Swan, SH, RL Kruse, L Fan, DB Barr, EZ Drobnis, JB Redmon, C Wang, C Brazil and JW Overstreet and the Study for the Future of Families Research Group. 2003. Semen quality in relation to biomarkers of pesticide exposure. Environmental Health Perspectives 111:1478-1484.

Press coverage:
Los Angeles Times
USA Today

In this paper, Swan et al. report that the risk of low sperm quality in Missouri men is strongly associated with pesticide exposures, especially alachlor, diazinon and atrazine. While the sample size is relatively small, the odds ratios revealed by their research are extraordinarily high and very unlikely to be the result of chance. The scientists conclude that exposures to these agricultural chemicals is most likely to be through contaminated drinking water.

This is the first time scientists have shown a link between levels of current, widely used pesticides and semen quality. Its importance is magnified even more because it is also the first to show an effect in the general population, not just in farmers or in pesticide workers.

The strongest association was with alachlor. Men with the highest alachlor levels were 30x more likely to have low sperm quality. This result was highly significant statistically, with the probability of it occuring by chance calculated to be less than 1 in 1000 (p = 0.0007).

What did they do? Swan's team obtained semen and urine samples from 426 men living in rural Missouri (208) and urban Minneapolis (218), each of whom was a partner of a pregnant woman. Thus by definition the men were fertile. The first stage of this analysis, describing the subjects and documenting geographic differences, was published earlier in 2003. Each of the semen samples was analyzed for sperm count (sperm per milliliter), sperm morphology (% normal sperm) and sperm motility (% motile sperm).

The current study carries that analysis farther. From the initial sample, Swan et al. selected a subset of the men with the highest (52 men, of them 25 from Missouri) and the lowest sperm counts (34 men, 25 from Missouri) and analyzed their urine for traces of 15 pesticides many in widespread use today.

They then used a variety of analyses to examine associations between pesticide level and sperm quality.

What did they find? Their first analysis showed that men in Missouri were exposed to a different suite of pesticides than those in Minneapolis. This is not surprising, as most of the Missouri men were from rural areas in the midst of agriculture, where most Minneapolis men were from an urban setting. And because of crop and climatological differences between Missouri and Minneapolis, the mix of pesticides applied agriculturally would be expected to differ between the samples also.

Five pesticides, including the herbicides alachlor and atrazine and the insecticide diazinon, were much more likely to be encountered at higher levels in Missouri than in Minneapolis.

They then turned their attention to links between pesticides and semen quality. The results were striking: Men with higher levels of alachlor, atrazine and diazinon were significantly more likely to have lower sperm counts than men with lower levels of pesticides in their urine. The risk of poor semen quality was elevated 30-fold with higher alachlor levels; 17-fold with IMPY (the diazinon metabolite), and 12-fold with higher atrazine levels. These odds ratios were highly statistically significant (p= 0.0007, p= 0.0004 and p= 0.01, respectively).

The results suggested reduced semen quality was linked to two other herbicides, metolachlor and 2,4-D, also, but the pattern was of only borderline statistical significance. Further research with a larger sample would be necessary to determine whether these associations were real. For one pesticide, acetochlor, the data suggested that greater exposure was linked to higher sperm count, but this too was of borderline significance.

They saw no link between exposure and sperm quality for DEET or malathion.

Based upon interview data, few of the men could have been exposed to the agricultural pesticides through occupation. Swan et al. conclude that the likely route of exposure is via drinking water. This is consistent with the fact that these pesticides are known to be common in drinking water sources in the mid-West, and that they are not removed by usual water treatment methods.

What does it mean? As noted above, this study is important because it links, for the first time, reduction in sperm quality to exposure to several pesticides in widespread use today, and it does so in a study of men selected from the general population instead of men likely to be exposed occupationally (e.g., farmers or pesticide workers). The exposures, moreover, are most likely through drinking water.

The associations are very strong for alachlor, diazinon and atrazine. Odds ratios of this magnitude are rarely seen in epidemiological studies of the general population, and are on par with the elevation in risk linking lung cancer to cigarette smoking. That association is now firmly established by many studies of many different types, a depth of proof that is still lacking for this new work. The magnitude of the odds ratios reported by Swan et al. should force public health authorities to take immediate note of these results, invigorate research into the issue, and stimulate evaluations of public health measures to be taken as scientific research continues.

What does this mean for couples attempting to conceive? This can't be answered definitively from the current study. The way that Swan et al. constructed their case and control groups was to draw them from partners of pregnant women, i.e., men who were confirmed to be fertile. By definition, the reductions in sperm quality in this men was insufficient to cause infertility--although whether it took longer for the partners of men with low sperm quality to become pregnant was not assessed.

To determine whether or not alachlor, atrazine and diazinon can reduce sperm count sufficiently to impair fertility will require repeating the study with a sample of men from couples seeking infertility treatment. In these couples, some cases of infertility will be due to male problems, others to female. Then when an analysis like Swan et al.'s is performed, sperm quality should span a range from high to low, including sufficiently low to impair fertility. The key question will be whether the cases of male infertility are associated with high exposures, perhaps exposures beyond the levels seen in this study. It is entirely plausible, for example, that by restricting their sampling to partners of pregnant women, Swan et al. inadvertently excluded men with even higher levels of pesticide exposure, if those caused complete fertility impairment.

Swan et al. conclude with an understatement: "If further study confirms these findings, the implications for public health and agricultural practice could be considerable." The findings are so striking that even without scientific certainty of causation, they pose an important challenge to public health authorities.