Authors
Swan SH, Kruse RL, Liu F, Barr DB, Drobnis EZ, Redmon JB, Wang C, Brazil, Overstreet JW, The Study for Future Families Research Group

Title
Semen quality in relation to biomarkers of pesticide exposure

Journal
Environmental Health Perspectives, Online June 18, 2003

Summary
The process of spermatogenesis, whereby spermatozoa are produced, is under close hormonal control. Therefore it is a potential target for endocrine disruption. Epidemiological studies have suggested that there are regional differences in semen quality but the role of the environment, if any, remains inconclusive. Occupational exposures to some pesticides have been shown to alter sperm counts and quality however, the mechanism remains uncertain. Previous studies in Columbia, Missouri have shown that men from agrarian areas have reduced sperm concentration and motility compare to men from urban areas (Los Angeles, CA, Minneapolis, MN and New York, NY). The current study was designed to test the hypothesis that pesticides used commonly in mid-Missouri and rarely used in urban counties, contributed to these observed differences in semen quality.

Subjects were all participants from The Study for Future Families. Men who were partners of pregnant women were recruited between 1999 and 2001 at prenatal clinics. Both partners completed a questionnaire and gave blood samples. The men received a physical exam and provided two semen samples (24 hours apart) and urine samples. Almost all semen samples were analyzed within 45 minutes of sampling and the urine was subsequently analyzed for 8 non-persistent pesticide metabolites commonly used in the Missouri area. After adjusting for abstinence time, cases were defined as men who had an average sperm concentration below the population mean and controls were men whose average sperm concentrations were above the population mean. In total, pesticide metabolite levels were obtained on 86 men, with 50 from rural Missouri (25 cases and 25 controls) and 36 from urban Minnesota (9 cases and 27 controls). The recruitment rates were both 27% for both Missouri and Minnesota.

Metabolites from the pesticides alachlor, diazinon, atrazine, metolochlor were detected more often in men from Missouri than Minnesota. Conversely, the pesticide metabolite 3,5,6-trichloropyridinol from the parent chemical chlorpyrifos/chlorpyrifos methyl, and 4-nitrophenol from the parent compound methyl parathion were detected more frequently in men from Minnesota.

In Missouri, the metabolites alachlor (p=0.0007), diazinon metabolite IMPLY (p=0.0004), atrazine (p=0.01) and acetochlor (p=0.04) were significantly higher in the urine of men with low semen quality than controls from Missouri. Yet, acetechlor levels were lower in cases than controls (p=0.04). The analysis was then repeated excluding 16 men with one or more risk factors for poor semen quality including: cryptorchidism, smoking, history of STD, history of infertility, recent fever, age over 40, and non-white race, but no substantial change was seen in the effect estimates. In Minnesota, no significant difference in pesticide metabolites was seen between cases and controls.

In further analysis, metabolite levels were categorized into two (low and high) or three (low, medium and high) groups. In Missouri, men with medium (OR=6.3, 95% CI=1.3-29.4) and high (OR=30, 95% CI=4.3-210) metabolite levels of alachlor were more likely to have lower semen quality compared to men with low levels. Similarly, an increased odds ratio was observed for medium (OR=10, 95% CI=2.0-49.2) and high (OR=16.7, 95% CI=2.8-98.0) levels of IMPLY, and men with high urine levels of atrazine (OR=11.3, 95% CI=1.3-98.9) compared to men with low levels of these same metabolites.

None of pesticides were significantly associated with any semen parameter (sperm concentration, morphology and motility) within the Minnesota group. However in the Missouri group, the herbicides alachlor and metoachlor were negatively associated with sperm morphology, and IMPLY with sperm concentration. However, statistical limitations (semen parameters were not normally distributed even after logarithmic transformation) limit the relevance of these findings.

Because the correlation among the metabolites was very high, the associations were not independent. Therefore it is likely the results may be due to multiple exposures. The authors examined concurrent exposures to multiple pesticides using a score ranging from 0 to 5. Within Missouri, exposure to 2-5 pesticide metabolites verses 0-1 significantly increased the odds of having poor semen quality (OR=3.4, 95% CI=2.0-25.5). A similar association was not found in Minnesota.

This is one of the first studies to demonstrate links between specific biomarkers of pesticide exposure and semen quality. However, these results should be viewed cautiously due in part to the low participation rate (27%), which may have introduced bias, and the imprecise nature of the effect estimates. As well, it is difficult to draw conclusions from epidemiological studies when there are multiple exposures to potentially toxic compounds that act through similar or divergent mechanisms. Nevertheless, given the wide-spread use of these pesticides in agriculture, similar studies are important to further investigate the link between pesticide exposure and semen quality.