Red Blood Cell Distribution Width-To-Albumin Ratio and 90-Day Mortality in Sepsis Patients with and Without Sepsis-Induced Coagulopathy: A Retrospective Observational Study Using the MIMIC-IV Database
Baochang Xie, Yu Cao, Yuanyuan Xiao, Diyang Zhu, Xiaoyu Guan, Guangming ChenBackground
The red blood cell distribution width-to-albumin ratio (RAR) has emerged as a novel prognostic biomarker in critically ill patients. However, its association with mortality in sepsis patients, particularly in the context of sepsis-induced coagulopathy (SIC), remains unclear.
Methods
This retrospective cohort study analyzed adult sepsis patients from the Medical Information Mart for Intensive Care IV (MIMIC-IV) version 3.1 database. RAR was calculated from the first laboratory measurements within 24 hours of intensive care unit (ICU) admission. The primary outcome was 90-day all-cause mortality. Cox proportional hazards regression, subgroup analyses, and restricted cubic spline (RCS) regression were performed. Discriminative ability was assessed using receiver operating characteristic (ROC) curves.
Results
Of 26,854 patients meeting sepsis criteria, 17,499 were excluded due to missing RDW or albumin data. A total of 9,206 sepsis patients were included, with 5,649 (61.4%) meeting SIC criteria. The overall 90-day mortality was 38.1% (absolute rates: 23.9% in the lowest RAR quartile to 55.5% in the highest). In the fully adjusted model (age, sex, SOFA, Charlson Comorbidity Index, infection source, vasopressor use, mechanical ventilation, renal replacement therapy, and lactate), each unit increase in RAR was associated with a 9% increase in mortality risk (hazard ratio [HR]=1.09, 95% confidence interval [CI]: 1.08-1.10, P<0.001). Compared with the lowest quartile, patients in the highest RAR quartile had a significantly higher mortality risk (HR=2.07, 95% CI: 1.91-2.25). The association remained consistent across subgroups. RCS analysis revealed a significant non-linear exposure-response relationship (P for non-linearity <0.001). DeLong testing confirmed that RAR (AUC=0.648) had statistically significantly higher discriminative ability than SOFA score (AUC=0.613, P<0.001) and SIC score (AUC=0.593, P<0.001) for 90-day mortality, though all AUC values indicated moderate discrimination.
Conclusions
Elevated RAR was independently associated with increased 90-day mortality after comprehensive covariate adjustment (highest vs. lowest quartile: HR=2.07, 95% CI: 1.91-2.25), with modest but statistically significant incremental discriminative ability over SOFA and SIC scores. As a biomarker derived from routine blood tests, RAR may complement existing tools for early risk stratification in sepsis. These findings are hypothesis-generating and warrant prospective validation.