Weidner, IS, H Møller, TK Jensen, and NE Skakkebæk. 1998. Cryptorchidism and hypospadias in sons of gardeners and farmers. Environmental Health Perspectives 106:793-796.

the molecular basis of cryptorchidism?
another epidemiological study of cryptorchidism

Weidner et al. use extensive health and occupational data from Danish population databases to test a prediction that maternal use of estrogenic or anti-androgenic pesticides may be a risk factor for cryptorchidism (undescended testes) and hypospadias in their sons. This prediction arises out of experiments with laboratory animals in which these two effects are reliably produced by in utero exposure to some estrogenic and anti-androgenic compounds (e.g, Gray and Kelce 1996).

In this study they find that a mother's involvement in gardening and farming is associated with increased risk of cryptorchidsm in her sons, but not of hypospadias. They observed no association with the father's occupation.

What did they do? Weidner et al. used data links in the Danish Population Register (which tracks all Danes through life) to the Danish National Patient Register and the Fertility Database at Statistics Denmark to construct a case-control study of the association between hypospadias and cryptorchidism and their parents' occupations during the year of conception. They classified the occupations as either in farming/gardening or not, reasoning that people in farming and/or gardening were likely to have an increased exposure to agricultural pesticides. Gardening included work in greenhouses, outdoor gardening, orchards, and nurseries, but not landscape gardening.

They included in their study all males born live in Denmark in the years 1983 to 1992 who were discharged from a Danish hospital with a diagnosis of hypospadias or cryptorchidism. They identified 6,177 cases with cryptorchidism and 1,345 cases with hypospadias. Ninety-two of the boys had both hypospadias and cryptorchidism; these were included independently in both data sets. A control group was selected at random from the entire population of boys born in Denmark between 1983 and 1992.

They tested separately the association between the mother's and father's' occupation and the likelihood of hypospadias or cryptorchidism. Statistical analyses of the data included included examination of a series of potential confounding variables, including year of birth, gestational age, parity, twin birth, a maternal history of previous stillbirth, parental age, nationality and professional status (self-employed, salaried employee, skilled worker, unskilled worker), and the occurrence of the same malformation in an older brother.

What did they find? Maternal involvement in gardening/farming was associated with an increased likelihood of cryptorchidism but not hypospadias:

What does it mean? Weidner et al. confirm one of the predictions of their hypothesis, that the boys of mothers engaged in activities that would increase the likelihood of exposure to estrogenic and/or anti-androgenic pesticides are more likely to be born with reproductive birth defects known to be caused in laboratory animals by these compounds.

The fact that the boy's father's occupation has no impact on risk is consistent with this interpretation because the posited mechanism is via in utero exposure.

Weidner et al. did not confirm the second prediction, that the risk of hypospadias should have increased for sons born to mothers engaged in gardening/farming. They conclude: "Either the exposure was insufficient, the association of hypospadias with gardening or farming was too weak to be detectable among the available number of cases potentially exposed to pesticides, or no such association truly exists in humans."

This study adds to the weight of evidence indicating that pesticide exposure may contribute to reproductive birth defects in boys in ways that are attributible to endocrine disruption. Data from animal experiments are very clear on this.

Epidemiological findings with people, however, are much less conclusive. This may be because endocrine disrupting effects are less powerful in people, but biologically this is implausible as the same biological mechanisms are at work. The more likely reason for the weakness of epidemiological findings is that they almost invariably contain strong biases against finding positive results, i.e., they are biased towards false negatives.

In this study, for example, the bias toward false negatives is built into the study design because of inherent limitations of what the investigators were able to measure. They had no direct measurement of exposure, and instead they had to depend on very general occupational categories. Within mothers involved in gardening/farming, the amount and timing of exposure almost certainly varied substantially. Some used pesticides with no estrogenic/anti-androgenic activity, or used compounds at times in development when the reproductive tract was not vulnerable.

When studies such as this actually find positive results, they should be taken very seriously.