A Risk Stratification Model for Adult Patients with Obstructive Sleep Apnea: Development and Evaluation

Despite emerging evidence that obstructive sleep apnea (OSA) may cause cardio-metabolic disturbances independently of known risk factors, the strength and significance of this association remains unclear. This thesis is comprised of three studies assessing the long term-consequences of OSA, and specifically the prognostic value of OSA-related variables for cardiovascular (CV) events, all-cause mortality and incident diabetes. Methods: The first study is a systematic review of longitudinal studies, from 1999 to 2011. Quality was assessed using published guidelines. Studies two and three linked a clinical sleep database to Ontario health administrative databases for the period of 1991 to 2011 to examine the relationship between OSA variables and a composite CV outcome and incident diabetes. For the latter, we assembled a cohort free of diabetes at baseline, as defined by health administrative data. Cox regression models were used to investigate the association between of OSA-related predictors and outcomes of interest, controlling for potential confounders. Study one identified significant relationships between OSA and all-cause mortality and composite CV outcome in men; associations with other outcomes remain uncertain. Among OSA-related variables, only apnea-hypopnea index (AHI) was a consistent predictor. Limitations of the clinically-based studies were small numbers of events, weak definitions of outcomes, and inconsistency in polysomnographic scoring criteria over time. Study two found that 1,172 (11.5%) of 10,149 participants experienced our composite CV event over a median follow-up of 68 months. In a fully adjusted model, the following OSA-related variables were significant independent predictors: time spent with oxygen saturation (SaO2) < 90%, sleep time, awakenings, periodic leg movements, heart rate, and presence of daytime sleepiness. In study three, of 8,678 cohort participants without diabetes at baseline, 1,017 (11.7%) developed incident diabetes over a median follow-up of 67 months. In fully-adjusted models, patients with AHI > 30 had a 30% higher hazard of developing diabetes than those with AHI < 5. Among other OSA-related variables, REM-AHI, and SaO2