Red Cell Distribution Width-to-Albumin Ratio as a Potential Biomarker for Short-Term Mortality Risk in Critically Ill Patients with Cerebral Hemorrhage: A Retrospective Study with Dual-Cohort Validation

Why It Matters

RAR provides a low‑cost, readily available biomarker that sharpens early risk stratification for ICH, potentially guiding treatment intensity and resource allocation. Its integration improves prognostic accuracy beyond established scoring systems, offering clinicians a more precise tool for critical‑care decisions.

Key Takeaways

• Higher RAR links to 17% increased ICU mortality risk.
• RAR improves AUC of APACHE II by up to 0.188.
• Seven‑variable model reaches 0.76 AUC in training set.
• External validation confirms model stability with 0.72 AUC.
• RAR combines inflammation, nutrition, stress markers for prognosis.

Pulse Analysis

Red cell distribution width‑to‑albumin ratio (RAR) merges two routine laboratory measures—RDW, reflecting erythrocyte size variability and systemic inflammation, and serum albumin, a marker of nutritional status and physiological stress. Because both components are inexpensive and universally available, RAR has emerged as a composite index that captures the interplay of inflammatory burden, malnutrition, and acute stress, all of which influence outcomes after intracerebral hemorrhage. Recent investigations have linked elevated RDW and low albumin separately to poorer neurologic recovery, but their combined ratio offers a more nuanced risk signal.

In the dual‑cohort retrospective analysis, researchers leveraged the large, publicly available MIMIC‑IV database and an external tertiary‑care cohort to validate RAR’s prognostic power. After rigorous multivariable adjustment, each unit increase in RAR corresponded to a 17% higher hazard of 28‑day ICU death and a 14% higher hazard of overall in‑hospital mortality. Importantly, integrating RAR into six conventional critical‑illness scores produced statistically significant AUC improvements, underscoring its additive value. A suite of five machine‑learning algorithms distilled seven key predictors—age, RAR, INR, total bilirubin, BUN, AST, and systolic blood pressure—into a logistic regression model that consistently outperformed traditional scores across internal and external validations.

For clinicians, the practical implication is clear: incorporating RAR into early assessment protocols could flag high‑risk ICH patients who may benefit from intensified monitoring, aggressive hemostatic therapy, or early transfer to specialized neuro‑ICU units. Nonetheless, the study’s retrospective nature, limited external sample size, and absence of neuroimaging variables temper its generalizability. Prospective, multicenter trials that blend RAR with radiologic metrics such as hematoma volume could solidify its role in precision neurocritical care and potentially reshape guideline‑driven risk stratification.