Epidemiologic studies have shown a correlation between antioxidant intake and coronary artery disease (CAD); however, the results of clinical trials have been inconsistent. We evaluated the effect of combined antioxidant supplementation on endothelial function and its correlation with change in low-density lipoprotein cholesterol (LDLC) oxidation in patients with established CAD. In a double-blind, placebo-controlled 12-week trial, 18 nonsmoking, nondiabetic patients (mean age 62.4 +/- 8.1 years) were randomized to receive placebo or antioxidant supplementation consisting of (a) 400 IU of vitamin E, 500 mg of vitamin C, and 12 mg of beta-carotene; or (b) 800 IU of vitamin E, 1000 mg of vitamin C, and 24 mg of beta-carotene daily. Endothelial function was evaluated on the basis of percent and absolute changes in brachial artery diameter in response to reactive hyperemia induced by occlusion-release. Baseline and 12-week values of LDL oxidation (measured on the basis of lag phase), endothelial function, dietary composition, serum antioxidants, and lipids were measured. We noted a significant between-group difference at 12 weeks for change in plasma concentrations of alpha-tocopherol, vitamin C, and beta-carotene between the placebo and antioxidant groups (p <.05). Both placebo and treatment groups demonstrated a significant improvement in lag phase; however, the treatment group achieved a greater, although nonsignificant, magnitude of change compared with the placebo group (181.3 +/- 177.8 minutes vs 80.6 +/- 63.0 minutes, P =.06). Within-group change in brachial reactivity from baseline to follow-up in the treatment group did not reach statistical significance (1.7% +/- 3.2% and 0.07 mm +/- 0.13 mm, P =.08 and P =.09, respectively), whereas an improved change in brachial reactivity was observed in the placebo group (2.2% +/- 1.9%, 0.09 mm +/- 0.06 mm, P <.05). No significant correlation was found between change in lag phase and change in endothelial function. On adjustment for confounders, antioxidant supplementation was found not to be a significant predictor of brachial reactivity. We conclude that antioxidant supplementation did not significantly alter brachial reactivity, despite significantly increased plasma levels of antioxidants and improved lag phase. These data should be confirmed in larger-scale trials and examined in studies evaluating individual dietary antioxidant supplementation.