Background: Guidelines for treating anemia in dialysis patients accept, as high-end range of serum ferritin useful to optimize erythropoietin therapy, values high as 500 to 900 microg/L, on the hypothesis that ferritin might be not representative of iron overload.
Methods: A superconducting quantum interference device (SQUID) was used to make direct noninvasive magnetic measurements of nonheme hepatic iron content in 40 dialysis patients treated with intravenous iron, and liver iron content was compared with biochemical markers of iron status.
Results: Only 12/40 (30%) patients showed normal hepatic iron content (SQUID <400 microg/g), while 32.5% had mild (400 to 1000 microg/g) and 37.5% severe (>1000 microg/g) iron overload, although 28/40 patients (70%) had serum ferritin below 500 microg/L. Among many parameters, hepatic iron content was only correlated with ferritin (r= 0.324, P= 0.04). The receiver operating characteristic (ROC) analysis showed the best specificity/sensitivity ratio to identify iron overload for ferritin >340 microg/L (W = 0.716). Multivariate logistic regression analysis demonstrated that an increase in serum ferritin of 100 microg/L and female gender were independent variables associated with moderate to severe hepatic iron overload: OR 1.71 (95% CI 1.10 to 2.67) and OR 10.68 (95% CI 1.81 to 63.15), respectively.
Conclusion: Hepatic iron overload is frequent in dialysis patients with ferritin below currently proposed high-end ranges, and the diagnostic power of ferritin in indicating true iron stores is better than presumed. Safety concerns should prompt a reevaluation of acceptable iron parameters, focusing on potential gender-specific differences, to avoid potentially harmful iron overload in a majority of dialysis patients, mainly females.