Abstract
Background
Glycemic variability is associated with increased risk for cardiovascular disease in patients with type 2 diabetes independent of glycosylated hemoglobin A1c (HbA1c) levels. Given the conflicting evidence on the effect of positive airway pressure (PAP) therapy for OSA on HbA1c, elucidating its effect on glycemic variability has value.
Research Question
Does the use of PAP therapy for OSA improve glycemic variability in patients with type 2 diabetes?
Study Design and Methods
A randomized controlled trial was conducted in 184 patients with type 2 diabetes and moderate-to-severe OSA. Participants received either 3 months of PAP therapy with lifestyle counseling or lifestyle counseling alone. End points included the SD of glucose levels along with other metrics derived from continuous glucose monitoring and self-monitoring of blood glucose.
Results
No differences were noted in either primary or secondary continuous glucose monitoring end points between the two groups. Average use of PAP therapy was 5.4 h/night (SD, 1.6). Exploratory analyses by sex showed significant differences in the primary and secondary outcomes. In female participants, PAP therapy was associated with improvement in the SD of glucose levels, with a mean difference in change between intervention and control groups of 3.5 mg/dL (P = .02). PAP therapy was also associated with lower post-dinner and bedtime glucose levels: 20.1 mg/dL (P < .01) and 34.6 mg/dL (P < .01), respectively.
Interpretation
PAP therapy did not improve glycemic control or variability in patients with moderate-to-severe OSA and type 2 diabetes. Exploratory analyses suggested that PAP therapy may improve glucose variability in female participants. Post-dinner and bedtime glucose levels were higher in those who did not receive PAP therapy.
Trial Registration
ClinicalTrials.gov; No.: NCT02454153; URL: www.clinicaltrials.gov
Key Words: glycemic variability, glucose, OSA, positive airway pressure
Graphical Abstract
Take-home Points.
Study Question: Does the use of positive airway pressure (PAP) therapy for OSA improve glycemic variability in patients with type 2 diabetes?
Results: In this randomized controlled trial of 184 patients with type 2 diabetes, no significant differences were seen in continuous glucose monitoring metrics in patients that received PAP therapy with lifestyle counseling compared with patients receiving lifestyle counseling only. Exploratory analyses showed that the SD and mean glucose levels from continuous glucose monitoring improved by –3.5 and –14.5 mg/dL, respectively, in female participants assigned to PAP therapy with lifestyle counseling compared with female participants assigned to lifestyle counseling only. Postprandial glucose levels were lower in female participants on PAP therapy with lifestyle counseling by 28.0 mg/dL after breakfast and by 20.0 mg/dL after dinner. Bedtime glucose levels also decreased by 34.6 mg/dL in female participants on PAP therapy with lifestyle counseling.
Interpretation: Glycemic variability and glycemic control did not improve in patients with type 2 diabetes after 3 months of positive airway pressure for OSA, but a response to PAP therapy was noted in glycemic variability and postprandial glucose levels in female but not male participants.
The role of OSA in altering glucose metabolism has been a topic of significant research. Cross-sectional and longitudinal studies show that, independent of factors such as age and obesity, OSA is associated with insulin resistance, glucose intolerance, and type 2 diabetes.1, 2, 3, 4 Experimental data from animal and human studies reveal that intermittent hypoxemia and recurrent arousals in OSA can independently alter glucose metabolism.5, 6, 7 Given the possibility of a causal adverse effect of OSA on glucose metabolism, a much sought-after question has been whether treatment with positive airway pressure (PAP) can favorably influence glycemic outcomes in patients with type 2 diabetes. Empirical evidence on the effects of PAP therapy on glycemic measures in type 2 diabetes has yielded mixed results. Initial studies showed improvements in glycemic control with PAP therapy for OSA in type 2 diabetes albeit with methodologic limitations, including nonrandomized study designs, small sample sizes, and lack of a control group.8, 9, 10 Findings from a few randomized controlled trials also have not demonstrated improvements in glycemic control with OSA therapy in type 2 diabetes.11, 12, 13 One of the major limitations in the available studies on the effects of PAP therapy on glycemic control is that most have used glycosylated hemoglobin A1c (HbA1c) as the primary outcome. Although the HbA1c test is of unquestionable clinical value, it integrates basal and postprandial glucose over the preceding few months and does not provide information on variations in daily glucose levels. Postprandial glucose excursions are now recognized as important determinants of vascular disease even in those exhibiting acceptable glycemic control.14,15 Continuous glucose monitoring (CGM) provides a simple approach for quantifying day-to-day glucose profiles and thus enables assessments of glucose variability. The other major limitation in the available data is that use of PAP therapy has been well below the conventional criteria for acceptable adherence (ie, ≥ 4 h/night for 70% of nights).16 In the largest randomized clinical trial of glycemic control and OSA to date,17 HbA1c was unchanged with PAP therapy, but only 45.8% of the participants were adherent to PAP therapy at 3 months. To address such limitations, the current study was designed to examine the effects of PAP therapy on CGM-derived measures of glucose variability while optimizing adherence. It was hypothesized that improved PAP adherence would favorably influence CGM metrics of glucose variability and HbA1c in patients with moderate-to-severe OSA and type 2 diabetes.
Study Design and Methods
Sample Recruitment
An open-label randomized control trial of 3 months of PAP therapy with lifestyle counseling vs lifestyle counseling alone in adults with type 2 diabetes and newly diagnosed OSA was conducted. Complete details of the study design have been previously reported.18 Briefly, participants were recruited from the general community in the Baltimore-Washington area. Exclusion criteria included age < 21 and > 75 years, pregnancy, ongoing therapy for OSA, use of insulin, participation in a weight loss program, change in glycemic medications in the previous 6 weeks, oral steroid use, and other sleep disorders (eg, restless legs syndrome, circadian rhythm disorder, insomnia). Screening procedures also included a point-of-care HbA1c measurement, a home sleep apnea test, and a run-in period with PAP therapy with a Philips Respironics System One RemStar autotitrating PAP machine. The DCA Vantage analyzer (Siemens) was used to determine HbA1c using 1 μl of blood from a fingerstick. The Apnealink monitoring device was used to screen for OSA. Oxyhemoglobin desaturation of at least 3% was used to determine the oxygen desaturation index, which was calculated as the number of oxygen desaturation events per hour of recording time. Only those participants with HbA1c ≥ 6.5% and an oxygen desaturation index ≥ 15 events/h were further considered for the PAP run-in phase. Minimum use of PAP therapy for 4 h/night for at least 3 consecutive nights was required for further eligibility. Sample size calculations were made on the basis of the magnitude of anticipated improvement in the primary outcome (24-h SD of glucose on CGM). A difference in the SD of glucose of 3.3 mg/dL on CGM between intervention and control groups after 3 months of PAP therapy was considered statistically significant. Assuming a type 1 error of 0.05 and power of 90%, a target sample size of 92 participants per group was needed. The protocol was approved by the Institutional Review Board on Human Research (IRB No. NA_00036672), and all participants provided written informed consent.19
Study Protocol and Outcomes
Participants were randomized 1:1 using a simple allocation randomization scheme to either PAP therapy with lifestyle counseling or lifestyle counseling alone. The random allocation sequence was independently generated by a computerized algorithm with block sequences of four and concealed until randomization. The block size used in generating the randomization sequence was unknown to those involved in the study. The study physicians enrolled participants and used a customized software platform to assign them to one of the two study arms. Data on PAP therapy use were recorded nightly. Seven to 14 days after randomization, the PAP device was set to the 95th percentile of pressure that eliminated apneas, hypopneas, and airflow limitation. Lifestyle counseling was provided to both groups and included information and educational brochures on nutrition and exercise. Counseling on good sleep hygiene practices, targeting 7 to 8 h of sleep, was provided, and adherence to medical therapy was emphasized. Participants were contacted weekly to address any study-related issues. If difficulties with PAP adherence or other study-related issues were identified, a study investigator contacted the participant to address the concern.
The Dexcom G4 Platinum CGM sensor was used to assess the primary outcome. The sensor was worn for up to 14 days before randomization to collect interstitial glucose levels every 5 min. Self-monitoring of blood glucose (SMBG) was conducted for 3 days, using a fingerstick (Abbott FreeStyle). SMBG values were assessed on awakening (fasting), 90 to 120 min after breakfast, right before lunch and dinner, 90 to 120 min after lunch and dinner, and at bedtime. The resulting values were used to assess meal-related glucose excursions. Participants were instructed to avoid snacks between meals on the days that 7-point SMBG data were collected. Use of acetaminophen was not allowed during the study as it interferes with CGM sensor performance. A second CGM sensor was placed at the 3-month visit and worn for up to 14 days. The Dexcom sensor was calibrated by participants at least bid using blood glucose values obtained by fingerstick. Participants were blinded to CGM results from the Dexcom sensor but were not blinded to SMBG values. Study participants were instructed not to change any glucose-lowering medications during the study. The primary outcome was change in the SD on CGM between baseline and 3 months. Secondary outcomes included changes in CGM-derived average glucose, HbA1c, premeal glucose, postmeal glucose, as well as bedtime glucose levels. There were no changes in glucose-lowering medications in either group over the 3-month period.
Statistical Analyses
Analyses of the primary outcome were conducted according to the intention-to-treat principle. Characteristics of the randomized sample were compared by assigned group, using analysis of variance for continuous variables and χ2 for categorical variables. Mean values and associated SDs for the various CGM metrics, HbA1c, and SMBG values were summarized at baseline and at 3 months. Multivariable mixed linear regression methods were used to assess whether the change in glycemic outcomes differed according to assigned group while accounting for baseline characteristics. In secondary analyses, a per-protocol assessment was conducted with those participants who were adherent to PAP therapy as defined by Medicare (ie, PAP use for ≥ 4 h each night for > 70% of nights). Exploratory analyses by sex and BMI (< 35 vs ≥ 35 kg/m2) were also conducted, using interaction terms between these variables and treatment groups in the multivariable mixed models as needed. A two-sided P value < .05 was used to indicate statistical significance. Finally, to further explore the impact of sex on the association between PAP therapy and glycemic outcomes, an interaction between sex and treatment group was examined. Stata version 17.0 (StataCorp) was used for all analyses.
Results
Sample Characteristics
A total of 768 patients were recruited between December 2014 and December 2019 and considered eligible for a home sleep apnea test after the initial screening (Fig 1). Of these, 262 proceeded to the run-in phase to assess ability to use PAP therapy. The final randomized sample consisted of 184 participants with 92 allocated to the PAP with lifestyle counseling group and 92 to the lifestyle counseling-only group. The mean age was 59.6 years (SD, 9.1), and male participants constituted 51.0% of the sample. Self-identified race, age, BMI, HbA1c, and other clinical measures were comparable between the two groups (Table 1). Oral hypoglycemic and statin use was also comparable. Prevalent cardiovascular disease (defined as reported history of stroke, congestive heart failure, coronary artery bypass graft surgery, myocardial infarction, and/or angina) was present in 10 participants (11%) in the lifestyle-only group and one participant (1%) in the PAP therapy group. OSA severity was comparable between PAP therapy and control groups. Eleven participants did not attend the 3-month visit. Of the 173 participants who attended the 3-month visit, nine were missing follow-up CGM data. Baseline CGM measurements for these participants were carried forward to the 3-month visit. The mean and median values for nightly PAP use were 5.4 h/night (SD, 1.6) and 5.5 h/night (interquartile range, 4.3-6.4). PAP adherence, defined as ≥ 4 h/night for at least 70% of nights, was 77% in the group.
Figure 1.
Enrollment, randomization, and treatment assignment. HbA1c = glycosylated hemoglobin A1c; ODI = oxygen desaturation index; PAP = positive airway pressure.
Table 1.
Demographic and Clinical Characteristics of Randomized Sample
| Characteristic | PAP Therapy (n = 92) | Lifestyle Counseling (n = 92) |
|---|---|---|
| Age,a y | 58.4 (8.6) | 60.8 (9.6) |
| Male, No. (%) | 52 (57%) | 42 (46%) |
| Race, No. (%) | ||
| White | 53 (58%) | 50 (54%) |
| Black | 29 (32%) | 34 (37%) |
| Other | 10 (11%) | 8 (9%) |
| Ever tobacco use, No. (%) | 39 (42%) | 38 (41%) |
| BMI,a kg/m2 | 33.8 (5.5) | 33.8 (5.8) |
| BMI category, No. (%) | ||
| < 25 kg/m2 | 3 (3%) | 7 (8%) |
| 25.0-29.9 kg/m2 | 17 (19%) | 16 (17%) |
| ≥ 30 kg/m2 | 71 (78%) | 69 (75%) |
| Point-of-care HbA1c,a % | 7.5 (1.0) | 7.5 (0.9) |
| Hypoglycemic medication use, No. (%) | 80 (87%) | 85 (92%) |
| Statin use, No. (%) | 32 (35%) | 26 (28.3%) |
| Prevalent CVD, No. (%) | 1 (1%) | 10 (11%) |
| Hypertension, No. (%) | 62 (67%) | 68 (74%) |
| Epworth Sleepiness Scale scorea | 10.1 (10.1) | 9.4 (5.1) |
| ApneaLink ODI,a events/h | 31.5 (18.5) | 28.2 (13.5) |
| PAP adherent,b No. (%) | 70 (77%) | … |
CVD = cardiovascular disease; HbA1c = glycosylated hemoglobin A1c; ODI = oxygen desaturation index; PAP = positive airway pressure.
Values reported represent mean (SD).
PAP adherent defined as nightly PAP ≥ 4 h on 70% or more of the nights.
Primary Analysis
The SD for glucose, derived from the Dexcom G4 CGM device, increased by an average of 0.8 mg/dL in the PAP therapy group and 1.1 mg/dL in the control group (Table 2) with a resulting difference (Δ) that was not statistically significant (mean Δ, –0.3 mg/dL; P = .81). Mean CGM glucose values increased by 4.3 mg/dL in the PAP therapy group vs 9.5 mg/dL in the control group with a resulting difference between the two groups of –5.2 mg/dL (P = .20). HbA1c increased by 0.2% in the PAP therapy group and 0.2% in the control group. The resulting difference (Δ) in HbA1c was also not statistically significant (mean Δ, 0.0%; P = .91). Results from SMBG demonstrated that, in the control group, pre-lunch, post-dinner, and bedtime glucose levels at 3 months were significantly higher compared with baseline by 16.1, 10.1, and 16.9 mg/dL, respectively (Table 3). In contrast, the PAP therapy group showed no differences between the baseline and 3-month SMBG glucose values. As a result, the pre-lunch, post-dinner, and bedtime glucose values between the PAP therapy and control groups differed by 14.7 mg/dL (P < .01), 10.7 mg/dL (P = .04), and 13.9 mg/dL (P = .01). No differences were noted in the number of daily average calories consumed or the amounts of daily carbohydrate, protein, or fat in either the PAP therapy or control groups over the 3 months (e-Table 1). There were no changes in the use of glucose-lowering medications in either group over the 3-month period.
Table 2.
Baseline and 3-Month CGM Values and HbA1c in PAP Therapy-Lifestyle Counseling Group vs Lifestyle Counseling-Only Group
| CGM Metric | PAP Therapy With Lifestyle Counseling |
Lifestyle Counseling Alone |
Difference |
P Valuea | ||||
|---|---|---|---|---|---|---|---|---|
| Baseline | 3 Mo | ΔP | Baseline | 3 Mo | ΔL | ΔP-L | ||
| SD, mg/dL | 37.9 (11.2) | 38.7 (11.3) | 0.8 (6.7) | 38.1 (11.2) | 39.2 (10.8) | 1.1 (8.6) | –0.3 (1.1) | .81 |
| MAGE, mg/dL | 67.7 (21.9) | 69.5 (20.5) | 1.8 (14.0) | 69.5 (20.8) | 71.4 (20.4) | 1.9 (18.4) | –0.1 (2.4) | .98 |
| Mean glucose, mg/dL | 150.4 (32.1) | 154.7 (40.1) | 4.3 (25.3) | 153.0 (34.7) | 162.5 (42.4) | 9.5 (29.3) | –5.2 (4.0) | .20 |
| HbA1c, % | 7.5 (1.0) | 7.7 (1.3) | 0.2 (0.7) | 7.5 (0.9) | 7.7 (1.2) | 0.2 (0.9) | 0.0 (0.1) | .91 |
Values represent 24-h SD and mean (SE). ΔP and ΔL calculated as: [3-mo – baseline] values in the PAP therapy-lifestyle counseling group vs lifestyle counseling-only group, respectively. ΔP-L= [ΔP – ΔL]. CGM = continuous glucose monitoring; HbA1c = glycosylated hemoglobin A1c; MAGE = mean amplitude of glycemic excursion; PAP = positive airway pressure.
P value comparing the change between the PAP therapy-lifestyle counseling group vs lifestyle counseling-only group (ΔP-L).
Table 3.
Baseline and 3-Month SMBG Values in PAP Therapy With Lifestyle Counseling Group vs Lifestyle Counseling-Only Group
| Glucose (mg/dL) | PAP Therapy With Lifestyle Counseling |
Lifestyle Counseling Alone |
Difference |
P Valuea | ||||
|---|---|---|---|---|---|---|---|---|
| Baseline | 3 Mo | ΔP | Baseline | 3 Mo | ΔL | ΔP-L | ||
| Breakfast | ||||||||
| Pre | 143.5 (4.2) | 148.9 (4.3) | 5.3 (3.4) | 141.2 (4.2) | 146.5 (4.3) | 5.3 (3.4) | 0.0 (4.8) | .99 |
| Post | 195.5 (4.3) | 190.9 (4.4) | –4.6 (3.5) | 190.9 (4.2) | 194.9 (4.4) | 4.0 (3.5) | –8.6 (5.0) | .08 |
| Lunch | ||||||||
| Pre | 135.6 (4.7) | 137.0 (4.8) | 1.4 (3.8) | 134.9 (4.6) | 151.0 (4.8) | 16.1 (3.8) | –14.7 (5.4) | < .01 |
| Post | 162.3 (4.7) | 164.7 (4.8) | 2.4 (3.8) | 162.6 (4.6) | 172.6 (4.8) | 10.0 (3.9) | –7.7 (5.5) | .16 |
| Dinner | ||||||||
| Pre | 133.0 (4.4) | 135.5 (4.5) | 2.5 (3.7) | 141.6 (4.3) | 148.0 (4.5) | 6.4 (3.6) | –3.9 (5.5) | .45 |
| Post | 174.6 (4.4) | 174.1 (4.5) | –0.5 (3.7) | 169.5 (4.4) | 179.6 (4.5) | 10.1 (3.7) | –10.7 (5.2) | .04 |
| Bedtime | 160.3 (4.8) | 163.3 (4.9) | 3.0 (3.9) | 159.3 (4.7) | 176.2 (4.9) | 16.9 (3.9) | –13.9 (5.5) | .01 |
Values represent mean (SE). ΔP and ΔL calculated as: [3-mo – baseline] values in the PAP therapy-lifestyle counseling group vs lifestyle counseling-only group, respectively; ΔP-L = [ΔP – ΔL]. PAP = positive airway pressure; SMBG = self-monitoring of blood glucose.
P value comparing the change between the PAP therapy-lifestyle group vs lifestyle-only group (ΔP-L).
Exploratory Analyses
Analyses were also conducted on the basis of per-protocol and complete case approaches. No differences were seen in the primary outcome of CGM-derived SD of glucose values with either approach. However, analyses by sex showed significant differences in both primary and secondary outcomes between male and female participants (Tables 4, 5). Baseline characteristics by sex are presented in e-Table 3. In female participants, there was an improvement (–0.4 mg/dL) in the primary outcome of CGM-derived SD in the PAP therapy group compared with a worsening of the SD by 3.1 mg/dL in the control group (Table 4), resulting in a mean difference of 3.5 mg/dL (P = .02). Mean glucose also increased by 14.0 mg/dL in female participants who were in the control group, whereas the mean glucose decreased by 0.5 mg/dL in the PAP therapy group. Although HbA1c did not change for either sex, SMBG profiles showed that female participants had a favorable response to PAP therapy (Table 5). Post-breakfast glucose levels were significantly lower in female participants in the PAP therapy group (–15.8 mg/dL) compared with the control group (12.2 mg/dL). This resulted in a post-pre (Δ) glucose different of 28.0 mg/dL (P < .01) with breakfast. A sex-based difference in the post-dinner and bedtime glucose values was also observed in female participants with a difference of 20.1 mg/dL (P < .01) and 34.6 mg/dL (P < .01), respectively. At both time points, there was an increase in glucose levels in the control group at 3 months. Post-dinner and bedtime levels were higher by 15.9 mg/dL and 25.3 mg/dL in female participants in the control group. In male participants, there were no significant changes in any of the SMBG parameters over the 3-month period.
Table 4.
Baseline and 3-Month CGM Values and HbA1c in PAP Therapy-Lifestyle Counseling Group vs Lifestyle Counseling-Only Group by Sex
| Metric | PAP Therapy With Lifestyle Counseling |
Lifestyle Counseling Alone |
Difference |
P Valuea | ||||
|---|---|---|---|---|---|---|---|---|
| Baseline | 3 Mo | ΔP | Baseline | 3 Mo | ΔL | ΔP-L | ||
| Females | ||||||||
| SD, mg/dL | 39.6 (11.3) | 39.3 (12.6) | –0.4 (5.5) | 34.9 (9.9) | 38.0 (11.1) | 3.1 (8.4) | –3.5 (1.5) | .02 |
| MAGE, mg/dL | 71.9 (22.6) | 70.9 (24.0) | –1.0 (14.1) | 64.6 (19.4) | 68.7 (20.1) | 4.1 (17.4) | –5.1 (3.3) | .12 |
| Mean glucose, mg/dL | 155.2 (33.8) | 154.7 (40.4) | –0.5 (22.0) | 147.2 (32.6) | 161.2 (40.1) | 14.0 (31.3) | –14.5 (5.6) | .01 |
| HbA1c, % | 7.5 (0.9) | 7.7 (1.3) | 0.3 (0.7) | 7.4 (1.0) | 7.6 (1.1) | 0.2 (0.7) | 0.1 (0.2) | .70 |
| Males | ||||||||
| SD, mg/dL | 36.5 (11.2) | 38.3 (10.3) | 1.7 (7.4) | 42.0 (11.4) | 40.8 (10.4) | –1.3 (8.3) | 3.0 (1.6) | .07 |
| MAGE, mg/dL | 64.4 (20.9) | 68.4 (17.6) | 4.0 (13.7) | 75.3 (21.3) | 74.5 (20.4) | –0.8 (19.4) | 4.8 (3.5) | .18 |
| Mean glucose, mg/dL | 146.7 (30.5) | 154.7 (40.3) | 8.0 (27.1) | 159.9 (36.2) | 164.0 (45.4) | 4.1 (26.0) | 3.9 (5.5) | .48 |
| HbA1c, % | 7.5 (1.0) | 7.6 (1.3) | 0.1 (0.8) | 7.6 (0.9) | 7.7 (1.4) | 0.1 (1.1) | 0.0 (0.2) | .99 |
Values represent 24-h SD and mean (SE). ΔP and ΔL calculated as: [3-mo – baseline] values in the PAP therapy-lifestyle group counseling vs lifestyle counseling-only group, respectively; ΔP-L= [ΔP – ΔL]. CGM = continuous glucose monitoring; HbA1c = glycosylated hemoglobin A1c; MAGE = mean amplitude of glycemic excursion; PAP = positive airway pressure.
P value comparing the change between the PAP therapy-lifestyle counseling group vs lifestyle counseling-only group (ΔP-L).
Table 5.
Baseline and 3-Month SMBG Values in PAP Therapy-Lifestyle Counseling Group vs Lifestyle Counseling-Only Group by Sex
| Glucose (mg/dL) | PAP Intervention Arm |
Lifestyle Counseling Arm |
Difference |
P Valuea | ||||
|---|---|---|---|---|---|---|---|---|
| Baseline | 3 Mo | ΔP | Baseline | 3 Mo | ΔL | ΔP-L | ||
| Females | ||||||||
| Breakfast | ||||||||
| Pre | 146.4 (6.4) | 150.8 (6.5) | 4.4 (4.9) | 140.1 (5.7) | 152.9 (5.9) | 12.8 (4.5) | 8.4 (6.7) | .21 |
| Post | 207.5 (6.5) | 191.7 (6.5) | –15.8 (5.1) | 180.6 (5.8) | 192.7 (6.0) | 12.2 (4.7) | –28.0 (7.0) | < .01 |
| Lunch | ||||||||
| Pre | 135.7 (7.0) | 136.8 (7.1) | 1.1 (5.9) | 124.8 (6.2) | 147.6 (6.4) | 22.9 (5.4) | –21.8 (8.0) | < .01 |
| Post | 166.9 (7.0) | 163.6 (7.1) | –3.3 (5.8) | 155.1 (6.2) | 163.9 (6.5) | 8.8 (5.4) | –12.2 (8.0) | .13 |
| Dinner | ||||||||
| Pre | 137.5 (6.3) | 137.5 (6.4) | 0.0 (5.4) | 140.5 (5.6) | 148.5 (5.8) | 8.0 (4.8) | –8.0 (7.2) | .27 |
| Post | 179.3 (6.3) | 175.0 (6.4) | –4.3 (5.3) | 162.7 (5.6) | 178.6 (5.8) | 15.9 (4.9) | –20.1 (7.3) | < .01 |
| Bedtime | 169.5 (6.7) | 160.5 (6.8) | –9.0 (6.0) | 153.1 (5.9) | 178.7 (6.1) | 25.3 (5.3) | –34.6 (8.0) | < .01 |
| Males | ||||||||
| Breakfast | ||||||||
| Pre | 141.8 (5.5) | 146.9 (5.7) | 5.2 (4.5) | 142.5 (6.0) | 139.5 (6.2) | –3.0 (4.9) | 8.1 (6.7) | .22 |
| Post | 186.7 (5.5) | 190.2 (5.7) | 3.4 (4.7) | 201.9 (6.1) | 196.9 (6.2) | –5.0 (5.0) | 8.4 (6.9) | .08 |
| Lunch | ||||||||
| Pre | 135.7 (6.1) | 136.6 (6.3) | 0.9 (5.0) | 147.3 (6.8) | 157.1 (6.9) | 9.8 (5.5) | –8.9 (7.5) | .23 |
| Post | 159.0 (6.1) | 165.1 (6.3) | 6.0 (5.1) | 171.0 (6.7) | 181.4 (6.9) | 10.4 (5.5) | –4.4 (7.5) | .56 |
| Dinner | ||||||||
| Pre | 130.5 (6.0) | 133.6 (6.2) | 3.1 (5.0) | 143.7 (6.5) | 147.7 (6.7) | 4.0 (5.4) | –0.9 (7.3) | .90 |
| Post | 170.9 (6.0) | 171.9 (6.2) | 1.0 (5.0) | 177.4 (6.6) | 182.4 (6.7) | 5.0 (5.4) | –4.0 (7.4) | .59 |
| Bedtime | 153.6 (6.6) | 164.1 (6.9) | 10.6 (5.1) | 166.0 (7.2) | 172.7 (7.5) | 6.8 (5.5) | 3.8 (7.5) | .61 |
Values represent mean (SE). ΔP and ΔL calculated as: [3-mo – baseline] values in the PAP therapy-lifestyle counseling group and lifestyle counseling group, respectively. ΔP-L = [ΔP – ΔL]. PAP = positive airway pressure; SMBG = self-monitoring of blood glucose.
P value comparing the change between the PAP intervention-lifestyle counseling group and lifestyle counseling-only group (ΔP-L).
Analyses by BMI group (< 35 vs ≥ 35 kg/m2) were also conducted and showed that neither BMI group had differences in CGM outcomes with the intervention. In patients with a BMI ≥ 35 kg/m2, 3 months of PAP therapy was associated with differences in SMBG values. PAP therapy had a beneficial effect on post-breakfast, pre-lunch, post-lunch, post-dinner, and bedtime glucose values with decreases as follow: 25.9 mg/dL (P = .01), 31.4 mg/dL (P < .01), 19.8 mg/dL (P < .02), 20.6 mg/dL (P < .01), and 27.7 mg/dL (P < .01), respectively (e-Table 2).
Discussion
The results of this randomized clinical trial demonstrate that PAP therapy for moderate to severe OSA does not improve glycemic variability as assessed by continuous glucose monitoring in patients with type 2 diabetes. The addition of PAP therapy to lifestyle counseling also was not associated with improvements in CGM-derived average glucose value or HbA1c. However, secondary analyses showed that postprandial hyperglycemia with the dinner meal and bedtime glucose levels worsened in those in the lifestyle-only group but remained stable in the PAP therapy group. Finally, exploratory subgroup analyses revealed sex-specific differences, with worsening of primary and secondary glycemic outcomes in female participants with lifestyle counseling only and improvements or no changes in female participants receiving PAP therapy and lifestyle counseling.
The current study extends the relatively limited and conflicted body of evidence on the potential influence of OSA on glycemic variability in type 2 diabetes. Although some studies have shown that CGM-derived metrics such as mean amplitude of glycemic excursion,20,21 nocturnal mean amplitude of glycemic excursion,20,22,23 and nocturnal coefficient of variance20 are higher in patients with OSA than in those without OSA,20,21,23,24 a randomized controlled trial did not find a difference in CGM-derived metrics after 3 months of PAP therapy in those with type 2 diabetes.25 However, the latter study included 72 participants with many receiving insulin therapy (61% in the PAP group and 69% in the control group), suggesting that most participants had advanced type 2 diabetes that is unresponsive to hypoglycemic agents. Furthermore, adherence was suboptimal with only 44% of the PAP therapy group having acceptable adherence.25 Uncontrolled studies using CGM or continuous blood glucose sampling before and after PAP therapy in patients with OSA with type 2 diabetes have shown improvements in overnight mean glucose levels26, 27, 28 and measures of glycemic variability.27 Although CGM-derived metrics improve with PAP therapy in most of the available studies using a pre-post design in patients with type 2 diabetes and OSA, there is substantial variation in the duration of follow-up across studies (range, 30-83 days) and sample sizes are modest (n = 20-40). Interestingly, in all the available studies, the beneficial effects of PAP therapy were most notable on CGM metrics derived specifically for the sleep period.8,26, 27, 28, 29, 30
Studies assessing the effects of PAP therapy on glycemic control (ie, HbA1c) have yielded conflicting results with some initial studies,8, 9, 10 including a randomized controlled study,31 suggesting that PAP therapy improves HbA1c in type 2 diabetes. As with studies on CGM, the duration of PAP varied between 3 and 6 months. In contrast, studies that have incorporated larger samples17,32 or longer duration of PAP therapy33 have failed to show an improvement in HbA1c even after 12 months of follow-up.33 In the largest randomized clinical trial to date, Shaw et al17 found no significant changes in HbA1c in type 2 diabetes with the institution of PAP therapy for moderate-to-severe OSA. It is worth noting that, similar to previous studies, PAP adherence in that trial was suboptimal with only 45% of those in the PAP group meeting the adherence criteria at 3 months with a mean nightly PAP use of 4.3 h/night. The limitation of suboptimal adherence that is common in previous studies was addressed in the current study, as a higher average duration of nightly use was achieved with 77% of the PAP therapy group meeting the conventional criteria for acceptable adherence (≥ 4 h/night on 70% of the nights). Despite the substantial efforts to improve adherence, the results presented here are consistent with findings from a meta-analysis of 581 patients enrolled across six randomized controlled trials.11 The lack of an improvement in HbA1c with PAP therapy in the current and previous studies may be attributable to the fact that most of the participants did not have extremely poor glycemic control given that their HbA1c values were between 6.5% and 8.5%. Further improvement in HbA1c in patients with adequate glucose control is likely to be difficult with PAP therapy even in the presence of moderate-to-severe OSA (ie, floor effect). In fact, two prior studies demonstrated a statistically significant decrease in HbA1c in patients with type 2 diabetes with PAP therapy, especially in patients with higher baseline HbA1c values (ie, ≥ 8.9%).8,34
An important observation from this study is that glucose values after dinner and at bedtime were markedly worse in the control group. The dinner meal is typically the largest meal of the day35,36 and, therefore, presents a substantial metabolic challenge compared with the other meals during the day. Post-dinner hyperglycemia is likely to be sustained up to bedtime, resulting in the elevated bedtime glucose levels noted in the current study. It is plausible that the beneficial effect of PAP therapy in stabilizing glucose values may be most apparent after a large glycemic load. Given the known detrimental consequences of large meal-related glycemic excursions, the findings reported herein suggest that PAP therapy for moderate-to-severe OSA has value in patients with type 2 diabetes by possibly dampening postprandial glucose increases. Exploratory analyses also revealed sex-specific differences in both primary (ie, CGM-derived SD) and several other outcomes (ie, average glucose on CGM and multiple SMBG metrics). Compared with the control group, female participants in the PAP therapy group showed a decrease in CGM-derived SD over the 3-month period. Mean glucose levels on CGM and several other SMBG metrics were higher in the lifestyle counseling-only group, but remained stable in those in the PAP therapy group. Sex has been recognized as an important mediator of outcomes in type 2 diabetes.37,38 Moreover, sex-based differences in OSA symptoms and sleep study characteristics are also well described.39,40 It is plausible that treatment of OSA has a differential, sex-based effect on metabolic outcomes in patients with type 2 diabetes. Similar to the current study, the Sleep Apnea Cardiovascular Endpoints (SAVE) trial also demonstrated a differential effect of PAP therapy in female participants with type 2 diabetes. Follow-up blood glucose levels at 6, 24, and 48 months were lower in female participants receiving PAP therapy vs those in the control arm.13
This study has several strengths that merit discussion. These include enrollment of a relatively large sample with inclusion of equal numbers of male and female participants and a significant number of Black patients, making the findings more generalizable. Second, the collection of multiple CGM-derived measures provided greater insight into the impact of PAP therapy on glycemic measure and the exploration of heterogeneity of treatment effects in particular subgroups. Third, data on caloric and nutrient intake were collected and were comparable between the two groups. Finally, adherence to PAP therapy in the current study was higher than in any of the previous studies, with 77% of the participants assigned to the PAP therapy group meeting the criteria for adequate adherence. There are, however, some limitations. Three months of PAP therapy may not be sufficient to induce changes in outcomes such as HbA1c. However, it may be sufficient to capture changes in CGM metrics, given that PAP duration in prior clinic studies demonstrated benefit in CGM metrics with significantly shorter durations. In addition, the current study did not include measures of insulin sensitivity or insulin resistance, which may show changes with OSA therapy not evident in CGM or SMBG profiles. Other limitations include a single-center design and having a selective population of PAP-adherent patients, given the requirements of the run-in period. These limitations notwithstanding, the current study provides significant insight on the association between OSA and short-term metrics of glycemic control in people with type 2 diabetes and highlights the need for larger and longer trials that are powered to detect heterogeneity of treatment effects across subgroups, using measures that are able to detect changes in glucose profiles over time.
Interpretation
In patients with type 2 diabetes and moderate-to-severe OSA, PAP therapy did not improve the primary outcomes of glycemic control or variability. However, exploratory analyses showed that PAP therapy may improve glucose variability in female participants. Larger randomized clinical trials are required to determine the metabolic benefit of PAP therapy in specific subsets of patients with OSA and type 2 diabetes.
Funding/Support
Supported by the following grants from the National Heart, Lung, and Blood Institute: HL117167, HL146709, and HL118414.
Financial/Nonfinancial Disclosures
None declared.
Acknowledgments
Author contributions: R. N. A. takes responsibility for the content of the manuscript, including data and analyses. N. M. P. and R. N. A. are responsible for project conception, project design, data acquisition and interpretation. M. R. R., D. W., and E. S. were responsible for data analysis and interpretation. All authors contributed to the development of the final manuscript.
Role of sponsors: The sponsor did not have any input or contributions in the development of the research or manuscript.
Additional information: The e-Tables are available online under “Supplementary Data.”
Supplementary Data
References
- 1.Lindberg E., Theorell-Haglow J., Svensson M., Gislason T., Berne C., Janson C. Sleep apnea and glucose metabolism: a long-term follow-up in a community-based sample. Chest. 2012;142(4):935–942. doi: 10.1378/chest.11-1844. [DOI] [PubMed] [Google Scholar]
- 2.Nagayoshi M., Punjabi N.M., Selvin E., et al. Obstructive sleep apnea and incident type 2 diabetes. Sleep Med. 2016;25:156–161. doi: 10.1016/j.sleep.2016.05.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.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(1):140–146. doi: 10.1111/j.1440-1843.2012.02267.x. [DOI] [PubMed] [Google Scholar]
- 4.Foster G.D., Sanders M.H., Millman R., et al. Obstructive sleep apnea among obese patients with type 2 diabetes. Diabetes Care. 2009;32(6):1017–1019. doi: 10.2337/dc08-1776. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Iiyori N., Alonso L.C., Li J., et al. Intermittent hypoxia causes insulin resistance in lean mice independent of autonomic activity. Am J Respir Crit Care Med. 2007;175(8):851–857. doi: 10.1164/rccm.200610-1527OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Louis M., Punjabi N.M. Effects of acute intermittent hypoxia on glucose metabolism in awake healthy volunteers. J Appl Physiol (1985) 2009;106(5):1538–1544. doi: 10.1152/japplphysiol.91523.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Stamatakis K.A., Punjabi N.M. Effects of sleep fragmentation on glucose metabolism in normal subjects. Chest. 2010;137(1):95–101. doi: 10.1378/chest.09-0791. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Babu A.R., 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(4):447–452. doi: 10.1001/archinte.165.4.447. [DOI] [PubMed] [Google Scholar]
- 9.Hassaballa H.A., Tulaimat A., Herdegen J.J., Mokhlesi B. The effect of continuous positive airway pressure on glucose control in diabetic patients with severe obstructive sleep apnea. Sleep Breath. 2005;9(4):176–180. doi: 10.1007/s11325-005-0033-y. [DOI] [PubMed] [Google Scholar]
- 10.Shpirer I., Rapoport M.J., Stav D., Elizur A. Normal and elevated HbA1C levels correlate with severity of hypoxemia in patients with obstructive sleep apnea and decrease following CPAP treatment. Sleep Breath. 2012;16(2):461–466. doi: 10.1007/s11325-011-0525-x. [DOI] [PubMed] [Google Scholar]
- 11.Labarca G., Reyes T., Jorquera J., Dreyse J., Drake L. CPAP in patients with obstructive sleep apnea and type 2 diabetes mellitus: systematic review and meta-analysis. Clin Respir J. 2018;12(8):2361–2368. doi: 10.1111/crj.12915. [DOI] [PubMed] [Google Scholar]
- 12.Feng Y., Zhang Z., Dong Z.Z. Effects of continuous positive airway pressure therapy on glycaemic control, insulin sensitivity and body mass index in patients with obstructive sleep apnoea and type 2 diabetes: a systematic review and meta-analysis. NPJ Prim Care Respir Med. 2015;25 doi: 10.1038/npjpcrm.2015.5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Loffler K.A., Heeley E., Freed R., et al. Continuous positive airway pressure treatment, glycemia, and diabetes risk in obstructive sleep apnea and comorbid cardiovascular disease. Diabetes Care. 2020;43(8):1859–1867. doi: 10.2337/dc19-2006. [DOI] [PubMed] [Google Scholar]
- 14.Gerich J.E. Clinical significance, pathogenesis, and management of postprandial hyperglycemia. Arch Intern Med. 2003;163(11):1306–1316. doi: 10.1001/archinte.163.11.1306. [DOI] [PubMed] [Google Scholar]
- 15.Monnier L., Lapinski H., Colette C. Contributions of fasting and postprandial plasma glucose increments to the overall diurnal hyperglycemia of type 2 diabetic patients: variations with increasing levels of HbA(1c) Diabetes Care. 2003;26(3):881–885. doi: 10.2337/diacare.26.3.881. [DOI] [PubMed] [Google Scholar]
- 16.Centers for Medicare & Medicaid Services Decision Memo for Continuous Positive Airway Pressure (CPAP) Therapy for Obstructive Sleep Apnea (OSA) (CAG-000932R2). 2008. https://www.cms.gov/medicare-coverage-database/details/nca-decision-memo.aspx?NCAId=204 2021.
- 17.Shaw J.E., Punjabi N.M., Naughton M.T., et al. The effect of treatment of obstructive sleep apnea on glycemic control in type 2 diabetes. Am J Respir Crit Care Med. 2016;194(4):486–492. doi: 10.1164/rccm.201511-2260OC. [DOI] [PubMed] [Google Scholar]
- 18.Rooney M.R., Aurora R.N., Wang D., Selvin E., Punjabi N.M. Rationale and design of the Hyperglycemic Profiles in Obstructive Sleep Apnea (HYPNOS) trial. Contemp Clin Trials. 2020;101 doi: 10.1016/j.cct.2020.106248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.National Institutes of Health Clinical Center Hyperglycemic Profiles in Obstructive Sleep Apnea: Effects of PAP Therapy (HYPNOS). NCT02454153. ClinicalTrials.gov. National Institutes of Health; 2015. https://clinicaltrials.gov/ct2/show/NCT02454153 Updated May 25, 2021.
- 20.Khaire S.S., Gada J.V., Utpat K.V., Shah N., Varthakavi P.K., Bhagwat N.M. A study of glycemic variability in patients with type 2 diabetes mellitus with obstructive sleep apnea syndrome using a continuous glucose monitoring system. Clin Diabetes Endocrinol. 2020;6:10. doi: 10.1186/s40842-020-00098-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Nakata K., Miki T., Tanno M., et al. Distinct impacts of sleep-disordered breathing on glycemic variability in patients with and without diabetes mellitus. PLoS One. 2017;12(12) doi: 10.1371/journal.pone.0188689. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hui P., Zhao L., Xie Y., et al. Nocturnal hypoxemia causes hyperglycemia in patients with obstructive sleep apnea and type 2 diabetes mellitus. Am J Med Sci. 2016;351(2):160–168. doi: 10.1016/j.amjms.2015.12.002. [DOI] [PubMed] [Google Scholar]
- 23.Peng C.S., Cao Y.A., Tian Y.H., Zhang W.L., Xia J., Yang L. Features of continuous glycemic profile and glycemic variability in patients with obstructive sleep apnea syndrome. Diabetes Res Clin Pract. 2017;134:106–112. doi: 10.1016/j.diabres.2017.08.010. [DOI] [PubMed] [Google Scholar]
- 24.Kurosawa H., Saisho Y., Fukunaga K., et al. Association between severity of obstructive sleep apnea and glycated hemoglobin level in Japanese individuals with and without diabetes. Endocr J. 2018;65(1):121–127. doi: 10.1507/endocrj.EJ17-0356. [DOI] [PubMed] [Google Scholar]
- 25.Banghoj A.M., Krogager C., Kristensen P.L., et al. Effect of 12-week continuous positive airway pressure therapy on glucose levels assessed by continuous glucose monitoring in people with type 2 diabetes and obstructive sleep apnoea; a randomized controlled trial. Endocrinol Diabetes Metab. 2021;4(2) doi: 10.1002/edm2.148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Mokhlesi B., Grimaldi D., Beccuti G., et al. Effect of one week of 8-hour nightly continuous positive airway pressure treatment of obstructive sleep apnea on glycemic control in type 2 diabetes: a proof-of-concept study. Am J Respir Crit Care Med. 2016;194(4):516–519. doi: 10.1164/rccm.201602-0396LE. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Pallayova M., Donic V., Tomori Z. Beneficial effects of severe sleep apnea therapy on nocturnal glucose control in persons with type 2 diabetes mellitus. Diabetes Res Clin Pract. 2008;81(1):e8–e11. doi: 10.1016/j.diabres.2008.03.012. [DOI] [PubMed] [Google Scholar]
- 28.Wei C.Y., He Z.M., Yan H., et al. Treatment effects of short-term continuous positive airway pressure on blood glucose control in type 2 diabetic patients with obstructive sleep apnea syndrome. Int J Gen Med. 2020;13:1567–1573. doi: 10.2147/IJGM.S280837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Dawson A., Abel S.L., Loving R.T., et al. CPAP therapy of obstructive sleep apnea in type 2 diabetics improves glycemic control during sleep. J Clin Sleep Med. 2008;4(6):538–542. [PMC free article] [PubMed] [Google Scholar]
- 30.Guo L.X., Zhao X., Pan Q., et al. Effect of continuous positive airway pressure therapy on glycemic excursions and insulin sensitivity in patients with obstructive sleep apnea-hypopnea syndrome and type 2 diabetes. Chin Med J (Engl) 2015;128(17):2301–2306. doi: 10.4103/0366-6999.163382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Martinez-Ceron E., Barquiel B., Bezos A.M., et al. Effect of continuous positive airway pressure on glycemic control in patients with obstructive sleep apnea and type 2 diabetes: a randomized clinical trial. Am J Respir Crit Care Med. 2016;194(4):476–485. doi: 10.1164/rccm.201510-1942OC. [DOI] [PubMed] [Google Scholar]
- 32.Donovan L.M., Rueschman M., Weng J., et al. The effectiveness of an obstructive sleep apnea screening and treatment program in patients with type 2 diabetes. Diabetes Res Clin Pract. 2017;134:145–152. doi: 10.1016/j.diabres.2017.10.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Torrella M., Castells I., Gimenez-Perez G., et al. Continuous positive airway pressure improves sleep quality, but not glycaemic control, in patients with poorly controlled long-standing type 2 diabetes. Diabetes Metab. 2017;43(6):547–549. doi: 10.1016/j.diabet.2017.08.006. [DOI] [PubMed] [Google Scholar]
- 34.Gharsalli H., Oueslati I., Sahnoun I., et al. Effect of continuous positive airway pressure on glycated hemoglobin in patients with type 2 diabetes and obstructive sleep apnea. Tunis Med. 2019;97(12):1345–1352. [PubMed] [Google Scholar]
- 35.Kant A.K., Graubard B.I. 40-year trends in meal and snack eating behaviors of American adults. J Acad Nutr Diet. 2015;115(1):50–63. doi: 10.1016/j.jand.2014.06.354. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Scheer F.A., Morris C.J., Shea S.A. The internal circadian clock increases hunger and appetite in the evening independent of food intake and other behaviors. Obesity (Silver Spring) 2013;21(3):421–423. doi: 10.1002/oby.20351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.McGill J.B., Vlajnic A., Knutsen P.G., Recklein C., Rimler M., Fisher S.J. Effect of gender on treatment outcomes in type 2 diabetes mellitus. Diabetes Res Clin Pract. 2013;102(3):167–174. doi: 10.1016/j.diabres.2013.10.001. [DOI] [PubMed] [Google Scholar]
- 38.Moura A.M., Antunes M., Martins S.O., Raposo J.F. A statistical model to identify determinants of glycemic control in patients with type 2 diabetes with different pharmacotherapeutic profiles. PLoS One. 2020;15(7) doi: 10.1371/journal.pone.0235376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.O’Connor C., Thornley K.S., Hanly P.J. Gender differences in the polysomnographic features of obstructive sleep apnea. Am J Respir Crit Care Med. 2000;161(5):1465–1472. doi: 10.1164/ajrccm.161.5.9904121. [DOI] [PubMed] [Google Scholar]
- 40.Redline S., Kirchner H.L., Quan S.F., Gottlieb D.J., Kapur V., Newman A. The effects of age, sex, ethnicity, and sleep-disordered breathing on sleep architecture. Arch Intern Med. 2004;164(4):406–418. doi: 10.1001/archinte.164.4.406. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.