Mukherjee, Sutapa MBBS, PhD; Rodrigues, Ema BA; Weker, Robert BA; Palmer, Lyle J. PhD; Christiani, David C. MD, MPH
Airborne polycyclic aromatic hydrocarbon (PAH) emissions are produced during the incomplete pyrolysis of organic material 1 and lead to the formation of gaseous compounds and particulate matter which may be absorbed through the respiratory and gastrointestinal tracts, as well as cutaneously. 2 Many epidemiologic studies have demonstrated that particulate exposure is associated with adverse health effects, including cardio-respiratory morbidity and mortality. 3–5 In addition, occupationally exposed individuals exposed to high levels of airborne PAH demonstrate increased rates of cancer, which can be attributed to the known carcinogenicity of several individual PAH. 6–10
Biological monitoring has been performed in several studies to document exposure to PAH in occupationally exposed workers. Various biomarkers of PAH exposure have been used, including metabolites in urine (benzpyrene, 1-hydroxypyrene, 1 urinary thioethers, 11 urinary mutagenicity 12 and PAH-DNA adducts. 13 The measurement of urinary 1-hydroxypyrene (1-OHP), a major metabolite of pyrene, has emerged as the most useful biomarker of PAH exposure in such studies for several reasons. 1,11,14,15 Firstly, pyrene has been demonstrated to be a significant component of PAH mixtures and its concentration within a certain industry remains relatively constant. 1,16,17 Secondly, pyrene is almost exclusively metabolized to 1-OHP, which accounts for approximately 90% of the total urinary excretion of pyrene in humans. 18,19 In addition, the measurement of urinary 1-OHP incorporates all PAH exposure routes, including inhalational and dermal absorption which have been shown to be important components of PAH exposure in occupational and other settings. 20,21
The majority of data reported on 1-OHP levels in the workplace have focused on heavily polluted areas, including coke ovens, electrode plants, and creosote impregnation plants. 11,22,23 In addition, some studies have characterized PAH biomarkers in workplaces assumed to have low to moderate PAH exposures for example, in the rubber industry 24 and among bitumen pavers. 25 This study describes for the first time urinary 1-OHP levels in boilermakers working at 2 sites, one with occupational particulate exposure predominantly from welding (metal fume) and the other with a mixed occupational particulate exposure from welding and contact with residual oil fly ash (ROFA).
Materials and Methods
The study was a repeated measures short-term prospective study at 2 collection sites, as described below.
The study was approved by the Institute Review Board of Harvard School of Public Health. The study population consisted of 41 boilermakers working at 2 work sites; either at an apprentice school or at an electric generating power plant during the overhaul of an oil-fired boiler in summer. At the apprentice school, apprentice boilermakers were taught how to weld mild steel, therefore the predominant occupational particulate exposure was to metal fume. The teaching schedule for apprentice boilermakers involved Saturday work between 9 am and 3 pm and preshift and a postshift urine samples were collected from subjects. A further urine sample was collected on Sunday morning from most subjects. Therefore, at the apprentice school the maximum number of urine samples collected was 3 samples per subject. Seven apprentice school subjects were studied on 2 separate occasions.
In contrast, at the power plant boilermakers were studied during 5 consecutive days of work (Monday to Friday), their hours were approximately 7 am to 5 pm, with a mixed chemical exposure to occupational particulates generated through welding and contact with ROFA. Boiler repair work in power plants has been associated previously with significant exposure to a complex mixture of toxic agents including ROFA, several transition metal oxides, nitrogen dioxide and ozone from welding emissions. 26–29 At the power plant, the maximum number of urine samples collected was 10 samples per subject since the power plant workers were followed for 5 consecutive days with a preshift and a postshift urine sample collected on each day.
At the apprentice school boilermakers were recruited from the welding school and their predominant exposure to occupational particulates was derived from welding mild steel. In contrast, at the power plant boilermakers were responsible for the repair or replacement of several large pieces of the interior wall of the oil-fired boiler that necessitated the use of acetylene torch cutting, carbon-stick and electric-arc welding. In addition, portions of the tubing used to circulate steam and water in the boiler were removed and replaced. Dark ash (presumed to be ROFA) coated the walls of the boiler and accumulated in the ash pit at the base of the boiler after high temperature combustion therefore these boilermakers are likely to be exposed to metal particulates through welding and to ROFA during boiler repair work.
All participants at both work sites completed a self-administered modified American Thoracic Society questionnaire with specific data collected on past medical history, respiratory symptoms, smoking and work exposure history. At the apprentice school spot urine samples were obtained from each subject pre- and postshift and on the next (Sunday) morning after working at the apprentice school with the maximum number of 3 urine samples from each individual. Seven individuals were assessed twice on 2 different occasions. At the power plant urine samples were collected from each subject pre-and postshift for 5 consecutive workdays leading to a maximum of 10 urine samples from each individual. However it was not possible for all subjects at both work sites to provide the maximum number of urine samples, therefore, some samples were not obtained from each individual. All urine samples were aliquotted and frozen at −20°C until 1-OHP determination. Urinary creatinine and cotinine were measured for each sample of urine (ESA Laboratories, Chelmsford, MA, USA). Urinary cotinine was analyzed using reverse phase high performance liquid chromatography (HPLC) with ultraviolet spectrophotometry detection and urinary creatinine was measured using spectrophotometry.
The analytical procedure to determine urinary 1-OHP levels by HPLC has been previously described. 14,23 Briefly, 10 mL urine aliquots were adjusted to pH 5.0 using 1.0 M acetic acid (LabChem Inc, Pittsburgh, PA), followed by the addition of 10 mL of 0.1M sodium acetate buffer (pH 5.0) and enzymatic hydrolysis using 25 μL of β-glucuronidase (Sigma, St Louis, MO). The samples were then left for 16 hours at 37°C in a shaking waterbath (100 rpm).
A solid-phase extraction (SPE) cartridge packed with C-18 reverse phase liquid chromatographic material (Sep-Pak C18 cartridge, Millipore, Waters, Milford, MA, USA) was used to extract the PAH metabolites from urine. The cartridge was primed with 5 mL methanol (J.T. Baker, Phillipsburg, NJ), followed by 10 mL of distilled water and then the hydrolyzed sample was passed through the cartridge at a flow rate of less than 1 mL/minute. Subsequently the cartridge was washed with 8 mL of distilled water. The retained solutes were eluted with 8 mL methanol. The eluate was evaporated at 40°C under a constant flow of ultra high purity nitrogen, and the residue dissolved in 2 mL methanol and stored at −80°C.
Reverse-Phase HPLC Analysis
1-OHP measurement was performed using an Agilent 1100 HPLC modular system consisting of a 4-solvent gradient pump (model number 1311A), an automatic sample injector (model number 1313A), vacuum degassing unit (model number 1322A) and a continuously variable, simultaneous multi-wavelength fluorescence detector (model number 1321A) interfaced to a Hewlett Packard computer loaded with Chemstation software. The column used to determine 1-OHP was a 250-4 Merck Lichrosphere (Agilent Technologies, Palo Alto, CA) PAH column thermostatted at 30°C. The mobile phase of 65% acetonitrile/35% ultrapure water (MilliQ, Millipore Corp, Billerica, MA) at a flow rate of 1 mL/minute gave a retention time of 5.8 minutes for 1-OHP. The fluorescence detector was set for an excitation wavelength of 240 nm and an emission wavelength of 390 nm.
Before analysis, a stock solution was prepared by dissolving reagent grade 1-OHP (Aldrich Chem Co, Milwaukee, WI, USA) in methanol. The stock solution was stored at −4°C and remained stable for 4 weeks. To calibrate each HPLC run, standard solutions ranging from 0 to 100 ng/mL of 1-OHP were prepared for each run by adding microliter aliquots to a sample of urine from healthy donors. These urinary control standards were used to minimize any matrix effect of urine. Linear standard curves were obtained with a correlation coefficient of 0.99 or higher and an intercept approaching 0 ng/mL. The limit of detection (LOD) was 0.034 ng/mL calculated as the mean + 3 standard deviations of the mean of the control. This value, based upon 10 healthy donors’ urine confirms previous data obtained from 19 other healthy donors.
All urine samples were analyzed in batches of 30 to 40. Standards were run at the beginning and end of the sequence and were repeated separately after 8 subject samples for additional quality assurance. Instrumental reproducibility was 11% and was established by calculating the coefficient of variation (CV) of repeated analyses of 10% of all samples. Finally, the calculated concentration of urinary 1-OHP was adjusted to the urinary concentration of creatinine (μg 1-OHP/g creatinine) to control for variation in urinary flow.
Statistical analysis investigated associations of urinary 1-OHP levels with the following potential explanatory variables: age, self-reported smoking status, urinary cotinine level, day of collection and assignment of daily exposure (ie, pre- or postshift). Analyses were performed separately for each worksite (the apprentice school and the power plant). Both urinary 1-OHP and cotinine levels were skewed with a long right-hand tail, and were loge transformed before analysis and were normally distributed after log transformation. Exposure status (high/low) on each day and smoking status (yes/no) were analyzed as binary variables. Assignment of high and low exposure was based on the time of day that the urine sample was collected, ie, a preshift sample was assumed to approximate low PAH exposure and a postshift sample was assumed to approximate higher PAH exposure. The day of measurement (day 1 to day 5) was analyzed as a categorical covariate. All other variables were analyzed as continuous.
Generalized estimating equations (GEE) 30 were used to investigate the bivariate and multivariate relationships and predictive value of the explanatory variables and covariates of interest with the loge(1-OHP) levels at each time point, and to appropriately adjust for the within-individual correlations present in these data. Both forward and backward stepwise procedures were used to select a set of independent predictors. The modeling included an investigation of the need for interaction or polynomial terms, and examination of the effect of observations with high regression leverage.
Analysis and data management was carried out using STATA v5 (Stata Corporation) and SPlus v2000 (Mathsoft Inc.). Formal statistical significance was defined at the conventional 5% level.
From the 2 work sites a total of 241 urine samples underwent 1-OHP analysis and these were collected from 41 individuals who were all male. The apprentice school subjects (n = 21) were younger with a mean age of approximately 32 years compared to the power plant workers (n = 20) where the mean age was approximately 45 years (Table 1). This age difference resulted primarily because the apprentice school is responsible for training new boilermakers and therefore these subjects were younger with a lower mean number of years in the trade; 3 years as compared to 20 years for the workers at the power plant. The percentage of current cigarette smokers at both work sites was approximately 50%.
Table 2 shows the geometric mean levels of 1-OHP (μg/g creatinine) at the apprentice school stratified according to self-reported smoking status. In subjects who were nonsmokers there was minimal change in urinary 1-OHP levels during the work shift however there was a slight decrease in 1-OHP levels by the next nonworkday morning. This coincided with a decrease in urinary cotinine levels by the next nonworkday morning suggesting that exposure to environmental tobacco smoke may have been a factor in determining the 1-OHP level in nonsmokers at the apprentice school (data not shown). In smoking subjects, there was a decrease in mean 1-OHP levels over the work shift but overall mean 1-OHP levels remained stable on the 1 workday and 1 nonworkday measured.
Multivariate GEE analysis indicated that age, smoking and exposure classification were not significantly associated with the measured 1-OHP level for the apprentice school subjects (Table 3). Similar models including the urinary cotinine level and day of collection as a covariate revealed similar results (data not shown).
Table 4 shows the geometric mean levels of 1-OHP (μg/g creatinine) of subjects at the power plant stratified according to self-reported smoking status. In nonsmoking subjects the 1-OHP levels doubled during the 5 day working week. In smokers there was a similar increase in 1-OHP levels during the working week, however the increase was not as marked as that observed in nonsmoking subjects. In smoking subjects, there was marked variation in urinary 1-OHP levels as reflected by the wider confidence intervals.
Figure 1 demonstrates the changes in loge (1-OHP levels) in nonsmokers and smokers during the course of the workweek. Importantly, the 1-OHP levels measured in smoking subjects were always higher than observed 1-OHP values from nonsmoking subjects. However, 1-OHP levels increased over the 5 days of the study in both smokers and nonsmokers.
Multivariate GEE analysis indicated that neither age nor exposure classification was significantly associated with the loge (1-OHP level) for the power plant subjects (Table 5). Similar models including the urinary cotinine level as a covariate revealed similar results (data not shown). However, smoking was a significant predictor of the loge (1-OHP level) (P = 0.009). Therefore, we decided to repeat the GEE analysis analyzing smokers and nonsmokers separately to explore this observation.
Within nonsmoking subjects, the assignment of exposure was shown to be a significant predictor of the loge (1-OHP level) (P = 0.03;Table 6). In contrast, the assignment of exposure based on pre- and postshift urine collection times was not a significant predictor of the loge (1-OHP level) in smokers (Table 6). In this group, loge (cotinine) was a significant predictor (P = 0.005) of the loge (1-OHP level).
This study has investigated urinary 1-OHP levels in an occupationally exposed population of boilermakers, assumed to have moderate exposure to PAH. We show in this study that boilermakers working at the power plant exposed to mixed occupational particulates generate increased levels of urinary 1-OHP, a biomarker of PAH exposure. The boilermakers welding steel at the apprentice school did not demonstrate increased 1-OHP levels during their workday. This suggests that PAH exposure in boilermakers is related to the specific job task of the boilermaker and is therefore likely to be dependent on worksite. This observation could have been further elucidated if PAH environmental monitoring (personal and/or area sampling) had been possible and it is to be hoped that these measuring devices will become more available and less expensive for use in such work sites in the future.
Our observation that PAH exposure in boilermakers is related to particular job tasks is important because it may result in preventative strategies to reduce PAH exposure. It has been shown that workers from industrial settings where airborne PAH levels are high, such as coke oven and aluminum industry workers, demonstrate excess rates of cancers. 6–10 Therefore, it is likely that long-term exposure to PAH in boilermakers similarly increases the risk of developing cancer in this worker population. Further epidemiologic studies to investigate causes of mortality in this and other moderate PAH exposure worker populations should be encouraged.
Other populations occupationally exposed to PAH demonstrate higher levels of urinary 1-OHP after exposure to PAH. For example, preshift asphalt pavers had 1-OHP levels of 1.35 μmol/mol creatinine compared to postshift levels of 1.76 μmol/mol creatinine, however the population was not stratified according to smoking status. 31,32 Similarly, coke-oven workers in a high exposure group had levels of 29 μg/g creatinine compared to postshift levels of 199 μg/g creatinine. 23 The 1-OHP levels reported in our study are significantly lower than those observed in these high PAH exposure populations. However, it can be observed from this study that during the course of the 5 day workweek in both smokers and nonsmokers there was an increase in urinary 1-OHP levels at the power plant. This is an important observation because the boilermaker work schedule tends to be seasonal and requires that they work for 2 to 3 months with few nonworkdays. This study suggests that cumulative PAH exposure probably occurs during these intensive work periods that may subject these individuals to increased risk of disease.
At the apprentice school pre- and postshift 1-OHP levels remained relatively stable in both nonsmokers and smokers (Table 2). The difficulty in detecting significant changes in the urinary 1-OHP level pre- and postshift in workers from the apprentice school may have arisen because pre- and postshift urine samples were collected only 6 hours apart because of the short workday at the apprentice school. The peak excretion of 1-OHP in the urine may have been missed since the half-life of 1-OHP has been previously determined to be approximately 18 hours 16 with a wide individual range of 6 to 35 hours. 17,32 Another explanation for this observation may be that apprentice school workers were significantly younger (mean age 32 years) with fewer years in the trade (mean of 2.9 years), compared to workers at the power plant (mean age of 45 years and mean of 20.3 years in the trade). This is supported by the observation that the efficiency of PAH detoxification enzymes decreases with increased age. 33 In addition, prior working exposures were different in apprentice school workers and power plant boilermakers. Similarly the GEE analysis to predict the loge (1-OHP level) from exposure classification was not statistically significant at the apprentice school, possibly because of the limited time between collection of urine samples, few repeated urine sample measures or the type of exposure, predominantly welding, which is likely to be a low source of PAH exposure.
In contrast, at the power plant 1-OHP levels increased in both smokers and nonsmokers during the course of the workweek (Table 4 and Fig. 1) and the overall change in 1-OHP levels was similar in smokers and nonsmokers during the workweek. In addition, urinary loge (1-OHP level) in nonsmokers could be predicted by assignment of exposure (high/low) depending on whether the sample was obtained pre- or postshift (Table 6). In contrast, GEE analysis at the power plant demonstrated that smoking was a significant predictor of the loge (1-OHP level) in smokers (Table 6). This suggests that the urinary 1-OHP level was a sensitive biomarker of PAH exposure in nonsmokers, but that interpretation of the 1-OHP level in smokers was difficult because of the overwhelming effect of smoking (which contains significant amounts of PAH) on the 1-OHP level. Importantly, even in smokers, the 1-OHP level increased during the week suggesting that occupational PAH exposure was contributing to the measured 1-OHP level. The urinary loge (1-OHP level) could also be predicted by the age of the worker which suggests that there are alterations in detoxification enzyme ability within different age groups (as mentioned above). 33 Another explanation for this result may be that age is acting as a surrogate marker for years in the trade since it is likely that long-term cumulative PAH exposure has multiple deleterious effects on an individual, including detoxification enzyme systems.
A major limitation of our study was that we were unable to control for dietary sources of PAH (eg, char-broiled meats), which have been demonstrated to significantly increase the urinary 1-OHP level. 34,35 This is likely to have influenced the results obtained at the apprentice school given that these subjects were only monitored for 24 hours and may explain the lack of association seen between the 1-OHP level and the classification of exposure and smoking status. However, at the power plant the use of multiple urine samples and the GEE method of analysis may have reduced the inter-individual variation due to diet in our study.
In summary, we have demonstrated for the first time that boilermakers exposed to occupational particulates are exposed to PAH, particularly those working with ROFA. We have shown that the measurement of the PAH biomarker, the urinary 1-OHP level, is most useful in nonsmokers. However, the use of a repeated measure study design allows the levels in smokers to be followed, and these levels continued to increase during the course of the workweek in both smokers and nonsmokers. This observation is important because several PAH are known to be carcinogenic and these workers (both nonsmokers and smokers) remain at increased risk for the development of cancer and other diseases. Better characterization of PAH exposure in other low to moderate PAH exposure groups should be encouraged through the use of PAH monitoring devices and measurement of the urinary 1-OHP level.
The authors acknowledge the assistance of the Boilermaker Local 29 Union, Dr Shannon Magari and Dr Jee Young Kim in conducting this study.
This study was supported by NIH grants ES09860, ES00002, and the Mickey Leland Urban Air Toxics Center. SM was supported by the Cottrell Fellowship from the Royal Australasian College of Physicians.
1. Jongeneelen FJ. Benchmark guideline for urinary 1-hydroxypyrene as biomarker of occupational exposure to polycyclic aromatic hydrocarbons. Ann Occup Hyg. 2001; 451: 3–13.
2. Kim H, Cho SH, Kang JW, et al. Urinary 1-hydroxypyrene and 2-naphthol concentrations in male Koreans. Int Arch Occup Environ Health. 2001; 741: 59–62.
3. Pope CA,3rd. Respiratory disease associated with community air pollution and a steel mill, Utah Valley. Am J Public Health 1989; 79 (5): 623–628.
4. Dockery DW, Pope CA,3rd, Xu X, et al. An association between air pollution and mortality in six U. S. cities. N Engl J Med. 1993; 32924: 1753–1759.
5. Pope CA,3rd, Thun MJ, Namboodiri MM, et al. Particulate air pollution as a predictor of mortality in a prospective study of U. S. adults. Am J Respir Crit Care Med 1995; 151(3 Pt 1): 669–674.
6. IARC. International Agency for Research on Cancer. IARC monographs on the evaluation of the carcinogenic risk of chemicals to humans: Polynuclear aromatic compounds. Part I. Vol. 32. Lyon, France, 1983.
7. IARC. International Agency for Research on Cancer (IARC) Polynuclear aromatic compounds. Industrial exposures Part 3. Vol. 34. Lyon, France, 1984.
8. IARC. International Agency for Research on Cancer (IARC) Polynuclear aromatic compounds. Bitumens, Coal tars and derived products, Shale oils and soots Part 4. Vol. 35. Lyon, France, 1985.
9. Bertrand JP, Chau N, Patris A, et al. Mortality due to respiratory cancers in the coke oven plants of the Lorraine coalmining industry (Houilleres du Bassin de Lorraine). Br J Ind Med. 1987; 448: 559–565.
10. Verma DK, Julian JA, Roberts RS, Muir DC, Jadon N, Shaw DS. Polycyclic aromatic hydrocarbons (PAHs): a possible cause of lung cancer mortality among nickel/copper smelter and refinery workers. Am Ind Hyg Assoc J. 1992; 535: 317–324.
11. Ferreira M, Jr., Buchet JP, Burrion JB, et al. Determinants of urinary thioethers, D-glucaric acid and mutagenicity after exposure to polycyclic aromatic hydrocarbons assessed by air monitoring and measurement of 1-hydroxypyrene in urine: a cross-sectional study in workers of coke and graphite-electrode-producing plants. Int Arch Occup Environ Health. 1994; 655: 329–338.
12. Clonfero E, Jongeneelen F, Zordan M, Levis AG. Biological monitoring of human exposure to coal tar. Urinary mutagenicity assays and analytical determination of polycyclic aromatic hydrocarbon metabolites in urine. IARC Sci Publ. 1990; 104: 215–22.
13. Whyatt RM, Jedrychowski W, Hemminki K, et al. Biomarkers of polycyclic aromatic hydrocarbon-DNA damage and cigarette smoke exposures in paired maternal and newborn blood samples as a measure of differential susceptibility. Cancer Epidemiol Biomarkers Prev. 2001; 106: 581–588.
14. Jongeneelen FJ, Anzion RB, Leijdekkers CM, Bos RP, Henderson PT. 1-hydroxypyrene in human urine after exposure to coal tar and a coal tar derived product. Int Arch Occup Environ Health. 1985; 571: 47–55.
15. Dor F, Dab W, Empereur-Bissonnet P, Zmirou D. Validity of biomarkers in environmental health studies: the case of PAHs and benzene. Crit Rev Toxicol. 1999; 292: 129–168.
16. Buchet JP, Gennart JP, Mercado-Calderon F, Delavignette JP, Cupers L, Lauwerys R. Evaluation of exposure to polycyclic aromatic hydrocarbons in a coke production and a graphite electrode manufacturing plant: assessment of urinary excretion of 1-hydroxypyrene as a biological indicator of exposure. Br J Ind Med. 1992; 4911: 761–768.
17. Jongeneelen FJ, van Leeuwen FE, Oosterink S, et al. Ambient and biological monitoring of cokeoven workers: determinants of the internal dose of polycyclic aromatic hydrocarbons. Br J Ind Med. 1990; 477: 454–461.
18. Boogaard PJ, van Sittert NJ. Exposure to polycyclic aromatic hydrocarbons in petrochemical industries by measurement of urinary 1-hydroxypyrene. Occup Environ Med. 1994; 514: 250–258.
19. Wu MT, Wypij D, Ho CK, et al. Temporal changes in urinary 1-hydroxypyrene concentrations in coke-oven workers. Cancer Epidemiol Biomarkers Prev. 1998; 72: 169–173.
20. VanRooij JG, De Roos JH, Bodelier-Bade MM, Jongeneelen FJ. Absorption of polycyclic aromatic hydrocarbons through human skin: differences between anatomical sites and individuals. J Toxicol Environ Health. 1993; 384: 355–368.
21. Quinlan R, Kowalczyk G, Gardiner K, Hale K, Walton S, Calvert I. Urinary 1-hydroxypyrene: a biomarker for polycyclic aromatic hydrocarbon exposure in coal liquefaction workers. Occup Med (Lond). 1995; 452: 63–8.
22. Elovaara E, Heikkila P, Pyy L, Mutanen P, Riihimaki V. Significance of dermal and respiratory uptake in creosote workers: exposure to polycyclic aromatic hydrocarbons and urinary excretion of 1-hydroxypyrene. Occup Environ Med. 1995; 523: 196–203.
23. Wu MT, Mao IF, Ho CK, et al. Urinary 1-hydroxypyrene concentrations in coke oven workers. Occup Environ Med. 1998; 557: 461–467.
24. Schoket B, Poirier MC, Mayer G, et al. Biomonitoring of human genotoxicity induced by complex occupational exposures. Mutat Res. 1999; 4452: 193–203.
25. Jarvholm B, Nordstrom G, Hogstedt B, et al. Exposure to polycyclic aromatic hydrocarbons and genotoxic effects on nonsmoking Swedish road pavement workers. Scand J Work Environ Health. 1999; 252: 131–136.
26. Williams N. Vanadium poisoning from cleaning oil-fired burners. Br J Ind Med. 1952; 9: 50–55.
27. Hauser R, Elreedy S, Hoppin JA, Christiani DC. Upper airway response in workers exposed to fuel oil ash: nasal lavage analysis. Occup Environ Med. 1995; 525: 353–358.
28. Woodin MA, Liu Y, Hauser R, Smith TJ, Christiani DC. Pulmonary function in workers exposed to low levels of fuel-oil ash. J Occup Environ Med. 1999; 4111: 973–980.
29. Woodin MA, Liu Y, Neuberg D, Hauser R, Smith TJ, Christiani DC. Acute respiratory symptoms in workers exposed to vanadium-rich fuel-oil ash. Am J Ind Med. 2000; 374: 353–363.
30. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986; 421: 121–130.
31. Jongeneelen FJ, Anzion RB, Scheepers PT, et al. 1-Hydroxypyrene in urine as a biological indicator of exposure to polycyclic aromatic hydrocarbons in several work environments. Ann Occup Hyg. 1988; 321: 35–43.
32. Strickland P, Kang D. Urinary 1-hydroxypyrene and other PAH metabolites as biomarkers of exposure to environmental PAH in air particulate matter. Toxicol Lett. 1999; 108 (2–3): 191–199.
33. van Lieshout EM, Peters WH. Age and gender dependent levels of glutathione and glutathione S- transferases in human lymphocytes. Carcinogenesis. 1998; 1910: 1873–1875.
34. Buckley TJ, Lioy PJ. An examination of the time course from human dietary exposure to polycyclic aromatic hydrocarbons to urinary elimination of 1-hydroxypyrene. Br J Ind Med. 1992; 492: 113–124.
35. Kang DH, Rothman N, Poirier MC, et al. Interindividual differences in the concentration of 1-hydroxypyrene-glucuronide in urine and polycyclic aromatic hydrocarbon-DNA adducts in peripheral white blood cells after charbroiled beef consumption. Carcinogenesis 1995; 165: 1079–1085.