Drs Gimovsky and Berghella report an open randomized trial of second-stage labor management where delivery is effected at 3 hours compared to 4 hours in nulliparous women with epidural analgesia. They report that waiting the extra hour significantly decreased the cesarean rate by about 50%.

A total of 78 women were randomized at 3 hours into the second stage, resulting in 37 women allocated to delivery at 3 hours vs 41 allocated to delivery at 4 hours. There were a total of 16 cesareans in the 3-hour group compared to 8 in those allocated to a 4-hour second stage. The 3-hour arm had 9 crossovers to the 4-hour group (5 physician and 4 patient decisions). There were 2 crossovers from the 4-hour group including 1 cesarean for fetal distress that actually should be an outcome of interest rather than a crossover. Analysis was by intent-to-treat, meaning that any cesareans done in the 9 crossover women originally intended to be delivered at 3 hours but instead managed as if they were in the 4-hour experimental group were counted in these 9 crossovers’ original 3-hour allocation group. Thus, cesareans, if any, in the 9 crossovers actually given an additional hour in the second stage would be counted as if they only had a 3-hour second stage.

If but 2 fewer cesareans had been allocated to the 3-hour delivery arm, ie, 14 instead of 16, the relative risk confidence interval becomes 0.24-1.09 rather than the reported 0.22-0.93. Thus, the statistical basis of The conclusion of Drs Gimovsky and Berghella that there were significantly fewer cesareans in the 4-hour experimental arm hinges on as few as 2 cesareans in an open study where physicians and patients could easily subvert group allocation as attested to by the unbalanced number of crossovers. Under the circumstances of this trial, a conclusion based on just 2 cesareans is too close for comfort.

We do not doubt that lengthening the 3-hour second stage would reduce the cesarean rate. But, at what cost? For >50 years, the crux issue has been infant safety that Drs Gimovsky and Berghella acknowledge was beyond the size of their study. Thus, their study does not address the important issue of infant safety and their conclusion about cesarean rates is flawed and therefore likely incorrect.