To evaluate whether the routinely collected data in birth defect registries could be used to assess association between medications and risk for congenital anomalies an "exposed case-only" design was performed. Twelve registries provided 18,131 cases exposed to a medication during the first trimester of pregnancy and with at least one major malformation. Odds ratios for malformations associated with maternal use of selected medications were computed. Among seven most commonly used medications very few significant associations with malformations were identified. Among fourteen potentially teratogenic medications several strong associations were found, including valproic acid with spina bifida, and insulin (as proxy for diabetes) with several types of cardiac defects. Finding known associations provides assurance on the validity of this approach, whereas identifying new associations provides a signal to be followed by confirmatory studies. Through this activity, international networks of birth defect registries can contribute with limited resources to post-marketing surveillance of the teratogenicity of medications.