Diagnostic performance of automated red cell parameters in predicting bone marrow iron stores.

OBJECTIVES

The aim of the study was to determine the diagnostic performance of novel automated red cell parameters for estimating bone marrow iron stores.

METHODS

The study was a retrospective single-centre study based on data from an automated haematology analyser and results of bone marrow iron staining. Red cell parameters were measured on a Sysmex XN-series haematology analyser. Bone marrow iron stores were assessed semiquantitatively by cytochemical reaction according to Perls.

RESULTS

The analysis included 429 bone marrow aspirate smears from 393 patients. Median age of patients was 67 years, 52 % of them were female. The most common indication for bone marrow examination was a plasma cell dyscrasia (n=104; 24 %). Median values of percentage of hypochromic and hyperchromic red blood cells (%HYPO-He, %HYPER-He), reticulocyte haemoglobin equivalent (RET-He) and microcytic red blood cells (MicroR) were statistically significantly different between cases with iron deplete and iron replete bone marrow. In a logistic regression model, ferritin was the best predictor of bone marrow iron stores (AUC=0.891), outperforming RET-He and %HYPER-He (AUC=0.736 and AUC=0.722, respectively). In a combined model, ferritin/MicroR index achieved the highest diagnostic accuracy (AUC=0.915), outperforming sTfR/log ferritin index (AUC=0.855).

CONCLUSIONS

While single automated red cell parameters did not show improved diagnostic accuracy when compared to traditional iron biomarkers, a novel index ferritin/MicroR has the potential to outperform ferritin and sTfR/log ferritin index for predicting bone marrow iron stores. Further research is needed for interpretation and implementation of novel parameters and indices, especially in the context of unexplained anaemia and myelodysplastic syndromes.