Advances in sleep research 睡眠的研究进展

It is well established that there are relationships of sleepdisordered breathing (especially obstructive sleep apnea), sleep architecture, sleep duration, and other sleep disorders with various metrics of insulin resistance and dysglycemia. However, our understanding of these relationships is still evolving, and is sometimes confusing and occasionally contradictory. Comparison of studies is hampered by the lack of uniform definitions of what is being measured, limitations of self-reporting, short duration of studies, and accounting for severity of co-morbidities. Four articles in this issue of the Journal of Diabetes provide significant contributions to our understanding of the relationship between sleep and diabetes. Zhang et al. studied the prevalence of obstructive sleep apnea (OSA) in a population of type 2 diabetics (T2DM) hospitalized primarily for poor glycemic control and examined the relationships between demographics and co-morbidities. The hospital is an opportunistic setting for diagnosing OSA and, indeed, screening for OSA may become part of guidelines for in-hospital management of T2DM (AACE/ACE). Our Scripps research team helped validate the convenient portable Apnea-Link device used by this study to measure OSA, and found it to be comparable to polysomnography (Erman et al.). The researchers were careful in accounting for as many variables as possible, including potential ascertainment bias in who consented to the study versus who did not. The overall OSA prevalence of 66.7% is consistent with prior studies, although the prevalence at different severities of OSA is difficult to compare across studies. Also consistent was the observation that, controlling for co-morbidities, OSA did not appear to have an independent association with coronary artery disease, hypertension, or diabetic microvascular complications. A new additional observation was that “lowest oxygen saturation was independently associated with the presence of PDR (proliferative diabetic retinopathy) and cerebral infarction”. This needs to be confirmed and better understood, but it does suggest that measurement of oxygen saturation is an important part of OSA assessment that currently is not universal. The Discussion section does an excellent job of summarizing the epidemiologic research in this area to date and the suggested mechanisms underlying the associations. In summary, this study demonstrates that Beijing hospital adult populations with T2DM are similar to those in other countries with regard to OSA and that the hospital setting can be very productive for screening individuals for OSA. Zheng et al. explore a subset of the REACTION study, which is commented on elsewhere in this journal. The study has the advantages of large size, over 18 000 subjects, direct glucose measurements, and a broad cross section of subjects with T2DM. It appears to be the largest such study conducted to date. Limitations include those of self-reporting sleep duration and quality, and especially the presence of sleep-disordered breathing. It is important to note that when self-reported snoring was factored in, the association between long sleep duration and worsened dysglycemia was no longer significant. Overall, this study is also an affirmation of prior observations that sleep duration has a U-shaped relationship to dysglycemia, with both shorter and longer duration sleepers having worse dysglycemia. Another interesting observation was that “the interaction between long sleep duration and TG seemed to be highly significant (P < 0.001), and the association of long sleep duration with HbA1 levels is attenuated if TG levels are adjusted, with P-value rising from 0.009 to 0.020.” This may relate to changes in insulin resistance, measures of which, such as HOMA-IR, were not performed. Zhu et al. offer novel and important insights into both the impact of sleep architecture and the relationship between sleep and dysglycemia in a pediatric and adolescent population. These populations have been grossly under-represented in sleep studies. Only three prior studies have examined sleep architecture in the pediatric population, each having methodological limitations that are overcome in this study. “This,” the authors state, “[appears to be] the first community-based study investigating the association between sleep architecture and glucose and/or insulin homeostasis in both normal and overweight children and adolescents.” The subjects were well characterized demographically and anthopometrically, and studied with the goldstandard polysomnography. This study demonstrated that “higher TST [total sleep time], SE [sleep efficiency], and Stage N3 [slow wave sleep] (% TST) were significantly correlated with lower 2-h glucose levels, higher insulin sensitivity and/or better β-cell function, whereas Stage N1 (% TST) and Stage N2 (% TST) had significant negative correlations with glucose tolerance capacity and/or insulin sensitivity.” The findings confirm our understanding of how sleep impacts glycemic indices, in bs_bs_banner