Purpose We report an association found in a surveillance study which systematically evaluated combinations of specific birth defects and drugs used in the first trimester of pregnancy. Method The database of a population‐based birth defects registry (birth years 1997–2007) was systematically screened for combinations of drugs and malformations that were disproportionately present compared to the rest of the database. Combinations with at least three exposed cases and a p < 0.01 (Fisher Exact test) were studied to analyse details of the malformation, timing of exposure, and additional case‐control analyses. Results Among the significant associations found, an association between maternal use of fluoxetine and infantile hypertrophic pyloric stenosis (IHPS) was of particular interest. In total 3/178 (1.7%) of the children with a HPS were exposed to fluoxetine in the first trimester compared to 8/4077 (0.2%) fluoxetine exposures among the children with other malformations (p = 0.009, OR = 8.7, 95%CI = 2.3–33.2). The three exposed cases were all isolated and fluoxetine was used in gestational weeks 4–8, 2–8 and ‐10–19, respectively. In additional case‐control analyses, using controls with a genetic disorder and after adjustment for maternal age and smoking in the first trimester of pregnancy, the adjusted odds ratio was 9.8 (95% confidence interval: 1.5–62.0). Conclusions Because we cannot rule out the possibility that the association between IHPS and fluoxetine is caused by chance, we encourage other investigators to study the association between IHPS and fluoxetine in their data. Copyright © 2010 John Wiley & Sons, Ltd.