Transferrin Saturation (TS) Is a Surrogate Marker for Iron Deficiency Anemia (IDA),

Iron deficiency anemia (IDA) is an important entity in clinical practice. Distinguishing between IDA and other kinds of anemia is challenging in many situations. Various laboratory indices have been studied in relation to IDA including red cell distribution width (RDW), mean cell volume (MCV), transferrin saturation (TS), total iron binding capacity (TIBC), serum ferritin and erythrocyte protoporphyrin (EPP).

We chose 8015 subjects in the National Health and Nutrition Examination Survey (NHANES) database between 2003–2008 who had the required hematologic values. After applying the Centers for Disease Control (CDC) definition for anemia, the number of anemic subjects became 627 (8%). After applying the IDA model as suggested by Cogswell et al using the log ratio of transferrin receptor to ferritin, 46.1% of the anemic subjects had IDA. Then we developed receiver operator (ROC) curves for MCV, RDW, erythrocyte protoporphyrin, TIBC, and percent transferrin saturation for those anemic subjects with and without iron deficiency. From these variables, the area under the curves (AUC), sensitivity, specificity, positive likelihood ratios (LR+), and negative likelihood ratios (LR-) along with their confidence intervals (CI) were calculated. Furthermore we investigated the performance of each available measure by sub-setting the subject group into variables. Data analysis was done using SAS (version 9.2), R version 9.2 with the ROCR library, and MedCalc (version 11.2; MedCalc Software, Mariakerke, Belgium). ROCR and MedCalc have slightly different but individually valuable approaches to ROC curve analysis. Confidence intervals for sensitivity, specificity, and the likelihood ratios were calculated.

For the 627 anemic subjects, median age was 26.1 year with 95% being females. Fifty four percent were African American (AA). The MCV showed an AUC of 0.666, sensitivity of 73.7, specificity of 58.8, LR(+) of 1.79, LR(−) of 0.45 with a criterion of <81.6. RDW showed an AUC of 0.803, sensitivity of 77.2, specificity of 69.2, LR(+) of 2.51, LR(−) of 0.33 with a criterion >13.9. For 535 subjects with anemia the EPP showed an AUC of 0.831, sensitivity of 66.4, specificity of 84.9, LR(+) of 4.41, LR(−) of 0.40 with a criterion >93. For 542 subjects with anemia the TIBC showed an AUC of 0.756, sensitivity of 80.3, specificity of 61.5, LR(+) of 2.09, LR(−) of 0.32 with a criterion >406. Finally for the same 542 subjects the TS showed an AUC of 0.869, sensitivity of 80.3, specificity of 78.8, LR(+) of 3.79, LR(−) of 0.25 with a criterion <10.4 [Figure 1]. Comparing the LR(+) and LR(−) between the indices in the setting of inflammation showed that the EPP is the least affected but all indicators are, to some extent, affected by inflammation. Comparing the LR(+) and LR(−) between the indices in the setting of pregnancy showed that TS is the strongest performer. Finally; comparing the LR(+) and LR(−) between the indices in the AA versus non-AA showed that AA ethnicity has a small impact on the LR(−) but reduces the clinical utility of the LR(+) somewhat.

Figure 1:

View largeDownload slide

ROCs for IDA indicators.

Figure 1:

View largeDownload slide

ROCs for IDA indicators.

Close modal

Alternative hematologic indices have diagnostic value in the diagnosis of IDA. Transferrin saturation (TS) is the most efficient alternative to ferritin in most circumstances. Erythrocyte proto-porphyrin (EPP) should be considered as an alternative tool in inflammatory conditions.

No relevant conflicts of interest to declare.