What birthweight percentile is associated with optimal perinatal mortality and childhood education outcomes?

Small for gestational age, defined as birthweight <10th percentile for gestational age, is known to be associated with clinically meaningful impairments in health and development. The effects of variation within the normal range of birthweight percentile on perinatal mortality and childhood education remain less well defined. We sought to quantify the association among birthweight percentile, perinatal mortality, and educational outcomes and to determine the optimal birthweight percentile for those outcomes in Aboriginal and non-Aboriginal Australian children. This was a retrospective cohort study. Perinatal data for all children born in the Northern Territory, Australia, from 1999 through 2008 were linked to measures of educational attainment at age 8-9 years. Multivariable analysis was used to determine the optimal birthweight percentile for low perinatal mortality and high reading and numeracy scores. The birth cohort contained 35,239 births (42% Aboriginal), of which 11,214 had linked and valid education records. Median birthweight percentile was 29.2 in Aboriginal infants and 44.0 in non-Aboriginal infants. The odds of perinatal mortality decreased by 4% with each 1-percentile increase birthweight percentile overall (adjusted odds ratio, 0.96; P = .000) and lowest mortality rates were at the 61st and 78th percentile in Aboriginal and non-Aboriginal infants, respectively. Although birthweights <10th percentile were associated with greatly increased odds of perinatal mortality, the increased risk extended well beyond this cut-off. Birthweight percentile was also positively correlated with scores in reading (P = .000) and numeracy (P = .000). In non-Aboriginal children, reading and numeracy scores peaked at the 66th percentile, but for Aboriginal children there was continuous benefit with increasing birthweight percentile. Birthweight percentile explained 1% of the variation in education outcomes, with much greater variation explained by other perinatal and sociodemographic factors. Birthweights between ...