It is well established that endothelial dysfunction is present in coronary artery disease (CAD), although few studies have determined the effect of training on peripheral conduit vessel function in patients with CAD. A randomized, crossover design determined the effect of 8 wk of predominantly lower limb, combined aerobic and resistance training, in 10 patients with treated CAD. Endothelium-dependent dilation of the brachial artery was determined, by using high-resolution vascular ultrasonography, from flow-mediated vasodilation (FMD) after ischemia. Endothelium-independent vasodilation was measured after administration of glyceryl trinitrate (GTN). Baseline function was compared with that of 10 control subjects. Compared with matched healthy control subjects, FMD and GTN responses were significantly impaired in the untrained CAD patients [3.0 +/- 0.8 (SE) vs. 5.8 +/- 0.8% and 14.5 +/- 1.9 vs. 20.4 +/- 1.5%, respectively; both P < 0.05]. Training significantly improved FMD in the CAD patients (from 3.0 +/- 0.8 to 5.7 +/- 1.1%; P < 0.05) but not responsiveness to GTN (14.5 +/- 1.9 vs. 12.1 +/- 1.4%; P = not significant). Exercise training improves endothelium-dependent conduit vessel dilation in subjects with CAD, and the effect, evident in the brachial artery, appears to be generalized rather than limited to vessels of exercising muscle beds. These results provide evidence for the benefit of exercise training, as an adjunct to routine therapy, in patients with a history of CAD.