Red cell distribution width measurement: what role does it have in heart failure?

Over the past decade, there has been a literal explosion of studies examining various prognostic biomarkers in patients with heart failure. Some of these biomarkers—such as the natriuretic peptides—directly reflect pathophysiologic processes in the diagnosis, while the prognostic links for other ‘‘heart failure biomarkers’’ remain less well-defined. In the article published in Revista Española de Cardiologia, Bonaque et al. present their data demonstrating the prognostic role of the red cell distribution width (RDW) in 698 subjects suffering from chronic heart failure. In adjusted analyses, RDW predicted mortality, as well as hospitalization in chronic heart failure, and reclassified risk beyond standard variables. The authors also identified a practical RDW cut-point of 15.4% (quite consistent with prior studies of RDW) as the upper reference limit for this curious ‘‘biomarker’’. Their study confirms the previously established association between RDW and morbidity and/or mortality in subjects suffering from the complete spectrum of cardiovascular diseases. After the initial publication by Felker et al. in 2007 reporting that RDW was fortuitously found to be a powerful independent predictor of prognosis in heart failure (and importantly, superior to other parameters of anemia including hemoglobin concentrations), RDW became the subject of interest of various clinical investigators working in the field. Across a broad range of cardiovascular diseases, including heart failure, RDW has been found to be a consistent, independent predictor of outcome. The key question at present is whether RDW should be routinely measured for patient management, as Bonaque et al. suggest. Recently, we proposed 4 conditions which we believe any novel biomarker in heart failure should fulfill in order to become clinically relevant and widely used; using this context, it is worthwhile and important to evaluate RDW in order to see whether it meets these conditions. Firstly, we suggested that the ‘‘statistical methods by which any novel biomarker is investigated should be thorough, including robust comparisons to other previously established biomarkers,