Ultrasound cervical measurement and prediction of spontaneous preterm birth in ICSI pregnancies: a prospective controlled study

A prospective controlled study was performed in which transvaginal ultrasound measurement of cervical length was compared in 222 twin ICSI pregnancies, 122 singleton ICSI pregnancies and 51 spontaneous singleton pregnancies. Preterm birth was defined as ≤34 weeks. Full data were obtained for 193 twin pregnancies (group A), 102 singleton pregnancies (group B) and 51 spontaneous singleton pregnancies (group C). Cervical length at midterm was not statistically different between the three groups: group A, 37.6 ± 7.1 mm; group B, 37.2 ± 7.2 mm; and group C, 39.2 ± 5.4 mm. The incidence of preterm birth was statistically different between groups: 30.5% in group A; 17.6% in group B; and 3.9% in group C (P = 0.011). The ROC curve for optimum cut-off of cervical length in prediction of preterm birth for group A was 38.05 mm, sensitivity 67%, specificity 50%, positive predictive value (PPV) 37.7, and negative predictive value (NPV) 78.1. For group B the data were 33.05 mm, sensitivity 50%, specificity 70%, PPV 34.6, and NPV 88.1. Contrary to the situation for spontaneous pregnancies, midtrimester cervical length measurement in ICSI singleton and twin pregnancies is not a predictor for preterm birth.