Background: Iron-catalyzed oxidative stress may be the primary mechanism for the pathogenesis of diseases related to iron excess. We hypothesized previously that certain markers of iron in bound form that are commonly used in epidemiologic studies might be inappropriate for investigating iron-related adverse health effects because oxidative stress requires iron in redox-active form. Methods: To study aspects of this hypothesis, we examined the association between levels of serum ferritin or the percentage of transferrin saturation (%TS) and levels of serum antioxidant vitamins and C-reactive protein (CRP). This cross-sectional analysis included 11 245 adults aged 20 years or older who participated in the Third National Health and Nutrition Examination Survey. Results: Adjusted concentrations of serum α-carotene, β-carotene, β-cryptoxanthin, and lycopene were inversely correlated with the serum ferritin concentration (P for trend < .01), even within the lower deciles of the serum ferritin. In contrast, the %TS was significantly and positively associated with β-cryptoxanthin, vitamin C, and vitamin E. In addition, the serum ferritin was positively associated but the %TS was strongly and inversely associated with the serum CRP (P for trend < .01). Conclusions: The serum ferritin and %TS showed contrasting associations with serum antioxidant vitamin levels and CRP although they have been used interchangeably in epidemiologic studies as markers of body iron. These results suggest that the %TS may not be a valid marker of exposure to iron-related oxidative stress. It appears that the serum ferritin is the preferred marker for assessment of clinical outcomes presumed to be caused by iron-related oxidative stress.