The effects of glucose-insulin-potassium (GIK) and placebo normal saline (S) infusion on treadmill-walking time to angina, ST depression, heart rate (HR), systolic blood pressure (SBP), rate pressure product (RPP), blood glucose (G), lactate (L) and free fatty acids (FFA) were studied in 14 non diabetic patients with exertional angina. For the whole group, the post-GIK walking time to angina (393 ± 33 sec, mean ± SEM) was greater than the values during control GIK (319 ± 20 sec, p < 0.02) and post-S infusion (334 ± 23 sec, p < 0.05), but circulatory and ST responses were similar in post-GIK and post-S studies. 7 of the 14 patients experienced significantly greater improvement in exercise tolerance following GIK (467 ± 39 sec) in comparison to control GIK (313 ± 29 sec, p < 0.001) and post-S infusion (334 ± 32 sec, p < 0.005) and exercised to a higher HR, SBP and RPP after GIK than after S infusion. At the onset of angina these patients had similar ST-segment depression before and after GIK but when ST segments were assessed after GIK at the same exercise duration when angina had occurred during the control and post-S studies, there was significantly less ST depression (p < 0.01). Of the remaining 7 patients exercise tolerance following GIK deteriorated in 3, remained unchanged in 2 and increased by 12 and 48 sec in 2 patients in comparison to post-S values. Comparison of post-GIK and post-S values for G, L and FFA for the whole group showed significantly lower resting values of FFA and post-exercise values of G following GIK infusion. The differences in clinical and circulatory responses between patients who improved and those who did not improve following GIK were not related to the angiographically determined severity of coronary artery disease or to GIK-induced metabolic changes. Results suggest that some patients with angina pectoris do benefit from GIK infusion but the response in a given patient to this therapeutic modality is unpredictable.