SUMMARY The linearity and sensitivity of the end-systolic pressure-volume (P-Ve,) relation to the inotropic state of the left ventricle were investigated in 11 patients with coronary heart disease and one patient with congestive cardiomyopathy. To minimize autonomic reflex responses, propranolol, 0.15 mg/kg, and atropine, 1 mg, were administered i.v. at the beginning of the study. Three ventriculograms were performed: at rest, after oral isosorbide dinitrate, 10 mg (systolic pressure decrease 15 mm Hg), and during infusion of methoxamine, 2 mg/min (systolic pressure increase > 10 mm Hg).The three points of the P-Ve, relation showed linearity (r 0.96). The relation between the slope k of the PVe, relation and the left ventricular ejection fraction at rest was best described by an exponential function (r = 0.94). The use of peak systolic pressure instead of end-systolic pressure showed equally good results. The intercept of the P-V,, line on the abscissa, which represents the theoretical end-systolic volume at zero pressure, was not related to the ejection fraction under control conditions. The P-Ve, relation in postextrasystolic beats was displaced toward the left (smaller end-systolic volumes) and became steeper.MYOCARDIAL fiber shortening is determined by preload, afterload and contractility, and end-systolic fiber length is inversely proportional to afterload.' In the isolated left ventricle, end-systolic fiber length is linearly related to end-systolic tension.2 In the experimental animal, the isolated left ventricle also shows a linear end-systolic pressure-volume (P-Ves) relationship that is not dependent on preload. The slope of the P-Ve, relation is determined by the contractile state and is increased by positive inotropic interventions.3Although the concept of the P-Ve, relationship has been applied for some time in animal experiments,1-4only recently has the P-Ve, relation been investigated as a means of assessing left ventricular function in man.Our study was undertaken to determine (1) whether the P-VVe, relation in man is linear; (2) the correlation between the slope k of the P-V,e relation and other variables of left ventricular function; (3) whether the theoretical volume at zero pressure (VO), i.e., the intercept of the P-Ves relation on the abscissa, reliably separates impaired from normal left ventricular func-