We thank Dr Averbach et al for their interest in our study. As we have highlighted in our article and in prior letters, we completely agree that we have found a weak association (odds ratio <2) between abortion and preterm birth and acknowledge the high risk of bias of the included studies. Most of the included studies did not control appropriately for confounders, and only 6 included parity, an important determinant of preterm delivery, as a potential confounder. Moreover, because women face stigma when reporting an induced abortion, patients in the case or control group could have omitted abortion from their medical history, which would lead to a high risk of recall bias.

Dr Averbach et al also cited a possible association between unintended pregnancies and preterm birth. They stated that “when women gained access to safe abortion…a decrease in preterm birth and neonatal mortality were observed.” We agree with this statement, as the social risk factors associated with preterm birth are the same risk factors associated with unintended pregnancies. In our article, we did not conclude that abortion is a risk factor for preterm delivery, but that prior surgical uterine evacuation could be. Indeed, we also found that women with prior medical abortion had a similar risk of spontaneous preterm birth compared with controls (28.2% vs 29.5%). Counseling women that, based on the available evidence, however poor, surgical uterine evacuation either for miscarriage or abortion may be an independent risk factor for preterm birth, may give women additional information when making a decision regarding medical or surgical management of unintended pregnancy or early pregnancy loss. Perhaps most importantly, the description of cervical preparation before uterine evacuation was limited, but cervical ripening (eg, with misoprostol) may indeed prevent any small increase in preterm birth associated with these procedures. Moreover, as there were no randomized controlled trials (RCTs) included in our meta-analysis, and no studies comparing prior medical with prior surgical abortion, we still call for future large clinical RCTs with long-term follow-up comparing surgical versus medical evacuation of the uterus.

We strongly believe that physicians and researchers should always be encouraged to practice based on the highest level of evidence, which comes from RCTs and meta-analyses of RCTs.