0
Original Research: Sleep Disorders |

Diabetes Mellitus Prevalence and Control in Sleep-Disordered BreathingSleep-Disordered Breathing and Diabetes: The European Sleep Apnea Cohort (ESADA) Study FREE TO VIEW

Brian D. Kent, MBBCh; Ludger Grote, PhD; Silke Ryan, PhD; Jean-Louis Pépin, PhD; Maria R. Bonsignore, PhD; Ruzena Tkacova, PhD; Tarja Saaresranta, PhD; Johan Verbraecken, PhD; Patrick Lévy, PhD; Jan Hedner, PhD; Walter T. McNicholas, MD, FCCP; on behalf of the ESADA collaborators
Author and Funding Information

From the School of Medicine and Medical Science (Drs Kent, Ryan, and McNicholas), University College Dublin and Pulmonary and Sleep Disorders Unit, St. Vincent’s University Hospital, Dublin, Ireland; Department of Sleep Medicine (Drs Grote and Hedner), Sahlgrenska University Hospital, Gothenburg, Sweden; Université Grenoble Alpes (Drs Pépin and Lévy), Institut National de la Santé et de la Recherche Médicale, and Centre Hospitalier Universitaire de Grenoble, Grenoble, France; Biomedical Department of Internal & Specialist Medicine (Dr Bonsignore), University of Palermo and National Research Council Institute of Biomedicine and Molecular Immunology, Palermo, Italy; Department of Respiratory Medicine (Dr Tkacova), Pavol Jozef Šafárik University in Košice and Louis Pasteur University Hospital, Kosice, Slovakia; Division of Medicine (Dr Saaresranta), Department of Pulmonary Diseases, Turku University Hospital and Sleep Research Unit, Department of Physiology, University of Turku, Turku, Finland; and Department of Pulmonary Medicine (Dr Verbraecken), Antwerp University Hospital and University of Antwerp, Antwerp, Belgium.

CORRESPONDENCE TO: Walter T. McNicholas, MD, FCCP, Pulmonary and Sleep Disorders Unit, St. Vincent’s University Hospital, Dublin 4, Ireland; e-mail: Walter.mcnicholas@ucd.ie


Part of this article has been published in abstract form (Kent BD, Grote L, Bonsignore MR, et al. Am J Respir Crit Care Med. 2012;185:A5379).

FUNDING/SUPPORT: The maintenance of the European Sleep Apnea (ESADA) study database is supported by unrestricted grants from ResMed and Philips Respironics (Koninklijke Philips N.V.). Drs Kent and Ryan are supported by grants from the Health Research Board, Ireland [HPF/2009/033], and Dr Grote is supported by grants from the Swedish Heart and Lung Foundation.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details.


Chest. 2014;146(4):982-990. doi:10.1378/chest.13-2403
Text Size: A A A
Published online

BACKGROUND:  OSA is associated with an increased risk of cardiovascular morbidity. A driver of this is metabolic dysfunction and in particular type 2 diabetes mellitus (T2DM). Prior studies identifying a link between OSA and T2DM have excluded subjects with undiagnosed T2DM, and there is a lack of population-level data on the interaction between OSA and glycemic control among patients with diabetes. We assessed the relationship between OSA severity and T2DM prevalence and control in a large multinational population.

METHODS:  We performed a cross-sectional analysis of 6,616 participants in the European Sleep Apnea Cohort (ESADA) study, using multivariate regression analysis to assess T2DM prevalence according to OSA severity, as measured by the oxyhemoglobin desaturation index. Patients with diabetes were identified by previous history and medication prescription, and by screening for undiagnosed diabetes with glycosylated hemoglobin (HbA1c) measurement. The relationship of OSA severity with glycemic control was assessed in diabetic subjects.

RESULTS:  T2DM prevalence increased with OSA severity, from 6.6% in subjects without OSA to 28.9% in those with severe OSA. Despite adjustment for obesity and other confounding factors, in comparison with subjects free of OSA, patients with mild, moderate, or severe disease had an OR (95% CI) of 1.33 (1.04-1.72), 1.73 (1.33-2.25), and 1.87 (1.45-2.42) (P < .001), respectively, for prevalent T2DM. Diabetic subjects with more severe OSA had worse glycemic control, with adjusted mean HbA1c levels 0.72% higher in patients with severe OSA than in those without sleep-disordered breathing (analysis of covariance, P < .001).

CONCLUSIONS:  Increasing OSA severity is associated with increased likelihood of concomitant T2DM and worse diabetic control in patients with T2DM.

Figures in this Article

OSA is a highly prevalent disorder characterized by recurrent episodes of upper airway collapse during sleep with associated airflow obstruction, intermittent hypoxemia, sympathetic excitation, and disrupted sleep quality. OSA affects at least 2% of women and 4% of men in the developed world and leads to excessive daytime sleepiness, diminished quality of life and cognitive functioning, and increased risk of road traffic accidents.1 Moreover, subjects with OSA are more likely to develop hypertension and experience cardiovascular morbidity and mortality, even following adjustment for confounding factors.24

A potential important driver of increased cardiovascular disease in subjects with OSA is metabolic dysfunction, and in particular impaired glycemic health.5 Specifically, an independent link between sleep-disordered breathing and impaired glucose tolerance or type 2 diabetes mellitus (T2DM) has been suggested by a number of population-level studies.610 However, discerning an independent effect of OSA on risk of T2DM is made more challenging by the key role obesity plays in both of these disorders.11 Data from community- and sleep laboratory-based populations have identified a significant cross-sectional relationship between OSA severity and prevalence of T2DM,6,7 but these studies have been limited by their comparative small size, and their reliance on patient self-reporting, administrative databases, or fasting plasma glucose measurement to establish a diagnosis of T2DM.

More recent data would suggest a strikingly high prevalence of undiagnosed T2DM in patients with suspected OSA and further suggest that reliance on measurement of fasting glucose levels fails to identify a substantial proportion of these subjects.12 The measurement of glycosylated hemoglobin (HbA1c) levels has been recommended as a stand-alone diagnostic test for T2DM, and facilitates the convenient, large-scale screening of at-risk populations.13 Moreover, use of HbA1c testing may identify a significantly greater burden of undiagnosed disease than fasting glucose measurement alone.12

Meanwhile, a number of case-control studies have examined the influence of sleep-disordered breathing on glycemic control in patients with diabetes, identifying a seemingly independent relationship despite adjustment for confounding factors.1416 However, there is a lack of population-level data evaluating how OSA severity may correlate with HbA1c levels in diabetic populations.

The European Sleep Apnea Cohort (ESADA) study is a multicenter, multinational study, prospectively recruiting subjects attending sleep laboratories across Europe to evaluate the relationship of sleep-disordered breathing with T2DM, metabolic dysfunction, and cardiovascular disease.17 The aim of the present study was to examine, in a cross-sectional analysis, the relationship between OSA severity and T2DM prevalence, using measurement of HbA1c levels as a screening tool for undiagnosed T2DM, alongside an evaluation of how sleep-disordered breathing may relate to glycemic control in diabetic subjects.

The ESADA study is a pan-European, multicenter, prospective study involving 24 sleep clinics across 15 European countries and Israel. The underlying design and investigative techniques used in ESADA have been discussed in detail previously.17 Briefly, ESADA was established to investigate the role of OSA in driving cardiovascular and metabolic morbidity and mortality, with the goal of prospectively evaluating a large cohort of subjects with suspected sleep-disordered breathing. ESADA uses a web-based collection platform to facilitate transfer of data from individual centers to the central database at the University of Gothenburg, Sweden.

Subjects

Subjects aged 18 to 80 years referred for assessment at any of the participating centers were considered eligible for enrollment in ESADA, unless they were receiving treatment of previously diagnosed OSA, had ongoing substance abuse, or had severe comorbidity portending limited life expectancy. At baseline, demographic, anthropometric, and clinical variables, including measured BMI, smoking history, comorbidities, and medication usage, were recorded for each patient. Participating subjects also provided a venous blood sample for assessment of lipid profile and HbA1c levels. An in-depth discussion of data collection policies and procedures is reported elsewhere.17 Subjects enrolled in ESADA from March 2007 to July 2012 were included in the present study. Research ethics committee approval for the study was obtained at each of the participating centers (e-Appendix 1). Informed consent was obtained from all participants.

Sleep Studies

Either cardiorespiratory polygraphy (PG) or full polysomnography (PSG) were performed according to local practice. All sleep data were manually edited according to protocol definitions before entry. PG recordings included a minimum of four recording channels (level 3 devices according to the American Academy of Sleep Medicine [AASM]),18 while PSG studies were performed and analyzed according to AASM criteria,19 utilizing a minimum of seven recording channels. Scoring of sleep studies in ESADA was performed in accordance with AASM 2007 rules, with sleep-disordered breathing assessed according to the apnea-hypopnea index (AHI) and with the oxyhemoglobin desaturation index (ODI), defined as the average number of apneas/hypopneas or the number of transient desaturations (≥ 4%) per hour of sleep for PSG and per hour of analyzed time for PG recordings.19 To enhance comparability between ESADA centers using PG and PSG recording, and to minimize inconsistencies between study centers using recommended and alternative AASM criteria for PSG scoring, ODI rather than AHI data were used to grade OSA severity in the present report. The Epworth Sleepiness Scale was used to assess subjective daytime sleepiness.20 Sleep study scoring in ESADA has been discussed in more detail previously.17

Outcome Measures and Statistical Analysis

The primary aim of this study was to investigate the likelihood of a concomitant diagnosis of T2DM across degrees of OSA severity. Severity of OSA was measured categorically according to ODI, with subjects with an ODI < 5 events/h classified as having no OSA, and serving as a reference group for the estimation of effect size. Mild, moderate, and severe OSA were defined as an ODI of 5 to 14.9, 15 to 29.9, and ≥ 30 events/h, respectively. Subjects were considered to be diabetic if they were using diabetic medications, had clinician-diagnosed T2DM, or had a HbA1c level ≥ 6.5% as measured in Diabetes Control and Compliance Trial accredited laboratories.13

Baseline patient characteristics across quartiles were compared using analysis of variance with post hoc Bonferroni analysis, Kruskall-Wallis, and Mann-Whitney U tests, and χ2 tests for parametric, nonparametric, and categorical variables, respectively. Logistic regression analysis was used to generate ORs and 95% CIs for prevalent T2DM according to OSA severity, following adjustment for study site, and demographic (race, age, sex) and anthropometric (BMI, neck circumference) variables, smoking status, average weekly alcohol consumption, comorbidities (coronary artery disease, congestive heart failure, hyperlipidemia, COPD, and psychiatric disease) and medication use (B-blockers, oral corticosteroids, and statins). Stratified analyses were performed to assess the relationship of OSA with T2DM prevalence in subpopulations categorized by sex and sleep study modality (PG vs PSG).

Guidelines from the American Diabetes Association suggest a HbA1c level of ≥ 7% constitutes suboptimal glycemic control.21 In an analysis confined to diabetic subjects, the relationship of OSA severity with poor diabetic control was examined by a similar logistic regression model to that described previously: ORs were generated across OSA severity categories, adjusting for demographic factors, obesity measures, study site, clinical variables, and medication prescription. Finally, adjusted mean HbA1c levels for OSA severity categories were calculated and compared by analysis of covariance with Bonferroni post hoc correction, with the analysis again confined to diabetic patients and adjusted for relevant confounding factors, including diabetic medication prescription.

Statistical analyses were performed using SPSS version 18.0 statistical software (SPSS Inc; IBM). A P value < .05 was considered statistically significant.

Population Characteristics

From March 2007 to July 2012, 12,636 subjects were enrolled in ESADA. A total of 6,616 of these had both a sleep study performed and a valid HbA1c level measured during the study period. Figure 1 summarizes the recruitment and evaluation process of these patients. Descriptive characteristics of the study population, stratified by OSA severity, are presented in Table 1. A total of 2,833 patients (42.8%) underwent PSG, the remainder PG. More severe OSA was associated with greater male predominance, increased prevalence and severity of obesity, and increasing burden of cardiovascular and metabolic disease, along with increased cardiovascular medication usage.

Figure Jump LinkFigure 1 –  Patient flow diagram. DCCT = Diabetes Control and Compliance Trial; ESADA = European Sleep Apnea Cohort; HbA1c = glycosylated hemoglobin.Grahic Jump Location
Table Graphic Jump Location
TABLE 1 ]  Patient Characteristics Stratified by OSA Severity

Data are expressed as percentage of total or mean (SD). CHF = congestive heart failure; DBP = diastolic BP; ESS = Epworth Sleepiness Scale; HTN = hypertension; IHD = ischemic heart disease; n/a = not applicable; OCS = oral corticosteroid; ODI = oxyhemoglobin desaturation index; RAS = renin-angiotensin system; SBP = systolic BP; Spo2 = peripheral oxyhemoglobin saturation.

Occurrence of T2DM in the ESADA Cohort

Prevalence of T2DM increased significantly as OSA severity increased. Of the overall cohort, 17.2% was diabetic, increasing from 6.6% in subjects without OSA to 28.9% in those with severe OSA, with a similar pattern seen in diabetic medication usage and mean HbA1c levels (Table 2). Reflecting this, in comparison with subjects with no sleep-disordered breathing, those with mild, moderate, and severe disease had unadjusted OR of 2.33 (95% CI, 1.85-2.93), 3.76 (95% CI, 2.98-4.75), and 5.74 (95% CI, 4.64-7.12), respectively, for prevalent T2DM (P < .001 for trend) (Table 3). In unadjusted analyses, an increased burden of T2DM was seen in subjects with more severe sleep-disordered breathing, regardless of sex or sleep study modality (Table 4).

Table Graphic Jump Location
TABLE 2 ]  Type 2 Diabetes Mellitus, Diabetic Medication Usage, and Mean HbA1c by OSA Severity

Data are expressed as percentage of cohort or mean (SD). HbA1c = glycosylated hemoglobin; T2DM = type 2 diabetes mellitus.

Table Graphic Jump Location
TABLE 3 ]  Crude and Adjusted ORs of Prevalent Diabetes Mellitus Stratified by OSA Severity

Data expressed as OR (95% CI). Model 1 = unadjusted model. Model 2 = adjusted for demographic factors (race, age, sex), average sleep length, active smoker, alcohol use, Epworth sleepiness scale, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Model 3 = adjusted as per model 2 with BMI and neck circumference.

Table Graphic Jump Location
TABLE 4 ]  Crude and Adjusted ORs of Prevalent Diabetes Mellitus By Subgroup Stratified by OSA Severity

Data expressed as OR (95% CI). Model 1 = unadjusted model. Model 2 = adjusted for demographic factors (race, age, sex), average sleep length, active smoker, alcohol use, Epworth sleepiness scale, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Model 3 = adjusted as per model 2 with BMI and neck circumference. PSG = polysomnography.

Multivariate Analyses of T2DM Prevalence

Multivariate logistic regression analysis, incorporating adjustment for demographic, anthropometric, and clinical factors was then performed. In the overall cohort, T2DM remained more common in subjects with more severe sleep-disordered breathing despite adjustment for study site, age, sex, race, smoking history, alcohol use, daytime sleepiness, BMI, neck circumference, comorbidities, and medication usage, with an OR of 1.87 (95% CI, 1.45-2.42; P < .001 for trend) among subjects with the most severe OSA (Table 3). Using a similar logistic regression model, subgroup analyses demonstrated a consistent relationship between OSA and prevalent T2DM among men and women, and irrespective of sleep study modality (Table 4).

Sensitivity Analysis

Subjects from the ESADA cohort who did not undergo measurement of HbA1c levels were of necessity excluded from this study. To evaluate the generalizability of our findings to the ESADA population as a whole, we undertook a similar analysis, with a more restricted definition of T2DM (self-report, clinician diagnosis, and/or diabetic medication usage), again using a multivariate logistic regression model. This yielded similar results to the primary analysis presented earlier with an OR of 1.38 (95% CI, 1.10-1.75) (P = .012 for trend) for prevalent T2DM in patients with severe OSA, following adjustment for the confounding effects of demographic, anthropometric, and clinical variables (Table 5).

Table Graphic Jump Location
TABLE 5 ]  Crude and Adjusted Odds of Prevalent Diabetes Mellitus by OSA Severity in Overall Study Cohort (N = 11,702)

Data expressed as OR (95% CI). Model 1 = unadjusted model. Model 2 = adjusted for demographic factors (race, age, sex), average sleep length, active smoker, alcohol use, Epworth sleepiness scale, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Model 3 = adjusted as per model 2 with BMI and neck circumference.

Relationship of OSA Severity With Diabetic Control

Among subjects with T2DM, increasing OSA severity predicted increased likelihood of suboptimal glycemic control (HbA1c ≥ 7%). When compared with nonapneic diabetic patients, those with severe OSA had an OR of 2.02 (95% CI, 1.11-3.66; P = .022 for trend) for poor diabetic control, following adjustment for study site, age, sex, race, smoking history, alcohol use, daytime sleepiness, BMI, neck circumference, comorbidities, and prescription of diabetic medications (Table 6). Similarly, adjusted mean HbA1c levels among diabetic patients increased with increasing OSA severity, from 6.76% (95% CI, 6.39-7.13) and 6.70% (95% CI, 6.34-7.05) in those with no or mild OSA to 6.88% (95% CI, 6.58-7.19) and 7.48% (95% CI, 7.18-7.79) in those with moderate or severe OSA, respectively (analysis of covariance, P < .001)(Fig 2).

Table Graphic Jump Location
TABLE 6 ]  Crude and Adjusted ORs of HbA1c ≥ 7% Among Diabetic Patients by OSA Severity

Data expressed as OR (95% CI). Model 1 = unadjusted model. Model 2 = adjusted for demographic factors (race, age, sex), average sleep length, active smoker, alcohol use, Epworth sleepiness scale, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Model 3 = adjusted as per model 2 with BMI and neck circumference. See Table 2 legend for expansion of abbreviation.

Figure Jump LinkFigure 2 –  Adjusted mean HbA1c levels among diabetic subjects stratified by OSA severity. Analysis adjusted for age, sex, race, study site, smoking history, alcohol use, average sleep length, BMI, neck circumference, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Analysis of covariance, P < .001. See Figure 1 legend for expansion of abbreviations.Grahic Jump Location

In the largest study in this field to date, to our knowledge, we observed a significant relationship between OSA severity and the likelihood of a coexistent diagnosis of T2DM in a multinational population of subjects attending sleep services across Europe, alongside evidence of worse glycemic control in diabetic subjects with more severe sleep-disordered breathing. These relationships persisted despite adjustment for important confounding factors such as age, obesity, smoking history, comorbidities, and medication prescription. Moreover, this held true regardless of sex or sleep study modality.

Important data from population studies performed in North America have suggested a similar relationship between sleep-disordered breathing and T2DM. Among 1,387 community-based subjects enrolled in the Wisconsin Sleep Cohort study, those with moderately severe OSA (AHI ≥ 15 events/h) had an OR of 2.30 (95% CI, 1.28-4.11) for prevalent T2DM.6 Similarly, an analysis of 2,149 patients attending Canadian sleep laboratories demonstrated an OR of 2.18 (95% CI, 1.22-3.89) for concomitant T2DM in subjects with severe OSA (respiratory disturbance index ≥ 30 events/h),7 although this effect appeared to be restricted to subjects with significant daytime sleepiness. However, both of these studies were limited by their reliance on physician diagnosis, patient self-report, administrative databases, or fasting glucose measurement to establish a diagnosis of T2DM, thereby potentially significantly underestimating overall T2DM prevalence, and both enrolled a substantially smaller patient cohort than ESADA. The present report also incorporates more thorough screening for undiagnosed T2DM, by including an HbA1c level ≥ 6.5% as one of the diagnostic criteria for T2DM in accordance with international guidelines.13

A stronger relationship was seen between severity of sleep-disordered breathing and prevalent T2DM in men than women. This may represent a statistical anomaly or a degree of selection bias, but may reflect a role for hormonal influences. Similarly, gender differences in adipose tissue distribution could lead to differing adipose tissue interactions with intermittent hypoxemia, systemic inflammation, and sympathetic nervous system activation induced by OSA.

While we observed a cross-sectional association between OSA severity and T2DM prevalence, a potential causative role for OSA in driving the development of T2DM is suggested by both population-level data and clinical trials of nocturnal CPAP (nCPAP) therapy. Among 544 subjects free of preexisting diabetes evaluated in a Yale sleep clinic, a diagnosis of OSA conferred a significantly increased risk of developing T2DM over a mean follow-up period of 2.7 years, despite adjustment for confounding demographic and clinical factors (HR 1.43 per quartile of OSA severity).9

There is a relative lack of population-level data from large cohorts examining the role played by sleep-disordered breathing in glycemic control among diabetic patients; the present report represents, to our knowledge, the first population-level data examining how a diagnosis of OSA may be associated with glycemic control in patients with diabetes. This relationship has, however, been examined by a number of case-control studies, which have generally suggested that increasing severity of untreated OSA independently predicts HbA1c levels.1416 Similarly, data from a randomized trial in Asian patients suggest that successful initiation of nCPAP therapy can significantly improve glycemic health.5,22 The degree of difference in adjusted mean HbA1c levels (0.72%) seen between the diabetic patients with no OSA and severe OSA in our cohort would potentially be clinically significant at a population level, with a 1% change in HbA1c level conferring a 21% increased risk of death among patients with diabetes in the landmark United Kingdom Prospective Diabetes Study (UKPDS).23 However, these data should still be treated with some caution, as reports from other authors have failed to demonstrate a similar relationship, or have found any beneficial effect from nCPAP to be confined to selected subgroups.24,25

This study has a number of limitations. First, these are cross-sectional data, and thus causality in the observed relationship between OSA and prevalent T2DM cannot be inferred. Second, these data are derived from a sleep clinic population, and should, therefore, only be applied to general populations with caution. While the multinational, multicenter nature of ESADA overall confers significant advantages, it is also an important limitation, largely due to intercenter differences in sleep study modality and analysis. We have attempted to reduce the impact of this limitation by using ODI to assess sleep-disordered breathing severity, with the aim of reducing interobserver and intercenter subjectivity in sleep study scoring, and by adjusting our statistical models for study site. However, we cannot discount the possibility that a residual confounding effect may survive in our data, and while the use of ODI may reduce intercenter discrepancies in sleep study scoring, heterogeneity is likely to persist between subjects undergoing PSG vs PG, with ODI in the polygraphy group likely underestimated compared with the PSG group. The use of ODI as a severity measure is also a limitation, as this represents only frequency of hypoxemic exposure, rather than the absolute severity of sleep-disordered breathing as measured by the AHI. While recent data would suggest that OSA-related factors other than AHI, in particular nocturnal hypoxemia, may be better predictors of subsequent cardiovascular mortality,26 AHI remains the preferred severity metric, both in a clinical and research context.

Furthermore, although our statistical models included BMI and neck circumference in an attempt to adjust for overall and central obesity, respectively, we did not have objective imaging measures of visceral adiposity, an important driver of insulin resistance and T2DM. Similarly, while we were able to adjust our analyses for current smoking status, we were not in a position to do so for overall tobacco exposure and consumption. Sleep restriction may contribute to the development of metabolic disease, and while we were able to incorporate patient-estimated usual sleep length into our statistical models, subjects did not undergo wrist actigraphy. Finally, although our analyses of glycemic control in patients with diabetes were adjusted for the prescription of diabetic medications, we were unable to control for the number or dosage of these, patient compliance with diabetic therapy, or indeed for important lifestyle-related variables such as daily activity and diet. These latter factors are clearly fundamental limitations of our data; thus, our observation that more severe nocturnal hypoxemia independently predicted worse control of diabetes should not be viewed as a guideline for practice, but should instead facilitate focused, patient-level studies in this field.

Challenging statistical research suggests that currently accepted thresholds for statistical significance may be associated with a high likelihood of type 1 errors, with the consequent reporting of spurious findings.27 Hence, it has been suggested that more stringent statistical cutoffs be applied to future research. Many of our analyses identified highly significant P values which would match these proposed new requirements, but a number did not (eg, the association between OSA severity and suboptimal diabetic control) (Table 6, model 3). Consequently, while we take reassurance from other authors reporting similar results in prior smaller cohorts,15 these data should be viewed with some caution, at least until reproduced in other large-scale studies.

In conclusion, in a large, multinational study evaluating consecutive subjects undergoing assessment of suspected sleep-disordered breathing, the diagnosis and severity of OSA was associated with an increased likelihood of a concomitant diagnosis of T2DM, despite adjustment for important confounding factors such as age, obesity, and tobacco use. Further, long-term follow-up studies from this cohort will address the role of OSA in the development of incident T2DM, while clinical and mechanistic studies exploring the impact of IH on adipose tissue function are warranted.

Author contributions: B. D. K. assumes responsibility for the content of the manuscript. B. D. K., L. G., S. R., J.-L. P., M. R. B., R. T., T. S., J. V., P. L., J. H., and W. T. M. contributed to study design, data collection, and manuscript preparation, and provided final approval for submission.

Financial/nonfinancial disclosures: The authors have reported to CHEST the following conflicts of interest: Dr Saaresranta received grants from Turku University Hospital (Governmental EVO Grant), the Jalmari and Rauha Ahokas Foundation, and the Finnish Antituberculosis Association Foundation and speaking fees from ResMed Finland. Dr Hedner has, on behalf of the ESADA study group, received grants for maintenance of the database from ResMed and Philips Respironics (Koninklijke Philips N.V.). He has provided advisory services to Almirall S.A., AstraZeneca, and Respicardia, Inc. Drs Kent, Grote, Ryan, Pépin, Bonsignore, Tkacova, Verbraecken, Lévy, and McNicholas have reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Collaborators: Ulla Anttalainen, PhD; Ferran Barbe, PhD; Ozen Basoglu, MD; Piotr Bielicki, MD; Pierre Escourrou, PhD; Cristina Esquinas, MPH; Ingo Fietze, PhD; Lynda Hayes, BSc; Marta Kumor, MD; John-Artur Kvamme, MD; Lena Lavie, PhD; Peretz Lavie, PhD; Carolina Lombardi, PhD; Oreste Marrone, MD; Juan Fernando Masa, PhD; Josep M. Montserrat, PhD; Gianfranco Parati, MD; Athanasia Pataka, PhD; Thomas Penzel, PhD; Robert Plywaczewski, MD; Martin Pretl, MD; Renata Riha, MD; Gabriel Roisman, PhD; Richard Schulz, MD; Pawel Sliwinski, MD; Richard Staats, MD; Paschalis Steiropoulos, PhD; Giedvar Varoneckas, PhD; Audrey Vitols, PhD; Heleen Vrints, MSc.

Role of sponsors: The sponsors played no role in data collection or analysis, or manuscript preparation, and had no access to study data.

Additional information: The e-Appendix can be found in the Supplemental Materials section of the online article.

AASM

American Academy of Sleep Medicine

AHI

apnea-hypopnea index

ESADA

European Sleep Apnea Cohort

HbA1c

glycosylated hemoglobin

nCPAP

nocturnal CPAP

ODI

oxyhemoglobin desaturation index

PG

cardiorespiratory polygraphy

PSG

polysomnography

T2DM

type 2 diabetes mellitus

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(17):1230-1235. [CrossRef] [PubMed]
 
McNicholas WT, Bonsigore MR; Management Committee of EU COST ACTION B26. Sleep apnoea as an independent risk factor for cardiovascular disease: current evidence, basic mechanisms and research priorities [published correction appears inEur Respir J.2007;29(3):614]. Eur Respir J. 2007;29(1):156-178. [CrossRef] [PubMed]
 
Marin JM, Carrizo SJ, Vicente E, Agusti AG. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet. 2005;365(9464):1046-1053. [CrossRef] [PubMed]
 
Punjabi NM, Caffo BS, Goodwin JL, et al. Sleep-disordered breathing and mortality: a prospective cohort study. PLoS Med. 2009;6(8):e1000132. [CrossRef] [PubMed]
 
Lam JC, Mak JC, Ip MS. Obesity, obstructive sleep apnoea and metabolic syndrome. Respirology. 2012;17(2):223-236. [CrossRef] [PubMed]
 
Reichmuth KJ, Austin D, Skatrud JB, Young T. Association of sleep apnea and type II diabetes: a population-based study. Am J Respir Crit Care Med. 2005;172(12):1590-1595. [CrossRef] [PubMed]
 
Ronksley PE, Hemmelgarn BR, Heitman SJ, et al. Obstructive sleep apnoea is associated with diabetes in sleepy subjects. Thorax. 2009;64(10):834-839. [CrossRef] [PubMed]
 
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(6):521-530. [CrossRef] [PubMed]
 
Botros N, Concato J, Mohsenin V, Selim B, Doctor K, Yaggi HK. Obstructive sleep apnea as a risk factor for type 2 diabetes. Am J Med. 2009;122(12):1122-1127. [CrossRef] [PubMed]
 
Priou P, Le Vaillant M, Meslier N, et al; IRSR Sleep Cohort Group. Independent association between obstructive sleep apnea severity and glycated hemoglobin in adults without diabetes. Diabetes Care. 2012;35(9):1902-1906. [CrossRef] [PubMed]
 
Bonsignore MR, McNicholas WT, Montserrat JM, Eckel J. Adipose tissue in obesity and obstructive sleep apnoea. Eur Respir J. 2012;39(3):746-767. [CrossRef] [PubMed]
 
Fitzgerald DB, Kent BD, Garvey JF, Russell A, Nolan G, McNicholas WT. Screening for diabetes mellitus in patients with OSAS: a case for glycosylated haemoglobin. Eur Respir J. 2012;40(1):273-274. [CrossRef] [PubMed]
 
American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2011;34(suppl 1):S62-S69. [CrossRef] [PubMed]
 
Tamura A, Kawano Y, Watanabe T, Kadota J. Obstructive sleep apnea increases hemoglobin A1c levels regardless of glucose tolerance status. Sleep Med. 2012;13(8):1050-1055. [CrossRef] [PubMed]
 
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(5):507-513. [CrossRef] [PubMed]
 
Pillai A, Warren G, Gunathilake W, Idris I. Effects of sleep apnea severity on glycemic control in patients with type 2 diabetes prior to continuous positive airway pressure treatment. Diabetes Technol Ther. 2011;13(9):945-949. [CrossRef] [PubMed]
 
Hedner J, Grote L, Bonsignore M, et al. The European Sleep Apnoea Database (ESADA): report from 22 European sleep laboratories. Eur Respir J. 2011;38(3):635-642. [CrossRef] [PubMed]
 
Standards of Practice Committee of the American Sleep Disorders Association. Practice parameters for the use of portable recording in the assessment of obstructive sleep apnea. Sleep. 1994;17(4):372-377. [PubMed]
 
Iber C, Ancoli-Israeli S, Chesson A, Quan S. The AASM Manual for the Scoring Of Sleep And Associated Events: Rules, Terminology And Technical Specifications. 1st ed. Westchester, IL: American Academy of Sleep Medicine, 2007.
 
Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14(6):540-545. [PubMed]
 
American Diabetes Association. Standards of medical care in diabetes—2013. Diabetes Care. 2013;36(suppl 1):S11-S66. [CrossRef] [PubMed]
 
Lam JC, Lam B, Yao TJ, et al. A randomised controlled trial of nasal continuous positive airway pressure on insulin sensitivity in obstructive sleep apnoea. Eur Respir J. 2010;35(1):138-145. [CrossRef] [PubMed]
 
Stratton IM, Adler AI, Neil HA, et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ. 2000;321(7258):405-412. [CrossRef] [PubMed]
 
Barceló A, Barbé F, de la Peña M, et al. Insulin resistance and daytime sleepiness in patients with sleep apnoea. Thorax. 2008;63(11):946-950. [CrossRef] [PubMed]
 
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(11):969-974. [CrossRef] [PubMed]
 
Kendzerska T, Gershon AS, Hawker G, Leung RS, Tomlinson G. Obstructive sleep apnea and risk of cardiovascular events and all-cause mortality: a decade-long historical cohort study. PLoS Med. 2014;11(2):e1001599. [CrossRef] [PubMed]
 
Johnson VE. Revised standards for statistical evidence. Proc Natl Acad Sci U S A. 2013;110(48):19313-19317. [CrossRef] [PubMed]
 

Figures

Figure Jump LinkFigure 1 –  Patient flow diagram. DCCT = Diabetes Control and Compliance Trial; ESADA = European Sleep Apnea Cohort; HbA1c = glycosylated hemoglobin.Grahic Jump Location
Figure Jump LinkFigure 2 –  Adjusted mean HbA1c levels among diabetic subjects stratified by OSA severity. Analysis adjusted for age, sex, race, study site, smoking history, alcohol use, average sleep length, BMI, neck circumference, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Analysis of covariance, P < .001. See Figure 1 legend for expansion of abbreviations.Grahic Jump Location

Tables

Table Graphic Jump Location
TABLE 1 ]  Patient Characteristics Stratified by OSA Severity

Data are expressed as percentage of total or mean (SD). CHF = congestive heart failure; DBP = diastolic BP; ESS = Epworth Sleepiness Scale; HTN = hypertension; IHD = ischemic heart disease; n/a = not applicable; OCS = oral corticosteroid; ODI = oxyhemoglobin desaturation index; RAS = renin-angiotensin system; SBP = systolic BP; Spo2 = peripheral oxyhemoglobin saturation.

Table Graphic Jump Location
TABLE 2 ]  Type 2 Diabetes Mellitus, Diabetic Medication Usage, and Mean HbA1c by OSA Severity

Data are expressed as percentage of cohort or mean (SD). HbA1c = glycosylated hemoglobin; T2DM = type 2 diabetes mellitus.

Table Graphic Jump Location
TABLE 3 ]  Crude and Adjusted ORs of Prevalent Diabetes Mellitus Stratified by OSA Severity

Data expressed as OR (95% CI). Model 1 = unadjusted model. Model 2 = adjusted for demographic factors (race, age, sex), average sleep length, active smoker, alcohol use, Epworth sleepiness scale, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Model 3 = adjusted as per model 2 with BMI and neck circumference.

Table Graphic Jump Location
TABLE 4 ]  Crude and Adjusted ORs of Prevalent Diabetes Mellitus By Subgroup Stratified by OSA Severity

Data expressed as OR (95% CI). Model 1 = unadjusted model. Model 2 = adjusted for demographic factors (race, age, sex), average sleep length, active smoker, alcohol use, Epworth sleepiness scale, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Model 3 = adjusted as per model 2 with BMI and neck circumference. PSG = polysomnography.

Table Graphic Jump Location
TABLE 5 ]  Crude and Adjusted Odds of Prevalent Diabetes Mellitus by OSA Severity in Overall Study Cohort (N = 11,702)

Data expressed as OR (95% CI). Model 1 = unadjusted model. Model 2 = adjusted for demographic factors (race, age, sex), average sleep length, active smoker, alcohol use, Epworth sleepiness scale, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Model 3 = adjusted as per model 2 with BMI and neck circumference.

Table Graphic Jump Location
TABLE 6 ]  Crude and Adjusted ORs of HbA1c ≥ 7% Among Diabetic Patients by OSA Severity

Data expressed as OR (95% CI). Model 1 = unadjusted model. Model 2 = adjusted for demographic factors (race, age, sex), average sleep length, active smoker, alcohol use, Epworth sleepiness scale, coronary artery disease, congestive heart failure, hyperlipidemia, COPD, psychiatric disease, and medication use. Model 3 = adjusted as per model 2 with BMI and neck circumference. See Table 2 legend for expansion of abbreviation.

References

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(17):1230-1235. [CrossRef] [PubMed]
 
McNicholas WT, Bonsigore MR; Management Committee of EU COST ACTION B26. Sleep apnoea as an independent risk factor for cardiovascular disease: current evidence, basic mechanisms and research priorities [published correction appears inEur Respir J.2007;29(3):614]. Eur Respir J. 2007;29(1):156-178. [CrossRef] [PubMed]
 
Marin JM, Carrizo SJ, Vicente E, Agusti AG. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet. 2005;365(9464):1046-1053. [CrossRef] [PubMed]
 
Punjabi NM, Caffo BS, Goodwin JL, et al. Sleep-disordered breathing and mortality: a prospective cohort study. PLoS Med. 2009;6(8):e1000132. [CrossRef] [PubMed]
 
Lam JC, Mak JC, Ip MS. Obesity, obstructive sleep apnoea and metabolic syndrome. Respirology. 2012;17(2):223-236. [CrossRef] [PubMed]
 
Reichmuth KJ, Austin D, Skatrud JB, Young T. Association of sleep apnea and type II diabetes: a population-based study. Am J Respir Crit Care Med. 2005;172(12):1590-1595. [CrossRef] [PubMed]
 
Ronksley PE, Hemmelgarn BR, Heitman SJ, et al. Obstructive sleep apnoea is associated with diabetes in sleepy subjects. Thorax. 2009;64(10):834-839. [CrossRef] [PubMed]
 
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(6):521-530. [CrossRef] [PubMed]
 
Botros N, Concato J, Mohsenin V, Selim B, Doctor K, Yaggi HK. Obstructive sleep apnea as a risk factor for type 2 diabetes. Am J Med. 2009;122(12):1122-1127. [CrossRef] [PubMed]
 
Priou P, Le Vaillant M, Meslier N, et al; IRSR Sleep Cohort Group. Independent association between obstructive sleep apnea severity and glycated hemoglobin in adults without diabetes. Diabetes Care. 2012;35(9):1902-1906. [CrossRef] [PubMed]
 
Bonsignore MR, McNicholas WT, Montserrat JM, Eckel J. Adipose tissue in obesity and obstructive sleep apnoea. Eur Respir J. 2012;39(3):746-767. [CrossRef] [PubMed]
 
Fitzgerald DB, Kent BD, Garvey JF, Russell A, Nolan G, McNicholas WT. Screening for diabetes mellitus in patients with OSAS: a case for glycosylated haemoglobin. Eur Respir J. 2012;40(1):273-274. [CrossRef] [PubMed]
 
American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2011;34(suppl 1):S62-S69. [CrossRef] [PubMed]
 
Tamura A, Kawano Y, Watanabe T, Kadota J. Obstructive sleep apnea increases hemoglobin A1c levels regardless of glucose tolerance status. Sleep Med. 2012;13(8):1050-1055. [CrossRef] [PubMed]
 
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(5):507-513. [CrossRef] [PubMed]
 
Pillai A, Warren G, Gunathilake W, Idris I. Effects of sleep apnea severity on glycemic control in patients with type 2 diabetes prior to continuous positive airway pressure treatment. Diabetes Technol Ther. 2011;13(9):945-949. [CrossRef] [PubMed]
 
Hedner J, Grote L, Bonsignore M, et al. The European Sleep Apnoea Database (ESADA): report from 22 European sleep laboratories. Eur Respir J. 2011;38(3):635-642. [CrossRef] [PubMed]
 
Standards of Practice Committee of the American Sleep Disorders Association. Practice parameters for the use of portable recording in the assessment of obstructive sleep apnea. Sleep. 1994;17(4):372-377. [PubMed]
 
Iber C, Ancoli-Israeli S, Chesson A, Quan S. The AASM Manual for the Scoring Of Sleep And Associated Events: Rules, Terminology And Technical Specifications. 1st ed. Westchester, IL: American Academy of Sleep Medicine, 2007.
 
Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14(6):540-545. [PubMed]
 
American Diabetes Association. Standards of medical care in diabetes—2013. Diabetes Care. 2013;36(suppl 1):S11-S66. [CrossRef] [PubMed]
 
Lam JC, Lam B, Yao TJ, et al. A randomised controlled trial of nasal continuous positive airway pressure on insulin sensitivity in obstructive sleep apnoea. Eur Respir J. 2010;35(1):138-145. [CrossRef] [PubMed]
 
Stratton IM, Adler AI, Neil HA, et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ. 2000;321(7258):405-412. [CrossRef] [PubMed]
 
Barceló A, Barbé F, de la Peña M, et al. Insulin resistance and daytime sleepiness in patients with sleep apnoea. Thorax. 2008;63(11):946-950. [CrossRef] [PubMed]
 
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(11):969-974. [CrossRef] [PubMed]
 
Kendzerska T, Gershon AS, Hawker G, Leung RS, Tomlinson G. Obstructive sleep apnea and risk of cardiovascular events and all-cause mortality: a decade-long historical cohort study. PLoS Med. 2014;11(2):e1001599. [CrossRef] [PubMed]
 
Johnson VE. Revised standards for statistical evidence. Proc Natl Acad Sci U S A. 2013;110(48):19313-19317. [CrossRef] [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).
Supporting Data

Online Supplement

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543