The study subjects were the coke oven workers in a steel plant in Taiyuan, northern China, where the workers are exposed to PAHs during the open-air coking process in their daily shift. Seventy-eight male workers selected from this coke oven plant were classified into two groups, COE-exposed and control, based on historic exposure data from mandatory regular air sample analysis. The exposed group consisted of 47 coke oven workers; 31 control subjects were workers of the distillery, maintenance sections, and offices in the same plant. The workers participating in the study had been employed for at least 10 years and were currently working in the coke oven plant. After providing their written informed consent to participate in this study, at enrollment all the individuals were interviewed; a questionnaire was used to elicit their health status, body weight and height, occupational history, smoking, and alcohol consumption.

The one-time blood (2 mL each) and urine samples (20 mL each) were collected from subjects at the end of the work shift. Fresh blood lymphocyte cultures were carried out for the MN analysis. Serum was separated by centrifugation and stored at −80°C for detection of GST activity. The urine samples were stored at −28°C before being analyzed for 1-OHP and 8-OHdG. The research protocol was approved by the Ethics and Human Subject Committee of Tongji Medical College.

We determined urinary 1-OHP using high-performance liquid chromatography (HPLC) according to ^{Li et al. (2003)}. We used 2 mL urine from each sample for the detection assay, in which 0.5 mL of sodium hydroxide (15 mol/L) was added, and the samples were incubated at 100°C for 3 hr in the dark. Then, 50 μL carbazole (100 μg/mL) was added to each sample as the internal standard. The samples were adjusted to pH 3–5 with hydrochloric acid (1 mol/L). Subsequently, 1-OHP was extracted from urine with 4 mL dichloromethane. The extracts were evaporated to dryness under vacuum before being dissolved in 0.3 mL HPLC solvent and analyzed by HPLC (Varian, Walnut Creek, CA, USA). The separation of 1-OHP was achieved using a reverse-phase C18 column (Spherisorb ODS2, 4.6 mm × 250 mm, 5 μm; Waters, Milford, MA, USA). The mobile phase used for isocratic elution of 1-OHP was composed of methanol and water (3:1). To quantify 1-OHP, we used a pump system (Varian-Prostar model 230; Varian, Walnut Creek, CA, USA), an autosampler (Varian-Prostar model 410), and a fluorescence detector (excitation wavelength 346 nm, emission wavelength 386 nm; Varian-Prostar model 360). Identification and quantification of 1-OHP were based on retention time and peak area measured using a linear regression curve obtained from internal standard solutions. The detection limit of urinary 1-OHP was 0.5 ng/mL; for measurements below 0.5 ng/mL, we used 0.35 ng/mL (70% of the detection limit) as the default. Urine 1-OHP concentrations were expressed as micromoles per mole creatinine.

The cleanup treatment of urine and analysis of urinary 8-OHdG were carried out according to a method described by ^{Germadnik et al. (1997)} with minor modifications. Briefly, urine samples were first centrifuged at 1,500 × g for 5 min to remove precipitates. Then, 2 mL supernatant was adsorbed twice on preconditioned cartridges (10 mL/500 mg; Bond Elut LRC C18 OH; Varian) and eluted twice with 50 mM potassium phosphate monobasic (KH₂PO₄; pH 7.5) containing 15% and 20% methanol, respectively. Subsequently, the eluate was evaporated to remove methanol, supplemented by HPLC solvent to bring a final volume of 1.5 mL, and analyzed by an HPLC system with electrochemical detector (Varian-Prostar model 370; Varian) equipped with a glassy-carbon working electrode operated at +0.8 V versus a silver/silver chloride reference electrode. The separation of 8-OHdG was carried out on a reverse-phase C18 column with 30°C column temperature. The mobile phase used for isocratic elution of 8-OHdG was composed of 50 mM KH₂PO₄ (pH 3.5), 1% methanol, 2.5% acetonitrile, 2 mM KCl, and 0.1 mM EDTA. The flow rate was 0.8 mL/min, and the limit of detection of urinary 8-OHdG was 1 ng/mL. We quantified 8-OHdG in urine by the peak area of measurement using the linear regression curve for standard solutions of 3.5, 14, 56, and 224 nmol/L. For measurements below 1 ng/mL, we used 0.5 ng/mL, half of the detection limit, as the default. The concentration of urinary 8-OHdG is presented as micromoles per mole creatinine.

We used a cytokinesis-block MN assay to measure lymphocytic MNs. Fresh blood lymphocyte cultures were set up by adding 0.5 mL whole blood to 4.5 mL RPMI-1640 medium supplemented with fetal calf serum (15%) and penicillin (100 IU/mL). Phytohemagglutinin (Sigma, St. Louis, MO, USA) was added to lymphocyte cultures at a final concentration of 20 μg/mL. After 44 hr of incubation, cytochalasin-B (Sigma) was added to the culture medium at a final concentration of 6 μg/mL to arrest cytokinesis. After a total incubation period of 72 hr, cells were harvested by centrifugation at 400 × g for 10 min and mild hypotonic treatment in 0.075 M KCl for 2–3 min at room temperature. The cell suspensions were again centrifuged at 400 × g for 10 min. The pellets were fixed twice in freshly prepared cold methanol/acetic acid (5:1) and dropped onto slides before staining with 10% Giemsa solution for approximately 10 min. For each sample, 1,000 binucleated cells and MNs in binucleated cells were examined, and the frequencies of BNMNs were assessed according to the criteria of ^{Kirsch-Volders et al. (2000)}. One reader blinded to the status of the subjects scored all slides.

We measured GST activity in serum, calculated as units per liter, using a GST colorimetric activity assay kit (Jiancheng Bio Company, Nanjing, China).

We performed the statistical analyses using SPSS software (version 12.0) for Windows (SPSS, Chicago, IL, USA). We examined the normal distribution of all data using the Shapiro-Wilk normality test to determine subsequent use of appropriate tests for statistical comparison. We used Mann-Whitney and Pearson chi-square tests to compare the demographics and lifestyle variables between the exposure and control groups. The mean values of 1-OHP, 8-OHdG, BNMNs, and GST were calculated after categorizing by smoking status (smokers and nonsmokers), and the data were reported as median and interquartile range because variables were not normally distributed. We used the Mann-Whitney test to compare values of biomarkers between the exposure and control groups. Spearman’s rank correlation coefficient was calculated to evaluate the relations between 1-OHP levels, GST activity, 8-OHdG concentration, and BNMN frequency. We performed multivariate logistic regression to calculate odds ratios (ORs) and 95% confidence intervals (CIs) to assess the impact of independent variables [occupational exposure, body mass index (BMI), smoking, alcohol drinking, and age] on dependent variables (8-OHdG, BNMNs, and GST). For all the tests, p < 0.05 was defined as significant with a two-sided test.

Table 1 shows the characteristics of study subjects by work site. Coke oven workers were 1 year older (mean age) than control subjects. We found no significant differences in employment time, BMI, or percentages of smokers and alcohol drinkers between the two groups.

As shown in Table 2, we found a significant difference in urinary 1-OHP between the exposure group and the control group (p = 0.033). As an internal exposure biomarker, body PAH burden was associated with coke oven exposure. Likewise, we observed that compared with the control group, coke oven workers had a significant increase of biologic effect biomarkers: 8-OHdG (p = 0.022), lymphocyte BNMN formation (p = 0.014), and serum GST activity (p < 0.001). Furthermore, the results appeared to be modified by smoking. For instance, statistically significant differences between the exposure group and the control group existed only in smokers for 1-OHP (p = 0.029), BNMNs (p = 0.002), and GST (p < 0.001) and in nonsmokers for 8-OHdG (p = 0.005). For both exposed and control groups, no statistical differences in these biomarkers were observed between smokers and nonsmokers. However, further multivariate analysis with adjustment for smoking was performed.

We observed a significant correlation in urinary 1-OHP and serum GST activities in the total population (exposed and controls) (p = 0.005) (Table 3) and in the exposed group (p = 0.002) (Figure 1A) but not in the control group (p = 0.686) (Figure 1B). However, we found no correlation between 1-OHP and other biomarkers, as shown in Table 3. Additionally, Table 3 presents a significant correlation between lymphocyte BNMN frequencies and serum GST activities (p = 0.009).

Multivariate logistic regression analyses were performed with adjustment for age and smoking to evaluate the possible determinants of increased 8-OHdG levels, BNMN frequency, and GST activity. We calculated OR values of occupational exposure, BMI, smoking, alcohol drinking, and age to compare their contribution with the measured biomarkers. As shown in Table 4, for all three biomarkers that were significantly different between exposure and control groups, the OR values of coke oven exposure were significantly associated with 8-OHdG (OR = 4.4; 95% CI, 1.4–13.7), GST (OR = 13.2; 95% CI, 3.9–44.3), and BNMNs (OR = 2.7; 95% CI, 1.0–7.4), although the association with the BNMN frequency was marginal (p = 0.047). Age was associated with an increase in the risk of BNMN formation (OR = 2.7; 95% CI, 1.0–7.3), but BMI was associated with a decrease in the risk of 8-OHdG adduct (OR = 0.3; 95% CI, 0.1–0.8).