Birth weight has been linked to the risk of developing childhood cancer, in particular childhood leukaemia. However, despite many childhood cancers having a male predominance and boys generally weighing more than girls at birth few studies have reported sex-specific associations. The relationship between birth weight and childhood cancer risk was examined using information from a national case-control study. Children (0-14 years) newly diagnosed with cancer in GB were ascertained between 1991 and 1996 (n=3651) and for comparison, controls matched on sex, month and year of birth were identified from primary care population registers (n=6337). Birth weights were obtained from the Office of National Statistics for all targeted subjects born in England and Wales. Overall, cases were, on average, 30 g heavier at birth than controls (p=0.003) with differences seen by cancer type; those diagnosed with hepatic tumours weighing around 500 g less than controls at birth (p<0.0001) and those with leukaemia being, on average, 50 g heavier than those without (p=0.001). An interaction between birth weight and sex was found for acute leukaemia (chi(2)=11.2, p=0.04) and when data were stratified by sex, an association between high birth weight and risk of ALL was seen with girls (>4000 g, OR 1.86, 95% CI 1.38-2.50, chi(2) for trend 20.2, p<0.0001). Our results support the hypothesis that birth weight is an important determinant for childhood cancer. In addition, the data are consistent with the notion that childhood leukaemia has a prenatal origin.