Amrinone was given to 14 patients with congestive heart failure as an intravenous bolus (1 mg/sec) at doses ranging from 0.5 to 3.5 mg/kg. Simultaneous determinations of cardiac index were made by thermodilution and of amrinone plasma concentration by high-performance liquid chromatography. A relationship between improvement in cardiac index and increasing plasma concentrations of amrinone was demonstrated for 13 of the 14 patients. The percentage increase in cardiac index correlated with amrinone plasma concentration (r = 0.81; p less than 0.001). Amrinone was given to four patients as an intravenous bolus dose of 1.5 mg/kg followed by a constant infusion of 10 micrograms/kg/min for 10 hr; simultaneous determinations of cardiac index and circulating levels of amrinone indicated that both declined after the initial rise. The plasma concentration of amrinone remained relatively constant during the infusion at about 1.7 micrograms/ml. In all cases, despite the relatively constant plasma levels there was a decline in cardiac index after about 4 to 5 hr of infusion, although the cardiac index remained above the baseline; during the constant infusion the cardiac index rose again and was maintained at a reasonably constant level for the last 3 hr. Seven patients received oral doses of amrinone of about 3 mg/kg, and simultaneous determinations of cardiac index and plasma concentration showed a relationship between amrinone level and rise in cardiac index (p less than 0.05). In 16 patients after amrinone orally sufficient blood samples were taken to estimate the apparent first-order terminal elimination t 1/2. The t 1/2 as estimated by log-linear regression ranged from about 3 to 15 hr; mean +/- SEM value was 8.3 (+/- 1.1) hr.