Ramos et al ^{Reference Ramos, St-André, Rey, Oraichi and Bérard1} report that the use of antidepressant medications by women during the first trimester of pregnancy is not associated with an increased risk for major congenital malformations in children. The authors have a good database to study this topic but have described and analysed it using a case–control framework. They assembled two cohorts, with and without exposure to antidepressants during pregnancy. They then observed the various outcomes in both groups. We calculated the relative risk (RR) for major congenital malformations following use of antidepressants during first trimester of pregnancy as 1.13 (95% CI 0.86–1.48) from their published data. Estimating such relative risk and population attributable risk (5.76%) would have bolstered their arguments, as a cohort design is superior to a case–control strategy.

However, we suggest caution in generalising these findings because of two important limitations that were not acknowledged in their paper. If antidepressants are associated with more spontaneous abortions and an increased number of minor congenital anomalies, their lack of association with major congenital anomalies will not imply safety. A previous meta-analysis of 3567 women established a significantly increased RR of 1.45 (95% CI 1.19–1.77) for spontaneous abortions following use of antidepressants during pregnancy. ^{Reference Hemels, Einarson, Koren, Lanctôt and Einarson2} Individual antidepressants such as selective serotonin reuptake inhibitors ^{Reference Pastuszak, Schick-Boschetto, Zuber, Feldkamp, Pinelli, Sihn, Donnenfeld, McCormack, Leen-Mitchell and Woodland3} and other newer antidepressants ^{Reference Djulus, Koren, Einarson, Wilton, Shakir, Diav-Citrin, Kennedy, Voyer Lavigne, De Santis and Einarson4,Reference Chun-Fai-Chan, Koren, Fayez, Kalra, Voyer-Lavigne, Boshier, Shakir and Einarson5} have led to more miscarriages when compared with unexposed control groups. As Ramos et al have included exclusively women who had their pregnancies ending in delivery, they do not add any information regarding spontaneous abortions.

In another study of 482 pregnant women, ^{Reference Chambers, Johnson, Dick, Felix and Jones6} fluoxetine caused significantly more prematurity (RR=4.8, 95% CI 1.1–20.8), more admissions to special care nurseries (RR=2.6, 95% CI 1.1–6.9) and worse neonatal adaptation (RR=8.7, 95% CI 2.9–26.6) after adjusting for all potential confounders. A total of 15.5% of infants exposed to fluoxetine had three or more minor congenital anomalies compared with 6.5% of infants who were not exposed to fluoxetine (P=0.03). ^{Reference Chambers, Johnson, Dick, Felix and Jones6} However, Ramos et al excluded minor congenital anomalies during case ascertainment without any explicit justification. Absence of association between use of antidepressants and major congenital malformations will not make a clinician confident to continue antidepressants during the first trimester of pregnancy if there are concerns over spontaneous abortions, prematurity and minor congenital anomalies. Hence, we encourage cautious interpretation of these findings as well as judicious use of antidepressants for women of reproductive age.