To estimate whether mid-trimester cervical length (CL) ≥ 25 mm in high-risk women is predictive of preterm birth.
Planned secondary analysis of the NICHD-sponsored cerclage trial. Women with documented prior spontaneous preterm birth (SPTB) at 17-336/7 weeks′ gestational age (GA) underwent serial cervical ultrasound scans initiated between 16 and 216/7 weeks. Fundal pressure-induced and spontaneous dynamic shortening were assessed to determine the shortest CL. The final scan was scheduled before 23 weeks. Women whose CL was < 25 mm were invited to consent for the intervention trial.
Of 1014 eligible women who began ultrasound screening, 318 had CL < 25 mm. Delivery dates were unavailable in an additional 24 (3.4%), leaving a study population of 672 for this analysis. The median (interdecile range) GA at the qualifying scan was 21.7 (19.1, 22.7) weeks, with a median CL of 34 (27, 42) mm. For comparison, the incidence of preterm birth < 35 wks was 42% in the 153 women who had CL < 25 mm and were randomized to no cerclage, as compared to 16% in this study population of women whose CL was ≥ 25 mm (p<0.0001). In a linear regression model, CL did not predict birth GA (p=0.15). Only in a Cox survival model was there a significant relationship between CL and time to birth (p=0.03), but this effect was null after controlling for maternal age. In logistic regression models, CL did not significantly predict PTB < 24, < 28, < 35, or < 37 weeks. Stratifying CL by 25-29 mm versus 30 mm or greater also had no significant predictive value for these 4 preterm birth cutoffs.
Women at high risk for recurrent preterm birth, but whose cervical length at < 23 weeks′ gestation remains ≥ 25 mm are still at increased risk of recurrent PTB (16% delivered < 35 weeks); however, CL measured before 23 weeks, whether considered on a continuum or stratified to consider CL′s near the 25 mm cutoff (25-29 mm), does not predict PTB.
© 2009 Mosby, Inc. Published by Elsevier Inc. All rights reserved.