The determinants of blood levels of carotenoids were previously investigated in small or select samples. The relations of serum lutein + zeaxanthin to possible diet, lifestyle, and physiological determinants in 7059 participants of the Third National Health and Nutrition Examination Survey (1988-1994), > or = 40 y old, were examined. In a fully adjusted, multiple linear regression model, lower serum lutein + zeaxanthin was significantly associated with smoking, heavy drinking, being white, female, or not being physically active, having lower dietary lutein + zeaxanthin, higher fat-free mass, a higher percentage of fat mass, a higher waist-hip ratio, lower serum cholesterol, a higher white blood cell count, and high levels of C-reactive protein (P < 0.05). The model explained 24% of the variation present in serum lutein + zeaxanthin for the current sample. The correlation between dietary and serum lutein + zeaxanthin was 0.17 and increased to 0.18 after adjusting for the effects of given covariates. Each 10% increase in dietary lutein + zeaxanthin was associated with a 1% increase in serum conditional on other terms in the model. Many factors that influence the level of serum lutein + zeaxanthin remain unknown.