Ethnicity-specific growth distributions for prediction of newborn morbidity

Birth weight distributions differ according to the ethnic origin of the mother. We aimed to determine whether using ethnicity-specific growth distributions would identify newborns at risk for adverse infant outcomes associated with small for gestational age (SGA) birth better than population-based distributions.

We examined 2647 singletons born to healthy non-smoking mothers at British Columbia Women's Hospital at 37 to 41 completed weeks' gestation. We compared the ability of ethnicity-specific growth distributions and population-based distributions to predict which infants were at increased risk of adverse outcomes associated with SGA, as well as extended length of stay in hospital, using crude and adjusted odds ratios.

The ethnicity-specific growth distributions were associated with an increased risk of adverse infant outcomes while the population-based distributions were not (adjusted odds ratio [aOR] 1.49 [95% CI 0.82 to 2.70] vs. aOR 0.88 [95% CI 0.48 to 1.64]). While both distributions predicted extended length of stay in hospital, this likely reflects clinicians' use of the population-based distribution to identify SGA infants.

The use of ethnicity-specific growth distributions will likely improve our ability to differentiate between babies who are pathologically small and those who are constitutionally small, and prevent misclassification of constitutionally small but healthy newborns born to mothers of Chinese or South Asian descent as SGA.