The directional response of human left ventricular stroke volume during exercise recovery is unclear. Stroke volume has been reported to increase and decrease over exercise values during early recovery. The confounding variable may be posture. With the use of pulsed Doppler ultrasound, we tested the hypothesis that there is a significant difference between seated and supine stroke index (SI) during passive recovery from seated ergometer exercise. Thirteen subjects aged 26 +/- 2 yr performed two seated cycle ergometer exercise tests to 70% of predicted maximum heart rate (HR). Recovery was supine on one test and seated on the other. Cardiac index (CI), HR, and SI were calculated during rest, exercise, and 10 min of recovery. At rest, SI and CI were significantly (P less than 0.01) less and HR significantly (P less than 0.01) greater when the subjects were seated than when they were supine. At the last exercise work load, no significant differences were found in any measured variable between tests. During recovery, supine SI was maximal 180 s postexercise (99 +/- 14 ml/m2) and exceeded (P less than 0.01) resting supine (81 +/- 14 ml/m2) and peak exercise (77 +/- 14 ml/m2) SI by 22 and 29%, respectively. Seated SI was constant at peak exercise levels for 2 min. Seated and supine recovery CI never exceeded exercise values. Systolic and diastolic blood pressure recovery curves were similar in the two postures. We conclude that posture significantly affects SI during recovery from submaximal seated exercise. These results have implications for choice of recovery posture after stress testing in cardiac patients where it is desirable to minimize ventricular loading.