Senat et al performed a well-designed randomized trial testing the synthetic progestogen, 17-hydroxyprogesterone caproate (17-OHPC). The investigators addressed a concern for insufficient dosing in prior studies by an injection of 1000 mg/wk. Unfortunately, 17-OHPC exposure was significantly associated with preterm birth at less than 32 weeks ( P < .01).

Phase III trials enrolling twins testing 17-OHPC have repeatedly demonstrated a greater risk for complications related to prematurity including: (1) a shorter latency to delivery or earlier gestational age at delivery or (2) a higher frequency of neonatal severe respiratory distress symdrome. In the most comprehensive pharmacodynamic study of 17-OHPC administration, Caritis et al identified a highly significant negative correlation between serum concentration and gestational age at delivery ( P = .001). Senat et al documented earlier birth, less than 32 weeks, was significantly related to a higher dose of 17-OHPC, possibly reflecting serum concentration. Therefore, this study’s findings are consistent with phase III, 17-OHPC studies in multiples: none have shown benefit, and all have demonstrated the potential for increased adverse outcomes.

Most concerning is the perinatal mortality data. Five antepartum or intrapartum deaths occurred with 17-OHPC exposure (5 of 164, 3%) vs 0 in controls (0 of 158, P = .06, Fisher exact test). When available data regarding antepartum, intrapartum, and neonatal deaths are combined (shown in Table 4), 9 of 158 of fetuses/neonates (6%) died after 17-OHPC treatment compared with 1 of 154 control participants (1%) ( P = .02, Fisher exact test). A composite outcome derived from stillbirth data and the most common neonatal complication, respiratory distress syndrome, also demonstrates harm related to 17-OHPC therapy, 55 of 160 (34%) vs 34 of 154 (22%) without treatment ( P = .016, χ ² test). A significant increase in composite morbidity and mortality has also been identified in multiples treated with 17-OHPC by metaanalysis with a number needed to harm of 31 (17-167). A partial agonist for a receptor will not stimulate the activities of a receptor identically to a full agonist, and when those activities are needed the most (eg, multiples), insufficient progesterone receptor–mediated activity can be deadly.

Evidence for harm is not absence of benefit. Because of the repeatedly observed mortality and morbidity associated with 17-OHPC exposure and the highly significant serum concentration-adverse response relationship, providers should be warned unequivocally not to prescribe 17-OHPC in multiples. A probable dose-dependent mechanism for harm exists, necessitating further investigation of safety. Finally, the synthetic 17-OHPC is not progesterone. Whether natural progesterone could benefit multiples remains a hopeful consideration, especially in those at risk for decidual dysfunction because these 2 distinct progestogens have very different safety profiles.