Sonographic cervical changes induced by Dilapan-S: a prospective observational study

To describe the sonographic cervical changes induced by Dilapan-S (DS) during cervical ripening at term and to explore their association with mode of birth.

This prospective observational study included singleton pregnancies with an unfavorable cervix (Bishop score ≤4) undergoing induction of labor at term (>37 weeks). DS was inserted under direct visualization and left in place for 12-16 h. Ultrasonographic assessment of cervical length (CL), posterior cervical angle (PCA), and fetal occiput position was performed before DS insertion and at rod removal. The exposures of interest were CL, PCA and fetal occiput position measured at the time of DS removal; the corresponding changes between insertion and removal (delta values) were examined as secondary exposures. The primary outcome was mode of birth, defined as vaginal birth (spontaneous or operative) versus cesarean delivery; vaginal birth was modelled as the event and cesarean delivery was the reference category. The secondary outcome was the induction-to-delivery interval, analysed in women who gave birth vaginally. Maternal age, body mass index (BMI), parity and Bishop score at DS removal were considered a priori as potential confounders.

Among 37 women who underwent cervical ripening with DS, 28 (75.7%, 95% CI 59.9-86.6) gave birth vaginally. All 11 multiparous women delivered vaginally, in every case by spontaneous vaginal delivery (100%, 95% CI 74.1-100), compared with 17 of 26 nulliparous women (65.4%, 95% CI 46.2-80.6; risk difference 34.6%, 95% CI 4.6-53.8; p = 0.036). The median induction-to-delivery interval was 27.0 h (IQR 24.8-30.0) among women who gave birth vaginally. All cervical sonographic parameters changed significantly after DS placement, while fetal head position did not. CL at DS removal was significantly shorter in women who delivered vaginally than in those who underwent cesarean section (21.3 vs 26 mm; median difference -6.1 mm, 95% CI -12.8 to -0.1; p = 0.047) and correlated significantly with the induction-to-delivery interval (r = 0.604, 95% CI 0.30-0.80; p = 0.0006). No other sonographic variable was significantly associated with the induction-to-delivery interval. In the prespecified multivariable model fitted with Firth penalised likelihood, cervical length at Dilapan-S removal showed a consistent association with vaginal delivery (adjusted OR 0.919; 95% CI, 0.807-0.989; p = 0.032), with similar findings after adjustment for parity (adjusted OR, 0.899; 95% CI, 0.777-0.991; p = 0.030) and in the cervical length-only model (adjusted OR, 0.923; 95% CI, 0.832-0.964; p = 0.004).

DS induces measurable sonographic changes in cervical length and posterior cervical angle. Shorter CL at the time of dilator removal was associated with vaginal birth and with a shorter induction-to-delivery interval. These findings should be regarded as hypothesis-generating rather than predictive; they require confirmation in larger studies.