Cervical length in asymptomatic twin pregnancies: prospective multicenter comparison of predictive indicators

Abstract Objective: To determine whether cervical shortening between 22 and 27 weeks predicts spontaneous preterm delivery before 34 weeks better than a single cervical length (CL) measurement at 22 or 27 weeks in asymptomatic twins.

Prospective 13-center study over a 2-year-period. CL was measured in 120 asymptomatic twin pregnancies. The area under the ROC curve was calculated for each parameter and the cutoff point providing the best diagnostic performance, sensitivity and specificity for predicting spontaneous delivery<34 weeks was defined too.

About 13/116 women gave birth before 34 weeks. The three ROC curves differed significantly at the 0.05 level. The best cutoff points were CL≤35 mm at 22 weeks, CL≤25 mm at 27 weeks and cervical shortening≥20%. For equal sensitivity values for each, specificity for CL≤25 mm at 27 weeks was 87.5%, significantly better.

The performance of cervical shortening for the prediction of preterm delivery of asymptomatic twins before 34 weeks does not differ from that of CL measurements at 22 or 27 weeks. The modest predictive value of CL at 22 weeks and of cervical shortening is an argument against recommending routine CL measurements.