Serum soluble transferrin receptor (sTfR) concentration has been evaluated in the diagnosis of iron deficiency in otherwise healthy individuals and in patients with rheumatoid arthritis, but has not been studied in a general population of patients with complicated clinical presentations. In this study, 145 anaemic patients with a variety of medical conditions undergoing diagnostic bone marrow aspiration for any reason were tested by a complete blood count, a panel of biochemical tests to evaluate iron status, bone-marrow aspirate iron stain, and serum sTfR concentration. Sixteen per cent lacked stainable iron in the marrow aspirate. All biochemical parameters differed significantly between patients with or without stainable marrow iron. The sTfR assay was significantly more sensitive but less specific than other iron status assays in identifying the absence of stainable iron. Logistic regression analysis demonstrated that only sTfR and ferritin contributed independently to the prediction of marrow iron status. Serum ferritin alone was highly specific but insensitive. A decision algorithm combining serum ferritin and sTfR was as sensitive as TfR and as specific as serum ferritin. The measurement of serum sTfR, especially in conjunction with serum ferritin, is a valuable addition to the existing methods for predicting the results of marrow aspirate iron stains.