The effect of general anesthesia induced with halothane and nitrous oxide on left ventricular (LV) contractility and diastolic mechanics was directly assessed in the chronically instrumented canine model. Seven dogs were instrumented with ultrasonic dimension transducers to measure LV diameter and micromanometers to measure LV transmural pressure. Contractility was assessed by the slope (EES) of the end-systolic pressure-diameter relationship PES = EES (LES - LD). Diastolic compliance was assessed by fitting end-diastolic pressure-dimension data to the exponential P = alpha (e beta epsilon L-1), where alpha and beta are nonlinear elastic coefficients. A new index (ID20) that identifies the dose of anesthetic necessary to depress the inotropic state by 20% was calculated to be 0.63% for halothane. Contractility was progressively decreased by halothane, with EES falling from 10.1 +/- 0.6 at control to 6.7 +/- 0.4 at 1% halothane and to 4.2 +/- 0.5 at 2% halothane (p less than 0.05 at each halothane level). However, similar levels of halothane did not significantly alter alpha and beta, nor did they significantly shift the exponential diastolic compliance curve from control. Seven patients who underwent coronary artery bypass grafting conducted under narcotic anesthesia showed a similar halothane-induced depression of contractility; 0.5% halothane decreased EES from 11.5 +/- 2.0 to 8.0 +/- 2.4 (p less than 0.01). Use of ID20 allows reclassification of anesthetic agents in accord with their myocardial depressant effects, which with halothane appears to be caused by decreased inotropism without alterations in diastolic chamber mechanics.