Background: The objective of this study was to evaluate
whether the red cell distribution width (RDW) is a significant risk factor for hospital mortality in critically ill
patients and to investigate whether RDW is a parameter
indicating inflammation, or a risk factor independent of
inflammation. Methods: We studied all patients admitted to a ten-bed mixed intensive care unit in the Netherlands between May 2005 and December 2011 for whom RDW was available, and who had not received a blood transfusion in the preceding three months. Inflammation was measured by C-reactive protein and leucocyte count. Analyses included correlation, logistic regression analysis, and receiveroperating characteristic (ROC) curves. Results: We included 2915 patients, of whom 387 (13.3%) did not survive to hospital discharge. In univariate analysis higher RDW values were associated with increased hospital mortality. In multivariate analysis RDW remained an independent risk factor for mortality after correction for APACHE II score, age, admission type and mechanical ventilation (odds ratio 1.04, 95% confidence interval 1.02-1.06, for each femtolitre of RDW). Adding RDW to APACHE II, however, increased the area under the ROC curve marginally (from 0.845 to 0.849, p<0.001). RDW was not correlated with C-reactive protein and leucocyte
count, refuting the hypothesis that the association between RDW and outcome is mediated through inflammation. Conclusion: In critically ill patients, the RDW on ICU admission was an independent predictor of mortality. Since RDW was not correlated with inflammation, the underlying mechanism of this association warrants further investigation.