Management of Twins With Short Cervix at a Single Institution: A Retrospective Review

In 2010, the incidence of preterm birth in twin gestations was 23% before 34 weeks and 58% before 37 weeks, while the overall incidence of PTB in all pregnancies was only 3.41% and 11.5%, respectively. Twin gestations are also at increased risk of being low birth weight infants.

Goldenberg et al demonstrated in 1996 that a cervical length less than 25 mm at 24 weeks gestation in twin pregnancies was the best predictor of preterm birth at <32 weeks, <35 weeks, and < 37 weeks. In 1999, Souka et al reported that in twin pregnancies with a cervical length less than 15 mm at 23 weeks gestation, the preterm birth rate <32 weeks was approximately 66%.

A Cochrane Review from 2014 concluded that placing a cerclage is not associated with a significant difference regarding perinatal deaths or neonatal morbidity, but this analysis was for all forms of cerclage in multiple gestations (physical examination-indicated, history-indicated, and ultrasound-indicated). It did also state "ultrasound-indicated cerclage appears to be associated with an increased risk for the composite adverse neonatal outcome, including respiratory distress syndrome, low and very low birthweight, given there is no evidence of a consistent subgroup effect, the observed results can most likely be attributed to chance."

In a 2015 retrospective cohort study of twin pregnancies with asymptomatic short cervix, Amanda Roman et al demonstrate similar results to those of the Cochrane Review for patients with CL <25 mm. However, they did show a trend towards prolonged pregnancy, decreased spontaneous PTB <34 weeks, and fewer NICU admissions for a subset of patients with CL < 15 mm that received a cerclage compared to control.

In February 2016, Goya et al investigated the role of the pessary for women with twin gestations and short cervix diagnosed via ultrasound. A significant reduction in the rate of preterm birth prior to 34 weeks was found in the pessary group (16.2% vs 39.4%), with no significant differences in neonatal morbidity or mortality.