Incidence of inflammatory myopathies in Victoria, Australia, and evidence of spatial clustering

Objective. To determine the incidence of idiopathic inflammatory myopathies (IIM) in Victoria, Australia, and look for evidence of space-time or spatial clustering. Methods. Cases of IIM diagnosed between 1989 and 1991 were identified by muscle biopsy and hospital discharge diagnosis review. Diagnosis was verified by medical record review and included if Bohan and Peter criteria for definite or probable disease were met. The pair-wise Euclidean distances between cases' residences were computed using grid references, and temporal distances were calculated between biopsy dates. The Mantel test for space-time clustering was computed. Each patient was also characterized by statistical local area (SLA) according to place of residence. For each SLA, the expected annual incidence of IIM was calculated, based upon its population distribution, and these were compared to the observed annual incidence. Confidence intervals for the true rate ratio (RR) for each SLA were calculated assuming a Poisson distribution, and the level of heterogeneity in the data was examined by calculation of a chi-squared for homogeneity. Results. Ninety-four cases met inclusion criteria for an annual incidence of 7.4 (95% CI 6.0-9.0) per million person-years. No space-time clustering was found (z = -0.434, p = 0.665), but there was evidence of spatial clustering. A total of 67 observed cases were distributed among 58 urban SLA. Four SLA had a greater than expected incidence of myositis (95% Poisson based CI excluded 1), accounting for 20 of the observed cases. Conclusion. The incidence of IIM in Australia is higher than most previous population based estimates. The finding of spatial clustering supports the hypothesis that environmental factors may be important in the pathogenesis of these diseases.