Early vaginal progesterone versus placebo in twin pregnancies for the prevention of spontaneous preterm birth: a randomized, double-blind trial.

Abstract

Background: In women with a singleton pregnancy and sonographic short cervix in midgestation, vaginal administration of progesterone reduces the risk of early preterm birth and improves neonatal outcomes, without any demonstrable deleterious effects on childhood neurodevelopment. In women with twin pregnancies the rate of spontaneous early preterm birth is 10-times higher than in singletons and in this respect all twins are at increased risk of preterm birth. However, six trials in unselected twin pregnancies reported that vaginal progesterone from mid-gestation had no significant effect on the incidence of early preterm birth. Such apparent lack of effectiveness of progesterone in twins may be due to inadequate dosage or treatment which is started too late in pregnancy. Objective: The Early vaginal progesterone for the preVention of spontaneous prEterm birth iN TwinS: A randomized, placebo controlled, double-blinded trial (EVENTS) was designed to test the hypothesis that, among women with twin pregnancies, vaginal progesterone at a dose of 600 mg per day from 11-14 until 34 weeks’ gestation, as compared with placebo, would result in a significant reduction in the incidence of spontaneous preterm birth between 24+0 and 33+6 weeks. Methods: The trial was conducted at 22 hospitals in England, Spain, Bulgaria, Italy, Belgium and France. Women were randomly assigned in a 1:1 ratio, to receive either progesterone or placebo and in the random-sequence generation there was stratification according to participating center. Primary outcome was spontaneous birth between 24 +0 and 33+6 weeks’ gestation. Statistical analyses were performed on an intention-to-treat basis. Logistic regression analysis was used to determine the significance of difference in incidence of spontaneous birth between 24+0 and 33+6 weeks’ gestation between the progesterone and placebo groups, adjusting for the effect of participating centre, chrorionicity, parity and method of conception. Prespecified tests of treatment interaction effects with chrorionicity, parity, method of conception, compliance and cervical length at recruitment were performed. A post hoc analysis using mixed effects Cox regression was used for further exploration of the effect of progesterone on preterm birth. Results: We recruited 1,194 women between May 2017 and April 2019; 21 withdrew consent and 4 were lost to follow up, which left 582 in the progesterone group and 587 in the placebo group. Adherence was good, with reported intake of ≥80% of the required number of capsules in 81.4% of the participants. After excluding births before 24 weeks and indicated deliveries before 34 weeks, spontaneous birth between 24+0 and 33+6 weeks occurred in 10.4% (56/541) participants in the progesterone group and in 8.2% (44/538) in the placebo group (odds ratio in the progesterone group, adjusting for the effect of participating center, chrorionicity, parity and method of conception, 1.35; 95% CI 0.88 - 2.05; p=0.17). There was no evidence of interaction between the effects of treatment and chorionicity (p=0.28), parity (p=0.35) method of conception (p=0.56) and adherence (p=0.34); however, there was weak evidence of an interaction with cervical length (p=0.08) suggestive of harm to those with cervical length ≥30 mm (odds ratio 1.61, 95% CI 1.01-2.59) and potential benefit for those with cervical length <30 mm (odds ratio 0.56; 95% CI 0.20-1.60). There was no evidence of difference between the two treatment groups for stillbirth or neonatal death; neonatal complications; neonatal therapy; and poor fetal growth. In the progesterone group there were 1.4% (8/582) women and 1.9% (22/1,164) fetuses with at least one serious adverse event; the respective numbers for the placebo group were 1.2% (7/587) and 3.2% (37/1,174) (p=0.80 and p=0.06, respectively). In the post hoc time to event analysis, miscarriage or spontaneous preterm birth between randomization and 31 +6 weeks’ gestation was reduced in the progesterone group relative to the placebo group (hazard ratio 0.23, 95% CI 0.08 - 0.69). Conclusions: In women with twin pregnancies universal treatment with vaginal progesterone did not reduce the incidence of spontaneous birth between 24+0 and 33+6 weeks’ gestation. Post hoc time to event analysis led to the suggestion that progesterone may reduce the risk of spontaneous birth <32 weeks in women with cervical length <30 mm and it may increase the risk for those with cervical length ≥30 mm.