The prediction of spontaneous preterm birth in women with threatened preterm labor using transvaginal ultrasound of the cervix

Abstract

Purpose: This study was to determine if cervical parameters, as assessed by transvaginal ultrasound, were predictive of spontaneous preterm birth in women following an episode of threatened preterm labor (TPL).Methods: Women with a singleton pregnancy (20–34 weeks gestation) and intact membranes were enrolled following admission for TPL. The transvaginal ultrasound assessment of the cervix was performed when uterine activity had ceased. Only the ultrasound relating to the episode of suspected preterm labor is reported.Results: Sixty women had a transvaginal scan performed at a gestation of 27.7 ± 3.7 weeks (mean ± SD) (range 20.1–33.7 weeks). Cervical parameters were significantly different in women who delivered preterm (shorter closed endocervical canal length –P < 0.0001; increased internal os dilatation –P < 0.0001) to the women who delivered at term. Twenty‐five women had an abnormal cervix (closed endocervical canal length <30 mm + internal os dilatation >5 mm). Of these, 16 (64%) delivered preterm (<37 weeks). In contrast, only one woman with a normal cervix delivered preterm (P < 0.0001; RR 22.4, 95% CI 3.17–158.1). Five women (20%) delivered within 1 day of their abnormal cervical ultrasound, seven (28%) in <7 days, and 10 (40%) in <14 days whereas no woman with a normal cervical ultrasound delivered within 14 days (sensitivity 100%; negative predictive value 100%). In a linear regression model with gestational age at delivery as the dependent variable, endocervical canal length had an R2 of 59.3% (P < 0.0001) which was not improved by the addition of any of the other parameter variables.Conclusions: Transvaginal ultrasound of the cervix can identify those women with an abnormal cervix following an episode of TPL who are at increased risk of preterm birth. It also identifies the women with normal cervices and therefore at low‐risk of preterm birth which potentially could reduce unnecessary interventions.