The aim of this study was to compare the direct effects of equivalent molar concentrations of sevoflurane (SEVO) and isoflurane (ISO) on electrophysiology, mechanical function, metabolism, and perfusion in isolated hearts, independent of neuronal, humoral, or hemodynamic influences. Three equimolar concentrations of SEVO or ISO were administered randomly in each of 14 guinea pig hearts perfused by the Langendorff technique. Spontaneous heart rate (HR), atrioventricular (AV) conduction time, left ventricular pressure (LVP), and coronary flow (CF) were measured directly. To differentiate a direct vasodilatory effect from an in direct metabolic effect due to autoregulation of CF, arterial and coronary sinus oxygen tension were measured continuously to calculate oxygen delivery (Do(2)), myocardial oxygen consumption (MVo(2)), percent O-2 extraction, and cardiac efficiency. Linear slope analysis (cardiac effect as a function of 0.1 mM anesthetic concentration) was used to compare anesthetic effects. Only AV time was increased more (P < 0.05) by ISO (+1.8 ms per 0.1 mM) than by SEVO (+1.1 ms per 0.1 mM). CF tended to be higher with ISO (+0.7 mL . g(-1). min(-1) per 0.1 mM) than SEVO (0.4 mL . g(-1). min(-1) per 0.1 mM) but this was not significant. HR (ISO, -1.4% per 0.1 mM; SEVO, -1.7% per 0.1 mh?), LVP (ISO, -5.8% per 0.1 mM; SEVO, -5.1% per 0.1 mM), and percent O-2 extraction (ISO, -6.1% per 0.1 mM; SEVO, -5.8% per 0.1 mM) were decreased similarly by both anesthetics and these effects were accompanied by proportional decreases in MVo(2), (ISO, -34% +/- 4%, SEVO, -37% +/- 6%) at the highest concentrations (0.53 mM). Neither anesthetic altered cardiac efficiency. SEVO and ISO produce equivalent negative inotropic, chronotropic, metabolic, and vasodilatory effects in a concentration-dependent manner in isolated hearts. Only AV conduction is delayed more by ISO than by SEVO.