Problem of using cases with genetic anomalies as a reference group in case-control studies on drug use and birth defects

Evaluation of the representativeness of a Dutch non-malformed control group for the general pregnant population: are these controls useful for EUROCAT?

PURPOSE

A case-control study is the most powerful design to test the risk of specific congenital malformations associated with a specific drug. However, malformation registries often lack non-malformed controls. For the Dutch EUROCAT, we collected a non-malformed control group: the 'Healthy Pregnant'. The aim of this study was to evaluate the representativeness of this control group for the general pregnant population in the northern part of the Netherlands.

METHODS

The Healthy Pregnant data set includes data from two midwife practices. The baseline characteristics of mother and child including smoking status, drinking status, body mass index, maternal age, educational level, exposures to several drugs for chronic diseases and pregnancy related symptoms were evaluated.

RESULTS

Compared with the general population, mothers in Healthy Pregnant group (n = 556) were from either low or high education level, were more likely to have a body mass index > 25 kg/m² (26% versus 22%, p = 0.08) or to smoke (19% versus 10%, p < 0.01) but were less likely to consume alcohol (20% versus 29%, p < 0.01). The use of drugs for chronic conditions was lower in Healthy Pregnant group. Furthermore, drugs for occasional use were prescribed less frequently, and a significant underreporting of children with a low birth weight and a short duration of gestation was found.

CONCLUSION

The Healthy Pregnant data set was not representative of the general pregnant population in the northern part of the Netherlands. Specifically, the exposure to (chronic) drugs was underestimated, possibly a result of second-line care on the basis of medical indication. Thus, continuous investigation of options for improvement of the Healthy Pregnant database is required.