Hauser, R, Z Chen, L Pothier, L Ryan and L Altshul. 2003. The relationship between human semen parameters and environmental exposure to polychlorinated biphenyls and p,p’-DDE. Environmental Health Perspectives, on line 19 May 2003.

Hauser et al. demonstrate a significant dose-response relationship between PCB 138 and sperm count, mobility and morphology. Higher exposures were associated with poorer semen quality. They also found suggestions of associations with other PCB congeners and with the sum of PCB congeners, but not as strong. Limited evidence of association between DDE levels and sperm motility emerged from their analysis.

What did they do? Hauser et al. determined sperm count, motility and % normal sperm in a sample of men who had come with their partner to a fertility lab at Massachusetts General Hospital for treatment of infertility. At the time of enrollment, it was not known whether the couple's infertility was due to male or female impairment.

Hauser et al. took blood the same day they obtained sperm samples, and analyzed the serum for p,p'-DDE and 57 different PCB congeners.

To examine the relationship between the contaminants and sperm parameters, they first determined which men had sperm characteristics indicating fertility problems. They used criteria established by the World Health Organization, based on sperm count, sperm motility and sperm morphology.

They then studied whether the likelihood that a man would have sperm characteristics beneath these reference criteria was related to the concentration of contaminants in his serum. They limited these analyses to 3 PCB congeners (118, 138, and 153), p,p'-DDE, and sums of PCB congeners.

To do this for a given chemical, they divided the sample of men into tertiles based on their measured contamination, and then calculated the odds ratio of subfertile sperm for each tertile. The odds ratios were corrected for abstinance time, age and smoking status.

What did they find? In the 212 participants, 46% had sperm measurements above the reference criteria for all three measurements. The remainder all had one or more sperm measurements beneath reference criteria: 19% with sperm count beneath 20 million per ml; 46% with less than 50% motile sperm; and 27% with less than 4% of sperm shaped normally.

Most of their sample (79%) was composed of Caucasian men. They averaged 36 yrs of age. Most (73%) had never smoked.

One PCB congener, PCB 138, showed a strong dose-response relationship with sperm motility and sperm morphology and a lesser (non-significant) trend for sperm count. Men with higher levels of PCB 138 were more likely to have sperm parameters beneath reference values. OR's calculated using men with all sperm values above reference level as base. Adapted from Hauser et al. 2003, Table 3.

Neither of the other individual PCB congeners (118 and 153) were associated with changes in odds-ratios. There was weak evidence for an association with DDE, with higher in the middle and highest quartiles.

In a multivariate analysis using sperm count as a continous variable (actual measurement) instead of as a dichotomous variable (above or below the reference value), Hauser et al. found a strong inverse relationship between PCB 138 and sperm count: the more PCB 138, the lower the sperm count(p < 0.008).This multivariate regression included age, abstinence time and smoking history as potential confounders.

Hauser et al. also combined the PCB measurements into groups based on their mode of action: estrogen-like PCBs, dioxin-like PCBs, and cytochrome C P450 enzyme-inducing PCBS. For this analysis, they summed the concentrations of congeners within each group and then determined, as above, whether the odds-ratios of different tertiles indicated an association with sperm parameters.

In this analysis, only enzyme-inducing PCBs revealed a dose-response pattern. Odds ratios for both poor motility and deformed morphology rose with PCB tertile. These trends were significant before adjustment for confounding variables, but not afterward, although the trends remained visible in the data after adjustment.

What does it mean? These results add weight to prior findings indicating that PCB levels may affect sperm parameters adversely, at levels experienced by many people. The nature of the current study design, however, prevents any conclusions about causality.

This research draws attention in particular the importance of analyzing different congeners separately, or in biochemically coherent subgroups, as it appears from these results that effects on sperm vary from congener to congener.

Thus the effort by Hauser et al. to examine together congeners that share modes of action ("dioxin-like," estrogenic, enzyme inducing) is an important advance. Refinements in this approach may allow detection of trends that in this first effort are not apparent.

According to the authors, this research is continuing, and additional publications will follow with larger samples and more sophisticated analysis.