We have demonstrated the feasibility of linking newborn blood spots, population-based cancer incidence data and birth certificate data. Incident cases of acute lymphocytic leukaemia and population-based controls were ascertained. We retrieved dried blood spot specimens, isolated and amplified DNA, and assayed the cancer susceptibility genes GSTT1 and GSTM1. The double null genotype was over-represented in the cases, consistent with previous reports based on other epidemiological methods. The design avoids issues of participation bias by cases and controls and can be used to investigate interactions of susceptibility genes and xenobiotics in semi-ecological studies. It can be useful for generating or testing hypotheses on associations of other paediatric illness and environmental contaminants.