Prognostic and incremental value of the red cell distribution width-to-albumin ratio for in-hospital mortality in acute respiratory failure.

BACKGROUND

Acute respiratory failure (ARF) is a life-threatening condition associated with substantial mortality in intensive care units. Biomarkers reflecting both inflammatory burden and nutritional status may enhance prognostic accuracy beyond conventional clinical scores. This study evaluated the independent and incremental prognostic value of the red cell distribution width-to-albumin ratio (RAR) for predicting in-hospital mortality in patients with ARF.

METHODS

In this retrospective cohort study, 1581 patients admitted to the intensive care unit with ARF between January 2023 and December 2025 were analyzed. Demographic characteristics, admission laboratory parameters, and APACHE II scores were recorded. Discriminatory performance was assessed using receiver operating characteristic (ROC) analysis. The incremental contribution of RAR beyond a baseline clinical model was evaluated using continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Internal validation was performed using bootstrap resampling.

RESULTS

Overall, 450 patients (28.5%) died during hospitalization. RAR values were significantly higher among non-survivors compared with survivors (6.00 vs. 4.57; p < 0.001). The area under the curve for RAR in predicting mortality was 0.711 (95% CI: 0.68-0.74). In multivariable analysis, RAR remained independently associated with mortality (OR 1.176; 95% CI 1.107-1.250; p < 0.001). Similar findings were observed in Cox proportional hazards models and in sensitivity analyses using 28-day mortality. Addition of RAR to the baseline model was associated with improved risk classification (continuous NRI 0.384; p < 0.001) and discrimination (IDI 0.014; p < 0.001).

CONCLUSIONS

RAR independently predicts in-hospital mortality in ARF and provides modest incremental prognostic value beyond conventional risk scores. Further external validation is warranted before routine clinical implementation.