Rationale: There is uncertainty about the effects of treating obstructive sleep apnea on glycemic control in patients with type 2 diabetes.
Objectives: To determine whether treatment of obstructive sleep apnea in patients with type 2 diabetes improves glycemic control.
Methods: In this trial, we randomized patients with type 2 diabetes and no previous diagnosis of obstructive sleep apnea, with a glycated hemoglobin level of 6.5–8.5%, and an oxygen desaturation index of 15 or more events per hour to positive airway pressure therapy or to usual care.
Measurements and Main Results: A total of 416 patients met the entry criteria as determined by each site and were randomized. Of the 298 participants who met centrally adjudicated entry criteria, no differences between the study groups were seen for change in glycated hemoglobin. Furthermore, there were no between-group differences when analyses were restricted to those with poorer baseline glycemic control, those with more severe sleep apnea, or those who were adherent to therapy. A greater fall in diastolic blood pressure occurred in the positive airway pressure group than in the usual care group (−3.5 mm Hg vs. −1.5 mm Hg; P = 0.07). This difference was significant in those who were adherent to positive airway pressure therapy (−4.4 mm Hg vs. −1.6 mm Hg; P = 0.02). There was a significant reduction in sleepiness in the positive airway pressure therapy group (P < 0.0001). Quality of life assessment revealed improvements in vitality, mental health, and mental component summary scores in the positive airway pressure therapy group.
Conclusions: This trial showed no effect of positive airway pressure therapy on glycemic control in patients with relatively well-controlled type 2 diabetes and obstructive sleep apnea.
Clinical trial registered with www.clinicaltrials.gov (NCT00509223).
Type 2 diabetes and obstructive sleep apnea are both common conditions, and observational evidence suggests that obstructive sleep apnea may directly worsen glucose metabolism. Clinical trials to date have been too short and too small to determine whether treating obstructive sleep apnea could improve metabolic control in people with type 2 diabetes.
This trial shows that treating obstructive sleep apnea had no effect on glycemic control in patients with relatively well-controlled type 2 diabetes and obstructive sleep apnea.
Type 2 diabetes and obstructive sleep apnea (OSA) are common conditions (1, 2). They frequently coexist within the same patient, to the extent that up to 86% of obese people with diabetes may have OSA (3). Observational studies have suggested that the presence of OSA increases the risk of subsequently developing type 2 diabetes independently of the effects of obesity (4), while animal and human studies have shown that intermittent hypoxemia and sleep fragmentation (two of the characteristics of OSA) can each have adverse effects on carbohydrate metabolism (5, 6). Recent data show that, among patients with type 2 diabetes, increasing severity of OSA is associated with worse glycemic control (7). Furthermore, observational studies have reported that treatment of OSA with positive airway pressure (PAP) was associated with improvements in glycemia in people with diabetes (8). A meta-analysis of randomized controlled trials 1–12 weeks in duration in people without diabetes reported a modest benefit of PAP therapy on insulin resistance (9), as did a similar, more recent 2-week trial in patients with prediabetes (10). Such evidence has led to suggestions that OSA directly impairs carbohydrate metabolism and that active screening for OSA among people with diabetes should be considered, as PAP therapy may improve glycemic control.
However, in the only randomized controlled trial (RCT) among people with diabetes, West and colleagues showed no positive effect of PAP therapy on glycemic control or insulin resistance in a 3-month study of 42 men (11). Because it may take several weeks for patients to acclimatize to PAP therapy, it is possible that the study was too short, as well as too small, to gain a full understanding of the potential effects of PAP therapy in people with diabetes and OSA. Given the high prevalence of OSA among people with type 2 diabetes and that much of this is undiagnosed, understanding the potential significance of PAP therapy for glycemic control is of paramount importance.
In this multicenter RCT, we tested the hypothesis that the addition of PAP therapy to usual care could lead to improved metabolic outcomes in people with type 2 diabetes and OSA.
We conducted an open-label, 6-month, randomized, parallel-group trial of PAP therapy in patients with type 2 diabetes and previously undiagnosed OSA who were attending hospital and specialist clinics in Australia and North America. Patients without a previous diagnosis of OSA were invited to be screened and were included if a domiciliary overnight sleep study (ApneaLink; ResMed, Sydney, Australia) showed an oxygen desaturation index (ODI) of 15 or more events per hour. Study participants were randomized (1:1) to usual care or to usual care plus PAP therapy. The study was approved by the ethics committees at each site, and all participants provided written informed consent.
Eligible participants were aged 18 years or older and had type 2 diabetes, glycated hemoglobin (HbA1c) greater than 6.5% and less than or equal to 8.5%, and an ODI of 15 or more events per hour. Exclusion criteria included severe OSA (apnea–hypopnea index [AHI] >70 events/h, oxygen saturation <70% for >2 min), previous PAP therapy use, transport-related occupation, motor vehicle collision related to sleepiness in the previous 5 years, insulin or glucagon-like peptide-1 receptor agonist treatment, unstable angina or uncontrolled hypertension (systolic blood pressure >150 mm Hg or, in subjects with microalbuminuria, systolic blood pressure >130 mm Hg; or diastolic blood pressure >80 mm Hg), body mass index (BMI) greater than 40 kg/m2 or past bariatric surgery, Cheyne–Stokes respiration, pregnancy, epistaxis, claustrophobia, or any condition that might impair adherence to PAP therapy.
Recruitment of 416 participants was initially based on an AHI of 15 events or more per hour. In the United States, in addition to those with an AHI of 15 or more events per hour, those with an AHI of 5–14 events per hour accompanied by excessive daytime sleepiness, impaired cognition, mood disorders, insomnia, hypertension, ischemic heart disease, or history of stroke could also be recruited. The AHI was determined at study sites using ApneaLink. However, because of variability in the definitions of AHI across the clinical sites, all records were subsequently reviewed by a single central scorer. The study entry was then based on an ODI of more than 15 events per hour (based on an oxygen desaturation of 3% or more). As a result, data from 118 participants were withdrawn (PAP group = 56; usual care group = 62), and 298 participants were analyzed.
Participants randomized to PAP therapy were instructed by a qualified sleep technician on the use of the device (S8 AutoSet Spirit II; ResMed) in autoadjusting mode (pressure settings between 5 and 20 cm H2O). PAP therapy use data were downloaded at each clinic visit to provide usage hours, leak, and residual AHI. All participants were provided with lifestyle counseling on a monthly basis throughout the trial. Usual care was provided by each participant’s usual physician, with changes to medication allowed if dictated by clinical needs. All participants were scheduled for monthly clinic appointments following a 1-week appointment to monitor glucose control and compliance with PAP therapy and to report adverse effects.
The primary endpoint was the change in HbA1c between baseline and 6 months. Secondary endpoints included changes between baseline and 6 months in the following parameters: fasting plasma glucose, blood pressure, quality of life (measured using the 36-item Short Form Health Survey), and daytime sleepiness (measured using the Epworth Sleepiness Scale [ESS]). HbA1c was measured at baseline and at 3 and 6 months at two central laboratories (in the United States and Australia) using the Bio-Rad VARIANT II device (Bio-Rad Laboratories, Hercules, CA). Samples were regularly compared between these core laboratories for quality and consistency. Clinic blood pressure was measured monthly using automated devices as the mean of three seated values after at least 5 minutes of rest.
Participants were randomly assigned one to one to usual care or to usual care plus PAP therapy. A computer-generated randomization schedule was produced separately for each study site, and sealed, sequentially numbered envelopes were generated, allocating eligible subjects to one of the two treatment groups. Participants and clinical staff were not masked to the allocation, but laboratory personnel and the sleep study scorer were masked.
The study was powered to detect a 0.5% reduction in mean HbA1c in the PAP therapy group compared with the usual care group for participants who were adherent to the study protocol (based on data showing that continuous positive airway pressure [CPAP] led to a fall in mean ± SD HbA1c from 7.8 ± 1.4% pre-CPAP to 7.3 ± 1.3% post-CPAP) (12). Participants in the PAP therapy group were considered adherent if they used their PAP device an average of 4 or more hours per night for at least 70% of the nights. In addition, participants in both groups must have had their HbA1c evaluated at baseline and at either Month 3 or Month 6 to be included in the adherent group. To obtain 80% power to detect this difference with a significance level of 0.05 (two tailed), and assuming dropout rates of 40% for the PAP therapy group and 10% for the usual care group (due to participant withdrawal or noncompliance), a sample size of approximately 197 participants per group was necessary.
Primary and secondary endpoint analyses were conducted on all randomized participants (the intention-to-treat group). In addition, analyses were prespecified to be conducted in the subgroup of those who were adherent (defined above). Baseline characteristics were compared between treatment groups using Student’s t test for continuous variables and χ2 or Fisher’s exact tests for categorical variables as appropriate. Wilcoxon rank-sum tests were also employed when a nonparametric test of continuous variables was warranted. For each endpoint, a repeated-measures mixed linear model was fit to the change from baseline, including treatment group, visit month, and a treatment-by-visit interaction as explanatory variables. In addition, regression models were generated to explore whether compliance and change in BMI had an effect on the study endpoints. Because change in BMI did have a significant effect on several of the study endpoints, least squares mean estimates are presented unadjusted and adjusted for change in BMI. Least squares mean estimates along with SDs and 95% confidence intervals are presented for Month 3 and Month 6 visits for each study endpoint. Missing values were handled using the last observation carried forward method. If data beyond the baseline visit were not available, no imputations were made and the data were considered missing. P values less than 0.05 were considered statistically significant. Analyses were generated using SAS version 9.2 software (SAS Institute, Cary, NC).
In total, 1,624 patients with type 2 diabetes were screened between September 2007 and May 2011, of whom 298 met all entry criteria and were randomized (Figure 1). Baseline data are described in Table 1 and show the two study groups to be well matched.
|Characteristic||Usual Care (n = 147)||PAP (n = 151)|
|Age, yr, mean (SD)||62.1 (9.0)||62.4 (9.1)|
|Male sex, n (%)||93 (63.3)||99 (65.6)|
|Ethnicity, n (%)|
|White||123 (83.7)||127 (84.1)|
|Black||8 (5.4)||11 (7.3)|
|Hawaiian/Pacific Islander||2 (1.4)||0|
|Asian||12 (8.2)||13 (8.6)|
|Other or unknown||2 (1.4%)||0|
|Body mass index, kg/m2, mean (SD)||32.6 (4.9)||33.4 (5.9)|
|Waist/hip ratio, mean (SD)||0.98 (0.07)||0.98 (0.08)|
|Duration of diabetes, yr, mean (SD)||7.9 (6.9)||8.4 (7.3)|
|Glucose-lowering therapy, n (%)|
|Lifestyle only||69 (46.9)||81 (53.6)|
|Monotherapy||48 (32.7)||38 (25.2)|
|Dual therapy||23 (15.7)||27 (17.9)|
|Triple therapy||7 (4.8)||5 (3.3)|
|Apnea–hypopnea index, events/h, mean (SD)||26.2 (12.9)||28.0 (14.1)|
|Oxygen desaturation index, events/h, mean (SD)||22.8 (12.9)||24.0 (13.3)|
|Time at SpO2 <90%, min, mean (SD)||59.0 (76.5)||60.5 (82.9)|
Mean PAP therapy use was 4.3 hours per night at the 3-month visit and 4.9 hours at the end of the trial (6-month visit). Among 113 participants in the PAP group with available data, mean residual AHI at 6 months was 6.2 events per hour. Among those in the PAP group, 45.8% and 61.3% were adherent to PAP therapy at 3 and 6 months, respectively; 93.9% of the usual care group were adherent to the study protocol. BMI remained relatively stable in the PAP therapy group as well as in the subgroup that was adherent to PAP therapy. BMI in the usual care group declined over the follow-up period (Figure 2).
No differences were detected between the two groups in regard to the change in HbA1c (Table 2) at the 3-month or 6-month visit. The change in HbA1c was also similar between the two groups after adjusting for the difference in BMI over time between the groups. Post hoc analyses showed that there were no between-group differences in the primary outcome when analyses were restricted to those with the following baseline features: HbA1c greater than 7.5% (PAP = 41 with HbA1c change −0.3%; usual care = 36 with HbA1c change −0.3%); or moderate to severe sleep apnea, defined as ODI of 20 or more events per hour (PAP = 97 with HbA1c change 0.0%; usual care = 111 with HbA1c change −0.1%); or ODI of 30 or more events per hour (PAP = 56 with HbA1c change −0.2%; usual care = 50 with HbA1c change −0.1%); or high ESS score, defined as greater than 10 (PAP = 60 with HbA1c change −0.2%; usual care = 47 with HbA1c change −0.1%). Furthermore, there was no difference in the primary outcome between those in the PAP group who were adherent to PAP therapy (n = 89) and those in the usual care group, nor was there a difference when all 416 initially randomized were included.
|Mean (SD)||Unadjusted Difference (95% CI)|
|Baseline||6 mo||Baseline||6 mo|
|Overall||7.3 (0.5)||7.1 (0.8)||7.3 (0.5)||7.2 (0.8)||−0.0 (−0.2 to 0.1)|
|Adherent*||7.2 (0.9)||7.1 (0.8)||7.3 (0.6)||7.3 (0.5)||−0.1 (−0.3 to 0.1)|
|Overall||8.9 (4.7)||8.0 (4.8)||10.0 (4.6)||6.4 (4.1)||2.9 (1.9–3.9)|
|Adherent||9.0 (4.8)||8.0 (4.8)||10.3 (4.3)||6.5 (4.0)||2.8 (1.7–3.9)|
|SBP, mm Hg|
|Overall||130.6 (14.6)||129.7 (15.9)||132.4 (13.9)||129.1 (14.3)||1.8 (−1.4 to 5.1)|
|Adherent||130.5 (14.3)||129.5 (16.0)||132.1 (13.8)||129.2 (14.1)||2.0 (−1.7 to 5.7)|
|DBP, mm Hg|
|Overall||77.3 (9.2)||75.7 (9.3)||76.7 (9.3)||73.6 (9.3)||1.9 (−0.2 to 4.1)|
|Adherent||77.1 (9.1)||75.5 (9.1)||77.4 (9.4)||73.0 (9.2)||2.9 (0.5–5.3)|
|SF-36, MCS score|
|Overall||50.9 (10.0)||51.8 (10.3)||50.0 (11.3)||52.6 (10.1)||−1.5 (−3.7 to 0.7)|
|Adherent||50.9 (9.8)||51.8 (10.3)||49.9 (12.0)||53.3 (10.2)||−2.6 (−5.1 to −0.2)|
|SF-36, VT score|
|Overall||47.8 (9.6)||50.1 (9.3)||47.4 (10.5)||52.2 (9.6)||−2.1 (−4.3 to 0.1)|
|Adherent||47.8 (9.5)||50.1 (9.3)||47.9 (10.4)||52.7 (10.0)||−2.5 (−4.9 to −0.1)|
|SF-36, MH score|
|Overall||51.2 (9.3)||51.7 (10.0)||50.1 (10.5)||52.8 (8.8)||−1.7 (−3.6 to 0.3)|
|Adherent||51.3 (9.1)||51.7 (10.0)||50.5 (10.5)||53.7 (8.5)||−2.8 (−4.9 to −0.7)|
|Fasting glucose, mg/dl|
|Overall||141 (32)||137 (29)||139 (31)||133 (32)||−2 (−9 to 5)|
|Adherent||141 (32)||137 (29)||137 (32)||133 (32)||−2 (−10 to 7)|
|Overall||32.6 (4.9)||31.4 (4.8)||33.4 (5.9)||33.1 (5.7)||−0.5 (−0.9 to −0.2)|
|Adherent||32.3 (4.8)||31.4 (4.8)||33.5 (6.0)||33.2 (5.9)||−0.7 (−1.0 to −0.4)|
There was a greater fall in diastolic blood pressure over 6 months in the PAP therapy group than in the usual care group (−3.5 mm Hg vs. −1.5 mm Hg; P = 0.07) (Table 2). This difference was greater in those who were adherent to PAP therapy (−4.4 mm Hg vs. −1.6 mm Hg; P = 0.02). After changes in BMI were accounted for, the difference in adjusted change in diastolic blood pressure was greater (−4.8 mm Hg vs. −1.0 mm Hg; P = 0.002). No differences were observed for systolic blood pressure.
Table 2 shows that daytime sleepiness, as measured using the ESS, improved significantly more over 6 months in the PAP therapy group than in the usual care group (P < 0.0001). Adjustments for change in BMI or restricting the analysis to the PAP-adherent group had no material impact on the magnitude of change in the ESS scores at the 3-month or 6-month visits. No statistically significant differences were evident in any of the eight 36-item Short Form Health Survey subscales. However, in the PAP-adherent subgroup, differences favoring PAP therapy were apparent for vitality (P = 0.04), for the mental health subscale (P = 0.01), and for the mental component summary score (P = 0.03). After the change in BMI was accounted for, the observed differences in vitality, mental health, and mental component summary scores changed little and remained significant.
A total of 10 participants (6.6%) had one or more serious adverse events in the PAP group, compared with 5 (3.4%) in the usual care group (P = 0.20) (Table 3). A total of 78 participants (51.7%) had one or more nonserious adverse events in the PAP group, compared with 66 (44.9%) in the usual care group (difference, 6.8%; 95% confidence interval, −4.6 to 18.1; P = 0.24). Hypoglycemia was recorded in 20.5% of the PAP group and 19.1% of the usual care group (P = 0.75). This included three severe adverse events (hypoglycemia requiring assistance from another person), one of which occurred in the PAP group.
|Event Term||Usual Care* (n = 147)||PAP* (n = 151)||Difference (95% CI)||P Value|
|GI distress||0 (0.0)||1 (0.7)||0.7% (−0.6 to 2.0)||1.00|
|Headache||1 (0.7)||0 (0.0)||−0.7% (−2.0 to 0.7)||0.49|
|Epistaxis||0 (0.0)||1 (0.7)||0.7% (−0.6 to 2.0)||1.00|
|Pain–musculoskeletal||1 (0.7)||1 (0.7)||0.0% (−1.9 to 1.8)||1.00|
|Pneumonia||0 (0.0)||1 (0.7)||0.7% (−0.6 to 2.0)||1.00|
|Unrelated surgery||2 (1.4)||2 (1.3)||0.0% (−2.7 to 2.6)||1.00|
|Other||1 (0.7)||4 (2.7)||2.0% (−0.9 to 4.9)||0.37|
|Total||5 (3.4)||10 (6.6)||3.2% (−1.7 to 8.2)||0.20|
This RCT shows that PAP therapy did not improve glycemic control in patients with relatively well-controlled type 2 diabetes and OSA. In a series of post hoc analyses restricting the population to those with worse glycemic control or more severe OSA at baseline, there was still no evidence of a positive effect of PAP therapy on glycemic control. Furthermore, even among those whose adherence to PAP therapy was best, no improvement in HbA1c was seen.
Despite a large body of observational evidence linking OSA with type 2 diabetes, the only other controlled trial of PAP therapy in type 2 diabetes had similar findings. West and colleagues showed no benefit of CPAP on glycemic control or insulin resistance in a 3-month RCT of 42 men, although mean CPAP use was only 3.6 hours per night (11). In an observational study of 59 community-based patients with type 2 diabetes and OSA, HbA1c was unchanged after 3 months of CPAP (13). We have extended these findings by showing no change in glycemic control in a study that included significantly more participants, ran for 6 months rather than 3 months, and was conducted in multiple centers.
There are several possible explanations for these findings. First, although many studies suggest OSA to be a risk factor for the development of glucose intolerance (14–17), far fewer show an association between glycemic control and OSA among people with established diabetes (18). Although some glycemic benefit has been reported in patients with diabetes and OSA commencing PAP therapy, this has been found in uncontrolled observational studies (8). Thus, it is possible that the impact of OSA is relevant mainly for the development of diabetes but not for control of established diabetes. Notably, a recent trial showed glycemic benefits for PAP therapy over 2 weeks among those with prediabetes (10). Second, it is possible that any benefits occur only among those with either poor glycemic control or severe OSA. With a mean baseline HbA1c and ODI of 7.3% and 23 events per hour, respectively, the population of the present trial was not generally in either of these categories. Subgroup analyses of those with worse disease control did not show benefit, but these analyses may have been underpowered and were done post hoc. Third, adherence to PAP therapy may have been insufficient to produce benefits. Indeed, in the study noted above showing benefits of PAP therapy in people with prediabetes (10), in-laboratory supervision ensured adherence of 8 hours per night. Grimaldi and colleagues(19) suggested that even 4 hours of PAP therapy per night may be inadequate, as they found that OSA during rapid eye movement sleep (which occurs mainly toward the end of the sleep period) was much more strongly associated with HbA1c than was OSA during the remaining sleep time. However, the improvements in sleepiness and in diastolic blood pressure suggest that the adherence was adequate for some effects of PAP therapy. Fourth, the usual care group lost weight, while weight remained stable in the PAP therapy group. While it is possible that BMI changes obscured the effects of CPAP therapy, post hoc analyses adjusted for differences in BMI still showed no differences between the groups. This difference in BMI is consistent with a recent report of PAP therapy promoting weight gain (20), though the mechanism of this phenomenon is unclear. The findings in relation to blood pressure were consistent with other data, but they were derived from one of the largest and longest trials to date in which the blood pressure effects of PAP therapy in type 2 diabetes have been reported. Our study showed a 2–mm Hg greater reduction in diastolic blood pressure in the PAP therapy group, which was of statistical significance in those who were adherent to PAP therapy. In a recent trial of nearly 300 participants, researchers also showed a fall in diastolic but not systolic pressure after 12 weeks of PAP therapy (21), and a meta-analysis of 32 RCTs showed that PAP therapy was associated with 2.6–mm Hg and 2.0–mm Hg reductions in systolic and diastolic blood pressure, respectively (22). Possible mechanisms through which PAP therapy may have a positive effect on blood pressure include reductions in catecholamine secretion (23), oxidative stress, and inflammation (24).
Several limitations of the present study need to be considered. The population was screened for the presence of OSA and so may not be representative of those presenting with clinical symptoms. The control group followed usual care rather than having sham PAP therapy. This design aspect might have been expected to increase the difference between the study groups, but it would not account for lack of difference in the primary outcome. Furthermore, evidence suggests that sham PAP is recognized as such by many study participants (25). The study may have been underpowered, but analyses including all 416 originally randomized subjects made no meaningful difference in any of the findings.
In summary, this clinical trial shows no benefit of PAP therapy on glycemic control among patients with relatively well-controlled type 2 diabetes and OSA. There were, however, improvements in sleepiness and diastolic blood pressure. Further studies should examine whether there are subgroups of patients for whom PAP therapy might improve metabolic control. Candidate subgroups for such studies would include those with prediabetes as well as those with poorly controlled diabetes and more severe OSA.
The authors thank Maureen Crocker and Lisa Erikli for their help with trial administration, monitoring, and coordination, as well as data management. The authors thank the following investigators for their work in conducting the trial: Richard Simpson, M.D., Eastern Clinical Research Unit, Melbourne, Australia; Timothy Bailey, M.D., Advanced Metabolic Care + Research Institute, Escondido, California; Joseph Barrera, M.D., Mission Internal Medical Group, Mission Viejo, California; Mark Kipnes, M.D., DGD Research, San Antonio, Texas; Matthew Davis, M.D., Rochester Clinical Research, Rochester, New York; Ulysses Magalang, M.D., The Ohio State University, Columbus, Ohio; Jeffrey Rosen, M.D., Clinical Research of South Florida, Coral Gables, Florida; Frédéric Sériès, M.D., Institut Universitaire de Cardiologie et de Pneumologie de Québec, Québec, Québec, Canada; Richard Bogan, M.D., SleepMed, Columbia, South Carolina; Wayne Vial, M.D., SleepMed of West Ashley, Charleston, South Carolina; and Charles Wells, M.D., SleepMed of Central Georgia, Macon, Georgia.
|1.||Guariguata L, Whiting DR, Hambleton I, Beagley J, Linnenkamp U, Shaw JE. Global estimates of diabetes prevalence for 2013 and projections for 2035. Diabetes Res Clin Pract 2014;103:137–149.|
|2.||Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328:1230–1235.|
|3.||Foster GD, Sanders MH, Millman R, Zammit G, Borradaile KE, Newman AB, Wadden TA, Kelley D, Wing RR, Sunyer FX, et al.; Sleep AHEAD Research Group. Obstructive sleep apnea among obese patients with type 2 diabetes. Diabetes Care 2009;32:1017–1019.|
|4.||Wang X, Bi Y, Zhang Q, Pan F. Obstructive sleep apnoea and the risk of type 2 diabetes: a meta-analysis of prospective cohort studies. Respirology 2013;18:140–146.|
|5.||Iiyori N, Alonso LC, Li J, Sanders MH, Garcia-Ocana A, O’Doherty RM, Polotsky VY, O’Donnell CP. Intermittent hypoxia causes insulin resistance in lean mice independent of autonomic activity. Am J Respir Crit Care Med 2007;175:851–857.|
|6.||Tasali E, Leproult R, Ehrmann DA, Van Cauter E. Slow-wave sleep and the risk of type 2 diabetes in humans. Proc Natl Acad Sci USA 2008;105:1044–1049.|
|7.||Kent BD, Grote L, Ryan S, Pépin JL, Bonsignore MR, Tkacova R, Saaresranta T, Verbraecken J, Lévy P, Hedner J, et al.; ESADA collaborators. Diabetes mellitus prevalence and control in sleep-disordered breathing: the European Sleep Apnea Cohort (ESADA) study. Chest 2014;146:982–990.|
|8.||Babu AR, Herdegen J, Fogelfeld L, Shott S, Mazzone T. Type 2 diabetes, glycemic control, and continuous positive airway pressure in obstructive sleep apnea. Arch Intern Med 2005;165:447–452.|
|9.||Iftikhar IH, Khan MF, Das A, Magalang UJ. Meta-analysis: continuous positive airway pressure improves insulin resistance in patients with sleep apnea without diabetes. Ann Am Thorac Soc 2013;10:115–120.|
|10.||Pamidi S, Wroblewski K, Stepien M, Sharif-Sidi K, Kilkus J, Whitmore H, Tasali E. Eight hours of nightly continuous positive airway pressure treatment of obstructive sleep apnea improves glucose metabolism in patients with prediabetes: a randomized controlled trial. Am J Respir Crit Care Med 2015;192:96–105.|
|11.||West SD, Nicoll DJ, Wallace TM, Matthews DR, Stradling JR. Effect of CPAP on insulin resistance and HbA1c in men with obstructive sleep apnoea and type 2 diabetes. Thorax 2007;62:969–974.|
|12.||Hassaballa HA, Tulaimat A, Herdegen JJ, Mokhlesi B. The effect of continuous positive airway pressure on glucose control in diabetic patients with severe obstructive sleep apnea. Sleep Breath 2005;9:176–180.|
|13.||Myhill PC, Davis WA, Peters KE, Chubb SA, Hillman D, Davis TM. Effect of continuous positive airway pressure therapy on cardiovascular risk factors in patients with type 2 diabetes and obstructive sleep apnea. J Clin Endocrinol Metab 2012;97:4212–4218.|
|14.||Ip MS, Lam B, Ng MM, Lam WK, Tsang KW, Lam KS. Obstructive sleep apnea is independently associated with insulin resistance. Am J Respir Crit Care Med 2002;165:670–676.|
|15.||Papanas N, Steiropoulos P, Nena E, Tzouvelekis A, Maltezos E, Trakada G, Bouros D. HbA1c is associated with severity of obstructive sleep apnea hypopnea syndrome in nondiabetic men. Vasc Health Risk Manag 2009;5:751–756.|
|16.||Punjabi NM, Beamer BA. Alterations in glucose disposal in sleep-disordered breathing. Am J Respir Crit Care Med 2009;179:235–240.|
|17.||Punjabi NM, Shahar E, Redline S, Gottlieb DJ, Givelber R, Resnick HE; Sleep Heart Health Study Investigators. Sleep-disordered breathing, glucose intolerance, and insulin resistance: the Sleep Heart Health Study. Am J Epidemiol 2004;160:521–530.|
|18.||Aronsohn RS, Whitmore H, Van Cauter E, Tasali E. Impact of untreated obstructive sleep apnea on glucose control in type 2 diabetes. Am J Respir Crit Care Med 2010;181:507–513.|
|19.||Grimaldi D, Beccuti G, Touma C, Van Cauter E, Mokhlesi B. Association of obstructive sleep apnea in rapid eye movement sleep with reduced glycemic control in type 2 diabetes: therapeutic implications. Diabetes Care 2014;37:355–363.|
|20.||Drager LF, Brunoni AR, Jenner R, Lorenzi-Filho G, Benseñor IM, Lotufo PA. Effects of CPAP on body weight in patients with obstructive sleep apnoea: a meta-analysis of randomised trials. Thorax 2015;70:258–264.|
|21.||Gottlieb DJ, Punjabi NM, Mehra R, Patel SR, Quan SF, Babineau DC, Tracy RP, Rueschman M, Blumenthal RS, Lewis EF, et al. CPAP versus oxygen in obstructive sleep apnea. N Engl J Med 2014;370:2276–2285.|
|22.||Fava C, Dorigoni S, Dalle Vedove F, Danese E, Montagnana M, Guidi GC, Narkiewicz K, Minuz P. Effect of CPAP on blood pressure in patients with OSA/hypopnea a systematic review and meta-analysis. Chest 2014;145:762–771.|
|23.||Hedner J, Darpö B, Ejnell H, Carlson J, Caidahl K. Reduction in sympathetic activity after long-term CPAP treatment in sleep apnoea: cardiovascular implications. Eur Respir J 1995;8:222–229.|
|24.||Dudenbostel T, Calhoun DA. Resistant hypertension, obstructive sleep apnoea and aldosterone. J Hum Hypertens 2012;26:281–287.|
|25.||Schwartz SW, Cimino CR, Anderson WM. CPAP or placebo-effect? Sleep 2012;35:1585–1586.|
This study was funded by ResMed Science Centre, ResMed Ltd. The work is partially supported by the Victorian OHS program. J.E.S. is supported by a National Health and Medical Research Council senior research fellowship. Employees of the sponsor were involved in the design and conduct of the study and provided logistical support during the trial, including preparation of the statistical analysis plan. All analyses, however, were performed by an independent consultant. The manuscript was prepared by the authors. The sponsor was permitted to review the manuscript and suggest changes, but the final decision on content was exclusively retained by the authors.
Author Contributions: J.E.S. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. L.W. and N.M.P. analyzed the data. J.E.S., N.M.P., and M.T.N. interpreted the data and drafted the manuscript. R.M.B., P.A.C., G.R.F., G.N.R., and P.Z.Z. contributed to redrafting of the manuscript and approved the final draft. All authors contributed to study design.
Originally Published in Press as DOI: 10.1164/rccm.201511-2260OC on March 1, 2016