With interest we read the Randomised Clinical Trial studying the capacity of 17-alpha-hydroxyprogesterone caproate (17OHP-C) to reduce preterm birth in women diagnosed with a shortened cervix during antenatal care. The study was discontinued for futility based on an interim analysis after recruitment of 105 participants.

The primary endpoint of the study was time from randomization to delivery, which was a mean of 76 days for the 17OHP-C arm vs 72 days for the control arm (difference 4 days, 95% confidence interval [CI] −9 to +17, P =0.48). Although the 37 week rates showed comparable results, women treated with 17OHP-C had lower 34 week (24% vs 30%, P = .51) and 32 week (14% vs 20% P = .44) delivery rates, thus jeopardizing the statement that there was no effect.

Previously Grobman et al. had also reported no effect from 17OHP-C on delivery rates <37 weeks (82 of 327 [25.1%] vs 80 of 330 [24.2%]) in a study that had been stopped after the recruitment of 657 participants. Similarly to the study of Winer et al., the study of Grobman et al showed reduced preterm delivery rates < 32 weeks (17OHP-C; 28 of 327 [8.6%] vs control; 32 of 330 [9.7%], relative risk [RR], 0.88 [95% CI, 0.54–1.43]) and < 28 weeks (17OHP-C; 15 of 327 [4.6%] vs control; 22 of 330 [6.7%], RR, 0.69 [95% CI, 0.36–1.30]) and a lower composite adverse neonatal outcome rate (7.0% vs 9.1%, RR, 0.77 [95% CI, 0.46–1.30]) compared with placebo.

We question whether we are stopping preterm birth trials too early. In women at risk for preterm birth, any delay in delivery, albeit earlier in pregnancy, may improve pregnancy outcomes. The pooled delivery rate of < 32 weeks from the 2 studies is 35 of 378 in the treatment group vs 43 of 384 in the control group, giving an RR of 0.82 (95% CI, 0.54–1.3). Although we by no means want to conclude that the data suggest effectiveness, we also are concerned that because of early stopping of the studies, we miss a potential effect of 17OHP-C.