The effects of desflurane and sevoflurane on the failing myocardium are still uncertain. We investigated the effects of different concentrations of sevoflurane, desflurane and halothane in dogs with pacing induced chronic heart failure. Global (left ventricular pressure, left ventricular dP/dt, Konigsbergtransducer) and regional myocardial function (systolic segment length shortening, ultrasonic crystals) were measured in chronically instrumented dogs with tachycardia induced severe congestive heart failure. Measurements were performed in healthy dogs and after induction of heart failure in the awake state and during anaesthesia with 0.75, 1.0, 1.25 and 1.75 minimum alveolar concentration (MAC) of halothane, sevoflurane or desflurane. The anaesthetics reduced dP/dtmax in a dose-dependent manner in healthy dogs (dP/dtmax decreased to 43-53% of awake values at 1.75 MAC). Chronic rapid left ventricular pacing increased heart rate and left ventricular end-diastolic pressure and decreased mean arterial pressure, left ventricular systolic pressure and dP/dtmax. The reduction in contractility was similar in the failing myocardium (to 41-50% of awake values at 1.75 MAC). Segmental shortening was reduced during anaesthesia by 50-62% after pacing compared with 22-44% in normal hearts. While there were similar effects of the different anaesthetics on diastolic function in healthy dogs, after induction of heart failure a more pronounced increase of the time constant of isovolumic relaxation and a greater decrease of dP/dtmin was observed with sevoflurane than with desflurane, indicating a stronger depression of diastolic function. While the negative inotropic effects of sevoflurane and desflurane were similar in normal and in the failing myocardium in vivo, desflurane led to a better preservation of diastolic function in the failing myocardium.