Longitudinal change in cervical length following vaginal or abdominal cervical cerclage: a randomised comparison.

BACKGROUND: Cervical cerclage has been shown to reduce the risk of recurrent spontaneous preterm birth in a high-risk patient population, however the mechanism is not well understood. Transabdominal cerclage is superior to low and high vaginal cerclage in reducing early spontaneous preterm birth and fetal loss in women with previous failed vaginal cerclage. Cervical length measurements are commonly used to monitor high-risk women and may explain the mechanism of success.

OBJECTIVE: Our objective was to evaluate the rate of change in longitudinal cervical length after randomised placement of low transvaginal, high transvaginal or transabdominal cerclage in women with a previous failed vaginal cerclage.

STUDY DESIGN: This was a planned analysis of longitudinal transvaginal ultrasound cervical length measurements from patients enrolled in MAVRIC, a randomised controlled trial comparing transabdominal cerclage or high transvaginal cerclage with low transvaginal cerclage. Cervical length measurements at specific gestational ages were compared over time and between groups, using generalised estimating equations fitted using the maximum likelihood random effects estimator. Cervical length measurements were also compared in women with transabdominal cerclage placed before and during pregnancy. Diagnostic accuracy of cervical length as a predictor of spontaneous preterm birth less than 32 weeks was explored.

RESULTS: We included 78 women who underwent longitudinal cervical length assessment (70% of the analysed cohort) with a history of failed cerclage, of whom 25 (32%) were randomised to low transvaginal cerclage, 26 (33%) to high transvaginal cerclage, and 27 (35%) to transabdominal cerclage. Abdominal cerclage was superior to low (p=0.008) and high (p=0.001) vaginal cerclage at maintaining cervical length over the surveillance period (14 to 26 weeks' gestation) (+0.08 mm/week, 95% CI -0.40 to 0.22, p=0.580). On average, cervical length was 1.8mm longer by the end of the 12-week surveillance period in women with transabdominal cerclage (+1.8 mm, 95% CI -7.89 to 4.30, p=0.564). High vaginal cerclage was no better than low cervical cerclage in the prevention of cervical shortening; the cervix shortened by 13.2 mm over 12 weeks in those with low vaginal cerclage (95% CI -21.7 to -4.7, p=0.002) and by 20 mm over 12 weeks in those with high vaginal cerclage (95% CI -33.1 to -7.4, p=0.002). Pre-conception transabdominal cerclage resulted in a longer cervix compared to those performed during pregnancy; this difference was significant after 22 weeks' gestation (48.5mm versus 39.6 mm, p = 0.039). Overall, cervical length was an excellent predictor of spontaneous preterm birth less than 32 weeks' gestation (ROC 0.92, 95% CI 0.82 to 1.00).

CONCLUSION: In women with a previous failed cervical cerclage, in the next pregnancy the cervical length in women treated with vaginal cerclage funnels and shortens over time, whereas there is preservation of cervical length in women who receive transabdominal cerclage. Cervical length remains longer in transabdominal procedures performed before, compared to during, pregnancy. Overall cervical length is an excellent predictor of spontaneous preterm birth in this cohort. Our findings may explain the mechanism of benefit for transabdominal cerclage, with its high placement better maintaining the structural integrity of the cervix at the level of the internal os.