Abstract
BACKGROUND: Red blood cell distribution width (RDW), a marker of erythrocyte size variability, has been associated with adverse outcomes in various inflammatory and critical illnesses, but the impact on long-term outcomes in critically ill patients remains unclear. We hence aimed to address the association between RDW and one-year mortality in critically ill patients through using a propensity score-based analysis. METHODS: The data from three databases were aggregated: the Medical Information Mart for Intensive Care IV (MIMIC-IV), the Salzburg Intensive Care database (SICdb), and the Taichung Veterans General Hospital (TCVGH) critical care database. Cox regression, propensity score matching (PSM), and weighting methods consisting of inverse probability of treatment weighting (IPTW), stabilized IPTW, as well as covariate balancing propensity score (CBPS) were used to identify the independent association between RDW and one-year mortality. RESULTS: A total of 39,699 critically ill patients were included in the analysis, with a one-year mortality rate of 27.3%. High RDW was significantly associated with increased one-year mortality (adjusted hazard ratio 1.980, 95% confidence interval (CI) 1.898-2.066, p < 0.001) after adjusting for covariates. The association was stronger among patients with milder forms of critical illnesses. The association between high RDW and mortality in PSM, IPTW, stabilized IPTW and CBPS populations were 1.751 (95% CI 1.627-1.885), 1.960 (95% CI 1.873-2.051), 1.962 (95% CI 1.875-2.052) and 1.946 (95% CI 1.845-2.053). CONCLUSION: Week-one RDW is a robust predictor of one-year mortality in critically ill patients. Our data suggest that RDW may serve as a prognostic tool for early risk stratification of long-term outcomes in critical care.