Elucidating Predictors of Obesity Hypoventilation Syndrome in a Large Bariatric Surgery Cohort

Rationale: Although understanding predictors of obesity hypoventilation syndrome (OHS), a condition associated with increased morbidity and mortality, is of key importance for risk prediction, existing characterization is limited. Objectives: We hypothesize that OHS patients referred for bariatric surgery have more severe obstructive sleep apnea and metabolic derangements compared with their eucapnic counterparts. Methods: A total of 1,718 patients undergoing polysomnography with end-tidal CO2 monitoring prior to bariatric surgery at Cleveland Clinic from September 2011 to September 2018 were included. OHS was defined by body mass index (BMI) >= 30 kg/m(2) and either polysomnography-based end-tidal CO2 >= 45 mm Hg or serum bicarbonate levels >= 27 mEq/L based on the updated European Respiratory Society guidelines. Unadjusted and multivariable logistic regression models (odds ratio; 95% confidence interval) were used to examine OHS predictors consisting of factors in domains of patient characteristics, polysomnography (cardiorespiratory and sleep architecture), laboratory, and metabolic parameters. Results: The analytic sample comprised 1,718 patients with the following characteristics: age of 45.3 +/- 12.1 years, 20.7% were male, BMI = 48.6 +/- 9 kg/m(2), and 63.6% were white individuals. OHS prevalence was 68.4%. Unadjusted analyses revealed a 1.5% increased odds of OHS (1.01; 1.00-1.03) per 1-unit BMI increase, 1.7% (1.02; 1.01-1.02) per 1% increase in sleep time Sa(O2) < 90%, 12% increase (1.12; 1.03-1.22) per 1-U increase in hemoglobin A1c, and 3.4% increased odds (1.03; 1.02-1.05) per 5-U increase in apnea-hypopnea index. The association of apnea-hypopnea index with OHS persisted after adjustment for age, sex, race, and BMI and its comorbidities (1.02; 1.01-1.04). Conclusions: OHS was highly prevalent in patients referred for bariatric surgery by more than two-thirds. Even after consideration of confounders including obesity, obstructive sleep apnea remained a strong OHS predictor, as were increasing age, male sex, nocturnal hypoxia, and impaired long-term glucose control. These findings can inform OHS risk stratification in bariatric surgery and set the stage for experimental studies to examine sleep-related respiratory and metabolic contributions to hypoventilation.