Аннтотация
© 2020 Elsevier B.V. Introduction: Vaginal birth after caesarean delivery is associated with better outcomes compared to repeat caesarean section. Accurate antenatal risk stratification of women undergoing a trial of labor after caesarean section is crucial in order to maximize perinatal and maternal outcomes. The primary aim of this study was to explore the role of antepartum ultrasound in predicting the probability of vaginal birth in women attempting trial of labor; the secondary aim was to build a multiparametric prediction model including pregnancy and ultrasound characteristics able to predict vaginal birth and compare its diagnostic performance with previously developed models based exclusively upon clinical and pregnancy characteristics. Methods: Prospective study of consecutive singleton pregnancies scheduled for trial of labor undergoing a dedicated antepartum ultrasound assessment at 36–38 weeks of gestation. Head circumference, estimated fetal weight cervical length, sub-pubic angle were recorded before the onset of labour. The obstetricians and midwives attending the delivery suite were blinded to the ultrasound findings. Multivariate logistic regression and area under the curve analyses were used to explore the strength of association and test the diagnostic accuracy of different maternal and ultrasound characteristics in predicting vaginal birth. Comparison with previously reported clinical models developed by the Maternal-Fetal Medicine Unit Network (Grobman's models) was performed using De Long analysis. Results: A total of 161women who underwent trial of labor were included in the study. Among them 114 (70.8 %) women had successful vaginal birth. At multivariable logistic regression analysis maternal height (adjusted odds ratio (aOR):1.24;9 5% Confidence Interval (CI)1.17−1.33), previous C-section for arrest labor (aOR:0.77; 95 %CI0.66−0.93), cervical dilation at admission (aOR:1.35; 95 %CI1.12−1.74), fetal head circumference (aOR:0.77; 5%CI0.43−0.89), subpubic angle (aOR:1.39 95 %CI1.11−1.99) and cervical length (aOR:0.82 95 % CI0.54−0.98) were independently associated with VBAC. A model integrating these variables had an area under curve of 0.839(95 % CI 0.710−0.727) for the prediction of vaginal birth, significantly higher than those achieved with intake (0.694; 95 %CI0.549−0.815; p = 0.01) and admission (0.732: 95 % CI 0.590−0.84; p = 0.04) models reported by Grobman. Conclusion: Antepartum prediction of vaginal birth after a caesarean section is feasible. Fetal head circumference, subpubic angle and cervical length are independently associated and predictive of vaginal birth. Adding these variables to a multiparametric model including maternal parameters improves the diagnostic accuracy of vaginal birth compared to those based only on maternal characteristic.