Because of evidence that increased body iron stores are associated with an increased risk of cancer, we examined iron status and cancer risk in the first National Health and Nutrition Examination Survey, a survey of more than 14,000 adults begun in 1971, with follow-up between 1981 and 1984. Among 242 men in whom cancer developed, the mean total iron-binding capacity was significantly lower (61.4 vs. 62.9 mumol per liter; P = 0.01) and transferrin saturation was significantly higher (33.1 vs. 30.7 percent; P = 0.002) than among 3113 men who remained free of cancer. The risk of cancer in men in each quartile of transferrin-saturation level relative to the lowest quartile was 1.00, 1.01, 1.10, and 1.37 (P = 0.02 for trend). The serum albumin level was significantly lower in men in whom cancer developed than in those who remained cancer-free. Among women, those in whom cancer developed did not have significantly lower total iron-binding capacity or higher transferrin saturation than those who remained cancer-free. However, a post hoc examination of 5367 women (203 with cancer) yielded a relative risk of 1.3 (95 percent confidence interval, 0.9 to 1.9) associated with a very high transferrin saturation (greater than or equal to 36.8 percent, a value in the highest quartile among men); in 5228 women with at least six years of follow-up (149 with cancer), the relative risk associated with transferrin saturation above this level was 1.5 (1.0 to 2.2). These results are consistent with the hypothesis that high body iron stores increase the risk of cancer in men. The possibility that a similar association exists in women requires further study.