In Reply: Drs Hou and Zhang note the limitation related to use of filled prescriptions as a proxy for drug exposure in our study. Because it is not known whether women actually used the drugs they purchased from the pharmacy, misclassification of exposure because of nonadherence may have introduced a bias toward the null. However, using a nationwide registry on filled prescriptions also has several advantages.
In view of the difficulties in performing studies of drug safety in pregnancy because of the relative rarity of both exposure and outcome, the use of nationwide registry data allows the identification of a sufficient number of exposed pregnancies to reach adequate power, at least for birth defects in aggregate. Furthermore, antivirals are prescribed for acute infections, and women who filled prescriptions for these drugs are therefore more likely to have used the drugs compared with (as an example) those who filled prescriptions for drugs prescribed for prophylactic indications. Moreover, people who fill prescriptions for drugs seem very likely to have used them since they have traveled to a pharmacy and purchased the drugs.
Björn Pasternak, Anders Hviid. Risk of Birth Defects After Use of Acyclovir, Valacyclovir, and Famciclovir in the First Trimester—Reply. JAMA. 2010;304(20):2242–2243. doi:10.1001/jama.2010.1702