Diagnostic Accuracy of Iron Deficiency Anemia in Elderly Patients with Inflammatory Conditions

Background/Objectives : Our aim was to evaluate the performance of the reticulocyte hemoglobin content (CHr) assay to diagnose iron deficiency anemia in a context of inflammation in an elderly population, by comparison with the gold standard. Methods: Patients over 18 years old, with anemia and biological inflammatory syndrome, were included consecutively. The accuracy of CHr was compared with that of the gold standard, soluble transferrin receptor/log ferritin. Analyses were performed according to age (subjects over and under 75 years of age) and included specificity, sensitivity, positive predictive value (PPV), negative predictive value (NPV), and the area under the ROC curve (AUC). Results: A total of 174 patients were included, with respectively 84 and 89 patients aged under or 75 years and over, the mean age being 72.5 years (SD 16.9). In the older population, the mean age was 84.6 years (SD 5.2), and 40% were female; the mean hemoglobin was 9.6 (SD 1.4) g/dl and mean CRP 89.7 (SD 72.9) mg/L. The optimal threshold value, determined by maximizing the Youden index, was 25.2pg, at which sensitivity and specificity reached 67% and 94%, respectively. The corresponding PPV and NPV were 62% and 95%. The AUC of the ROC curve was 0.81 with 95% CI: [0.65-0.97], suggesting good diagnostic accuracy. Conclusions: In an elderly population with anemia in a context of inflammation, assessing iron deficiency by including CHr is simple, inexpensive, rapid and practical, as it can be measured as a by-product of a routine complete blood count at little incremental cost, and good NPV can exclude a diagnosis of IDA.