Background: Enteroviruses, in particular Coxsackie B4, have been implicated in the aetiology of Type 1 diabetes mellitus, but the epidemiological evidence has not been systematically evaluated.
Methods: Systematic review of evidence from published controlled studies of the relationship between Coxsackie B virus serology and incident or prevalent Type 1 diabetes mellitus. Studies were identified through a Medline search (1966 to 2002), supplemented by references from identified papers and hand search of relevant journals. All studies (full papers, abstracts or letters) with data adequate for calculation of unadjusted odds ratios (with 95% confidence intervals) for Type 1 diabetes mellitus in relation to Coxsackie B virus serology were included.
Results: The review included 26 case-control studies; no cohort study met the inclusion criteria. Odds ratios for Type 1 diabetes mellitus in serology-positive vs. serology-negative subjects ranged from 0.2 to 22.3. For Coxsackie B (any serotype) 7/13 studies had point estimates significantly greater than 1.0 (P < 0.05). For Coxsackie B3, Coxsackie B4 and Coxsackie B5-specific assays, 1/11, 6/17 and 1/11 studies, respectively, had point estimates significantly greater than 1.0. Summary odds ratios were not calculated because of doubts about the validity of individual study estimates, heterogeneity between studies, and the possibility of publication bias.
Conclusions: The results of these studies are inconsistent and do not provide convincing evidence for or against an association between Coxsackie B virus infection and Type 1 diabetes mellitus. Better designed studies using effective assays are needed to resolve this important issue.