In this study, we examined whether chronic severe diabetes may affect ischaemic and post-ischaemic regional myocardial dysfunction in vivo in the dog. Diabetes was chemically induced in randomized animals and major metabolic alterations were observed confirming the severity and chronicity of the diabetes. After 70 days, halothane-anaesthetized dogs underwent a 20-min coronary occlusion, followed by reperfusion. During ischaemia, global left ventricle function (dP/dtmax) was more altered (P<0.005) in diabetics ( n=10) than in controls (n=10), whereas area-at-risk (29+/-2.5% of the left ventricle in diabetics v 32.4+/-1.9% in controls) and ischaemic subendocardial myocardial blood flow (radioactive microsphere technique, 0.11+/-0.02 v 0.10+/-0.03 ml/min/g) were similar. During reperfusion, both groups developed significant (P<0.05) regional myocardial dysfunction (somomicrometry, 41+/-14% of baseline in controls and 66+/-8% in diabetics), whereas the difference between groups was not significant. No dog of either group developed myocardial cell necrosis on tissue histology. Multivariate analyses, including the severity of prior ischaemia and the occurrence of ventricular fibrillation as covariables, confirmed that myocardial stunning was not increased in diabetics, although ischaemia was clearly less-well-tolerated in diabetic dogs as global (dP/dtmax) as well as regional myocardial function were significantly (P<0.05) more altered in diabetics during ischaemia. Whilst alteration of arachidonate and cholesterol metabolism may partly explain this apparent paradox, further studies are required to resolve this issue.