Risk stratification in cardiac surgery is a relevant issue. Many mortality risk models are currently available [1,2], and they are used to assess riskof operativemortality, and toprovide internal and/or external benchmark comparisons. However these risk models, such as the European System for Cardiac Operative Risk Evaluation (EuroSCORE) [2], have shown several limitations, including poor performance in the elderly and overestimated mortality, especially among isolated coronary artery bypass grafting (CABG) patients. Therefore there is a need to identify new risk factors in this setting which are expected to be simple enough for clinical use, available for all patients and cost-effective. Complete blood cell count parameters, such as haemoglobin, have been shown to predict outcomes in patients with cardiovascular disease [3,4] but none of themhas been incorporated into surgical risk scores. In particular, the preoperative haemoglobin has been investigated as a risk factor of early death but discordant results were reported. Red blood cell distribution width (RDW), routinely reported in automated complete blood counts, is a numerical measure of the variability in size of circulating erythrocytes, and has been emerging as a strong predictor of adverse events for several categories of patients [5]. However, the role of RDW in predicting mortality of CABG patients has not been investigated yet. To assess predictive power of RDW on both early and late mortalities following CABG, prospectively collected data from the surgical database of two European institutions (Sapienza University of Rome, Italy and Papworth Hospital NHS, United Kingdom) have been retrospectively analysed by authors whom have certified that they comply with the Principles of Ethical Publishing in the International Journal of Cardiology. The primary end points of the study were early (within 30 day) death and all-cause late death after hospital discharge. All-cause death is the most robust and unbiased index because no adjudication is required, thus avoiding inaccurate or biased documentation and clinical assessments. Furthermore, to deeply investigate the relationship between RDWandmortality, cause of deaths were recorded and a sub-analysis of cardiac-related late mortality was performed (any sudden, cardiac, or unknown death was considered a cardiac-related death; stroke as well was considered as cardiac death). Participants were classified into 4 groups using the RDW quartile values. Differences in baseline clinical characteristics among groups were examined by 1-way analysis of variance for continuous variables and χ test for categorical variables. Kaplan–Meier analysis of survival was performed and estimates were compared with logrank test. Potential confounding factors were derived among variables included into EuroSCORE [1,2], on the basis of a literature review and a clinical plausibility. We used 4 stepwise logistic and Cox proportional hazards models to assess the associationsbetween thebaselineRDWquartiles and risks of early and late mortalities. Estimates were calculated as crude association between outcomes and RDW quartiles (Model 1), adjusting for demographic risk factors showing significant correlation at univariate analysis (Model 2), adding haemoglobin levels (Model 3), and adding other biochemical factors (Model 4). All statistical analyses were performed using SPSS v. 20.0 (SPSS, Chicago, IL).