Arrhythmia due to ischaemia and/or reperfusion is an important problem, especially in open heart surgery and for patients with ischaemic heart diseases undergoing non-cardiac surgery. This experimental study is planned to evaluate the effects of midazolam on ischaemia and/or reperfusion-induced arrhythmias by comparison with thiopentone in two sets of experiments (n = 20 for every group). In total ischaemia-reperfusion experiments, hearts were perfused in constant pressure conditions. In the control group, after a stabilisation period, perfusion was totally stopped for 30 min and the hearts were reperfused for 10 min. For the experimental groups, hearts were pretreated for 5 min with either 10(-6) mol l(-1) midazolam or 10(-5) mol l(-1) thiopentone before total ischaemia and reperfused for 10 min with the same concentrations of the drugs. In low-flow ischaemia-reperfusion experiments, hearts were perfused at a constant flow of 10 ml g(-1) heart per min initially. In the control group, after a stabilisation period, perfusion rate was decreased successively 1 ml g(-1) heart per min for 10 min (mild ischaemia) and to 0.2 ml g(-1) heart per min for another 10 min (severe ischaemia). The ischaemic hearts were then reperfused for an additional 10 min at a flow rate of 10 ml g(-1) heart per min. Electrogram recordings were evaluated before ischaemia and at the 5th and 10th min of mild ischaemia, severe ischaemia and reperfusion. Midazolam, 10(-6) mol l(-1), or thiopentone, 10(-5) mol l(-1), were added to the perfusion solution in the midazolam and thiopentone groups, respectively. In these two groups, hearts were perfused according to perfusion rates mentioned above in the control group. As a commonly used i.v. anaesthetic, thiopentone was arrhythmogenic for hearts exposed to ischaemia-reperfusion by increasing ventricular premature beat (% incidences for control, midazolam and thiopental groups in the 10th min of reperfusion were 25, 15 and 65 in total ischaemia-reperfusion experiments, P < 0.01) and ventricular tachycardia (respective % incidences were 0, 5, 25, P < 0.05) incidences, but in our experiments we found out that the new agent midazolam does not have more arrhythmia incidence than the respective control group in any criteria evaluated. None of the agents exerted atrioventricular conductance abnormalities. So we conclude that midazolam is a safe agent for ischaemic hearts and might also be antiarrhythmic and the mechanism of action of this effect remains to be further investigated. (C) 1998 The Italian Pharmacological Society.