To the Editor On behalf of the entire writing team, we are writing in reply to 2 letters recently published in response to our systematic review and meta-analysis of pregnancy and delivery outcomes after exposure to antidepressant medication.1
In their letter, Tufanaru and Jureidini2 questioned our conclusion of “no significant association” between antidepressant exposure and spontaneous abortion. We fully agree that the distinction between statistically significant or not significant at P = .055 is arbitrary. However, we disagree that a statistically significant finding necessarily implies a clinically meaningful finding. Indeed, the point we emphasized in our discussion regarding preterm birth and low birth weight was that despite being statistically significant, these associations were likely not clinically significant considering their very small magnitude.1 In the context of the limited and poor-quality data linking spontaneous abortion to prenatal antidepressant exposure, it would be premature to warn the public of an increased risk for spontaneous abortion associated with prenatal antidepressant exposure. The comparison of our data on antidepressant exposure with data on smoking is totally inappropriate, considering that antidepressants are therapeutic while smoking is only hazardous. As such, a much lower threshold for establishing risk is clearly appropriate to advance public education regarding the dangers of smoking. Although we disagree that broad public education is appropriate at this time, we feel strongly that women making decisions about antidepressant use during pregnancy should have access to all relevant information to guide their decision. Indeed, making the best-quality information available to women and their physicians and other health care providers, such as midwives, is the goal of our research in this area.