defects that could be traced to consumption of the drug mebendazole. This was a retrospective analysis of birth outcomes in that region, not a double-blind study that followed the course of pregnancies and monitored maternal behavior and drug consumption. It was found that 2.5% of women (sample = 407) who, against medical advice, had taken the anti-hookworm treatment mebendazole in the first trimester of pregnancy gave birth to babies with major congenital defects. While the association between first-trimester mebendazole and incidence of congenital defect was measurable, the drug was not considered significantly but instead relatively risky. The authors explain:
The safety of a drug is best tested in a prospective, randomised, double-blind placebo-controlled cohort study, but for obvious ethical reasons, this cannot be achieved in groups of pregnant women without very good evidence supporting a lack of teratogenicity (De Silva & Others, 1999, p. 1145).
Apart from the constituents of a given substance suspected of being a teratogen, the teratogenic effect itself may be highly idiosyncratic from fetus to fetus, or from host womb to womb, varying with the substance and its behavior per se, the targeted organism, and/or the stage of development. For example, in the Sri Lanka study, women who had consumed mebendazole during pregnancy as a group were found to have had relatively higher birth weight babies than their nonconsuming counterparts. That does not, however, necessarily prove that mebendazole had health-giving properties. De Silva, et al., cite the increased attention to prenatal health care more generally as a possible explanation for the phenomenon.
Reflecting the indeterminacy of definition implied in Mosby's and problems with sample formation and study design, Thomas's catalogue (1998) of teratogens divides them into three general categories: unlikely, proven, and possible. This method of categorization allows Thomas to ...