Predicting sleep apnea worsening and cardiometabolic risk without treatment

PURPOSE: The natural history of untreated obstructive sleep apnea (OSA) and its cardiometabolic consequences are not fully characterized. We aimed to identify predictors of apnea severity progression and incident cardiometabolic comorbidities in untreated patients, comparing continuous and categorical measures of apnea severity. METHODS: We retrospectively studied 538 adults (419 men, 119 women) who underwent two attended overnight polysomnographic evaluations separated by 3.8 ± 2.7 years and declined OSA treatment. Changes in apnea severity were assessed using continuous change in apnea-hypopnea index (ΔAHI) and categorical transitions across standard severity thresholds. Multivariable linear regression was used to identify predictors of ΔAHI, and multivariable logistic regression was used to identify predictors of incident hypertension, diabetes, and cardiovascular disease. RESULTS: Apnea severity and nocturnal hypoxemia worsened significantly over time. In linear regression, worsening ΔAHI was independently associated with higher baseline body mass index (BMI), greater weight gain, more severe baseline hypoxemia, and longer sleep latency, whereas higher baseline AHI was associated with smaller subsequent increases. In contrast, categorical progression to moderate or severe OSA was primarily predicted by baseline BMI. Incident hypertension and diabetes were independently associated with age, baseline BMI, and duration of follow-up, but not with baseline AHI or ΔAHI. Incident cardiovascular disease was independently associated with baseline apnea severity and lower sleep efficiency. Model discrimination ranged from acceptable to good (C-statistics 0.75-0.86). CONCLUSION: In untreated patients, OSA commonly worsens over time, with progression influenced by obesity, hypoxemia, and sleep quality. Continuous measures of apnea severity capture physiologic progression more sensitively than categorical classifications. Cardiometabolic consequences differ by outcome, with obesity predominating for hypertension and diabetes and apnea severity and sleep fragmentation more closely linked to cardiovascular disease.