Despite the implication of oxidative stress in CVD development, neither blood levels of nor supplementation with any one individual antioxidant is consistently related to CVD incidence and death. Thus, the relationship between diet-derived antioxidants and CVD remains unclear. However, many analyses consider the predictive value of only a few antioxidant markers in isolation, which may not reflect overall antioxidant status. We developed an antioxidant score and tested the hypothesis that it would be a more robust predictor of CVD death than concentrations of single antioxidants in aging adults. Quintiles of 10 serum antioxidant concentrations (α-carotene, β-carotene, β-cryptoxanthin, lutein, lycopene, vitamins A, C, E, retinyl esters and selenium), collected at baseline from 7094 (weighted N = 58,875,272) adults aged 50+ years enrolled in the Third National Health and Nutrition Examination Survey (1988-1994) were used as predictors in a time-to-death analysis using Cox regression modeling. Forty-two percent (N=3695, weighted to 26189208) of the sample was deceased at 12-18 year follow-up, with 1362 deceased due to CVD. In models adjusted for known demographic, behavioral and biological risk factors, concentration quintiles of the individual antioxidants α-carotene, lycopene, selenium and retinyl ester were significantly associated with CVD death (ps for trend ranged from 0.015 to 0.042, HRs of quintile 1 vs 5 ranged from 1.34 to 1.52). A stepwise regression method was used to identify the best combination of candidate markers, and an antioxidant score was computed using only those 6 markers (vitamins C and E, lycopene, selenium, α and β-carotene). Quintiles of the antioxidant score were also significantly associated with risk of death (p for trend = 0.004, HR for Q1 vs. Q5 = 1.88) and in a more predictive manner than a score using all 10 antioxidants (p for trend = 0.016, HR for Q1 vs. Q5 = 1.69). After controlling for traditional risk factors, an antioxidant score including vitamins C and E, lycopene, selenium, α and β-carotene was more strongly associated with rates of CVD death than individual antioxidants in a large nationally representative sample of aging adults. An antioxidant score may more closely reflect oxidative status than measures of single dietary antioxidants, and therefore be more useful when assessing risk of CVD death.