Association of Prognostic Nutritional Index With Mortality in High-Risk OSA Individuals and with Disease Severity in Clinical OSA Patients

Background: Obstructive sleep apnea (OSA) shows marked variability in cardiovascular and mortality risk. The Prognostic Nutritional Index (PNI), derived from serum albumin and lymphocyte count, reflects nutritional and inflammatory status, but its relevance in OSA populations remains unclear. Methods: We analyzed 10,229 adults at risk for OSA from NHANES 2005-2008 and 2015-2018, with mortality follow-up through 2019 (median 5.08 years). Survey-weighted Cox models examined associations between PNI and all-cause and cardiovascular mortality. Dose-response relationships were assessed using restricted cubic splines, and discrimination was evaluated with time-dependent ROC analyses. Subgroup and sensitivity analyses assessed robustness. Additionally, 555 polysomnography-confirmed OSA patients were analyzed cross-sectionally to examine associations between PNI and disease severity. Results: During follow-up, 1025 all-cause and 287 cardiovascular deaths occurred. Higher PNI was associated with lower all-cause (adjusted HR 0.59, 95% CI 0.47-0.75) and cardiovascular mortality (HR 0.57, 95% CI 0.36-0.89), with significant dose-response trends and nonlinear associations for all-cause mortality. Associations were consistent across subgroups and sensitivity analyses. PNI showed better discrimination for mortality than its individual components (12-month AUC 0.74). In the clinical cohort, higher PNI was associated with reduced odds of severe OSA (OR 0.864, 95% CI 0.833-0.896) and lower apnea-hypopnea index. Conclusion: Lower PNI is independently associated with higher mortality risk among individuals at risk for OSA and with greater disease severity in confirmed OSA. These findings highlight the potential clinical relevance of nutritional and immune status in OSA populations. However, the temporal and causal relationships between PNI, OSA severity, and mortality require further investigation.