AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 58:764–772 (2015)

Biomarkers for Polycyclic Aromatic Hydrocarbons and Serum Liver Enzymes Young-Sun Min, MD, PhD,1 Hyun-Sul Lim,

MD, PhD,

1

and Heon Kim, MD, PhD2

Background Limited evidence suggests that human liver toxicity is associated with exposure to polycyclic aromatic hydrocarbons (PAHs). Methods The association of urinary PAH metabolites with serum liver enzymes was tested among 288 workers at a petrochemical plant, using a general linear model (GLM) and multiple logistic regression. Results Urine 2-naphthol levels were positively correlated with serum AST after adjustment for covariates in GLM. Comparing third tertile versus first tertile of 2-naphthol levels, the odds ratios (OR) were elevated for abnormal serum AST levels [OR ¼ 4.1 (95%CI 1.6–10.2)] and abnormal serum ALT levels [OR ¼ 2.4 (95%CI 1.2–4.9)]. Conclusions Although confounding by alcohol intake was not completely ruled out, our findings demonstrate an association between PAHs exposure and elevation in serum liver enzymes. Urinary 2-naphthol is a biomarker of exposure to PAHs that is associated with liver toxicity. Am. J. Ind. Med. 58:764–772, 2015. ß 2015 Wiley Periodicals, Inc. KEY WORDS: polycyclic aromatic hydrocarbon; PAHs; 2-naphthol; toxicity; liver; biomarker

INTRODUCTION Polycyclic aromatic hydrocarbons (PAHs) are ubiquitous environmental pollutants that are formed by pyrolysis or incomplete burning of fuels, organic substances such as coke, coal tar and pitch, cigarette smoke, and meats [Harrison, 2007]. There are several hundreds of PAHs which occur as complex mixtures rather than as individual compounds in industrial workplaces and living environments. As it is difficult to assess the effects in humans of individual PAHs, PAHs are treated as a group for risk assessment [Stellman and Guidotti, 2007]. Naphthalene and

1 Department of Preventive Medicine, Dongguk University College of Medicine, Gyeongju-si, South Korea 2 Department of Preventive Medicine, College of Medicine and Medical Research Institute, Chungbuk National University, Cheongju-si, South Korea  Correspondence to: H. Lim, Department of Preventive Medicine, Dongguk University College of Medicine 123, Dongdae-ro, Gyeongju-si, Gyeongsangbuk-do 780-714, South Korea. E-mail: [email protected]

Accepted 27 March 2015 DOI 10.1002/ajim.22463. Published online 5 May 2015 in Wiley Online Library (wileyonlinelibrary.com).

ß 2015 Wiley Periodicals, Inc.

pyrene are common compound in many PAHs mixture. For workers exposed to PAHs, monitoring such as 2-naphthol (metabolite for naphthalene) and 1-hydroxypyrene (1-OHP; metabolite for pyrene) in urine can provide an accurate measurement of exposures [Kim et al., 1999a; 2001]. In Korea, workplaces with exposure to PAHs from volatile naphthalene, coal tar pitch, and carbon black must by law provide exposed workers with a special health examination, including biological monitoring and industrial environmental workplace monitoring [Kim et al., 1999b]. Almost all Korean workers are exposed to levels below the American Conference of Governmental Industrial Hygienists (ACGIH)-recommended standard for PAHs in workplace air. Therefore, acute exposure is rare and the main concern from PAHs toxicity is chronic effects. PAHs may affect respiratory, gastrointestinal, renal, and dermatologic disorders. However, as most studies have focused on cancer research, there are few available studies of other health effects [Stellman and Guidotti, 2007]. Likewise, the cascade of events leading from exposure of PAHs to the human liver toxicity is poorly studied. Studies on animals have found exposure to PAHs associated with hepatomegaly or increasing serum liver enzymes [Yoshikawa et al., 1985].

Biomarkers for PAHs and Serum Liver Enzymes

Some human case reports noted elevated levels of aspartate aminotransferase (AST) associated with naphthalene exposure [Ojwang et al., 1985; Kurz, 1987]. Wu et al. [1997] and Chen et al. [2006] reported elevated serum liver enzymes in coke oven workers. Thus, there is limited evidence suggesting that liver toxicity is associated with PAHs when considering the metabolism of PAHs. Biomarkers can be classified into markers of exposure, effect, and susceptibility [Perera and Weinstein, 1982]. The transaminases are sensitive indicators of liver cell injury and are most helpful in recognizing toxin-induced liver injury [Hoofnagle and Ghany, 2011]. However, they are very nonspecific. Many researchers have studied g-glutamyl transferase (GGT) as marker of general oxidative stress [Lee and Jacobs, 2009a]. 2-naphthol and 1-OHP have been widely studied as exposure biomarkers, but there is little research on associated health effects. If PAHs metabolites reflect exposure to PAHs, there is likely an association with abnormal liver enzymes due to PAHs. Therefore, the authors conducted this study to evaluate the relationship of PAHs biomarkers and serum liver enzymes among the workers in a chemical factory.

MATERIALS AND METHODS Study Population A petrochemical plant, which was the research site, produces coal tar, carbon black, phthalic anhydride, plasticizer, carbomer, crude light oil, and HCP (hexachlorophene). Personal air sampling and analysis for PAHs (benzene-soluble coal tar pitch fraction) were conducted using the National Institute for Occupational Safety and Health method 5515 for all processes [NIOSH, 1994]. The geometric mean of the coal tar process was highest at 10 mg/m3 (ACGIH workplace air standard: 8-hr. time-weighted average of 200 mg/m3). Three hundred and thirty-six workers are employed at the study company. We selected 288 subjects for analysis after applying the following exclusion criterion: 16 workers with hepatobiliary system disease (including hepatitis B carrier), 10 absentee workers, 12 female workers, and 10 workers with taking medication (hepatotoxic drugs). All participants provided written informed consent prior to participation in this study, and all the processes were reviewed and approved by the Institutional Review Board at Dongguk University College of Medicine.

765

advance of the broader survey. The questionnaire included general questions pertaining to: age, drinking habits (including the amount of intake, kind of alcohol, and amount of drinking weekly), smoking, amount of exercise (physical activities enough to produce perspiration), past medical history, and present medication. The medical history and present medication were assessed by the examining physicians. Blood samples were collected after an overnight fast. An automated chemistry analyzer (AU1000, Olympus, Japan) was used for serum AST, alanine aminotransferase (ALT), and GGT analysis.

Analysis for PAHs Metabolite Urine samples were collected at the end of the work day and kept at 20°C until analyzed [Kim et al., 2001]. The analysis of urinary 2-naphthol and 1-OHP was performed as described by Kim et al. [1999] and Jongeneelen et al. [1987], respectively. In a dark room, 1 ml of the urine samples was buffered with 100 ml of 0.2 M sodium acetate buffer (pH 5.0) and hydrolyzed enzymatically with 10 ml of b-glucuronidase with sulfatase activity (Sigma G-8076, St. Louis, MO) for 16 hr. at 37°C in a shaking water bath. After hydrolysis, 1.5 ml of acetonitrile was added, and the samples were centrifuged at 10,000 g for 10 min. 1-OHP (361518, Sigma Aldrich, St. Louis, MO) and 2-naphthol (N-1250, Sigma Aldrich, St. Louis, MO) were used as the standard reagents. A high-performance liquid chromatography (HPLC) system consisting of a pump (Waters 600E, Millipore, Milford, MA) and a variable fluorescence detector (RF-10AxL, Shimadzu, Kyoto, Japan) was used. A 150 mm-long reverse phase column (TSK gel ODS-80, Tosoh, Tokyo, Japan) was used for 1-OHP analysis, and a 250 mm-long reverse phase column (J’sphere ODS-H80, YMC, Wilmington, NC) was used for 2-naphthol analysis. The mobile phase was 60% acetonitrile for 1-OHP and 38% acetonitrile for 2-naphthol. The flow rate was 1 ml/min. The excitation/emission wavelengths used in the detection of 1-OHP were 242/ 288 nm and for 2-naphthol 227/355 nm. The detection limits, set at three times the standard deviation of a blank or unspiked urine sample divided by the slope, were 0.13 ng/ml for 2-naphthol and 0.20 ng/ml for 1-OHP. The measured values of 1-OHP and 2-naphthol were corrected with the creatinine (Cr) concentration for the urinary output. The Cr concentration was measured using Jaffe’s reaction.

Classification of Variables Questionnaire and Medical Examination A questionnaire survey was given to assess general characteristics. Interviews, health examinations, and sample collections were conducted onsite at the company. A selfadministered questionnaire was distributed 2 weeks in

Body mass index (BMI) was calculated by dividing the weight by the height in meters squared. Amount of alcohol intake was calculated using the kind of alcohol, alcohol proof, and weekly average intake volume over 1 month prior to the survey (gram per week). Amount of exercise was the

766

Min et al.

average self-reported frequency per week (minute per week). PAHs exposure level in the workplace was classified into clerical and production workers according to job type. When used as a continuous variable, 2-naphthol and 1-OHP, AST, ALT, and GGT were logarithm transformed to normalize the right skewed distributions. Other variables are used without transformation. When used as a categorical variable or for stratification, exposure was categorized as first tertile (T1), second tertile (T2), and third tertile (T3) by tertiles of 1-OHP and 2-naphthol. Covariates were dichotomized. Alcohol intake was divided into >196 g/week and 196 g/week following a U.S. recommendation in a set of international drinking guidelines [ICAP, 2003]. Smoking was classified into smoker and non-smoker (never-smoker and ex-smoker). Amount of exercise and BMI were dichotomized at the median. The non-healthy behavior group (NHBG) included the following: alcohol intake, >196 g/week (NHBG-A); current smoker; BMI, >23.9 kg/m2; exercise, 100 min/week. The healthy behavior group (HBG) included the following: alcohol intake, 196 g/week (HBG-A); non-smoker; BMI, 23.9 kg/m2; exercise, >100 min/week.

according to ROC curves, indicating the best cut-off values for the prediction of liver disease in men. Likewise, independent variables following the above categorization criteria were used to adjust the model. For the linear trends, the number of 2-naphthol tertiles was used as a continuous variable and tested on each model. The multiplicative and additive models of interaction were used to examine the interaction between 2-naphthol and alcohol intake. Multiplicative interaction was assessed by modeling the 2-naphthol– alcohol intake product term in the logistic regression. Also, we calculated the relative excess risk due to interaction (RERI), the proportion attributable to interaction (AP), and synergy index (S) to measure the interaction on an additive scale with methods proposed by Andersson et al. [2005], and their 95% confidence intervals using the delta method [Hosmer and Lemeshow, 1992]. In the absence of additive interaction, RERI and AP are equal to 0; S is equal to 1. For statistical analysis, SPSS 20.0 for Windows (SPSS, Inc., an IBM Company, Chicago, IL) was used. P < 0.05 was considered statistically significant.

RESULTS Statistical Methods The Mann–Whitney U test was used to compare production jobs and clerical work. The association of PAHs metabolites (categorical) with liver enzymes (AST, ALT, and GGT; continuous) was tested using a general linear model (GLM). Therein, we adjusted for alcohol consumption (categorical), cigarette smoking (smoker, non-smoker), amount of exercise (categorical), age (continuous), and BMI (continuous). Duration of employment (continous) was used as a variable in place of age. We analyzed the stratified GLM using the above categorization criteria to check for an interaction between 2-naphthol and other variables. For this, the stratification criterion variable was used as a continuous variable in each stratum. b-coefficient and p for trend were calculated by linear contrast test in GLM. Multiplicative interaction was assessed by modeling the 2-naphthol covariates product term in the GLM. To compare subjects who had abnormal liver function index values in tertiles of 2-naphthol, we used multiple logistic regression models. The dependent variables (AST, ALT, and GGT) were divided into normal/abnormal using three reference values, as below. The Korea Occupational Safety & Health Agency’s (KOSHA) reference values (AST 40 IU/L, ALT 40 IU/L, and GGT 62 IU/L) were set by the occupational environmental medicine experts’ agreement [KOSHA, 2013]. The authors could not locate the protocol for McPherson and Pincus’ [2011] reference values (AST 33 IU/L, ALT 36 IU/L, and GGT 40 IU/L). Kim et al. [2004] proposed reference values (AST 31 IU/L and ALT 30 IU/L)

Distributions of the study subjects categorized as clerical and production workers are shown in Table I. The concentration of 1-OHP in production workers was significantly higher than in clerical workers (P < 0.05). There was no difference in the concentration of 2-naphthol between production and clerical workers. The general characteristics of subjects by tertile groups of PAHs metabolites are shown in Table II. Subjects with high urinary concentrations of 2naphthol were more likely to be current smokers compared with those with low urinary concentrations. Age, alcohol intake, BMI, and amount of exercise were not associated with urinary 2-naphthol or 1-OHP concentrations. Among PAHs metabolites, only 2-naphthol had a positive trend with serum AST after adjustment for alcohol intake, smoking, BMI, exercise, and age. 1-OHP had no association with liver enzymes (Table III). The model with duration of employment substituted for age showed similar results as Table III (data not shown). There were statistically significant interactions between 2-naphthol and alcohol intake in association with liver enzymes (Table IV). When study subjects were stratified by alcohol intake (196 g/week), smoking (current smoking status), BMI (median), and the amount of exercise (median), 2-naphthol consistently showed a positive association with serum liver enzymes in the NHBG. However, 2-naphthol showed no association with serum liver enzymes in the HBG. Multiple logistic regression showed that when liver enzymes were divided by KOSHA cut-off values [KOSHA, 2013], there were no significant differences in the distribution and trend based on tertiles of 2-naphthol (model 1).

Biomarkers for PAHs and Serum Liver Enzymes

TABLE I. Distribution of Demographic Factors, Urine PAHs Metabolites, and Serum Liver Enzymes in Study Population

Variables Current smokera Age (years)b Duration of employment (years)b BMI (kg/m2)b Alcohol intake (g/week)b Exercise (min/week)b AST (U/L)c ALT (U/L)c GGT (U/L)c 2-Naphthol (mmole/mole  Cr)c 1-Hydroxypyrene (mmole/mole  Cr)c

Production workers (N = 236)

Clerical workers (N = 52)

128 (54.2) 41.4 (6.0) 16.8 (5.8) 23.9 (2.5) 132.7 (144.9) 153.7 (173.9)d 24.5 (1.3) 24.6 (1.6) 33.1 (1.9) 2.86 (6.65) 0.12 (3.58)e

27 (51.9) 40.9 (6.8) 15.3 (7.1) 24.4 (3.2) 123.6 (110.1) 96.8 (99.7) 24.1 (1.4) 23.9 (1.7) 33.4 (1.9) 2.53 (5.56) 0.08 (2.44)

AST, aspartate aminotransferase; ALT, alanine aminotransferase; GGT, g-glutamyl transferase. a Number (%) b Mean (standard deviation) c Geometric mean (geometric standard deviation) d P< 0.05 by student t-test. e P< 0.05 by Mann-Whitney U test.

Model 2 adopting McPherson and Pincus [2011] cut-off values, showed a higher distribution in the subgroup with abnormal AST by increasing of 2-naphthol. In model 3, with the cut-off values proposed by Kim et al. [2004], comparing third tertile versus first tertile of 2-naphthol levels, the odds ratios (OR) were elevated for abnormal serum AST levels [OR ¼ 4.1 (95%CI 1.6–10.2)] and abnormal serum ALT levels [OR ¼ 2.4 (95%CI 1.2–4.9)] (Table V). There were no associations between GGT and 2-naphthol in any models. There was no evidence of additive or multiplicative interaction of 2-naphthol and alcohol intake for the increased risk of abnormal liver enzymes in model 2 (Table VI).

767

DISCUSSION In this study, 2-naphthol and 1-OHP were lower than other studies for similar workers [Ovrebo et al., 1995; Yang et al., 1999]. This is presumably because the overall exposure level was probably lower than in similar workplaces. This may reflect a general trend for improved industrial hygiene and exposure control in the Korean industry. Detailed explanations for the dose–response relationship between urinary 2-naphthol (or 1-OHP) and PAHs exposure status in the workplace were discussed in a separate paper [Lee et al., 2014]. In the selection of subjects, females were excluded because their numbers were few and unrepresentative. Subjects taking medication or having hepatobiliary system disease did not fit the eligibility criteria because elevated liver enzymes in such people is associated more with drug or disease than environmental toxins or alcohol. Alcohol intake classification is based on international drinking guidelines. Bellentani et al. [1997] proposed 30 g of ethanol per day as a risk threshold for alcohol induced liver damage, but the number of high intake subjects was small when this criterion was applied in our study. In the results, we observed that only serum AST was positively associated with urinary 2-naphthol among all workers. There was no significant association or trend of ALT and GGT with 2-naphthol, even though AST was highly correlated with ALT (crude correlation coefficient ¼ 0.69; data not shown) and GGT (crude correlation coefficient ¼ 0.51; data not shown). However, there was interaction between urinary 2-naphthol and alcohol intake on serum liver enzyme level in the GLM analysis (Table IV). In the GLM with stratified alcohol intake, 2-naphthol in NHBG-A was associated with increase of all liver enzymes, while there was no association for HBG-A. These results suggested that the liver toxicity of PAHs seems to be synergistic above a certain alcohol intake threshold. As the amount of alcohol intake of T3 (tertiles of 2-naphthol) is less than T2 in NHBGA, this result suggests that the main factor of increased liver enzymes is not only alcohol in NHBG-A.

TABLE II. General Characteristics of the Study Subjects byTertiles of 2-Naphthol and1-OHP Variables Tertiles of 2-naphthol

a

Tertiles of1-hydroxypyrenea

T1 T2 T3 T1 T2 T3

Age (years)

Alcohol intake (g/week)

BMI (kg/m2 )

Exercise (min/week)

Current smoker (N)

42.3  6.1 40.6  6.2 40.9  6.1 41.9  5.3 41.0  6.6 41.0  6.6

113.1  127.8 131.9  137.9 148.1  150.1 133.4  158.4 115.9  107.8 143.7  146.2

23.9  2.2 23.7  2.7 24.2  3.0 24.0  2.8 23.7  2.8 24.2  2.3

129.2  141.5 159.5  181.6 141.6  167.7 125.6  140.5 150.0  154.5 154.7  193.7

24 (25.0%) 59 (61.5%) 72 (75.0%)b 41 (42.7%) 52 (54.2%) 62 (64.6%)b

T1, first tertile; T2, second tertile; T3, third tertile. a Total of 288 subjects are divided into three groups of 96 subjects by tertiles of 2-naphthol or1-hydroxypyrene. b Ptrend < 0.05 by x2 for trend test.

768

Min et al.

TABLE III. Adjusteda Geometric Mean (Standard Error) of Serum Liver Enzymes byTertiles of Urinary 2-Naphthol and1-OHP Tertiles

2-Naphthol (mmole/mole  Cr) AST (U/L) ALT (U/L) GGT (U/L) 1-Hydroxypyrene (mmole/mole  Cr) AST (U/L) ALT (U/L) GGT (U/L)

T1 (N ¼ 96)

T2 (N ¼ 96)

T3 (N ¼ 96)

2.56 23.9 (1.0) 24.2 (1.0) 36.5 (1.1) 0.08 25.6 (1.0) 25.0 (1.0) 38.4 (1.1)

T1196 g/week) 2-Naphthol (mmole/moleCr) A 2.14 (73) B 2.51 (23) Alcohol intakeb A 55.7  54.8 B 303.7  109.3 AST (U/L) A 22.4 (1.0) B 24.3 (1.1) ALT (U/L) A 23.3 (1.1) B 22.1 (1.1) GGT (U/L) A 29.8 (1.1) B 39.8 (1.2) Smoking (C: never-smoker and ex-smoker, D: current smoker) 2-Naphthol (mmole/moleCr) C 1.34 (44) D 5.53 (51) Alcohol intake C 135.4  178.8 D 155.6  151.1 AST (U/L) C 23.3 (1.0) D 23.9 (1.0) ALT (U/L) C 23.4 (1.1) D 22.9 (1.1) GGT (U/L) C 32.8 (1.1) D 40.4 (1.1) BMI (E: 23.9 kg/m2, F: >23.9 kg/m2) 2-Naphthol (mmole/mole  Cr) E 2.62 (48) F 2.28 (48) Alcohol intake E 99.4  125.9 F 120.8  123.4 AST (U/L) E 23.7 (1.0) F 24.3 (1.0) ALT (U/L) E 22.5 (1.1) F 26.2 (1.1) GGT (U/L) E 32.0 (1.1) F 41.3 (1.1) Exercise (G: >100 min./week, H: 100 min./week) 2-Naphthol (mmole/mole  Cr) G 2.41 (49) H 2.58 (47) Alcohol intake G 124.2  145 H 105.7  108 AST (U/L) G 24.9 (1.0) H 23.2 (1.0) ALT (U/L) G 24.8 (1.1) H 24.1 (1.1) GGT (U/L) G 34.7 (1.1) H 39.0 (1.0)

T2 (N)

T3 (N)

T1

Biomarkers for polycyclic aromatic hydrocarbons and serum liver enzymes.

Limited evidence suggests that human liver toxicity is associated with exposure to polycyclic aromatic hydrocarbons (PAHs)...
119KB Sizes 5 Downloads 8 Views