This study tests the hypothesis that naloxone, but not its quarternary salt, naloxone methiodide (which does not enter the central nervous system), improves recovery from myocardial stunning in conscious dogs. Twenty dogs were chronically instrumented for measurement of heart rate, left atrial, aortic and left ventricular pressure (LVP), LV dP x dtmax(-1) and myocardial wall thickening fraction (WTF). Regional myocardial blood flow was determined with coloured microspheres. Occluder around the left anterior descending artery (LAD) allowed induction of reversible LAD ischaemia. Each of the 20 dogs underwent two LAD ischaemic challenges. Experiments (performed on separate days, in crossover fashion) were: (i) 10 min of LAD occlusion after application of naloxone 63 microg kg(-1) or naloxone methiodide 63 microg kg(-1) and (ii) occlusion without naloxone or naloxone methiodide. WTF was measured at baseline and until complete recovery occurred. LAD ischaemia significantly reduced LAD WTF with (mean (SD) 54 (15)% lower than baseline) and without naloxone (55 (16)% lower than baseline), without significant haemodynamic differences. Between I to 30 min of reperfusion, WTF was significantly higher with naloxone (P < 0.05). There was no difference in WTF with or without naloxone methiodide. We conclude that naloxone improved recovery from myocardial stunning in conscious dogs, and that this was centrally mediated.