To describe associations between birthweight and infant, child and early adult mortality from natural causes in a remote Australian Aboriginal community against a background of rapidly changing mortality due to better health services.

Cohort study of 995 people with recorded birthweights who were born between 1956 and 1985 to an Aboriginal mother in a remote Australian Aboriginal community. Participants were followed through to the end of 2006.

Rates of natural deaths of infants (aged 0 to < 1 year), children (aged 1 to < 15 years) and adults (aged 15 to < 37 years), compared by birth intervals (1956–1965, 1966–1975 and 1976–1985 for infants and children, and 1956–1962 and 1963–1969 for adults) and by birthweight.

Birthweights were low, but increased over time. Deaths among infants and children decreased dramatically over time, but deaths among adults did not. Lower birthweights were associated with higher mortality. Adjusted for birth interval, hazard ratios for deaths among infants, children and adults born at weights below their group birthweight medians were 2.30 (95% CI, 1.13–4.70), 1.78 (95% CI, 1.03–3.07) and 3.49 (95% CI, 1.50–8.09), respectively. The associations were significant individually for deaths associated with diarrhoea in infants, with cardiovascular and renal disease in adults, and marginally significant for deaths from pulmonary causes in children and adults.

The striking improvements in infant and child survival over time must be applauded. We confirmed a predisposing effect of lower birthweights on deaths in infants and children, and showed, for the first time, an association between lower birthweights and deaths in adults. Together, these factors are probably contributing to the current epidemic of chronic disease in Aboriginal people, an effect that will persist for decades. Similar phenomena are probably operating in developing countries.