Mortality rates of very preterm infants may vary considerably between healthcare facilities depending on the neonates’ place of inclusion in the cohort study. The objective of this study was to compare the mortality rates of live-born extremely preterm neonates observed in two French tertiary referral hospitals, taking into account the occurrence of neonatal death both in the delivery room and in the neonatal intensive care unit (NICU).

Retrospective observational study including all pregnancy terminations, stillbirths and live-born infants within a 22- to 26-week ^0/6 gestational age range was registered by two French level 3 university centers between 2009 and 2013. The mortality rates were compared between the two centers according to two places of inclusion: either the delivery room or the NICU.

A total of 344 infants were born at center A and 160 infants were born at center B. Among the live-born neonates, the rates of neonatal death were similar in center A (54/125, 43.2%) and center B (33/69, 47.8%; P=0.54). However, neonatal death occurred significantly more often in the delivery room at center A (31/54, 57.4%) than at center B (6/33, 18.2%; P<0.001). Finally, the neonatal death rate of live-born very preterm neonates admitted to the NICU was significantly lower in center A (25/94, 26.6%) than in center B (27/63, 42.9%; P=0.03).

This study points out how the inclusion of deaths in the delivery room when comparing neonatal death rates can lead to a substantial bias in benchmarking studies. Center A and center B each endorsed one of the two models of preferential place of neonatal death (delivery room or NICU) detailed in European studies. The reasons behind the two different models and their impact on how parents perceive supporting their neonate need further investigation.