The meta-analysis by Dr Saccone and colleagues concludes that surgical abortion “is an independent risk factor” for subsequent preterm birth. The authors found a weak association (odds ratios [OR], 1.44; 95% confidence interval, 1.09–1.90) between abortion and preterm birth, but we question whether this association is causal. We agree with the discussion of study limitations and will highlight several key points. First, the reported associations all had OR <2. Not only do bias and confounding often account for weak associations, but OR exaggerate true relative risk. Second, most studies included failed to adjust for important known confounders such as prior preterm birth, race, smoking, and short interpregnancy interval. Third, many studies had case-control designs, and recall bias has been shown to have a powerful impact in case-control studies of abortion, exaggerating negative outcomes of abortion.

Even if some of the reported association is causal, the attributable risk of preterm birth following abortion is very small. When women continue unintended pregnancies, however, they may be at increased risk of preterm birth in that pregnancy. One systematic review found an association between unwanted pregnancies and preterm birth with an OR magnitude similar to the findings presented here (OR, 1.50; 95% confidence interval, 1.41–1.61). When women gained access to safe abortion in Oregon, a decrease in preterm birth and neonatal mortality were observed. Access to abortion also has clear social and economic benefits for women and families, likely affecting future pregnancy outcomes and preterm births.

The data presented are insufficient to support counseling women that abortion is a risk factor for preterm birth or to warrant the large and expensive randomized trials to further evaluate this association as proposed by the authors. We suggest funding would be better spent on interventions known to prevent preterm birth: prenatal care, contraception, and smoking cessation, for example.