Differences in polysomnography predictors for hypertension and impaired glucose tolerance

Rationale: It is unclear which indexes of physiologic stress obtained from polysomnography best predict sleep apnea-related outcomes. We assessed the relationships between various indexes of sleep-associated physiologic stress with 2 outcomes: hypertension and impaired glucose tolerance. Methods: Three hundred ninety-four participants in the Cleveland Family Study underwent overnight polysomnography, blood pressure measurements, and an oral glucose tolerance test. Graphical techniques and generalized estimating equations for logistic models were used to quantify the relationship between polysomnography indexes and to estimate the odds of each outcome, adjusting for age, sex, race, and body habitus. Results: Of the sample, 31% had hypertension, and 32% had impaired glucose tolerance. The odds of hypertension increased approximately 20% per 5-unit increase in arousal index (odds ratio 1.22; 95% confidence interval 1.06, 1.41). Weaker associations of hypertension with the apnea-hypopnea index and oxygen-saturation variables were seen. In contrast, the strongest index associated with impaired glucose tolerance was time spent at an oxygen saturation of < 90%; individuals with at least 2% of time spent at a saturation level less than 90% had 2.33 times (95% CI 1.38, 3.94) the odds of impaired glucose tolerance. Conclusions: These data are consistent with disparate pathways mediating hypertension and impaired glucose tolerance. Vascular responses may be more directly related to sympathetic surges and arousals, whereas metabolic sequelae may be mediated more by hypoxic stress. One single index from polysomnography may not be adequate to fully predict the myriad health outcomes associated with sleep apnea.