Mortality rates are considerably higher in chronic ischemic heart disease (IHD) patients with non-insulin-dependent diabetes mellitus (NIDDM) than in those who are nondiabetics. The relationship between different types of antihyperglycemic pharmacological therapy and mortality rate in this NIDDM population is uncertain. We aimed to examine the survival in NIDDM patients with IHD using various types of oral antidiabetic treatments over a 5-year follow-up period. The study sample comprised 11,440 patients with a previous myocardial infarction and/or stable anginal syndrome, aged 45–74 years, who were screened, but not included in the Bezafibrate Infarction Prevention study. Among them, 9,045 were nondiabetics and 2,395 diabetics. The diabetic patients were divided into four groups on the basis of their therapeutic regimen at screening: diet alone (n = 990), sulfonylureas (n = 1,041), metformin (n = 78) and a combination of a sulfonylurea and metformin (n = 266). All NIDDM groups were similar with regard to age, gender, hypertension, smoking, heart failure, angina and prior myocardial infarction. Crude mortality rate was lower in the nondiabetic group (11.21 vs. 21.8%; p < 0.001). In the diabetic group, mortality was 18.5% for patients on diet alone, 22.5% for those on sulfonylureas, 25.6% for patients on metformin, and 31.6% for the combined sulfonylurea/metformin group (p < 0.01). When analyzing age-adjusted mortality rate and actuarial survival curves, the lowest mortality was found in patients on diet alone and the highest in patients on metformin (alone or in combination with sulfonylureas). After adjustment for variables connected with long-term prognosis, the use of metformin was associated with increased relative risk (RR) for all-cause mortality of 1.42 (95% CI 1.10–1.85), whereas the use of sulfonylureas alone was not [RR 1.11 (95% CI 0.90–1.36)]. NIDDM patients with IHD using metformin, alone or in combination with sulfonylureas, exhibited a significantly increased mortality. Until the results of problem-oriented prospective studies on oral control of NIDDM will be available, alternative therapeutic approaches should be investigated in these patients.