Risk analysis of the association between red blood cell distribution width-to-peripheral oxygen saturation ratio and poor prognosis in critical patients with obstructive sleep apnea.

BACKGROUND

Obstructive sleep apnea (OSA) is prevalent among critically ill patients and associated with increased mortality. This study investigates the red blood cell distribution width (RDW)-to-peripheral oxygen saturation ratio (RSR) as a predictor of all-cause mortality in critically ill OSA patients using the MIMIC-IV database.

METHODS

We extracted clinical data from 3237 critically ill OSA patients in MIMIC-IV (version 3.1). The Boruta algorithm was employed for feature selection, and patients were stratified into quartiles based on RSR values. Primary and secondary outcomes were all-cause mortality at 365-day and 30-day, respectively. Kaplan-Meier survival analysis, restricted cubic splines (RCS), and Cox regression were applied to analyze RSR's relationship with outcomes. ROC curves compared RSR with classical prognostic scores, and Delong test assessed AUC differences.

RESULTS

Higher RSR levels were associated with increased 365-day and 30-day mortality (Kaplan-Meier, log-rank P < 0.01). Cox regression confirmed higher mortality risk in the highest RSR quartile. RCS analysis revealed an S-shaped relationship, with inflection points at 19.89 for both outcomes. Below the inflection point, higher RSR levels increased 30-day mortality risk by 30.68 % (HR 1.31, 95 %CI: 1.23-1.38) and 365-day mortality by 35.30 % (HR 1.35, 95 %CI: 1.30-1.41). For 365-day mortality, RSR outperformed SOFA and OASIS (Delong test P < 0.01), and was comparable to SAPS II, APS III, and Charlson scores (Delong test P > 0.05).

CONCLUSION

RSR significantly associates with all-cause mortality in critically ill OSA patients, particularly those with elevated RSR, serving as a promising prognostic indicator for short- and long-term mortality risk assessment.