Spermicidal contraceptives and poor reproductive outcomes: the epidemiologic evidence against an association

Abstract

The first widely publicized report of an association between spermicidal contraception and congenital malformations and spontaneous abortion had considerable impact on obstetric practice. A large number of more recent epidemiologic studies have generally failed to support the earlier finding, and it is concluded that no such association has been demonstrated. The available evidence precludes the need for additional regulation of spermicidal contraception.

PIP: This article reviews the 14 known epidemiologic studies that have considered an association of spermicide use with congenital malformations, spontaneous abortions, or low birth weight, and presents some observations on how epidemiologic inquiry relates to regulatory policy. 5 case control studies and 5 prospective studies examined congenital malformations. Some studies examined selected congenital diagnoses whereas others reported for all malformations. The 1 case-control study reporting a positive association had only 16 cases and represented a select group of newborn infants with Down's syndrome and congenital heart defects. The time of exposure was simply "during the period before this pregnancy". The only prospective study reporting an association can be criticized for the extremely broad measure of spermicide use (a prescription filled within 600 days of delivery), the low number of malformations observed, the lack of a common etiology in the assemblage of malformations identified as being associated with spermicides, and inadequate control for confounding variables. Of the 4 prospective studies examining associations of spermicides and risk of miscarriage, 1 that found a positive association failed to adjust for potentially confounding factors. A reported association between spermicide use and spontaneous abortion identified in a recent analysis of the National Survey of Family Growth data suffered from several major difficulties including comparing women exposed to spermicides close to conception with those exposed after conception; asking women about their pregnancies over a period of almost 4 years, enhancing the chance of biased recall; and using a fixed observation period so that women with a history of multicple spontaneous abortions were likely to be overrepresented. A better-designed cohort study with 32,123 cases found no association between spermicide use and spontaneous abortion. 2 research groups that studied low birth weight and spermicide use found that the proportion of clinically low birth weight infants was not significantly influenced by spermicide usage, or found no association at all. It is quite common in epidemiology for the 1st published study of an association to suggest a particular association, and it is probably no coincidence that the 1st and most widely publicized spermicide study was a positive one. A number of large studies specifically designed to test the spermicide malformation hypothesis are currently in the data collection stage. The risks and benefits of a product itself and the risks and benefits of regulation both need to be considered for regulatory purposes. Regarding the decision on whether to label spermicide contraceptives as a potential hazard, the products have social benefits and the epidemiological evidence for an association with poor reproductive outcome is very weak.