Objective: We examined whether the route of delivery for near-term (> or = 34 weeks' gestation) twins, as candidates for vaginal delivery, affected neonatal and infant mortality rates. We further evaluated whether these mortality rates were modified by fetal presentation.
Methods: A population-based retrospective cohort study based on the matched multiple births data in the USA (1995-97) was performed. Analyses were restricted to non-malformed liveborn twins delivered at (> or = 34 weeks' gestation. Twins with breech-breech and breech-vertex presentations were excluded, since they are not candidates for vaginal delivery. Neonatal mortality rates (death within the first 27 days) and post-neonatal mortality rates (death between 28 and 365 days) per 1000 twin live births, by route of delivery and fetal presentation, were derived. The associations between neonatal mortality, post-neonatal mortality and the route of delivery for vertex-breech versus vertex-vertex presentations were expressed based on relative risks (RR) and 95% confidence intervals (CI) derived from logistic regression models based on the method of generalized estimating equations.
Results: Of the 177,622 twins analyzed, 87% (n = 154,531) presented as vertex-vertex. Fifty-five per cent (n = 97,692) of twins were both delivered vaginally, 41% (n = 72,825) were both delivered by Cesarean section and, of the remaining 4% (n = 7,105), the first twin was delivered vaginally and the second by Cesarean section. Twins with vertex-breech presentations delivered by Cesarean-cesarean sections, as well as those with vertex-vertex presentations delivered vaginally, had the lowest neonatal mortality rate (1.6 per 1000 live births). The highest neonatal mortality rate in the vertex-breech pairs occurred with vaginal-Cesarean deliveries (2.7 per 1000 live births). Among twins with vertex-vertex presentations, twins delivered via the vaginal-Cesarean route experienced the highest neonatal mortality (3.8 per 1000 live births). The RR for neonatal mortality in this group was 2.24 (95% CI 1.35, 3.72) compared with twins both delivered vaginally.
Conclusion: Route of delivery and fetal presentation both confer an impact on twin infant mortality rates. Strategies to reduce discordant routes in complicated vaginal deliveries may lead to improved neonatal survival.