27, 28, 29, 30, 31, 32). The presence of reactive metabolites of BD has also been indirectly confirmed through measurements of hemoglobin adducts, i.e., hemoglobin N-(2-hydroxy-3-butenyl)-valine and N-(1-hydroxy-3-butenyl)-valine adducts and THBVal, in previous studies (12, 13, 20, 27, 28, 29, 33, 34, 35) as well as in the present investigation (14).

Overall, data regarding metabolism of BD in humans are still scarce, and many issues have not yet been investigated thoroughly. The influence of genetic polymorphisms in BD-metabolizing enzymes, i.e., cytochrome P450s, mEH, GSTs, and ADH is one of these issues. The activity of these enzymes may be responsible for individual differences in metabolic activation and detoxification reactions and may ultimately modulate the effects of BD exposure (36).

Until today, some investigators have studied the influence of genetic polymorphisms of GSTT1 and GSTM1 on cytogenetic damage (SCEs, CAs, and MN) induced in vitro in human lymphocytes exposed to BD epoxides (37, 38, 39, 40, 41, 42, 43). The results of these studies indicate a role of the polymorphisms in these genes on genotoxic endpoints. In humans occupationally exposed to BD, the evidence of an influence of genetic polymorphisms on biomarkers of BD is controversial, with some studies indicating higher levels of genetic damage in GSTT1 and/or GSTM1 null subjects (27, 32, 44, 45, 46, 47), whereas in other studies no influence has been found (13, 28).

The aim of the present study was to evaluate the influence of genetic polymorphisms of enzymes involved in BD metabolism in humans, i.e., CYP2E1, mEH, GSTM1, GSTT1, GSTP1, and ADH3, on the excretion of urinary MI and the level of THBVal in a cohort of workers exposed to low levels of BD during production of BD monomer and polymers and in a group of unexposed subjects.

The research was carried out in a technologically advanced production plant located in Italy, where BD is produced by extractive distillation and rectification of the oil C4 fraction and polymerized by addition to other monomers (styrene, acrylonitrile, and isoprene). Because the production processes were enclosed, general environmental emissions were limited.

Forty healthy Italian male workers, all Caucasians, were enrolled in the study. Thirty subjects were occupationally exposed to BD in three processes: BD monomer production (n = 10), BD copolymerization to produce cis-polybutadiene (n = 10), and styrene-BD polymerization to produce rubber (n = 10). The workers were involved in routine surveillance work, mainly performed via computer from the control room located in proximity of each production facility. Among the BD-exposed workers, 13 were smokers.

For comparison, 10 male subjects without occupational exposure to BD were recruited among clerks working in the administrative department of the plant. There were no smokers among these subjects. The studied groups are presented in Table 1.

At the beginning of the study, workers received information about the aim of the research, and written informed consent, according to the Helsinki declaration and later amendments, was obtained. For BD-exposed workers, air samples were collected during the work shift (6 a.m. to 2 p.m.), by personal samplers equipped with charcoal cartridges positioned in the respiratory zone. At the end of the 8-h work shift, the subjects were interviewed by an occupational health physician, who submitted a questionnaire to gather information about lifestyle, smoking habits, medical history, and occupational activity. From each subject, venous blood samples and one spot urine sample were collected. All samples were coded, frozen at −20°C, and delivered to the various laboratories, where analyses were performed without knowledge of their origin.

Personal exposure to airborne BD, as time-weighted average, was assessed only for BD-exposed workers. Collection of samples and analysis were performed essentially according to the National Institute of Occupational Safety and Health 1024 analytical method (48). Briefly, BD was adsorbed on activated charcoal contained in a glass cartridge (400 mg in the front section and 200 mg in the back section), connected to a personal pump. The pump was set at 50 ml/min, so that an air volume of ∼24 liters was pumped during the shift. For the analysis, BD was desorbed from charcoal by carbon disulfide and analyzed by GC/MS in the EI single ion monitoring mode. The limit of detection for BD was estimated as 1 μg/m³.

For the determination of urinary MI, a GC/MS-EI procedure was developed, based mainly on a method published previously (49). MI was synthesized as described by Sabourin et al. (49). The quantification was performed in the presence of the deuterated analogue, N-(d₃-acetyl)-S-(3,4-hydroxybutyl)-l-cysteine prepared from N-(d₃-acetyl)-l-cysteine, as internal standard. Briefly, urine samples were acidified with trifluoroacetic acid to pH 2 and spiked with a known amount of internal standard. Analytes were purified on C₁₈ reverse phase silica cartridges, derivatized with bis(trimethylsilyl)trifluoroacetamide and analyzed by GC/MS-EI. Quantification was based on the ratio between the chromatographic peak area of the analyte and the relative internal standard, registering single ions: m/z 452 [MI molecular ion − 15]⁺, qualifier, and 308, MI quantifier, and m/z 455 [internal standard molecular ion − 15]⁺, qualifier, and 311, internal standard quantifier. The estimated detection limit was 0.15 mg/g creatinine. Reproducibility of the assay within the series was in the range 3–6%.

The isolation of globin was carried out according to Bader et al. (50). The determination of THBVal was performed according to a procedure described previously (35). Briefly, a modified Edman degradation was used to cleave THBVal from hemoglobin and produce the pentafluorophenyl thiohydantoin derivative. After addition of ¹³C₅-THBVal-pentafluorophenyl thiohydantoin as internal standard, the mixture was purified on C₁₈ reverse phase silica cartridges. The hydroxyl groups were acetylated with acetic anhydride, and the derivatized mixture was analyzed by high resolution GC/MS-NCI. Quantification was based on the ratio between the chromatographic peak areas of the analyte and the internal standard registering the fragments m/z 534.1084 and m/z 539.1254, respectively. The limit of detection of the procedure was 10–15 pmol/g globin (14).

DNA was prepared from WBCs. After lysis and proteinase digestion, the samples were subjected to a modified salting out procedure (51). DNA was then isolated after precipitation with ethanol.

For CYP2E1, the RsaI polymorphism (c1/c2) in the 5′-flanking region (*5) of the gene was analyzed by PCR/RFLP as described previously (52). The G₋₃₅T (5′-flanking region) polymorphism was determined by PCR/RFLP as described by Fairbrother et al. (53). The insertion (96 bp) polymorphism in a repeat region of the promoter of CYP2E1 was determined by PCR, using 5′-TGG TAC ATT GTG AGA CAG TG-3′ as the forward primer and 5′-ATA CGG GAA CAC CTC GTT TG-3′ as the reverse primer (54), yielding fragments of 633 bp (six repeats; CYP2E1*1C) and 729 bp (eight repeats; CYP2E1*1D; Ref. 55).

The deletion polymorphisms of glutathione transferases GSTM1 and GSTT1 were determined by PCR according to methods described previously (56, 57, 58). β-Actin (primers from Stratagene) was used as positive control to verify the presence of amplifiable DNA. The PCR method used for the detection of individuals lacking the GSTM1 and GSTT1 genes (the null genotypes) do not differentiate between the heterozygous and homozygous carriers of the functional gene.

Analyses of the GSTP1 polymorphisms resulting in an Ile to Val substitution at residue 104 in exon 5 and an Ala to Val substitution at residue 113 in exon 6, were performed as described previously (59).

For epoxide hydrolase (mEH), the Tyr/His polymorphism in exon 3 (amino acid 113) was analyzed by allele-specific PCR as described previously (60). The His/Arg polymorphism in exon 4 (amino acid 139) was determined by PCR/RFLP as described by Hassett et al. (60). On the basis of current knowledge (60) of the in vitro functional expression of the variant alleles at residue 113 (exon 3, His₁₁₃ slow allele) and at residue 139 (exon 4, Arg₁₃₉ rapid allele), three predicted mEH enzymatic activity levels were assigned. Low activity individuals were homozygous for His₁₁₃ and homozygous for His₁₃₉ (HH + HH), homozygous for His₁₁₃ and heterozygous for His₁₃₉ (HH + HR), or heterozygous for His₁₁₃ and homozygous for His₁₃₉ (YH + HH). High-activity individuals were homozygous for Tyr₁₁₃ and homozygous for Arg₁₃₉ (YY + RR), homozygous for Tyr₁₁₃ and heterozygous for Arg₁₃₉ (YY + HR), or heterozygous for Tyr₁₁₃ and homozygous for Arg₁₃₉ (YH + RR). The remaining genotypes were classified as having intermediate mEH activity.

The genetic polymorphism in exon 8 of the ADH γ subunit (ADH3) was determined by PCR/RFLP. Primers 321 and YC351 (61) were used to yield a 145-bp fragment, which was subsequently cleaved by the restriction enzyme SspI. ADH3*1 contains A (Ile) and ADH3*2 contains G (Val) at the polymorphic position. The γ1 subunit (ADH3*1) has a higher catalytic activity (V_max) than the γ2 subunit (ADH3*2; Ref. 62).

The statistical analyses were carried out using SPSS for Windows statistical package. Nonparametric procedures were used to compare groups [Mann-Whitney t test (to compare two subgroups); Kruskas-Wallis H test (to compare three subgroups); and χ² test]. To test the correlation between variables, multiple linear regression and multifactor ANOVA were performed. The variables were logarithmically transformed and used after having verified that transformed variables were normally distributed (Kolmogorov-Smirnov and χ² test). Adduct levels were correlated to “metabolic score” (the number of polymorphic genotypes that individually increases adduct levels) using nonparametric Spearman correlation. P = 0.05 was considered significant.

Selected characteristics of BD-exposed and unexposed workers are reported in Table 1. The two groups differ mainly in age, with those exposed being younger than unexposed subjects, and in smoking habits, with 13 smokers in the exposed group and none among unexposed.

In Table 2, data on personal exposure to airborne BD during the shift, as time-weighted average (only for BD-exposed subjects), excretion of MI in a spot urine sample collected at the end of the shift, and the levels of THBVal in the investigated subjects are reported. Subjects were first divided by exposure into BD exposed (n = 30) and unexposed workers (n = 10); slightly higher levels of MI and THBVal were observed in BD-exposed workers compared with unexposed subjects, but the differences were not statistically significant. When subjects were divided according to smoking habits into smokers (n = 13) and nonsmokers (n = 27), it was observed that cigarette smoke positively influenced the formation of adducts (P = 0.027). Further subdivision of subjects according to their exposure and smoking status led to three groups: unexposed subjects (all nonsmoking, n = 10), nonsmoking BD-exposed workers (n = 17), and smoking BD-exposed workers (n = 13). Although no differences were found either in the BD exposure or in the level of urinary MI among the three groups, a significant or borderline difference was observed in the levels of THBVal comparing smoking BD-exposed workers with nonsmoking unexposed (P = 0.049) or with nonsmoking BD-exposed workers (P = 0.065). Different levels of THBVal were also detected in workers divided by type of processes: BD monomer production workers (n = 10) showed higher THBVal levels (median, 44.7; minimum to maximum, 30.3–61.3 pmol/g globin) in comparison with styrene-BD polymerization workers (n = 10; THBVal, 40.9; minimum to maximum, 17.7–48.4 pmol/g globin) and BD copolymerization workers (n = 10; THBVal, 32.9; minimum to maximum, 24.9–43.9 pmol/g globin; P = 0.028) as discussed previously by Begemann et al. (14).

Among all subjects (n = 40) and among the BD-exposed workers only (n = 30), levels of urinary MI and THBVal were significantly correlated (r = 0.333, P = 0.035 and r = 0.45, P = 0.013, respectively). The highest correlation between MI and THBVal was found among the 13 workers that were also current smokers (r = 0.621, P = 0.024). Among BD-exposed workers (n = 30), correlations of MI and THBVal levels to BD exposure showed comparable but statistically nonsignificant correlation (r = 0.345, P = 0.062 and r = 0.338, P = 0.068, respectively).

No differences in the frequencies of genotypes and predicted phenotypes between BD-exposed and unexposed workers were observed with the exception of ADH3 (P = 0.046; χ² test was used for comparison). The observed genotypes are in agreement with allele frequencies reported previously for other European populations (63).

Because levels of THBVal and MI in urine were not significantly different between exposed and unexposed subjects, the total study group (n = 40) was used for analyses of the effects of genotypes on the internal dose biomarkers. The results, reported in Table 3, showed higher median levels of THBVal in subjects lacking GSTM1 (P = 0.017) or GSTT1 (P = 0.005). Moreover, an influence of the G₋₃₅T polymorphism in the 5′-flanking region of CYP2E1 on THBVal levels was observed (P = 0.073 when GG versus GT + TT subgroups were compared). No significant influence of genetic polymorphisms on urinary MI levels was found, although a borderline influence was noticed for GSTP1 (A¹¹³V; P = 0.066).

A combined polymorphism analysis was performed by grouping the investigated subjects based on the combination of polymorphic genotypes that had an influence on the adduct levels, i.e., CYP2E1(G₋₃₅T), GSTT1, and GSTM1. The results, reported in Table 4, showed a statistically significant difference in THBVal levels (P = 0.011) among the groups. The highest adduct levels were observed in subjects with the combination GG in CYP2E1(G₋₃₅T), GSTM1 null, and GSTT1 null, whereas the lowest adduct levels were observed in subjects with the combination TT or GT in CYP2E1(G₋₃₅T), GSTM1 positive, and GSTT1 positive. A further approach to illustrate the effects of combined genotypes was also attempted by calculating, in each subjects, a metabolic score, as the number of polymorphic genotypes that individually increase adduct levels, i.e., GG in CYP2E1(G₋₃₅T), GSTM1 null, and GSTT1 null. In Fig. 1, the adduct level distribution in groups with different metabolic scores is shown. A statistical comparison, performed with the Kruskal-Wallis H test, showed significant differences in the levels of adducts among the groups (P = 0.002). A positive correlation was found between the metabolic score and THBVal levels (Spearman ρ = 0.580, P = 0.0001).

Considering all parameters influencing adducts, i.e., smoking, CYP2E1(G₋₃₅T), GSTM1, and GSTT1 polymorphisms, a model for multiple linear regression analysis was set (Y = α + β₁X₁ + β₂X₂ + β₃X₃ + β₄X₄). The parameters given above were considered as independent variables (X_i) and THBVal as the dependent variable (Y). In Table 5, the result of the statistical analysis is reported. The model was highly significant (P < 0.001) and explains ∼50% of the observed variability (R² = 0.487). Each of the independent variables contributed significantly (P < 0.05) to the level of THBVal. Further attempts to enlarge the model with other variables, i.e., BD exposure or other polymorphic genotypes did not improve the model.

The study investigated the role of genetic polymorphisms on biomarkers of exposure to BD in a cohort of workers used in BD production and polymerization and in a group of unexposed subjects. The personal exposure to airborne BD in workers ranged from 4 to 201 μg/m³. These BD levels are considerably lower than the current occupational exposure limit values recommended or established by the different regulatory agencies (ranging from 1 to 34 mg/m³; Refs. 1, 2) and are comparable with those characterizing indoor and outdoor contamination attributable to traffic emissions and cigarette smoke. The dose of BD inhaled during a work shift from the subjects under study (4–201 μg/m³), assuming an inhalation rate of 1 m³/h, ranged from 32 to 1600 μg. At the lower end of this range doses are comparable with those assumed for nonsmoking, non-occupationally exposed subjects that inhale BD as a contaminant of urban air (BD ranging from 0.4 to 12 μg/m³), whereas at the upper end, doses are comparable with those inhaled by subjects smoking a moderate number of cigarettes/day; in fact, BD amounts as high as 75 and 361 μg/cigarette have been reported in main stream and side-stream smoke, respectively (3).

Although the external BD exposures were low, it was possible to detect the urinary mercapturic acid MI and the hemoglobin adducts THBVal of BD in all of the samples analyzed, both from BD exposed and unexposed subjects. No differences were found in the levels of these biomarkers of BD exposure between the two groups. The lack of differences in the biomarkers is probably attributable to the low exposure in BD workers as well as to the presence of significant amounts of MI and THBVal in subjects without occupational exposure to BD (Table 2). A further division of workers by type of production (BD monomer production, BD copolymerization to produce cis-polybutadiene, and styrene-BD polymerization to produce rubber) showed some exposure-related differences in THBVal among subgroups as discussed previously by Begemann et al. (14) and summarized above (see “Results”).

The levels of MI and THBVal found in the present investigation were compared with data from previous studies (21, 27, 28, 29, 30, 31, 32). It is noteworthy that all of the studies reported background levels of MI and THBVal in unexposed subjects. The sources of BD exposure in the general population (i.e., traffic emissions and cigarette smoking) are not high enough to generate these high background levels of MI and THBVal. Because 3-butene-1,2-diol is a possible precursor of both MI (via conjugation with glutathione) and THBVal (after oxidation to BDE), a dietary and/or endogenous source of this molecule may be responsible for the formation of background levels of MI and THBVal (29). 3-Butene-1,2-diol is a small and hydrophilic molecule, and an endogenous formation from, for example, catabolism of carbohydrates, provides a realistic hypothesis. Further investigations to understand the origin of background levels of MI, as well as THBVal, are needed. The high background levels and the associated variability limit the specificity of these biomarkers and therefore their usefulness, at least at low exposures of BD. The background levels of MI in the present study, conducted on Italian subjects, were higher than reported previously in the United States, in Czechs, and in Chinese subjects (21, 27, 28, 29, 30, 31, 32). For THBVal, the levels measured in the present investigation were similar to those observed in United States and China controls but were lower than those of Czech controls (27, 28, 29, 35). At the moment, the reason for this in unclear, but both differences in diet and differences in the analytical procedures used may be of importance.

Although when comparing exposed and unexposed subjects no statistically significant differences were found in internal dose biomarkers, significant correlations were found between urinary MI and THBVal levels. The observed correlations between a short-term exposure index (MI) and an index that accumulates over 3–4 months (THBVal), generally not expected, may suggest a nearly constant BD exposure over time in the investigated cohort.

The significant influence of GSTT1 and GSTM1 genotypes on THBVal formation, with higher levels measured in subjects with null genotypes, is in accordance with the role of GST in the detoxification of BD metabolites leading to adducts. The fact that GSTT1 is expressed in a high amount in the erythrocytes, the presumed site of hemoglobin adducts formation, may have contributed to the strong effect of the GSTT1 null genotype on the THBVal levels. An influence on THBVal levels was also suggested for a polymorphism of CYP2E1, with lower levels (P = 0.073) in subjects having at least one allele carrying T in position −35 in the 5′-flanking region. This polymorphism was first reported by Fairbrother et al. (54), and in vitro expression experiments indicated that the variant allele (T; CYP2E1*7B) might be associated with increased transcriptional activity of the CYP2E1 gene (54). Apparently, the findings in the present study are contradictory to what would be expected based on the in vitro experiments.

The only polymorphism that showed a suggested influence on the urinary excretion of MI in the present study was GSTP1(A¹¹³V). The median MI level in five heterozygous individuals was 73% higher than in the remaining subjects. Because the difference was only borderline significant (P = 0.066) and a functional consequence of this polymorphism has not been established clearly, this observation is hard to evaluate and may be a finding of no biological significance.

Many chemicals such as BD are metabolized in several steps, and the metabolic capacity of an individual is dependent on the sum of all of these steps. Combinations of polymorphisms, rather than a single polymorphism, are therefore expected to more clearly influence the metabolism and thereby also the biomarkers of internal dose. In the present study, it was found that different combinations of genotypes CYP2E1(G₋₃₅T), GSTM1, and GSTT1 were associated with different adduct levels and that the levels of THBVal correlated with the metabolic score, i.e., the number of polymorphic genotypes that individually increases the adduct levels (see Table 4 and Fig. 1).

In Table 6, the studies that evaluated the influence of genetic polymorphisms on biomarkers of BD exposure and/or effect in humans are summarized. The former studies suggest that polymorphisms in the genes for GSTT1 or GSTM1 may influence the level of SCE and/or CA (32, 44, 45) in BD-exposed subjects, although some observations are contradictory (45). Analogously, in a recent study, the polymorphism in GSTM1 was found to influence the level of BD-induced DNA adducts (46). Most of these findings are in agreement with results of in vitro studies showing increased SCEs, CAs, and MN in human lymphocytes from donors with GSTT1 and GSTM1 null genotypes after treatment with BD epoxides (37, 38, 39, 40, 41, 42, 43). On the other hand, no effects of GSTM1 and GSTT1 on a wide panel of biomarkers, including THBVal, were found in Chinese workers exposed to high levels of BD during polymerization (28). In a study on Czech workers in which several genetic polymorphisms were investigated, a weak influence of GSTM1 on the excretion of urinary metabolites was noted, whereas no effects were observed on the other investigated biomarkers, including THBVal (27). Finally, an influence of mEH on HPRT somatic mutations in workers exposed to BD during polymerization has been reported (47).

To this varied picture the results of the present investigation add further support for a role of GSTM1 and GSTT1 polymorphisms in the metabolism of BD and suggest that also the polymorphism in CYP2E1(G₋₃₅T) may be of importance for the formation of reactive species suspected to play a role in the carcinogenicity of BD. In fact, for the first time, the influence of genetic polymorphism on the in vivo formation of BD hemoglobin adducts in humans was observed. Moreover, our results indicate that the analysis of combined polymorphisms may be the key to a better understanding of the role played by genetic polymorphisms in enzymes involved in the metabolism of BD.

We thank Dr. Alfonso Gelormini, the staff at the Occupational Health Department of Enichem, and the management of the plant for the wide collaboration given. We are indebted to the investigated subjects who volunteered in the study.