Association between nocturnal hypoxemic burden and glucose metabolism.

Abstract

To evaluate the association between a novel integrated event-based and hypoxemia-based parameter of polysomnography (PSG), hypoxemic load or HL100, and fasting blood glucose (FBG) and hemoglobin A1c (HbA1c) levels. Adult patients, who underwent an in-lab PSG at the University of Iowa Hospitals and Clinics with FBG or HbA1c levels, were included. Event-based parameter and hypoxemia-based parameter data were derived. HL100, defined as the integrated area of desaturation between the 100% oxygen saturation and the measured saturation levels during sleep divided by the total sleep time, was calculated by Python software. Demographic data and glycemic parameters within 1year prior to PSG (FBG and HbA1c) were retrieved from chart review. Spearman correlation analysis and stepwise backward regression analysis were performed to determine independent predictors of FBG and HbA1c levels. Of the 467 patients who underwent an in-lab PSG, 218 had FBG levels, 84 had HbA1c levels, and 118 had both values. All event-based and hypoxemia-based parameters, including HL100, were significantly correlated to FBG and HbA1c levels. Stepwise backward regression analyses, adjusted for age, sex, body mass index, and diabetes status, revealed that log HL100 was significantly related to FBG (B = 23.9, p = 0.010), but none of log event-based or hypoxemia-based parameters were found to be significantly related HbA1c levels. HL100 was shown to be an independent predictor of FBG in this cohort, implying that any degree of desaturation below 100% could adversely affect glucose metabolism. HL100 may be useful for interpretation of sleep studies, risk stratification, and patient management purposes in the future.