Single transvaginal sonographic measurement of cervical length early in the third trimester as a predictor of preterm delivery

Abstract

To evaluate the accuracy of a single sonographic measurement of cervical length early in the third trimester as a predictor of preterm delivery. A total of 771 women attending the antenatal clinic at the Maharaj Nakorn Chiang Mai Hospital between January 1, 1990, and November 30, 1993, with singleton gestations, cervical competence, accurate dates of last menstrual period, and gestational ages between 28-30 weeks, were recruited into the study. Forty-one were excluded, ten because of induced preterm delivery, 24 because of measurement problems, and seven because they were lost to follow-up. We analyzed the remaining 730 women. Ninety-one (12.5%) women ended with preterm births and the remaining 639 (87.5%) delivered at term. The mean (+/- standard deviation) cervical lengths of the term and preterm groups were statistically significantly different at 37 +/- 5 and 34 +/- 6 mm, respectively (P < .001). The likelihood ratio of cervical length at various cutoff points was calculated. The appropriate cutoff point based on the receiver operating characteristic curve (35 mm) was associated with a significantly increased likelihood of preterm delivery (20 versus 7%) and was detected in two-thirds of preterm births. This cutoff point gave a sensitivity and specificity of 65.9 +/- 5.1% (95% confidence interval [CI]) and 62.4 +/- 5.2% (95% CI), respectively. A single transvaginal sonographic measurement of cervical length at 28-30 gestational weeks can be used to predict the risk of preterm delivery, using a cutoff point of 35 mm, but its cost-effectiveness should be assessed further.