To examine the relationship between hepatic iron stores and plasma ferritin concentration in individuals treated with red cell transfusion and iron chelation therapy, 37 patients with sickle cell anemia and 74 patients with thalassemia major were studied. In each patient, hepatic iron stores were measured by an independently validated noninvasive magnetic method, and plasma ferritin was determined by immunoassay. The correlation between hepatic iron and plasma ferritin was significant both in patients with sickle cell anemia (R = 0.75, P < 0.0001) and in those with thalassemia major (R = 0.76, P < 0.0001). Regression analysis showed no significant difference between the two groups in the linear relationships between hepatic iron stores and plasma ferritin. Considering all 111 transfused patients as a group, the coefficient of correlation between hepatic iron stores and plasma ferritin was highly significant (R = 0.76, P < 0.0001). Regression analysis found that variation in body iron stores, as assessed by magnetic determinations of hepatic iron, accounted for only ∼57% of the variation in plasma ferritin, suggesting that the remainder was the result of other factors, such as hemolysis, ineffective erythropoiesis, ascorbate deficiency, inflammation, and liver disease. The 95% prediction intervals for hepatic iron concentration, given the plasma ferritin, were so broad as to make a single determination of plasma ferritin an unreliable predictor of body iron stores. Variability resulting from factors other than iron status limits the clinical usefulness of the plasma ferritin concentration as a predictor of body iron stores. © 1993 Wiley-Liss, Inc.