We commend Suhag et al for their systematic review and metaanalysis of randomized trials on vaginal progesterone for maintenance tocolysis. However, we take issue with the included articles and especially about authors’ conclusions. We have published the largest trial on vaginal progesterone in women with preterm labor. The 4P trial was performed in Switzerland (9 centers) and in Argentina (20 centers) and evaluated 379 women (193 received progesterone, and 186 received placebo). Preterm birth occurred in 42.5% of women in the progesterone group vs 35.5% in the placebo group (relative risk, 1.2; 95% confidence interval, 0.93–1.5). There were also no differences about delivery at <32 and <34 weeks of gestation and neonatal morbidity. When indicated preterm deliveries were excluded, there was a higher risk of spontaneous preterm deliveries in the progesterone group (relative risk, 1.4; 95% confidence interval, 1.04–1.88; P = .03). In a secondary analysis of the 4P trial that was presented at Society of Maternal Fetal Medicine (San Diego, CA; 2015), we showed that, in Switzerland, as compared with placebo, progesterone use increased the risk for spontaneous delivery within 14 days and preterm birth at <37 weeks of gestation.

Suhag et al might argue that the 4P trial was not included in their review because progesterone was given within 24 hours of the start of acute tocolysis and therefore could not be considered as “maintenance” tocolysis. The study medication was initiated after stabilization of the patient (diminution/arrest of uterine contractions with acute tocolysis) and continued until 36 weeks of gestation or delivery, if it happened before. There were only 9 women who delivered within the first 48 hours. Analysis that included only women who were stabilized for at least 48 hours showed the same result (ie, progesterone did not decrease the risk of preterm delivery).

Regarding the studies included in the review of Suhag et al, they were all small (between 52 and 183 cases), and only 2 were compared with placebo. Metaanalysis is of great help in clinical practice, especially when large studies are not available. The decision to exclude trials should be based on strict inclusion criteria. We do not understand why Suhag et al decided not to include the largest randomized trial on vaginal progesterone in women with preterm labor to be published until now.

On the basis of existing literature, progesterone should not be given as maintenance tocolysis in women with preterm labor. Benefits have not been found, and harms have not been excluded. New large, randomized, and placebo-controlled studies will be soon published and will help in drawing conclusions.