Geographic density of landfill sites and risk of congenital anomalies in England.

OBJECTIVE To investigate the risk of congenital anomalies in relation to an index of geographic density of landfill sites across 5x5 km grid squares in England. METHODS 2 km zones were constructed in a geographical information system around 8804 landfill sites, including 607 that handled special (hazardous) wastes, and intersected with postcode coordinates of over 10 million births (136,821 with congenital anomalies), 1983-98. A landfill exposure index was calculated to represent the geographic density of landfill sites within 2 km of births for each 5x5 km grid square, calculated separately for landfill sites handling special, and non-special or unknown, waste. For each group of landfills, the index was classified into four categories of intensity, and risks for the second, third and top categories were compared to the bottom category, comprising areas with no such landfill sites within 2 km (index of zero). We used hierarchical logistic regression modelling in a Bayesian framework, with adjustment for potential confounding. RESULTS For special waste sites, adjusted odds ratios were significant for the third category of the landfill exposure index for all anomalies combined and cardiovascular defects (OR 1.08 (95% credible interval 1.02 to 1.13) and 1.16 (1.00 to 1.33), respectively) and for hypospadias and epispadias for the third and top categories (OR 1.11 (1.02 to 1.21) and 1.12 (1.02 to 1.22), respectively). After adjustment, there were no excess risks in relation to sites handling non-special or unknown waste types. CONCLUSIONS There was a weak spatial association between risk of certain congenital anomalies and geographic density of special (hazardous) waste sites at the level of 5x5 km grid squares. Exposure pathways and mechanisms to help interpret these findings are not well-established.