Objective: To compare the effectiveness of expectant management (EM), cervical cerclage (CC) and vaginal progesterone (VP) in decreasing the rate of spontaneous preterm birth in twin gestations with midtrimester cervical shortening.
Study design: This is a retrospective cohort study comparing pregnancy outcomes of twin gestations with midtrimester cervical shortening, defined as a cervical length (CL) on routine transvaginal ultrasound between 15 weeks 0 days and 24 weeks 6 days gestation of <2.5 cm, managed with either EM, CC or VP. Women were categorized by final management strategy. Primary outcome was gestational age at delivery. Secondary outcomes included latency period (defined as number of weeks between a diagnosis of cervical shortening and delivery), gestational age at delivery <32 weeks, mode of delivery, perinatal death, neonatal birthweight and rate of chorioamnionitis. Subanalysis of women with a CL < 1.5 cm was also performed. Logistic regression was used to identify predictors of delivery <32 weeks, controlling for potential confounders.
Results: Between January 2006 and July 2016, 64 pairs of twins with midtrimester cervical shortening were identified, 18 managed with EM (28.1%), 29 CC (45.3%), and 17 VP (26.6%), 52 of which had information regarding delivery outcomes. 90.4% of women delivered prematurely (<37 weeks). Women in the CC group were diagnosed with cervical shortening at a significantly earlier mean gestational age (CC 20.6 ± 1.7 weeks versus EM 22.2 ± 2.9 weeks and VP 22.2 ± 2.0 weeks, p = .02) and had a shorter mean cervical length at the time of diagnosis (CC 1.18 ± 0.7 cm vs. EM 1.56 ± 0.7 and VP 1.95 ± 0.6, p = .002), as compared to those in the EM and VP groups. There was no difference in gestational age at delivery (EM 30.9 ± 5.2 weeks, CC 30.4 ± 4.9 weeks and VP 32.4 ± 4.1 weeks, respectively) or any of the secondary outcomes listed above. Women with a CL <1.5 cm delivered significantly earlier than those with a cervical length ≥1.5 cm (28.4 ± 4.7 weeks vs. 33.2 ± 3.6 weeks, p = .0001). After adjusting for potential confounders, cervical length <1.5 cm, not the management strategy, was the predictor of PTB before 32 weeks in this twin population [AOR 6.56 (95% CI 1.78, 24.20), p = .005].
Conclusion: Twin pregnancies with midtrimester cervical shortening are at high risk for preterm delivery, and outcomes were similar regardless of management strategy. Large prospective trials are needed to evaluate the effect of different management strategies for cervical shortening in twins.
Keywords: Twin gestation; cerclage; cervical shortening; preterm birth; vaginal progesterone.