Duty, SM, MJ Silva, DB Barr, JW Brock, L Ryan, Z Chen, RF Herrick, DC Christiani and R Hauser 2003. Phthalate Exposure and Human Semen Parameters. Epidemiology 14:269 –277. [note]

Two other phhtalate-sperm studies:
In India
In Boston

Laboratory experiments with rodents have shown repeatedly that phthalate exposures can adversely affect sperm quality. Critics have argued that these results are not relevant to humans, either because of differences in reproductive physiology or because environmental exposures experienced by people are beneath those causing effects in rodents.

Conducted by scientists from the Centers for Disease Control, Harvard School of Public Health, the Dana Farber Cancer Institute and Harvard Medical School, this study dramatically undermines those critics by showing that men with higher phthalate levels have reduced sperm counts, lower sperm motility and more deformed sperm. The levels observed in the men are well within the range of exposures seen in the American public. Different types of phthalates had different impacts.

What did they do? Duty et al. obtained sperm and urine samples from 168 men who were partners in infertile couples coming for treatment to an infertility clinic at Massachusetts General Hospital. Participants also responded to a questionnaire about lifestyle factors and medical history. Responses were used to control for potentially confounding variables. It was not known at the time of enrollment whether the couple's infertility was due to male or female complications.

From semen samples the research team determined sperm concentration (sperm count), sperm motility and the percentage of deformed sperm in samples. Urine analysis yielded measurements of the concentrations of 8 different phthalate metabolites. Of those 8, 5 were included in the statistical analysis. The other 3 were not because in most men levels were quite low, indeed often undetectable.

The 5 metabolites measured were:

What did they find? In analyzing the men's sperm, Duty et al. 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.

Of the 168 men studied, 54% had sperm characteristics lower than WHO criteria on at least one of the variables measured: 17% had sperm counts beneath 20 million per ml, 44% had less than 50% motile sperm and 26% had less than 4% normally-shaped sperm. Some men were suboptimal on more than one parameter.

Most of the men were white (77%) and had never smoked. They averaged 34.6 years of age.

To study the relationship between phthalate levels and sperm characteristics, Duty et al. carried out two types of analysis.

In the first, they divided the men into two groups based on sperm parameters (a high group, composed of men with sperm parameters higher than the median, vs. a low group, i.e., men beneath the median) and two groups based on phthalate measurements (a high group, composed of men with phthalate parameters higher than the median, vs. a low group). They then determined for different sperm measurements whether men in the high sperm group were more or less likely to be in a high phthalate group.

The results of this analysis showed that men with high MBP levels were more than twice as likely to have low sperm motility (adjusted odds ratio= 2.4; 95% CI = 1.1 to 5.0). High MBP levels were also associated with reduced sperm concentration (OR= 2.4) and higher percentage of deformed sperm (OR = 1.7).

Duty et al. also found associations between MMP and sperm morphology and MBzP and sperm count, motility and morphology.

In the second round of analysis, Duty et al. looked for dose response relationships between sperm parameters and phthalate exposure. To do this, they divided the men into tertiles based on phthalate levels and then after adjusting for confounding variables, .

In this analysis, three trends emerged. Men with higher MBzP or higher MBP were more likely to have lower sperm count. Men with higher MBP were more likely to have lower sperm motility.

What does it mean? These results demonstrate associations between phthalate levels commonly experienced by the American public and impaired sperm quality. They are consistent with laboratory results studying the effects of phthalates in animals, although they suggest that humans may be more sensitive than rodents.

A key point is that there is nothing out of the ordinary for the phthalate levels observed in the patients studied by Duty et al. The levels are somewhat lower than those reported in national surveys by the US CDC. Indeed, while the comparison is complicated because of differences in how the samples were composed, the CDC data indicate that American men may often carry levels 2-3x those reported in this study. This would suggest that if the associations revealed by Duty et al. are causal in nature, many American men have sperm parameters affected by phthalate exposure.

As always, there are ways that this study can be criticized.

• Measurements of sperm parameters are difficult; single samples of sperm may not present an accurate snapshot of a man's reproductive condition.
• One time measurements of phthalates may not be representative of a man's exposures over time, because phthalates have a relatively short half-life in the human body.
• The men in the sample may not be representative of the population... although there is no reason to suggest they are not representative of men in couples with fertility problems.

Crucially, the first two of these potential weaknesses would work against finding significant relationships within the data, because they introduce random measurement error into the system. There is no plausible reason why measurement errors of the sort caused by these limitations would produce be expected to yield a false positive. Thus if anything, the underlying pattern of association between phthalate exposure and sperm parameters is likely to be stronger than that reported in this article.

In an editorial accompanying this article and a related article in Epidemiology, Univ. Missouri reproductive epidemiologist Dr. Shanna Swan notes that the evidence is now good, albeit not certain, that environmental agents, even at low levels, can impair semen quality. Teasing apart and proving conclusively the contribution of any given agent will be especially difficult because of the wide array of chemicals to which people are regularly exposed, each of which has been linked by laboratory animals to adverse effects on sperm condition.

Her comments also place these new results in the broader context of debate over whether broad declines in sperm count have occurred, regionally or globally. This debate, kicked off by an analysis of global trends carried out by a team of Danish scientists, ultimately led to the conclusion that no firm conclusion could be reached even though existing historical data were consistent with a decline, because of the limitations of a retrospective analysis of data gathered using different methods for different purposes over a long period of time. This realization forced scientists to shift their focus away from long-term trends and toward geographic variation, because while the past could not be reconstructed, current geographic patterns could be examined.

Such studies are currently underway, for example, a recent report of large differences in sperm count among men from different regions of the United States. Work on geographic differences, coupled with work like the current paper by Duty et al., should shed light on what is causing the geographic patterns and then allow firmer conclusions to be reached about changes over time.

[note] Epidemiology prevents direct links to individual abstracts of published articles on its website. The journal's home page is: www.epidem.com. The abstract and (for subscribers) full text of this article can be found by browsing through published issues of the journal.