Background: Iron overload such as that in idiopathic hemochromatosis is a well-established, albeit rare, cause of non-insulin-dependent diabetes mellitus (NIDDM). Most patients with NIDDM have no recognized cause of their disease. Investigators have proposed that subclinical iron overload may cause diabetes mellitus in a substantial number of patients with NIDDM. Objective: The aim of the current study was to evaluate hepatic iron stores in autopsy specimens from a group of community residents with NIDDM. Methods: Fifteen patients with NIDDM and 17 age-matched control subjects were identified from a review of medical records of deceased residents of Olmsted County, Minnesota. Formalin-fixed liver tissue was analyzed for iron concentration by flameless atomic absorption spectrophotometry, and distribution of hepatic iron was determined histochemically. Results: No significant difference was found in either the distribution or the mean amount of hepatic iron between the diabetic and the control group (1,303 versus 1,349 mu g Fe/g dry weight; P = 0.87). Thus, the mean difference was -46 mu g Fe/g dry weight (confidence interval, -631 to 540). Conclusion: Because hepatic iron quantification is the definitive means of assessing total body iron stores, our results suggest that NIDDM is typically not associated with a substantial level of iron overload.