First Twin in Breech Presentation and Neonatal Mortality and Morbidity According to Planned Mode of Delivery
Diane Korb1, François Goffinet, Florence Bretelle
1Université de Paris, Epidemiology and Statistics Research Center/CRESS, INSERM, INRA, the Department of Obstetrics and Gynecology, Robert Debré Hospital, APHP, and Port-Royal maternity unit, Cochin Hospital, APHP, Paris, the Department of Gynecology-Obstetric and Reproductive Medicine, Hôpital La Conception/Hôpital Nord, AP-HM, and Aix Marseille Univ, UM 63, CNRS 7278, IRD 198, INSERM 1095, Marseille, Inserm, UMR1027, Equipe SPHERE, the Université de Toulouse III, UMR1027, and CHU Toulouse, Pôle de gynécologie obstétrique, Hôpital Paule de Viguier, Toulouse, the Obstetrical Care Unit, Hôpital Couple Enfant, Grenoble University Hospital, CS 10217, 38043, Grenoble, the Department of Obstetrics and Gynecology, Bordeaux University Hospital, Bordeaux, and the Department of Obstetrics and Gynecology, CHU de Rouen, Université de Rouen, Rouen, France.
Objective:
To compare neonatal mortality and morbidity of first twins according to the planned mode of delivery when the first twin is in breech presentation, in a country where planned vaginal delivery is an option.
Methods:
This is a planned secondary analysis of the JUMODA (JUmeaux MODe d'Accouchement) cohort, a national prospective population-based study of twin deliveries conducted in 176 French hospitals. We analyzed pregnancies with first twins in breech presentation and applied the inclusion criteria of the Twin Birth Study (except the criterion for first-twin presentation): both fetuses alive, with a birth weight between 1,500 g and 4,000 g, at or after 32 0/7 weeks of gestation. The primary outcome was a composite of neonatal mortality and morbidity. We used multivariate Poisson regression models to control for potential confounders and propensity score analyses, that is, matching and inverse probability of treatment weighting to control for indication bias.
Results:
Among the 1,467 women with a breech-presenting first twin included in this analysis, 1,169 (79.7%) had planned cesarean and 298 (20.3%) planned vaginal births, of whom 185 (62.1%) delivered both twins vaginally. The neonatal mortality and severe morbidity rate for first twins was 1.7% (5/298) in the planned vaginal and 1.9% (22/1,169) in the planned cesarean delivery groups (crude relative risk [RR] 0.90, 95% CI 0.34-2.34). Planned vaginal delivery was not associated with higher neonatal mortality and morbidity than planned cesarean delivery, regardless of the statistical method used: adjusted RR 0.71, 95% CI 0.27-1.86; RR 0.61, 95% CI 0.20-1.83 after matching for propensity score; RR 0.63, 95% CI 0.23-1.74 with inverse probability of treatment weighting. Analyses of neonatal mortality and morbidity of second twins yielded similar results.
Conclusion:
Although our sample size precluded a robust assessment for small differences in outcomes between planned cesarean and planned vaginal delivery in twin pregnancies in which the first twin was in breech presentation, in our cohort planned vaginal delivery was not associated with higher neonatal mortality and morbidity for either twin.
