In this paper we present a method for determining the iron saturation of ferritin as a possible independent predictor of iron stores. Serum ferritin was purified by immunochemical precipitation, and could be completely recovered from serum without any contamination from transferrin. The iron content of the precipitated ferritin was determined by flameless atomic absorption spectrophotometry (FAAS) and the ferritin-iron saturation was calculated using the original serum ferritin concentration. The intra- and inter-assay variation coefficients were 4.2% and 13.4% respectively. The first results with this assay indicate that serum ferritin contains a considerable amount of iron. Furthermore the results show that the iron saturation of ferritin in patients with acute phase response is significantly lower than the saturation found in healthy volunteers (19.3% and 24.3% respectively). These results suggest a possible role for the ferritin-iron saturation in the assessment of iron stores in patients suffering from acute phase response. In addition, the considerable amount of iron in ferritin suggests the need to revise the physiological role of this substance in relation to the serum iron homeostasis.