Non-insulin-dependent diabetic (NIDDM) patients show a high incidence of cardiovascular disease, with greater risk of recurrent myocardial infarction and a less favourable clinical outcome than non-diabetic patients. The majority of NIDDM patients are treated with sulphonylurea (SU) derivatives. In the 1970's the University Group Diabetes Program concluded that tolbutamide treatment caused increased cardiovascular mortality; the study, which led to curtailment of oral antidiabetic treatment in the USA, was received with scepticism in Europe. Later criticism of its methodology reduced the impact of the study; however, the question of the safety of SU in NIDDM patients with cardiovascular disease has been re-opened in the face of new experimental data. The heart and vascular tissues do have prerequisites for SU action, i.e. SU receptors and ATP-dependent K+ (K+ATP) channels. These channels play an important role in the protection of the myocardium against ischaemia-reperfusion damage, and their closure by SU could lead to amplified ischaemic damage. Here we review evidence from animal and human studies for deleterious SU effects on ischaemia-induced myocardial damage, either by direct action or through diminished cardioprotective preconditioning. Closure of K+ATP channels by SU can lead to reduction of post-infarct arrhythmias; the drug has also been claimed to improve various atherosclerosis risk factors. The evidence for these beneficial effects of SU is also reviewed. We look at the major difficulties that hamper transfer of information from experimental studies to clinical decision-making: a) The affinity of SU for heart K+ATP channels is orders of magnitude lower than for beta-cell channels; is it reasonable to expect in vivo cardiac effects with therapeutic 'pancreatic' SU doses? b) Most studies utilized high doses of acutely administered SU; are effects similar in the chronic steady-state of the SU-treated diabetic patient? c) Convincing SU effects have been demonstrated in acutely induced ischaemia by acutely administering the drug; do such effects persist in the clinical situation of gradually progressive ischaemia? d) Ischaemia and modification of K+ATP channel activity induce complex events, some with opposing effects; what is the net result of SU action, and do different SU derivatives lead to different outcomes? e) In the chronic (and hence clinically relevant) situation, how can direct (deleterious or beneficial) SU effects be separated from beneficial effects mediated by the metabolic action of the drug? Only large prospective clinical studies, making use of advanced technology for assessment of cardiovascular function, can answer these questions. Millions of NIDDM patients are treated with SU derivatives; many are in the age group where cardiovascular risks are extremely high. The question of whether SU derivatives are beneficial or deleterious for these patients must finally be settle unequivocally.