Diabetes mellitus increases the rate of atrioventricular block after a myocardial infarction; right bundle branch block is more common than expected in the diabetic outpatient population. We decided to assess whether diabetic subjects were more likely to need permanent cardiac pacemaker insertion. Data from patients in Leicestershire who had undergone permanent cardiac pacemaker insertion during a 4-year period were analysed. The capture recapture technique was used: Hospital Activity Analysis data as capture, ward admission book and case note confirmation as recapture. Diabetes coding accuracy was confirmed from the central register and from an analysis of 100 sets of case notes. The number of these patients with diabetes was then recorded. Data for the reference population was obtained using a survey of a market town in Leicestershire. There were 688 patients, most (79%) aged over 65 years. Our analysis was performed on this group. Of this age group undergoing pacemaker insertion. 11.1% was diabetic. In the same age group, 8.3% of the controls were diabetic: a relative risk of 1.34 (P < 0.01, 95% confidence interval 1.25-1.44). This relative risk is likely to be an underestimate. The aetiology of this excess risk is uncertain: it is most likely due to ischaemic heart disease, but microangiopathy or increased cholinergic sensitivity may play a role.