Repaglinide is a prandial glucose regulator indicated for management of type 2 diabetes. This post-marketing study used the observational cohort technique of prescription-event monitoring (PEM) to monitor safety of repaglinide prescribed in primary care in England. Patients were identified from dispensed prescriptions issued by general practitioners (GPs) between December 1998 and January 2001. Demographic and clinical event data were collected from questionnaires posted to GPs at least six months after the date of first prescription for each patient. The cohort consisted of 5731 patients [median age 60 (IQR 51-68), 49.9% male]. Event incidence densities (IDs) [no. 1st reports/1000 patient-months of exposure] were calculated for all events reported. The most frequently recorded clinical events in the first month were diarrhoea (ID(1) 10.3), malaise/lassitude (ID(1) 8.1) and nausea/vomiting (ID(1) 7.9). The most frequently reported reason for stopping was 'not effective' (647), with the most common clinical reasons being diarrhoea (60), malaise/lassitude (55) and intolerance (54). One hundred and thirteen adverse drug reactions (ADRs) were reported, with the most frequently specified being diarrhoea (10), abdominal pain (10) and nausea/vomiting (9). We concluded that repaglinide is generally well tolerated when used in general practice in England and did not identify any serious unrecognised adverse events.