Diabetes, coronary heart disease and sulphonylureas—not the final word

severe a subsequent, more prolonged episode. The sulphonylurea glibenclamide has been shown to block both ischaemic and pharmacological preconditioning in various animal species and tissues including human cardiac muscle. In humans therapeutic concentrations of glibenclamide inhibit diazoxide (a KATP opener) induced vasodilatation in human forearm inferring an in vivo action of the drug. In addition, ischaemic pre-conditioning during coronary angioplasty is inhibited by glibenclamide. These observations have prompted reviews seeking to clarify the possible clinical consequences of KATP channel blockade by sulphonylureas. Klamann et al. have provided important and reassuring evidence that glibenclamide treatment does not affect in-hospital mortality of type II diabetic patients following acute myocardial infarction. Over a 6·5 year period (January 1991 to June 1997) all patients admitted with acute myocardial infarction were evaluated retrospectively and divided into four groups; non-diabetics, type II diabetics newly diagnosed on admission, type II diabetics taking glibenclamide and type II diabetics not taking sulphonylureas. Of 607 patients identified less than 1% were excluded because of lack of information on the diabetic state. Eight patients with type 1 diabetes were excluded. Mortality, creatinine kinase levels, atrial and ventricular arrhythmias and patient characteristics including possible confounding factors were recorded. Diabetic patients had a higher in-hospital mortality but there was no difference between type II diabetics treated with or without glibenclamide. In addition and consistent with previous reports the authors report a tendency towards lower creatine kinase increments in known diabetics regardless of the type of treatment. This possibly relates to the more diffuse coronary atherosclerosis in diabetics with less myocardium dependent on the occluded vessel. This study leaves open the possibility that more sulphonylurea-treated diabetic patients die before hospital admission. This is an important caveat as almost a half of all patients with acute myocardial infarction do not reach hospital and this figure may be even higher in the diabetic population. Can these apparent conflicting findings be reconciled? On the one hand, important evidence of potentially adverse effects on KATP channel activity in the myocardium and vasculature and on the other, no demonstrable effect on in-hospital mortality and See Eur Heart J 21: 3, pages 220–229 for the article to which this Editorial refers