Search

Aims: Sleeping oxygen saturation (SaO 2 ) and sleep stage duration have been linked with prediabetic alterations but the pathogenic pathways are not well understood. This study of insulin sensitive and resistant adults examined the effect on postprandial metabolic regulation of repeated mixed-meal challenges of different carbohydrate loading. The aim was to examine whether the relationship between lower sleeping oxygen saturation (SaO 2 ) and poorer fasting and postprandial metabolic function may be linked with reduced slow wave sleep (SWS) and rapid eye movement (REM) duration, independent of age, sex and total adiposity., Methods: The 24 men and women, aged 25-54 years, had no diabetes or other diagnosed conditions, were evaluated with polysomnography to derive indices of SaO 2 and sleep architecture. In addition, an OGTT and two 14-h serial mixed-meal tests were administered over 3 successive in-patient days. The carbohydrate content of the mixed-meals was manipulated to compare a standard-load day with a double-load day (300 vs. 600 kcal/meal). Quantitative modeling was applied to derive β-cell glucose sensitivity (β-GS), early insulin secretion rate sensitivity (ESRS), and total postprandial insulinemia (AUC INS )., Results: Analyses showed that, for the 14-h tests, the SaO 2 relationship with metabolic outcomes was associated significantly with percent time spent in REM but not SWS, independent of age, sex and total adiposity. Specifically, indirect pathways indicated that lower SaO 2 was related to shorter REM duration, and shorter REM was respectively associated with higher β-GS, ESRS, and AUC INS for the 300- and 600-load days (300 kcal/meal: β = -8.68, p < .03, β = -8.54, p < .002, and β = -10.06, p < .008; 600 kcal/meal: β = -11.45, p < .003, β = -11.44, p < .001, and β = -11.00, p < .03)., Conclusion: Sleeping oxygen desaturation and diminished REM duration are associated with a metabolic pattern that reflects a compensatory adaptation of postprandial insulin metabolism accompanying preclinical diabetic risk., (Copyright © 2018 Elsevier Inc. All rights reserved.)

The statistical distributions of bout lengths for the different (macro) sleep states (wake, N1, N2, N3, and REM sleep) are essential to understanding whether any memory-free subcomponent ("micro state") is involved in the organization of sleep. Micro state detection can be prevented by the fusion of data including various sources of variability, in particular by the differences in sleep architecture between individuals, along sleep time (or nighttime), or between different nights. In this analysis, a mathematical model of sleep was adopted to disentangle these features and advance the understanding of the dynamics and mechanisms of sleep and its states. The analysis involved 116 primary insomnia patients taking placebo before going to bed and undergoing polysomnography for one night. The individual sequences of macro sleep states had been previously modeled with a mixedeffect nonhomogeneous modified Markov chain model, from which individual conditional probability distributions for the bout durations were derived in this analysis as functions of sleep time. The probability distributions, affected by neither subject, night-time, nor multiple-night pooling, substantially changed at ¼ and ¾ sleep time, had modified exponential shape, and were best described as the sum of one to four exponentials, depending on the sleep state. The time constants and proportions of bouts contributing to each exponential were similar in the different subjects, changing over sleep time. Variability in bout durations thus indicated the presence of multiple memory-free sleep subcomponents whose mean residence times and access probabilities could be identified and shown to be consistent among the studied subjects. [ABSTRACT FROM AUTHOR]