Bone marrow hemosiderin iron was assessed in 48 patients with alcoholic, and in 34 patients with nonalcoholic liver disease (53 men, 29 women, median age 55 years, range 18-84) and correlated to serum (S)-iron status markers (iron, transferrin, ferritin), as well as to histochemical hepatocyte iron and chemical liver iron content. In a control group of 53 healthy subjects (23 men, 30 women, median age 28 years, range 18-90) marrow hemosiderin iron and iron status markers were evaluated as well. Among liver patients, the marrow iron grade was higher in men than in women (p = 0.03). Correlations were found between marrow iron and histochemical liver iron (rho = 0.38, p = 0.0001) as well as chemical liver iron (rho = 0.33, p = 0.01). Marrow iron was correlated to S-ferritin (rho = 0.53, p = 0.0001), mean red cell volume (rho = 0.34, p = 0.003), and S-transferrin (rho = -0.24, p = 0.02). Alcoholics had a higher marrow iron grade than nonalcoholics (p = 0.001) and controls (p = 0.0001). Among controls, the marrow iron grade was likewise higher in men than in women (p = 0.01). Correlations were found between marrow iron and ferritin (rho = 0.64, p = 0.0001), transferrin saturation (rho = 0.56, p = 0.001), transferrin (rho = 0.53, p = 0.001), S-iron (rho = 0.37, p = 0.01), and hemoglobin in women (rho = 0.38, p = 0.05). The results indicate that alcoholics either have increased marrow hemosiderin iron stores, or display a redistribution of iron in reticuloendothelial cells from soluble ferritin-bound iron to insoluble hemosiderin iron. Among patients with absent marrow hemosiderin iron, 81% had absent hepatocyte hemosiderin iron as well. Among patients with absent hepatocyte hemosiderin iron, 23% had absent and 77% normal or increased marrow hemosiderin iron. Therefore, in patients with iron depletion, assessment of marrow hemosiderin iron yields more relevant information of iron status than assessment of hepatocyte hemosiderin iron.