[Likelihood ratio and ROC curve in evaluation of iron parameters for diagnosing iron deficiency].

OBJECTIVE

To evaluate the value of iron parameters in diagnosing iron deficiency (ID).

METHODS

Cross-sectional study was performed for diagnostic tests. Ninety consecutive patients with anemia including iron deficiency anemia without chronic diseases (36 cases), chronic diseases (54) were divided into chronic diseases without ID(ACD) (23 cases) and chronic diseases with ID(CDID) (31) by bone marrow iron staining. The exclusion criteria included hemolytic anemia, deficiency of Vitamin B12 or folic acid, blood transfusion, taking iron preparations within one month and hematological malignancies. By using absence of bone marrow iron as gold standard of ID, we compared the diagnostic powers in the diagnosis of iron deficiency for sTfR, serum iron, total iron binding capacity, serum ferritin (SF), transferrin saturation and the presence of hypochromic red cells by analyzing the likelihood ratio, the area under ROC(AUCROC).

RESULTS

AUCROC for sTfR in determining ID in groups IDA + CDID vs. ACD, IDA vs. ACD and CDID vs. ACD were 0.9 (95% CI: 0.82-0.98), 0.96(0.9-0.99) and 0.84 (0.72-0.96) respectively, AUCROC for SF in above groups 0.87 (95% CI: 0.77-0.87), 0.94(0.86-0.99) and 0.77(0.63-0.91) respectively. AUCROC for other iron parameters in determining iron deficiency in chronic diseases was lower than 70%.

CONCLUSION

ROC and LR are useful tools for evaluating iron parameters in diagnosing iron deficiency. TfR is a best parameter in determining iron deficiency in chronic diseases than SF and other iron parameters.