Document 7q2mE6nyJ7NXRrN5a5EQwDmB

ENVIRONMENTAL ehp HEALTH PERSPECTIVES ehponline.org Occupational Exposure to Benzene and Chromosomal Structural Aberrations in the Sperm of Chinese Men Francesco Marchetti, Brenda Eskenazi, Rosana H Weldon, Guilan Li, Luoping Zhang, Stephen M Rappaport, Thomas E Schmid, Caihong Xing, Elaine Kurtovich, and Andrew J Wyrobek http://dx.doi.org/10.1289/ehp.1103921 Online 15 November 2011 National Institutes of Health U.S. Department of Health and Human Services Page 1 of 26 Occupational Exposure to Benzene and Chromosomal Structural Aberrations in the Sperm of Chinese Men Francesco Marchetti1*, Brenda Eskenazi2, Rosana H Weldon2, Guilan Li3, Luoping Zhang2, Stephen M Rappaport4, Thomas E Schmid1, Caihong Xing3, Elaine Kurtovich2, and Andrew J Wyrobek1 1 Life Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA 2 Center for Environmental Research and Children's Health (CERCH), School of Public Health, University of California, Berkeley, CA, USA 3 National Institute of Occupational Health and Poison Control, Chinese Center for Disease Control and Prevention, Beijing, China 4 Center for Exposure Biology, School of Public Health, University of California, Berkeley, CA, USA *Present Address: Environmental Health Sciences and Research Bureau, Health Canada, Ottawa, ON, Canada Corresponding author: Brenda Eskenazi, Ph.D., Center for Environmental Research and Children's Health (CERCH), School of Public Health, UC Berkeley, 1995 University Ave, Suite 265, Berkeley, CA 94704, USA. Phone: 510-642-3496 Fax: 510-642-9083 Email: eskenazi@berkeley.edu Running head: Benzene and Sperm Chromosomal Damage Page 2 of 26 Key Words: benzene, chromosome 1, germ cells, sperm Fluorescent in situ hybridization, structural aberrations Acknowledgements: This research was supported by the National Institute of Environmental Health Sciences, National Institutes of Health and the Environmental Protection Agency (NIEHS IAG Y01-ES-8016-5, P42 ES04705, and NIEHS R03 ES015340-02). Work performed in part under the auspices of the US DOE by the Lawrence Livermore National Laboratory under contract W-7405-END-48 and Lawrence Berkeley National Laboratory under contract DE-AC02-05CH11231. We thank Debby Cabreros (LLNL) and Suzanne Young (UCB) for their contributions to this research. Ideas expressed in this manuscript are those of the authors and do not necessarily reflect the official views of the funders or institutions. S.M.R. has received consulting and expert testimony fees from law firms representing plaintiffs' cases involving exposure to benzene and has received research support from the American Petroleum Institute and the American Chemistry Council. All other authors declare that they have no competing financial interests. Abbreviations: AC.AC.M Duplications of the 1cen region AC.AC.M.M Numerical aberrations including diploidy/disomy of chromosome 1 AC.C.M Breaks in 1q12 ACM alpha, classical, midi A.C.M Breaks between 1q12 and 1cen AC.M.M Duplications of 1p36.3 ACO Deletions of 1p36.3 2 Page 3 of 26 CI Dels DF Dups FISH GM GSD IRR LOD OOM OOO OSHA p50 PEL ppm ppb SD MA US Confidence Interval Deletions Detection Frequency Duplications Fluorescence in situ hybridization Geometric mean Geometric standard deviation Incidence-Rate Ratio Limit of detection Deletions of the 1cen region Chromosome 1 nullisomy Occupational Safety and Health Administration Median Permissible exposure limit Part per million Part per billion Standard deviation trans,trans-Muconic acid United States 3 Page 4 of 26 ABSTRACT Background: Benzene is an industrial chemical that causes blood disorders including acute myeloid leukemia. We previously reported that occupational exposures near the US OSHA permissible exposure limit (PEL; 8 hour) of 1 ppm was associated with sperm aneuploidy. Objective: To investigate whether occupational exposures near 1 ppm increase the incidence of sperm carrying structural chromosomal aberrations. Methods: We applied a sperm fluorescence in situ hybridization assay to measure frequencies of sperm carrying partial chromosomal duplications or deletions of 1cen or 1p36.3 or breaks within 1cen-1q12 among 30 benzene-exposed and 11 unexposed workers in Tianjin, China as part of the China Benzene and Sperm Study (C-BASS). Exposed workers were categorized into low-, moderate- and high-exposure groups based on urinary benzene (medians: 2.9, 11.0 and 110.6 g/L, respectively). Median air benzene concentrations in the three exposure groups were 1.2, 3.7, 8.4 ppm, respectively. Results: Adjusted incidence rate ratios (IRR) and 95% confidence intervals (CI) for all structural aberrations combined were 1.42 (1.10-1.83), 1.44 (1.12-1.85) and 1.75 (1.36-2.24) and for deletion of 1p36.3 alone were 4.31 (1.18, 15.78), 6.02 (1.69, 21.39), and 7.88 (2.21, 28.05) for men with low-, moderate- and high-exposure, respectively, compared to unexposed men. Chromosome breaks were significantly increased in the high-exposure group (IRR=1.49, 95% CI =1.10-2.02). Conclusions: Occupational exposures to benzene were associated with increased incidence of chromosomally defective sperm raising concerns for worker infertility, spontaneous abortions, as well as mental retardation and inherited defects in their children. Our sperm findings point to benzene as a possible risk factor for de novo 1p36 deletion syndrome. Because chromosomal aberrations in sperm can arise from defective stem cells/spermatogonia, our findings raise concerns that occupational exposure to benzene may have persistent reproductive effects in formerly exposed workers. 4 Page 5 of 26 INTRODUCTION Benzene is an important industrial chemical with more than 2 billion pounds produced every year in the United States (US). Low-level exposures (<5 ppb) to benzene are widespread in the US population, primarily from smoking, gasoline fumes and vehicle emissions (Hricko 1994). Early epidemiological cohort studies found that benzene is associated with an increased risk of leukemia at high levels (around 10 ppm average or 40 ppm-years) (Hayes et al. 1997; Yin et al. 1996; Yin et al. 1987a; Yin et al. 1987b) while more recent studies found excess leukemia risk associated with levels of exposure as low as 0.8-1.6 ppm or 2-4 ppm-years of cumulative exposure (Glass et al. 2003; Glass et al. 2004; Hayes et al. 2001). Benzene is hematotoxic and in a large study of more than 400 workers almost all blood cell counts were significantly decreased, even in individuals exposed to <1 ppm benzene (mean (SD) of exposed=0.57 ppm (0.24) and unexposed0.04 ppm) in (Lan et al. 2004). Thus, benzene is a highly regulated chemical with the Permissible Exposure Limit (8-hour time-weighted average) in the US set at 1 ppm by the Occupational Safety and Health Administration (OSHA Occupational Safety and Health Standards 1987). While there has been significant international progress in reducing occupational exposure to benzene, workers in some countries still experience levels of benzene well above the US PEL (Liang et al. 2005). Specific chromosomal aneuploidies and aberrations implicated in leukemia have been detected in the blood cells of benzene-related leukemia patients as well as in healthy benzeneexposed workers, suggesting that these abnormalities precede and may be a potential mechanism underlying benzene-induced leukemia (Zhang et al. 2002; Zhang et al. 2011; Zhang et al. 2005). Recent findings from our group and others suggest that occupational exposure to benzene induces aneuploidies in sperm (Li et al. 2001; Liu et al. 2000; Zhao et al. 2004), even in workers 5 Page 6 of 26 exposed at or below the US PEL (Xing et al. 2010). In addition, in Chinese workers exposed to high doses of benzene (>10 ppm), exposure appeared to increase terminal duplications and deletions for chromosome 1, and increase centromeric duplications and deletions for chromosome 1 in sperm (Liu et al. 2003). There are no data yet available on whether benzene exposures near US PEL have detrimental effects on chromosomal structural aberrations in sperm. In humans, ~0.6% of newborns carry constitutive chromosomal abnormalities in the form of aneuploidies or structural aberrations (Jacobs et al. 1989; Shelby et al. 1993). Although, the incidence of children born with chromosomal aberrations is lower than for aneuploidies, these types of aberrations have ~80% paternal contribution and severe health consequences (Chandley 1991; Crow 2001). Inherited chromosomal aberrations of paternal origin can arise in spermatogonia, meiotic cells or in post-meiotic cells by different mechanisms. Sperm carrying partial chromosomal duplications and deletions in sperm can arise from stem cells carrying reciprocal translocations (Van Hummelen et al. 1997). In healthy men, sperm with chromosomal breaks are more prevalent than sperm with partial chromosomal duplications and deletions (Sloter et al. 2000) and both these classes of aberrations have been shown to increase with age (Sloter et al. 2004). The postmeiotic period of spermatogenesis appears to be very sensitive to the induction of DNA lesions that can lead to DNA strand breaks directly in sperm or after fertilization (Marchetti and Wyrobek 2005; Olsen et al. 2005). Unrepaired or misrepaired DNA damage in sperm can result in chromosomally abnormal offspring with a variety of heritable disease and dysmorphologies (Marchetti and Wyrobek 2005). The objective of our study was to investigate the prevalence of sperm carrying structural chromosomal abnormalities in a group of Chinese workers exposed to levels of benzene near the 6 Page 7 of 26 US PEL. We utilized the sperm ACM-FISH assay that employs DNA probes specific for three regions of chromosome 1 (Sloter et al. 2000). A major advantage of the ACM assay over previous sperm FISH assays (Li et al. 2001; Liu et al. 2000; Liu et al. 2003; Zhao et al. 2004) is that, in addition to partial chromosomal duplications and deletions, which are induced in spermatogonia or meiosis, it detects chromosomal breaks that are induced during postmeiosis. METHODS Study Population and Design Details of the study design, recruitment, and exposure assessment have been previously reported (Xing et al. 2010). All men who participated in the China Benzene and Sperm Study (CBASS) were recruited from factories in Tianjin, China. Briefly, exposed men were recruited from three factories that used benzene-containing glues to manufacture shoes, paper bags and sandpaper. Unexposed men were recruited from a meat packing plant and an ice cream manufacturing factory, both of which had no history of benzene use. Eligibility criteria included: age of 18-50 years, employment at the factory for at least one year, and no history of cancer or vasectomy. Informed consent was obtained from eligible men prior to enrollment in the exposure assessment phase of the study. All study materials were approved by the Committees for the Protection of Human Subjects at the University of California, Berkeley, Lawrence Livermore National Laboratory, Lawrence Berkeley National Laboratory and the Tianjin Occupational Disease Hospital (Tianjin 3rd Municipal Hospital) under an IRB authorization agreement with the National Institute of Occupational Health and Poison Control, Chinese Center for Disease Control and Prevention. Questionnaires and other study materials were developed in English, translated to Mandarin and reverse-translated to English. 7 Page 8 of 26 Exposure assessment consisted of personal passive-air badge monitoring (3M Organic Vapor Monitor, model 3500) and spot urine samples. Both sample types were collected on two different occasions, approximately one month apart. We measured benzene in air samples (ppm) and urine samples (g/L) for both exposure groups. In addition, we measured muconic acid (MA) in the urine samples from exposed men. Details of benzene measurements in air and urine samples were published elsewhere (Xing et al. 2010). The limits of detection (LOD) were 0.2 ppm for air benzene, 0.016 g/L for urinary benzene and 10 mg/L for urinary trans, transmuconic acid. ACM Sperm FISH analysis Semen samples were collected an average and standard deviation (SD) of 3.7 (2.2) days after the second urine collection from 78 men (34 exposed and 44 unexposed) of the 96 men that originally enrolled in the study. Semen samples were aliquoted and stored at -80C without preservative. ACM sperm FISH was performed on a subset of men: 10 men from each exposed group (low, moderate, and high) and 11 men from the unexposed group. At the time men were selected for the ACM analysis the urinary MA data were not yet available. Therefore, the 30 exposed men were selected based on their measured concentrations of urinary benzene such that 10 men were from the lowest tertile of urinary benzene concentration (low), 10 from the middle tertile (moderate) and 10 from the highest tertile (high). These three groups of men were then frequency-matched on age and smoking status in the three months prior to semen collection (yes/no) with unexposed men such that age distribution and smoking habits were balanced in each of the four groups. 8 Page 9 of 26 The ACM assay detects both structural (including breaks, duplications and deletions of the telomeric and centromeric regions of chromosome 1) and numerical abnormalities using probes for three repetitive sequence regions of chromosome 1: 1cen (D1Z5, alpha or A), 1q12 (pUC1.77, classical or C), and 1p36.3 (D1Z2, midi or M). Semen samples were thawed to room temperature, gently mixed with a Pasteur pipette, and a 5 mL aliquot was smeared onto a glass microscope slide and allowed to air dry for two days. The sperm decondensation and hybridization procedures were as previously described (Sloter et al. 2000) except that the 1cen alpha satellite probe was derived from a purified plasmid clone, pSD21-1 (Waye et al. 1987) and directly labeled with tetramethylrhodamine-6-dUTP (Roche Applied Science, Indianapolis, IN) with the use of the GIBCO BRL Nick Translation System (Life Technologies, Inc., Gaithersburg, MD). Slides were scored on a Zeiss Axioplan fluorescence microscope (Carl Zeiss) using a 100x Plan-NEOFLUAR Ph3 objective and equipped as previously described (Van Hummelen et al. 1996). Samples from 41 men (30 exposed and 11 unexposed) were analyzed by a single expert scorer who was blind to exposure status using the ACM sperm FISH assay and strict scoring criteria (Sloter et al. 2000). Slides were de-identified and coded by a technician other than the scorer. Then 5000 sperm were scored from the top half of the slide. Slides were then re-coded and another 5000 sperm were scored from the bottom half of the slide for a total of approximately 10,000 sperm per slide. For quality control, the two sets of scores per participant for the aggregate types of anomalies such as total structural anomalies and total breaks were compared using Cochran's equal-proportions test and data were accepted only if the p value was greater than 0.05. In this study, no slide failed to meet this criterion. 9 Page 10 of 26 Statistical methods Demographic characteristics for the unexposed and exposed men were tabulated and differences between the groups were examined using chi-square or Fisher's exact tests (exact test was used if categories contained fewer than 5 observations). The correlations were high between the benzene levels measured on the two personal passive-air monitors and the benzene levels in the two urine samples (Spearman >0.7, p<0.0005 for each pair); hence, the geometric mean (GM) of the two measurements was used as the summary value for each measure of exposure (air, urinary benzene and urinary MA) for each participant. We then derived the geometric mean, geometric standard deviation (GSD), and the percentiles of the distribution for each measure of exposure among participants classified as unexposed, exposed, and low-, moderate-, or highexposed. All 11 unexposed men and one low-exposed man had air benzene concentrations below the LOD. The air benzene concentration for the low-exposed man was imputed as the limit of detection divided by the square root of 2 (LOD/2) in order to calculate the geometric mean and standard deviation (GSD) for air benzene among men in the low-exposed group (Hornung and Reed 1990). The dependent variables from the sperm ACM FISH assay include: duplications of 1p36.3, deletions of 1p36.3, duplications of the 1cen region, deletions of the 1cen region, breaks in 1q12, breaks between 1q12 and 1cen and numerical aberrations including diploidy/disomy of chromosome 1 and chromosome 1 nullisomy. Because only one chromosome is examined in this assay, diploidy cannot be distinguished from disomy of chromosome 1. Counts of some anomalies were combined to summarize total duplications (1p36.3 + 1cen), deletions (1p36.3 + 1cen), breaks (breaks in 1q12 + breaks between 1q12 and 1cen), structural aberrations (all duplications, deletions and breaks), and numerical aberrations (diploidy/disomy + nullisomy). 10 Page 11 of 26 The detection frequency of each anomaly (% of men with at least one anomaly per 10,000 sperm for each outcome) and the overall median, mean and range of each anomaly/10,000 sperm were determined for each exposure group. Separate standard negative binomial models were used to estimate incidence rate ratios (IRR) and 95% confidence intervals (CI) for the frequency of each ACM outcome (counts/10,000 sperm) among low-, moderate- and high-exposed men compared with unexposed men. Zero-inflated negative binomial models were not deemed appropriate due to the small sample sizes even though some ACM outcomes had a moderate to large number of zero counts. We considered as potential confounders variables that were associated with benzene exposure, semen quality or genetic damage in the literature (see Supplemental Material Table 1). These included age (continuous), abstinence (continuous and 5 days vs. >5 days), body mass index (<18.5, 18.5-24.9, and 25-29.9 kg/m2), history of chronic disease (includes self-reported history of high blood pressure, other diseases of the heart or blood vessels, tuberculosis, lung disease, anemia, other blood diseases, diabetes, thyroid diseases, other hormonal diseases, stomach ulcers or other diseases of the GI tract, hepatitis, liver disease, epilepsy or other neurological disorders, or other chronic diseases), education (middle school or less vs. high school or more, use of hot baths in the three months prior to semen collection (yes/no), hours spent biking (< vs. >0.5 hrs/day), and consumption of tea, cola, multivitamins, alcohol (yes/no in the three months prior to semen collection), and fruits and vegetables per day (> or < the median of 3.6 times eaten per day). Covariates were included in models if they were associated with exposure and outcomes (at p<0.2) using separate bivariate negative binomial models or if the coefficients changed by more than 10% upon removal of the covariate. Although the exposure categories were frequency-matched on age and smoking in the three months prior to semen collection (yes/no), 11 Page 12 of 26 these two variables were included in all models to control for any residual confounding. For simplicity, the set of covariates that was associated with more than 50% of the ACM outcomes (age, smoking, alcohol consumption, and history of chronic disease) was applied to all models. All others were excluded from models after determining whether inclusion changed coefficients by more than 10%. Models were not adjusted for education because this variable was collinear with exposure such that men with lower education had higher exposure. Abstinence was not associated with exposure and was only associated with one ACM outcome. Inclusion of abstinence as a covariate did not change coefficients; therefore it was not retained in models. Associations with p-values below 0.05 were considered statistically significant. Lastly, we used negative binomial models to perform a trend analysis (separately for each outcome) where the urinary benzene explanatory variable was coded as 0 for unexposed, 1 for low-exposed, 2 for moderate-exposed and 3 for high exposed. Trend analysis models were adjusted by age, smoking, alcohol consumption, and history of chronic disease and are reported as ptrend. RESULTS The population characteristics of the exposed and unexposed men analyzed in the present study (see Supplemental Material Table 1) were similar with the exception of education, such that unexposed men were more likely to have completed high school than exposed men (55% compared with 10%, p=0.006). Overall, men smoked on average 10 9 cigarettes per day in the three months prior to semen collection; because men were frequency matched on whether they smoked in the three months prior to semen collection, this variable did not differ by exposure group (73% and 77% of unexposed and exposed men smoked within the previous 3 months, 12 Page 13 of 26 respectively). Duration of abstinence was similar in the two groups (mean SD: 6.4 5.4 vs. 6.6 3.7 days). Two unexposed men reported a doctor-diagnosed fertility problem, yet one had fathered a child. Semen quality and the percentages of men who had below normal values for sperm concentration, sperm count, semen volume, and % sperm motility according to World Health Organization criteria (W.H.O. 1992) did not differ significantly by exposure group (see Supplemental Material Table 2). All three measures of exposure (air benzene, urinary benzene and urinary MA) were highly correlated among exposed men (Spearman>0.8, p<0.0005 for each pair). Median urinary benzene measurements were 2.9, 11.0 and 110.6 g/L, respectively for men in the low-, moderate- and high- exposed groups compared to 0.1 g/L for unexposed men (Table 1). Median air benzene concentrations were 1.2, 3.7 and 8.4 ppm among the low-, moderate- and highexposed groups respectively, compared to <LOD for all unexposed men. Five men in the lowexposed group and three men in the moderate-exposed group had air benzene measurements below the US PEL of 1 ppm. Finally, median concentrations of MA were 1.7, 7.9 and 13.4 mg/L for low-, moderate- and high-exposed men, respectively. MA was not measured in unexposed men. Structural aberrations (1p36.3 and 1cen duplications and deletions, breaks in 1q12 and breaks between 1p12 and 1cen), and numerical aberrations (diploidy/disomy or nullisomy) of chromosome 1 were found in sperm of all men participating in the study, and counts/10,000 sperm increased with exposure category (Table 2). Two ACM outcomes, 1cen deletions and nullisomy were rare and were excluded from further analyses. Occupational exposure to benzene was associated with chromosomal structural aberrations, but not numerical abnormalities (Table 3). Adjusted IRRs for all structural 13 Page 14 of 26 aberrations combined were 1.42 (95%CI=1.10-1.83), 1.44 (1.12.-1.85) and 1.75 (1.36-2.24) in the low-, moderate- and high-exposed men compared to unexposed men. Among the structural aberrations, associations increased with increasing categorical benzene exposure for combined 1p36.3 duplications and deletions (p for trend <0.005), but not the 1cen region (p for trend=0.19). Compared to unexposed men, adjusted IRRs for 1p36.3 duplication were 2.31 (95%CI=1.03-5.18), 2.13 (0.95-4.79) and 3.12 (1.39-6.97) for the low-, moderate- and highexposed men, respectively. Associations were stronger for 1p36.3 deletion, with adjusted IRRs of 4.31 (95%CI=1.18-15.78), 6.02 (1.69-21.39), and 7.88 (2.21-28.05) for the low-, moderateand high-exposed groups compared with unexposed men. Results were similar when benzene exposure was categorized according to tertiles of urinary MA (see Supplemental Material, Table 3). Benzene exposure was also associated with chromosomal breaks, but the association with all breaks combined was significant for the high-exposure benzene group only (IRR=1.49, 95% CI=1.10-2.02) (Table 3). This association appeared to be specific for breaks within the 1q12 region (IRR=2.23, 95% CI=1.22-4.09 for high-exposure compared with no exposure), while there was no clear association with breaks between the 1cen and 1q12 region. An analysis comparing the 8 men exposed to <1 ppm of air benzene, the current U.S. PEL, to the 11 unexposed men indicated significant associations with 1p36.3 duplications and deletions [IRR 3.36 (1.66-6.81)], 1cen duplications and deletions [3.38 (1.62-7.09)], and all structural aberrations (duplications, deletions and breaks) combined [1.53 (1.17-1.99)]. However, total breaks were not associated with exposure (data not shown). No significant associations between benzene exposure and numerical abnormalities were observed (Table 3). 14 Page 15 of 26 DISCUSSION We previously reported evidence that occupational exposure to benzene, even at levels near the OSHA PEL of 1 ppm, was associated with aneuploidy for chromosomes X, Y and 21 in sperm (Xing et al. 2010). Here, in a subsample of the same study population, we report that occupational exposure to benzene was also positively associated with structural aberrations of chromosome 1 in sperm. Even men occupationally exposed to benzene in air at levels below 1 ppm had higher frequencies of sperm with structural chromosomal aberrations than unexposed controls. We did not find associations with aneuploidy of chromosome 1; however, the ACM assay does not effectively distinguish between disomy 1 and diploidy (Sloter et al. 2000). Therefore, this result should not be interpreted to mean that benzene does not induce aneuploidy for chromosome 1 in sperm. Our finding of increasing associations in sperm with duplications and deletions of chromosome 1 with increasing exposure is consistent with previous findings in men who were exposed to higher levels of benzene (mean 10.81 ppm vs. 5.57 ppm among all exposed in our study) (Liu et al. 2003). More importantly, our study suggests that benzene may induce chromosomal damage in sperm at low doses. Because the ability to detect this kind of damage in sperm by FISH is unique to the ACM assay, no previous data are available for comparison; however, men exposed to high concentrations of benzene (mean = 22.12 ppm) were found to have increased DNA breaks in sperm using the single cell gel electrophoresis assay (Song et al. 2005). The clastogenic effects of benzene are thought to result from a combination of DNA adduct formation, oxidative damage and inhibition of topoisomerase II activity (Whysner et al. 2004). 15 Page 16 of 26 Duplications and deletions of 1q36.3 and 1q12 detected by our assay occur before the two meiotic segregation divisions. Two plausible mechanisms of origin are possible. First, benzene may induce breaks during meiosis, which result in random segregation of the broken fragments during meiotic chromosome segregation. Second, benzene may generate DNA damage in stem cells/spermatogonia that are misrepaired into balanced chromosomal rearrangements (reciprocal translocations or inversions involving 1p). In the first case, duplications and deletions represent events induced during a ~25 day window, the duration of meiosis in humans (Adler 1996), which occurs between ~40 and ~65 days before semen collection. In the second case, chromosomal rearrangements would form quadrivalents during meiosis I and generate unbalanced gametes carrying partial chromosomal duplications or deletions (Van Hummelen et al. 1997). Sperm carrying duplications and deletions generated by this mechanism would persist for as long as the stem cell carrying the balanced rearrangement survives in the seminiferous epithelium, which could be the entire reproductive lifetime of the individual. This mechanism thus represents persistent benzene-induced sperm damage that can extend far beyond the end of the exposure window. Further research will be needed to discriminate between duplications and deletions that are generated through breaks induced in meiosis or through chromosomal rearrangements in stem cells/spermatogonia. Studies of men who are no longer exposed to benzene but have a history of occupational exposure to benzene could help distinguish between these two mechanisms. Chromosomal breaks detected by the ACM sperm FISH assay likely originated during the postmeiotic phase of spermatogenesis, a window of ~5 weeks before semen collection (Adler 1996). This is a period known to be sensitive for the induction of DNA damage and chromosomal breaks (Marchetti and Wyrobek 2008). However, we hypothesize from our results 16 Page 17 of 26 that occupational exposure to benzene must reach a threshold before a noticeable number of breaks are induced and that exposure to levels of benzene near the US PEL is not likely to induce elevated numbers of breaks during this time window. This finding supports the conclusion that the majority of duplications and deletions observed in the low exposure groups are most likely the results of accumulated chromosomal rearrangements in stem cells/spermatogonia rather than breaks induced during the meiotic cell cycle that produced the sampled sperm. Therefore, the effects of occupational exposure to benzene may not be reversible because stem cells with chromosomal rearrangements would continue to produce a lifetime of unbalanced gametes at every cycle of spermatogenesis even after the end of occupational exposure to benzene (i.e., a worker leaves that job). The ACM assay provides a direct assessment of chromosomal damage for the portion of the genome contained within the 1p36 to 1q12 regions, which represents 4-5% of the haploid genome (Mendelsohn et al. 1973). Regardless of exposure status, duplications and deletions of 1p36.3 were more common than duplications and deletions of 1cen and, as a whole, duplications and deletions were 4-5 times less common than breaks. Also, the frequencies of sperm with duplications were slightly higher that the frequencies of sperm with deletions for the same chromosomal region. These findings are in agreement with previous studies of healthy men (Bosch et al. 2003; Sloter et al. 2004; Sloter et al. 2000) and support the accuracy of the scoring criteria utilized in this study. Our study identifies occupational exposure to benzene as a possible risk factor in the etiology of the 1p36 deletion syndrome. Monosomy of 1p36 (terminal deletion of chromosome 1) is an established cause of mental retardation and major congenital malformations in children with an incidence of 1-2 children per 10,000 (Shaffer and Heilstedt 2001; Zenker et al. 2002). In 17 Page 18 of 26 our study, adjusted IRRs for 1.36.3 deletions were 4.31, 6.02 and 7.88 for the low-, moderateand high- exposed groups, respectively. The relationship between benzene exposure and 1p36 deletion appears to be stronger than the association between sperm aneuploidy and benzene exposure (where adjusted IRRs for disomy X, the strongest association observed, were 2.0 and 2.8 in low- and high-exposed men, respectively versus unexposed men) observed in the same population (Xing et al. 2010). This supports the notion that the clastogenic activity of benzene is stronger than its aneugenic activity (Chen et al. 1994). The ACM assay provides an indirect assessment of the damage in the rest of the genome on the assumption that the area targeted by the ACM assay is, on average, representative of the rest of the genome. This assumption is supported by studies that have found good concordance between the frequencies of sperm with chromosomal aberrations as detected by the human sperm/hamster-egg technique, which detects aberrations in all sperm chromosomes (Rudak et al. 1978), and the frequencies obtained by extrapolating to the whole genome the frequencies of sperm with structural aberrations as detected by the ACM assay (Sloter et al. 2000). Extrapolating the ACM data to the whole genome yields estimated frequencies of sperm with chromosomal structural aberrations of 2.6-3.2%, 3.4-4.3%, 3.6-4.5% and 4.3-5.3% for unexposed, low-, moderate- and high-exposed groups, respectively. A limitation of our study is that occupational exposure to benzene was monitored only twice during a ~1 month period before semen sample collection. Because some of the damage measured by our assay may have occurred at an earlier time point than that monitored by our exposure assessment, it remains possible that exposure may have been misclassified. But we consider this to be of minimal impact. To limit misclassification of exposure, participants were required to have worked at the same factory for at least a year to be enrolled in the study. In 18 Page 19 of 26 addition, there was a very strong correlation among all measurements of exposure (air benzene, urinary benzene, and urinary MA) and between the two times of exposure assessment (Xing et al. 2010) suggesting that the measured exposure levels represented the usual workplace exposure level for each man. Due to the nature of the ACM assay, we performed many tests of association for all of the different outcomes and groups of outcomes. Thus, it is possible that some of our findings may be due to random chance. A final limitation is that recruitment from an occupational setting may predispose our population to biases such as the "healthy worker survival effect" whereby exposed men who were eligible for recruitment may have relatively low susceptibility to the effects of benzene (Garcia and Checkoway 2003). This bias may have underestimated associations. In addition, there may be uncontrolled confounding due to inherent differences between the exposed and unexposed populations. CONCLUSIONS We found that that occupational exposure to benzene, even at levels at or below the US PEL, was associated with chromosomal abnormalities in sperm that have been associated with infertility, spontaneous abortions, and with mental retardation and inherited defects in children. Our findings suggest that benzene may be a possible risk factor for de novo 1p36 deletion syndrome and raise the possibility that occupational exposure to benzene may induce persistent chromosomal rearrangements in stem cells/spermatogonia that would continue to generate unbalanced gametes throughout the reproductive life of the exposed men even after they are no longer exposed to benzene. These results, as well as our previous findings (Xing et al. 2010), indicate a need for more studies of the potential effects of low exposures ( 1ppm) to benzene, especially chromosomal abnormalities in the germ line of exposed men, and a re-evaluation of the current occupational exposure limit of 1ppm by US OSHA. 19 Page 20 of 26 REFERENCES Adler ID. 1996. Comparison of the duration of spermatogenesis between male rodents and humans. Mutat Res 352(1-2):169-172. Bosch M, Rajmil O, Egozcue J, Templado C. 2003. Linear increase of structural and numerical chromosome 9 abnormalities in human sperm regarding age. Eur J Hum Genet 11(10):754-759. Chandley AC. 1991. On the parental origin of de novo mutation in man. J Med Genet 28(4):217223. Chen H, Rupa DS, Tomar R, Eastmond DA. 1994. Chromosomal loss and breakage in mouse bone marrow and spleen cells exposed to benzene in vivo. Cancer Res 54(13):3533-3539. Crow JF. 2001. The origins, patterns and implication of human spontaneous mutations. Nature Reviews Genetics 1:40-47. Garcia AM, Checkoway H. 2003. A glossary for research in occupational health. J Epidemiol Community Health 57(1):7-10. Glass DC, Gray CN, Jolley DJ, Gibbons C, Sim MR, Fritschi L, et al. 2003. Leukemia risk associated with low-level benzene exposure. Epidemiology 14(5):569-577. Glass DC, Gray CN, Jolley DJ, Sim MR, Fritschi L. 2004. Benzene Exposure and Leukemia. Epidemiology 15(4):510-511. Hayes RB, Songnian Y, Dosemeci M, Linet M. 2001. Benzene and lymphohematopoietic malignancies in humans. American Journal of Industrial Medicine 40(2):117-126. Hayes RB, Yin SN, Dosemeci M, Li GL, Wacholder S, Travis LB, et al. 1997. Benzene and the dose-related incidence of hematologic neoplasms in China. Chinese Academy of Preventive Medicine--National Cancer Institute Benzene Study Group. J Natl Cancer Inst 89(14):1065-1071. Hornung RW, Reed LD. 1990. Estimation of Average Concentration in the Presence of Nondetectable Values. Applied Occupational and Environmental Hygiene 5(1):46-51. Hricko A. 1994. Rings of controversy around benzene. Environ Health Perspect 102(3):276-281. Jacobs P, Hassold T, Harvey J, May K. 1989. The origin of sex chromosome aneuploidy. Prog Clin Biol Res 311:135-151. 20 Page 21 of 26 Li X, Zheng LK, Deng LX, Zhang Q. 2001. Detection of numerical chromosome aberrations in sperm of workers exposed to benzene series by two-color fluorescence in situ hybridization [Chinese]. Yi Chuan Xue Bao 28(7):589-594. Liang YX, Wong O, Armstrong T, Ye XB, Miao LZ, Zhou YM, et al. 2005. An overview of published benzene exposure data by industry in China, 1960-2003. Chemico-biological interactions 153-154:55-64. Liu S, Zheng L, Deng L, Tang G, Zhang Q. 2000. Detection of numerical chromosome aberrations in sperm of workers exposed to benzene series by two-color fluorescence in situ hybridization [Chinese]. Zhonghua Yu Fang Yi Xue Za Zhi 34(1):17-19. Liu XX, Tang GH, Yuan YX, Deng LX, Zhang Q, Zheng LK. 2003. Detection of the frequencies of numerical and structural chromosome aberrations in sperm of benzene series-exposed workers by multi-color fluorescence in situ hybridization [Chinese]. Yi Chuan Xue Bao 30(12):1177-1182. Marchetti F, Wyrobek AJ. 2005. Mechanisms and consequences of paternally-transmitted chromosomal abnormalities. Birth Defects Res C Embryo Today 75(2):112-129. Marchetti F, Wyrobek AJ. 2008. DNA repair decline during mouse spermiogenesis results in the accumulation of heritable DNA damage. DNA Repair (Amst) 7(4):572-581. Mendelsohn ML, Mayall BH, Bogart E, Moore DH, 2nd, Perry BH. 1973. DNA content and DNA-based centromeric index of the 24 human chromosomes. Science 179(78):11261129. Olsen AK, Lindeman B, Wiger R, Duale N, Brunborg G. 2005. How do male germ cells handle DNA damage? Toxicol Appl Pharmacol 207(2 Suppl):521-531. OSHA (Occupational Safety and Health Administration). 1987. Occupational Safety and Health Standards. Regulations-29 CFR, Standard number 1910.1028. Rudak E, Jacobs PA, Yanagimachi R. 1978. Direct analysis of the chromosome constitution of human spermatozoa. Nature 274(5674):911-913. Shaffer LG, Heilstedt HA. 2001. Terminal deletion of 1p36. Lancet 358 Suppl:S9. Shelby MD, Bishop JB, Mason JM, Tindall KR. 1993. Fertility, reproduction, and genetic disease: studies on the mutagenic effects of environmental agents on mammalian germ cells. Environ Health Perspect 100:283-291. 21 Page 22 of 26 Sloter E, Nath J, Eskenazi B, Wyrobek AJ. 2004. Effects of male age on the frequencies of germinal and heritable chromosomal abnormalities in humans and rodents. Fertil Steril 81(4):925-943. Sloter ED, Lowe X, Moore ID, Nath J, Wyrobek AJ. 2000. Multicolor FISH analysis of chromosomal breaks, duplications, deletions, and numerical abnormalities in the sperm of healthy men. Am J Hum Genet 67(4):862-872. Song B, Cai ZM, Li X, Deng LX, Zheng LK. 2005. Effect of benzene on sperm DNA [Chinese]. Zhonghua Nan Ke Xue 11(1):53-55. Van Hummelen P, Lowe XR, Wyrobek AJ. 1996. Simultaneous detection of structural and numerical chromosome abnormalities in sperm of healthy men by multicolor fluorescence in situ hybridization. Hum Genet 98(5):608-615. Van Hummelen P, Manchester D, Lowe X, Wyrobek AJ. 1997. Meiotic segregation, recombination, and gamete aneuploidy assessed in a t(1;10)(p22.1;q22.3) reciprocal translocation carrier by three- and four-probe multicolor FISH in sperm. Am J Hum Genet 61(3):651-659. Waye JS, Durfy SJ, Pinkel D, Kenwrick S, Patterson M, Davies KE, et al. 1987. Chromosomespecific alpha satellite DNA from human chromosome 1: hierarchical structure and genomic organization of a polymorphic domain spanning several hundred kilobase pairs of centromeric DNA. Genomics 1(1):43-51. Whysner J, Reddy MV, Ross PM, Mohan M, Lax EA. 2004. Genotoxicity of benzene and its metabolites. Mutat Res 566(2):99-130. W.H.O. (World Health Organization). 1992. WHO laboratory manual for the examination of human semen and sperm-cervical mucus interaction. 3rd ed. Cambridge, UK: Cambridge University Press. Xing C, Marchetti F, Li G, Weldon RH, Kurtovich E, Young S, et al. 2010. Benzene exposure near the U.S. permissible limit is associated with sperm aneuploidy. Environ Health Perspect 118(6):833-839. Yin SN, Hayes RB, Linet MS, Li GL, Dosemeci M, Travis LB, et al. 1996. An expanded cohort study of cancer among benzene-exposed workers in China. Benzene Study Group. Environmental Health Perspectives 104 Suppl 6(24):1339-1341. 22 Page 23 of 26 Yin SN, Li GL, Tain FD, Fu ZI, Jin C, Chen YJ, et al. 1987a. Leukaemia in benzene workers: a retrospective cohort study. British Journal of Industrial Medicine 44(2):124-128. Yin SN, Li Q, Liu Y, Tian F, Du C, Jin C. 1987b. Occupational exposure to benzene in China. British Journal of Industrial Medicine 44(3):192-195. Zenker M, Rittinger O, Grosse KP, Speicher MR, Kraus J, Rauch A, et al. 2002. Monosomy 1p36--a recently delineated, clinically recognizable syndrome. Clin Dysmorphol 11(1):43-48. Zhang L, Eastmond DA, Smith MT. 2002. The nature of chromosomal aberrations detected in humans exposed to benzene. Crit Rev Toxicol 32(1):1-42. Zhang L, Lan Q, Guo W, Hubbard AE, Li G, Rappaport SM, et al. 2011. Chromosome-wide aneuploidy study (CWAS) in workers exposed to an established leukemogen, benzene. Carcinogenesis 32(4):605-612. Zhang L, Lan Q, Guo W, Li G, Yang W, Hubbard AE, et al. 2005. Use of OctoChrome fluorescence in situ hybridization to detect specific aneuploidy among all 24 chromosomes in benzene-exposed workers. Chem Biol Interact 153-154:117-122. Zhao T, Liu XX, He Y, Deng LX, Zheng LK. 2004. Detection of numerical aberrations of chromosomes 7 and 8 in sperms of workers exposed to benzene series by two-color fluorescence in situ hybridization [Chinese]. Zhonghua Yi Xue Yi Chuan Xue Za Zhi 21(4):360-364. 23 Page 24 of 26 Table 1. Distributions of benzene exposure measurements (air, urinary benzene, and urinary muconic acid (MA)) among benzeneexposed and unexposed workersa. Exposure measurement N GM (GSD) min p10 p25 p50 p75 p90 max Urinary benzene (g/L) Unexposed Exposedc 11 0.1 30 14.9 (1.3) 0.1 0.1 0.1 0.1 0.2 0.2 0.2 (5.0) 0.8 2.4 4.3 11.0 59.0 120.5 617.0 Low-exposed 10 2.8 (1.7) 0.8 1.4 2.4 2.9 4.3 5.0 5.4 Moderate-exposed 10 11.6 (1.6) 6.7 7.0 7.2 11.0 21.0 21.3 21.3 High-exposed 10 102.4 (2.1) 48.5 49.2 59.0 110.6 123.1 373.9 617.0 Total 41 4.1 (12.8) 0.1 0.1 0.2 5.4 21.3 114.9 617.0 Air (ppm) Unexposed Exposedb,c Low-exposedb Moderate-exposed High-exposed Totalb 11 -30 2.8 10 1.0 10 3.0 10 7.6 41 1.3 -(3.6) (2.6) (3.4) (2.2) (5.6) <LOD <LOD <LOD 0.6 2.0 <LOD <LOD 0.7 0.3 0.7 2.8 <LOD <LOD 1.0 0.7 0.8 4.4 <LOD <LOD 3.3 1.2 3.7 8.4 1.3 <LOD 7.0 2.1 7.0 14.1 4.7 <LOD 16.3 3.1 17.2 19.4 11.6 <LOD 23.6 4.1 23.6 20.2 23.6 Muconic acid (mg/L) Unexposedd Exposedc 0 -30 5.3 -- -- -- -- -- -- -- -- (3.3) 0.8 1.2 1.9 6.6 13.8 25.2 40.9 Low-exposed 10 1.7 (1.6) 0.8 0.9 1.2 1.7 2.7 3.2 3.3 Moderate-exposed 10 5.4 (3.1) 1.1 1.3 1.9 7.9 14.4 19.1 23.7 High-exposed 10 15.6 (1.8) 6.5 7.6 11.2 13.4 26.6 34.0 40.9 Total 30 5.3 (3.3) 0.8 1.2 1.9 6.6 13.8 25.2 40.9 a Number of men (N), Geometric Mean (GM), Geometric Standard Deviation (GSD), percentiles (p) and range of concentrations (min, max) among men. Urine samples and personal air measurements were obtained from each man at two time points approximately one month apart, and the GMs of the paired measurements were used to calculate summary statistics. b To estimate the GM and GSD, values below the LOD (all 11 unexposed men and one low-exposed man) were imputed as LOD/sqrt(2). c Exposure categories are based on tertiles of urinary benzene concentrations among exposed men (low: 0.8-5.4, moderate: 6.7-21.3, high: 48.5-617.0 g/L). d MA was not measured in the unexposed group. 24 Page 25 of 26 Table 2. Median (p50), mean, range of anomaly frequencies (counts/10,000 sperm)a and percent of men with at least one abnormality detected as determined by ACM sperm FISH of chromosome 1, stratified by benzene exposure group. Unexposed (N=11) Low-exposed (N=10) Moderate-exposed (N=10) High-exposed (N=10) % men % men % men % men with with with Outcome anomalyb p50 Mean Range anomaly p50 Mean Range anomaly p50 Mean with Range anomaly p50 Mean Range Total anomalies 100 22.9 22.613.0- 42.6 100 27.9 28.216.9- 50.3 100 29.9 30.0 17.9- 48.8 100 33.2 34.815.9- 48.8 Total structural aberrations 100 12.9 12.8 6.9- 19.0 100 15.4 17.111.0- 26.6 100 19.4 17.9 11.0- 23.9 100 21.4 21.211.9- 32.9 Total 1p36.3 & 1cen dup/del 73 2.0 2.3 0.0- 7.9 100 4.5 4.9 1.0- 10.0 100 6.0 5.4 2.0- 10.9 100 6.0 5.9 3.0- 9.0 Total dups 64 1.0 1.9 0.0- 6.9 100 2.5 3.7 1.0- 10.0 100 3.5 3.9 1.0- 8.9 100 4.0 3.9 1.0- 6.0 Total dels 36 0.0 0.4 0.0- 2.0 60 1.0 1.2 0.0- 4.0 80 1.5 1.5 0.0- 3.0 100 1.0 2.0 1.0- 5.0 Total 1p36.3 dup/del 64 1.0 1.3 0.0- 5.0 100 2.5 3.4 1.0- 9.0 90 3.5 3.3 0.0- 5.0 100 4.0 4.3 2.0- 8.0 1p36.3 dup 64 1.0 1.0 0.0- 4.0 70 1.5 2.2 0.0- 6.0 80 2.0 1.9 0.0- 4.0 100 2.5 2.6 1.0- 6.0 1p36.3 del 27 0.0 0.3 0.0- 1.0 60 1.0 1.2 0.0- 4.0 80 1.5 1.4 0.0- 3.0 80 1.0 1.7 0.0- 4.0 Total 1cen dup/del 55 1.0 1.1 0.0- 3.0 60 1.5 1.5 0.0- 4.0 80 1.5 2.1 0.0- 6.9 90 1.0 1.6 0.0- 4.0 1cen dup 45 0.0 0.9 0.0- 3.0 60 1.5 1.5 0.0- 4.0 70 1.5 2.0 0.0- 6.9 60 1.0 1.3 0.0- 4.0 1cen del 9 0.0 0.2 0.0- 2.0 0 0.0 0.0 0.0- 0.0 10 0.0 0.1 0.0- 1.0 30 0.0 0.3 0.0- 1.0 Total breaks 100 8.9 10.5 5.8- 18.0 1q12c 100 3.0 4.0 1.0- 13.0 1cen-1q12d 100 5.0 6.5 3.9- 14.0 100 12.4 12.2 8.0- 19.7 70 3.5 4.1 0.0- 11.0 100 8.9 8.2 3.0- 11.0 100 12.9 12.5 7.0- 17.9 100 4.5 4.9 1.0- 9.0 100 7.9 7.7 3.0- 11.0 100 14.4 15.3 8.9- 23.9 100 7.9 7.7 3.0- 12.9 100 6.5 7.6 3.9- 13.8 Total numerical aberrations 100 8.0 9.5 1.0- 27.7 100 9.5 10.7 2.0- 22.7 100 10.5 11.6 5.0- 25.9 100 13.8 12.7 3.0- 20.9 disomy 1 100 8.0 9.4 1.0- 27.7 100 9.0 10.6 2.0- 22.7 100 10.5 11.6 5.0- 25.9 100 13.4 12.6 3.0- 20.9 nullisomy 1 9 0.0 0.1 0.0- 1.0 10 0.0 0.1 0.0- 1.0 0 0.0 0.0 0.0- 0.0 10 0.0 0.1 0.0- 1.0 Othere 18 0.0 0.3 0.0- 2.0 40 0.0 0.4 0.0- 1.0 40 0.0 0.4 0.0- 1.0 50 0.5 0.9 0.0- 4.0 a Sperm structural and numerical defects were determined using the ACM sperm FISH assay of chromosome 1. Frequencies per 10,000 sperm counted are reported. Median and mean frequencies include all participants and men without a detected anomaly were assigned a value of zero. b i.e., percent of men with at least one sperm with this defect per 10,000 sperm analyzed; c breaks within 1q12; d breaks between 1cen and 1q12; e other is defined as all anomalies that are not detailed above. 25 Page 26 of 26 Table 3. Adjusteda associations between urinary benzene exposureb and sperm ACM-FISH outcomes. Outcome Total anomalies Low vs. unexposed Moderate vs. unexposed IRR (95% CI) p-value IRR (95% CI) p-value 1.33 (1.00, 1.75) 0.05 1.38 (1.05, 1.82) 0.02 High vs. unexposed IRR (95% CI) p-value 1.69 (1.28, 2.23) <0.005 ptrendf <0.005 Total structural aberrations 1.42 (1.10, 1.83) 0.01 1.44 (1.12, 1.85) 0.01 1.75 (1.36, 2.24) <0.005 <0.005 Total 1p36.3 & 1cen dup/del 2.22 (1.35, 3.66) <0.005 2.53 (1.56, 4.11) <0.005 2.96 (1.81, 4.85) <0.005 <0.005 Total duplications 2.19 (1.15, 4.17) 0.02 2.31 (1.23, 4.33) 0.01 2.51 (1.31, 4.82) 0.01 0.01 Total deletions 2.66 (0.91, 7.76) 0.07 3.68 (1.31, 10.33) 0.01 5.25 (1.90, 14.55) <0.005 <0.005 Total 1p36.3 dup/del 2.72 (1.43, 5.19) <0.005 2.92 (1.54, 5.54) <0.005 4.07 (2.16, 7.66) <0.005 <0.005 1p36.3 duplications 2.31 (1.03, 5.18) 0.04 2.13 (0.95, 4.79) 0.07 3.12 (1.39, 6.97) 0.01 0.01 1p36.3 deletions 4.31 (1.18, 15.78) 0.03 6.02 (1.69, 21.39) 0.001 7.88 (2.21, 28.05) <0.005 <0.005 Total 1cen dup/del 1.67 (0.70, 4.01) 0.25 2.07 (0.91, 4.73) 0.08 1.75 (0.71, 4.27) 0.22 0.19 1cen duplications 1cen deletionsc 2.00 (0.69, 5.82) 0.20 2.47 (0.87, 6.97) 0.09 1.75 (0.57, 5.44) 0.33 -- -- -- -- -- -- -- -- -- 0.28 -- Total breaks 1q12d 1cen-1q12e 1.24 (0.91, 1.69) 1.29 (0.68, 2.45) 1.22 (0.86, 1.74) 0.18 0.44 0.26 1.20 (0.89, 1.63) 1.27 (0.69, 2.34) 1.17 (0.83, 1.67) 0.24 0.44 0.37 1.49 2.23 1.14 (1.10, 2.02) (1.22, 4.09) (0.79, 1.63) 0.01 0.01 0.49 0.02 0.01 0.56 Total numerical aberrations 1.21 (0.71, 2.07) 0.49 1.32 (0.78, 2.25) 0.30 1.59 (0.92, 2.76) 0.10 0.10 disomy 1 nullisomy 1c 1.22 (0.71, 2.10) 0.47 1.35 (0.79, 2.30) 0.28 1.61 (0.92, 2.81) 0.10 -- -- -- -- -- -- -- -- 0.10 -- Other 1.40 (0.28, 6.92) 0.68 1.34 (0.28, 6.47) 0.72 2.95 (0.71, 12.30) 0.14 0.13 a Multivariable negative binomial models were used to estimate Incidence Rate Ratios (IRRs). IRRs represent comparisons of counts/10,000 sperm. All models were adjusted for age, smoking or alcohol consumption in the three months prior to semen collection and history of any chronic disease. b Urinary benzene concentrations (summarized by the GM of the two measurements) among the exposed were divided into three groups with ten men in each. Statistical models compared each exposure group with the unexposed group. cModels were not constructed due to low detection frequency. d breaks within 1q12; e breaks between 1cen and 1q12; f ptrend is the p-value given by an adjusted negative binomial regression model where the urinary benzene explanatory variable was coded as 0 for unexposed, 1 for low-exposed, 2 for moderate-exposed and 3 for high-exposed. 26