Using serial invasive hemodynamics in concert with noninvasive radionuclide monitoring of left ventricular (LV) function, 20 patients with unstable angina were evaluated during incremental infusion of intravenous nitroglycerin. Of 20 patients, 17 demonstrated a favorable hemodynamic response and dose responses could be determined for individual patients. There was excellent agreement between hemodynamic and noninvasive radionuclide measurements, and patient responses could be inferred from the radionuclide data alone. Cardiac output improved by 29% (mean 4.7 +/- 1.2 to 6.0 +/- 1.3), LV ejection fraction increased an average of 0.11 (0.39 +/- 0.14 to 0.50 +/- 0.16) and diastolic function as assessed by peak filling rate improved from 1.80 +/- 0.60 end-diastolic volumes/s to 2.70 +/- 0.90. Changes in systolic blood pressure and heart rate were not predictive of hemodynamic response. The dose of nitroglycerin necessary to produce maximal hemodynamic benefit was variable (mean 98 micrograms/min, range 48 to 144). In 7 patients, nitroglycerin caused excessive decreases in pulmonary arterial wedge pressure and adverse hemodynamics that corrected with intravenous fluids, allowing continued administration of intravenous nitroglycerin with improvement in hemodynamic status. In 15 patients, peak systolic pressure--end-systolic volume relations were assessed to define possible changes in LV contractility induced during nitroglycerin infusion. In 12 of these patients, this relation was linear, suggesting altered loading rather than augmented contractility as nitroglycerin's mechanism of action. In 3 patients a downward and rightward shift of the systolic relation was seen, suggesting that significant underloading with nitroglycerin was associated with depressed contractility.