We disagree with Clark and colleagues.1 Given that the use of nonsteroidal anti-inflammatory drugs (NSAIDs) is on an acute basis, even when used for chronic conditions, women who took the time to go to the pharmacy most likely took at least one pill. Glover and colleagues2 have shown that most pregnant women who fill a prescription take it, and Daniels and colleagues3 have shown that self-reporting measures of medication use do not provide accurate measurements when compared with electronic monitoring, as was done in our study.4
Data on maternal smoking and body mass index (BMI) were available only for a subset of women included in our registry. However, these variables would need to confound the effect to bias our results. For smoking and BMI to be confounders, they would need to be associated with both outcome and exposure. Delaney and colleagues5 have shown that smoking and BMI are not associated with use of NSAIDs that are prescribed or over the counter.
The registry includes data on the use of prescription medication during gestation. We looked at over-the-counter ibuprofen obtained with a prescription; however, the number of women who used over-the-counter ibuprofen without a prescription could not be mea-sured. There is no reason to believe that women who miscarried used over-the-counter ibuprofen without a prescription differently than those who did not. Hence, this led to nondifferential misclassification, and our results are underestimates of the true effects.
The aim of our study was to empower women by informing them and their physicians of the potential risks of medication use during pregnancy.