Predictive value of cervical length in women with twin pregnancy presenting with threatened preterm labor.

OBJECTIVE To assess the accuracy and determine the optimal threshold of sonographic cervical length (CL) for the prediction of preterm delivery (PTD) in women with twin pregnancies presenting with threatened preterm labor (PTL). METHODS This was a retrospective study of women with twin pregnancies who presented with threatened PTL and underwent sonographic measurement of CL in a tertiary center. The accuracy of CL in predicting PTD in women with twin pregnancies was compared with that in a control group of women with singleton pregnancies. RESULTS Overall, 218 women with a twin pregnancy and 1077 women with a singleton pregnancy, who presented with PTL, were included in the study. The performance of CL as a predictive test for PTD was similar in twins and singletons, as reflected by the similar correlation between CL and the examination-to-delivery interval (r, 0.30 vs 0.29; P = 0.9), the similar association of CL with risk of PTD, and the similar areas under the receiver-operating characteristics curves for differing delivery outcomes (range, 0.653-0.724 vs 0.620-0.682, respectively; P = 0.3). The optimal threshold of CL for any given target sensitivity or specificity was lower in twin than in singleton pregnancies. However, in order to achieve a negative predictive value of 95%, a higher threshold (28-30 mm) should be used in twin pregnancies. Using this twin-specific CL threshold, women with twins who present with PTL are more likely to have a positive CL test, and therefore to require subsequent interventions, than are women with singleton pregnancies with PTL (55% vs 4.2%, respectively). CONCLUSION In women with PTL, the performance of CL as a test for the prediction of PTD is similar in twin and singleton pregnancies. However, the optimal threshold of CL for the prediction of PTD appears to be higher in twin pregnancies, mainly owing to the higher baseline risk for PTD in these pregnancies.