Re-Evaluating Positive Airway Pressure Indications in Obstructive Sleep Apnea: A Cross-Classification Analysis of Polysomnographic Severity and the Baveno Phenotypic Classification

Background/Objectives: Obstructive sleep apnea (OSA) is traditionally diagnosed via polysomnography (PSG) using the apnea–hypopnea index (AHI). However, AHI alone may inadequately reflect the clinical heterogeneity, symptom burden, and cardiometabolic consequences of the disease. The Baveno classification incorporates symptom severity and comorbidity burden into a multidimensional framework. We aimed to evaluate the agreement between PSG-staging and the Baveno classification, to investigate the impact of reclassification on positive airway pressure (PAP) treatment recommendations. Methods: This retrospective study included 395 adults with OSA diagnosed via PSG between January 2016 and June 2024. Patients were categorized according to PSG severity (mild, moderate, and severe) and Baveno phenotypes (A, B, C, D). PAP therapy indication was defined as AHI ≥ 15 events/hour in the PSG-based classification and as Baveno groups B-D in the multidimensional approach. Agreement between the systems was assessed using Cohen’s kappa coefficient. Results: PAP therapy was indicated in 252 patients (63.8%) via PSG and in 268 patients (67.8%) via the Baveno classification. Overall agreement between the systems was 63.5%, (κ = 0.191) and increased with disease severity, reaching 87.1% with severe OSA. Overall, 144 patients (36.5%) were reclassified, including 80 patients newly identified as candidates for PAP therapy and 64 patients reclassified as not requiring PAP therapy according to the Baveno framework. BMI showed only a weak correlation with AHI overall and no significant association in women. Conclusions: Combined use of PSG and the Baveno approach provides more individualized PAP treatment decisions, particularly among clinically heterogeneous patients and those within intermediate severity categories.