Aims Birth weight above the 90th percentile on growth charts based on population norms, or large for gestational age (LGA), are associated with adverse perinatal outcome. We compared the association between neonatal mortality, morbidity and LGA as determined by currently used centiles based on population standards (method A) and customised birth weight centiles (method B).
Methods Cohort study of 6125 consecutive neonates, labelled as LGA using both methods (A) and (B), the latter calculated by adjusting for maternal height, booking weight, parity, ethnic origin, gestational age at delivery and sex. The outcomes were analysed in function of the method used to diagnose LGA.
Results There were 1295 LGA infants (prevalence of 11.3%) with the majority (94.7%) not born to diabetic mothers and 90.4% born at term. Method A alone identified 44.5%, method B alone 11.3% and both methods concurred in identifying 44.2% of these infants. After correcting for prematurity, LGA infants diagnosed by method B alone (not identified by standard population method) did not have an increase in morbidity or mortality except for congenital anomalies (p = 0.01).
Conclusion With the majority of LGA infants being born from non-diabetic mothers and with 11.3% not correctly identified by using population-based growth standards, these infants would therefore not be considered at risk of LGA-associated complications.