Specific defects in innate and adaptive immune function have been identified in diabetic patients in a range of in vitro studies. However, the relevance of these findings to the integrated response to infection in vivo remains unclear, especially in patients with good glycaemic control. Vaccine efficacy seems adequate in most diabetic patients, but those with type 1 diabetes and high glycosylated haemoglobin levels are most likely to exhibit hyporesponsiveness. While particular infections are closely associated with diabetes, this is usually in the context of extreme metabolic disturbances such as ketoacidosis. The link between glycaemic control and the risk of common community-acquired infections is less well established but could be clarified if infection data from large community-based observational or intervention studies were available. The relationship between hospital-acquired infections and diabetes is well recognized, particularly among post-operative cardiac and critically ill surgical patients in whom intensive insulin therapy improves clinical outcome independent of glycaemia. Nevertheless, further research is needed to improve our understanding of the role of diabetes and glycaemic control in the pathogenesis and management of community- and hospital-acquired infections. Copyright (c) 2006 John Wiley & Sons, Ltd.