Space-time clustering analyses of type 1 diabetes in children from north-east England: support for an infectious aetiology?

BACKGROUND: The aetiology of type 1 diabetes in children is uncertain. A number of recent studies have suggested an infectious aetiology. It has been postulated that an infectious agent may be involved. Support for this hypothesis may be provided by a finding of space-time clustering. The aims of this study were: (i) to determine whether there was space-time clustering in cases of childhood diabetes from north-east England; and to test for differences in space-time clustering: (ii) due to age at diagnosis; (iii) between the sexes and (iv) between levels of residential population density. METHODS: We studied incidence of type 1 diabetes diagnosed in children aged 0-14 years and diagnosed during the period 1990-2007. All cases were resident in a defined geographical region of north-east England (Northumberland, Newcastle upon Tyne and North Tyneside). We applied a second-order procedure based on K-functions to test for global clustering. Locations were residential addresses at time of diagnosis. Tests were repeated using nearest neighbour thresholds to allow for variable population density, providing the primary result for each analysis. Differences between sexes and between levels of population density were assessed. RESULTS: We analysed 457 cases of type 1 diabetes. Overall, there was marginally significant evidence of global space-time clustering (P = 0.089). There was statistically significant clustering amongst pairs of cases that contained at least one female (P = 0.017), but not amongst pairs of cases that contained at least one male (P = 0.190). Furthermore, there was significant clustering amongst pairs of cases that contained at least one from a more densely populated area (P = 0.044), but not amongst pairs of cases that contained at least one from a less densely populated area (P = 0.226). CONCLUSION: Although the analyses have only found marginally significant evidence of global space-time clustering for cases of type 1 diabetes diagnosed in north-east England, there were two notable findings. First, there was evidence of clustering amongst females and secondly clustering was confined to cases from more densely populated areas. These findings are consistent with a possible aetiological involvement of an infectious agent.