We would like to thank Dr O’Brien for his comments regarding our recent article entitled “Prevention of preterm delivery by 17 alpha-hydroxyprogesterone caproate (17P) in asymptomatic twin pregnancies with a short cervix: a randomized controlled trial.”

We noted that the intent-to-treat analysis in our study showed that treatment with 17P was associated with a significant increase in the preterm delivery rate before 32 weeks: 29% (24 of 82) vs 12% (10 of 83) in the control group. This increased risk was even more obvious when the analysis was restricted to the evaluated cases (ie, those not lost to follow-up): 29% (24 of 82) vs 8% (6 of 79) in the control group ( P = .0005) before 32 weeks and 40% (33 of 82) vs 24% (19 of 79) in the control group ( P = .03) before 34 weeks.

We would also like to note the potential role of castor oil in these negative findings. Because the design of our study was an open-label, randomized controlled trial, our control group did not receive any placebo, and in particular no castor oil, usually used as placebo in other trials. On the other hand, castor oil was the vehicle used in our 17P. A tocographic study long ago observed an increase in contractility in the human myometrium after castor oil exposure. Our study therefore renews previously expressed concerns of the uterotonic effects of castor oil, especially in that we used high doses of 17P (500 mg repeated twice weekly). The recent study reported by O’Sullivan et al in the Journal strengthens our doubts: they demonstrated in vitro that exposure of human myometrial preparations to castor oil results in enhanced oxytocin-induced contractility.

Therefore, we cannot rule out the possibility that the potential harmful effect of 17P in our experimental group was linked to high maternal concentrations of castor oil that increased contractility in patients at high risk because of a twin pregnancy and short cervix, especially at early gestational ages. Our results, like those of O’Sullivan et al, raise once again the role and validity of the use of castor oil as the vehicle for 17P.

Furthermore, beyond our randomized trial, no study has been specifically designed to investigate the effect of any type of progesterone in reducing preterm birth in asymptomatic twin pregnancies with a short cervix. There is insufficient evidence, therefore, according to the recommendation of the SFMF clinical guideline, to recommend the use of any progestogen agent in women with asymptomatic twin pregnancies and a short cervix. In light of the literature and our data, we agree with Dr O’Brien that 17OHP is ineffective in the prevention of prematurity for twin pregnancy and may be potentially harmful for fetuses. Currently only natural progesterone should be assessed in the prevention of preterm birth in twin pregnancy, and its use should be carefully restricted to randomized trials, with special attention given to the evaluation of the tolerance of this drug.

Finally, in routine practice, we think that providers should be warned not to prescribe 17P or any other progestogen agent in twin pregnancies.