The Association Between Obstructive Sleep Apnea Defined by 3 Percent Oxygen Desaturation or Arousal Definition and Self-Reported Cardiovascular Disease in the Sleep Heart Health Study

Stuart F. Quan, M.D.^1,2

Rohit Budhiraja, M.D.¹

Sogol Javaheri, M.A., M.D., M.P.H.¹

Sairam Parthasarathy, M.D.²

Richard B. Berry, M.D.³

¹Division of Sleep and Circadian Disorders, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; ²Department of Medicine, University of Arizona College of Medicine, Tucson, AZ; ³Division of Pulmonary, Critical Care, and Sleep Medicine, University of Florida, Gainesville, FL

Editor's Note: Click here to see an accompanying editorial.

Abstract

Background: Studies have established that OSA defined using a hypopnea definition requiring a >4% oxygen desaturation (AHI4%) is associated with cardiovascular (CVD) or coronary heart (CHD) disease. This study determined whether OSA defined using a hypopnea definition characterized by a >3% oxygen desaturation or an arousal (AHI3%A) is associated with CVD/CHD.

Methods: Data were analyzed from 6307 Sleep Heart Health Study participants who had polysomnography. Self-reported CVD included angina, heart attack, heart failure, stroke or previous coronary bypass surgery or angioplasty. Self-reported CHD included the aforementioned conditions but not stroke or heart failure. The association between OSA and CVD/CHD was examined using logistic regression models with stepwise inclusion of demographic, anthropometric, social/behavioral and co-morbid medical conditions. A parsimonious model in which diabetes and hypertension were excluded because of their potential to be on the causal pathway between OSA and CVD/CHD also was constructed.

Results: For CVD, the odds ratios and 95% confidence intervals for AHI3%A >30/hour were 1.39 (1.03-1.87) and 1.45 (1.09-1.94) in the fully adjusted and parsimonious models. Results for CHD were 1.29 (0.96-1.74) and 1.36 (0.99-1.85). In participants without OSA according to more stringent AHI4% criteria but with OSA using the AHI3%A definition, similar findings were observed.

Conclusion: OSA defined using an AHI3%A is associated with both CVD and CHD. Use of a more restrictive AHI4% definition will misidentify a large number of individuals with OSA who have CVD or CHD. These individuals may be denied access to therapy, potentially worsening their underlying CVD or CHD. 

Introduction

Obstructive sleep apnea (OSA) is a common disorder characterized by recurrent episodes of either complete upper airway collapse (apneas) or partial collapse (hypopneas) during sleep. A number of large studies have established that OSA is a risk factor for the development of hypertension and cardiovascular disease (CVD) as well as higher mortality; individuals with more severe OSA are at greater risk (1-3). The most commonly used metric of OSA severity is the apnea hypopnea index (AHI). However, there is controversy regarding the definition of the AHI. In 2012, the American Academy of Sleep Medicine (AASM) recommended that the hypopnea definition include any decrease in airflow by at least 30% from the baseline with an oxyhemoglobin desaturation of at least 3%, or an arousal from sleep (4). However, several payors including the Centers for Medicare and Medicaid Services (CMS) continue to require a more stringent hypopnea definition necessitating a 4% or greater decrease in oxygen saturation (5) despite evidence documenting a relationship between the AASM recommended standard and daytime sleepiness (6). The resistance to universal acceptance of the AASM criteria is based in part on the lack of evidence that 3% desaturations or arousals have an adverse cardiovascular impact. This reluctance to adopt a more inclusive definition of sleep apnea has restricted access to OSA treatment for many patients (7). Therefore, determining if there is relationship between OSA characterized by at least 3% drop in saturation or an arousal from sleep and CVD may assist in identification of persons at risk for CVD, allow greater access to care and potentially improve other health-related outcomes.

Using the database from the Sleep Heart Health Study, a large well-characterized community based cohort that had undergone polysomnography, the current study aimed to determine the association between the AASM recommended definition of the AHI which incorporates hypopneas with at least a 3% desaturation or an arousal (AHI3%A) and self-reported CVD and coronary heart disease (CHD) in middle-aged and older adults. In addition, we sought to ascertain whether there was an association between CVD or CHD and OSA severity among individuals who were not identified as having OSA using the more restrictive standard of requiring at least a 4% oxygen desaturation irrespective of an arousal (AHI4%), but were classified as having OSA by the AHI3%A definition. We hypothesized that increasing OSA severity represented by the AHI3%A would be associated with a greater likelihood of having prevalent CVD or CHD, and that persons who were not identified as having OSA using the AHI4% criteria would have a higher likelihood as well.

Methods

This study analyzed data obtained from the Sleep Heart Health Study (SHHS) which was a prospective multicenter cohort study designed to investigate the relationship between OSA and cardiovascular diseases in the United States. The study’s rationale and design have been published elsewhere (8). Briefly, 6,441 subjects, 40 years of age and older were recruited starting in 1995 from several ongoing “parent” cardiovascular and respiratory disease cohorts that were initially assembled between 1976 and 1995 (9). These cohorts included the Offspring Cohort and the Omni Cohort of the Framingham Heart Study in Massachusetts; the Hagerstown, MD, and Minneapolis, MN, sites of the Atherosclerosis Risk in Communities Study; the Hagerstown, MD, Pittsburgh, PA, and Sacramento, CA, sites of the Cardiovascular Health Study; 3 hypertension cohorts (Clinic, Worksite, and Menopause) in New York City; the Tucson Epidemiologic Study of Airways Obstructive Diseases and the Health and Environment Study; and the Strong Heart Study of American Indians in Oklahoma, Arizona, North Dakota, and South Dakota. Because of sovereignty issues, 134 participants from the Arizona cohort of the Strong Heart Study withdrew consent. Analyses were performed on the remaining 6307 participants. The SHHS was approved by each site’s institutional review board for human subjects’ research, and informed written consent was obtained from all subjects at the time of their enrollment.

Polysomnography and Home Visit

Participants underwent overnight in-home polysomnograms using the Compumedics Portable PS-2 System (Abbottsville, Victoria, Australia) administered by trained technicians (10). The home visits were performed by two-person, mixed-sex teams in visits that lasted 1.5 to 2 hours. Participants were asked to schedule the visit so that it would occur approximately two hours prior to their usual bedtime. At the time of the home visit, an inventory of each participant’s medications was made. In addition, a health interview was completed that ascertained the presence of several health conditions. Questionnaires that were completed included the SHHS Sleep Habits Questionnaire which incorporated the Epworth Sleepiness Scale (ESS) (11) and the Medical Outcomes Study SF-36 (12). Blood pressure was measured manually in triplicate in a seated position after 5 minutes of rest (13). The average of the second and third measurements was used for this analysis. Body weight was obtained using a digital scale.

The SHHS recording montage consisted of electroencephalogram (C4/A1 and C3/A2), right and left electrooculogram, a bipolar submental electromyogram, thoracic and abdominal excursions (inductive plethysmography bands), airflow (detected by a nasal-oral thermocouple (Protec, Woodinville, WA), oximetry (finger pulse oximetry [Nonin, Minneapolis, MN]), electrocardiogram and heart rate (using a bipolar electrocardiogram lead), body position (using a mercury gauge sensor), and ambient light (on/off, by a light sensor secured to the recording garment). Sensors were placed, and equipment was calibrated during an evening home visit by a certified technician. After technicians retrieved the equipment, the data, stored in real time on PCMCIA cards, were downloaded to the computers of each respective clinical site, locally reviewed, and forwarded to a central reading center (Case Western Reserve University, Cleveland, OH). Comprehensive descriptions of polysomnography scoring and quality-assurance procedures have been previously published (14). In brief, sleep was scored according to guidelines developed by Rechtschaffen and Kales (15). Strict protocols were maintained to ensure comparability among centers and technicians. Intra-scorer and inter-scorer reliabilities were high (14).

The apnea hypopnea index (AHI) was calculated for each participant using two definitions of hypopnea, the AASM recommended definition [AHI3%A] and the AASM acceptable [CMS] definition [AHI4%]. For AHI3%A, hypopneas were identified if the amplitude of a measure of flow or volume (detected by the thermocouple or thorax or abdominal inductance band signals) was reduced discernibly (at least 30% lower than baseline breathing) for at least 10 seconds, did not meet the criteria for apnea and the event was associated with either a 3% oxygen desaturation from baseline or terminated with electroencephalographic evidence of an arousal. For AHI4%, hypopneas were identified if the aforementioned reduction in flow or volume occurred and the event was associated with a 4% oxygen desaturation from baseline. In both cases, an apnea was defined as a complete or almost complete cessation of airflow, as measured by the amplitude of the thermocouple signal, lasting at least 10 seconds.

Outcome Assessment

Self-reported CVD and CHD were the outcomes of interest for this analysis and were obtained from the standardized health interview performed at the time of each participant’s polysomnography home visit. Participants were asked if they had ever been told by a doctor that she or he had angina, heart attack, heart failure, or stroke and if the participant had ever undergone coronary bypass surgery or coronary angioplasty. Prevalent CVD was defined as a positive response to one or more of the aforementioned conditions or procedures. Prevalent CHD was defined as an affirmative response to the same questions with the exclusion of responses to the presence of heart failure or stroke.

Covariates

Selection of potential covariates was based on previous studies documenting an association with either CVD or CHD. These included various demographic (e.g., sex, race/ethnicity, education, marital status), anthropometric (e.g., height, weight and blood pressure [BP]), social/behavioral (e.g., smoking history, alcohol use, sleep duration, quality of life) indices as well as plasma lipids (cholesterol, high density lipoprotein [HDL], triglycerides), several diseases (depression, hypertension, diabetes) and spirometry.

The following definitions were used for those covariates that were not primarily recorded. Body mass index (BMI) was calculated as weight (kg)/height (m²). The ankle arm index (AAI) was computed as the ratio of blood pressure at the ankle to that in the arm. Waist to hip ratio was the waist divided by hip circumferences. Hypertension was defined as a self-report of hypertension or the use of anti-hypertensive medications. Diabetes was considered present if it was self-reported by the participant or if there was use of oral hypoglycemic agents or insulin. Depression was defined as present if the participant indicated on the SF-36 that he/she was feeling “blue” or “down” for at least “a good bit of the time” for the previous 4 weeks, or he/she was using antidepressant medications. Insomnia was defined as often or almost always having “trouble falling asleep”, “waking up during the middle of the night and having difficulty getting back to sleep” or “waking up too early in the morning and being unable to get back to sleep”. Sleepiness was assessed by the ESS as well as by self-report of being excessively sleepy during the day most or almost all of the time.

Statistical Analyses

Mean and standard deviation, and percentages were used to provide an overall description of the data used in the analyses. Unadjusted differences were compared using Student’s t test or c². For both definitions of the AHI, each participant’s AHI was assigned to one of 4 OSA severity categories: Normal (AHI <5 /hour), Mild (AHI ≥5 and <15 /hour), Moderate (AHI ≥15 and < 30/hour) and Severe (AHI ≥30/hour).

Missing data was present in 4.8% of observations and were felt to be missing at random. Inasmuch as using a “complete case analysis” would result in exclusion of a significant number of participants from our analyses with a consequent reduction in statistical power, multiple imputation using the multiple imputation by chained equation (MICE) package in R was employed to generate replacement values. Comparison of the imputed to the original dataset did not identify any outliers in the imputed dataset and means of the same variables between datasets were comparable.

To reduce the number of relevant predictors, overfitting of models, reduce potential collinearity and minimize prediction error, a Least Absolute Shrinkage and Selection Operator (lasso) regression was performed for both outcome variables using the glmnet package in R. This resulted in an analytic dataset for CVD that consisted of the following: age, sex, race/ethnicity, BMI, AAI, diastolic BP, smoking, SF-36 physical component summary (PCS), SF-36 general health rating (GenHlth), SF-36 ability to perform vigorous activity (VigActiv), hypertension, diabetes, depression and HDL. For CHD, the analytic dataset consisted of the following: age, sex, race/ethnicity, BMI, diastolic BP, smoking, PCS, GenHlth, VigActiv, hypertension, diabetes, triglycerides and HDL.

For the entire cohort, logistic regression using SPSS v27 (Armonk, NY) was used to generate increasingly complex models of the relationship between either CVD or CHD and severity of OSA adjusting for the covariates identified using the lasso regression. For CVD, after the unadjusted model, models were generated for the sequential addition of demographic factors (age, sex, race/ethnicity), anthropometric factors (BMI, AAI, diastolic BP), social/behavioral characteristics (smoking, PCS, GenHlth, VigActiv) and diseases/conditions (hypertension, diabetes, depression, HDL). For CHD after the unadjusted model, the corresponding sequential models were demographic factors (age, sex, race/ethnicity), anthropometric factors (BMI, diastolic BP), social/behavioral characteristics (smoking, PCS, GenHlth, VigActiv) and diseases/conditions (hypertension, diabetes, triglycerides, HDL). Because of the possibility that adjustment for a hypertension and a diabetes indicator would be “overadjustment” (i.e., adjustment for a variable on a causal pathway), we excluded both hypertension and diabetes from the final set of covariates in additional analyses and these are referred to as “parsimonious models.” Lastly, sensitivity analyses were performed in which the natural log of AHI3%A was used instead of categorial levels of that factor in the above models.

Associations between both CVD and CHD, and OSA severity were further analyzed in the subgroup of participants who were not classified as having OSA based on AHI4% criteria but were classified as OSA using AHI3%A criteria. The moderate and severe categories were combined because of the small number of cases in the severe OSA category. Otherwise, the modelling approaches employed were identical.

In Tables 2-5, odds ratios and 95% CI are presented versus the reference level of AHI <5 /hour. P values refer to the overall significance of the model with respect to OSA severity. Odds ratios, 95% CI and P values in Table 6 refer to AHI3%A expressed as the continuous factor lnAHI3%A+0.1 (0.1 added to mitigate 0 values of lnAHI3%A).

Results

Table 1 shows the univariate association of potential risk factors or characteristics with the presence of CVD or CHD.

Table 1. Univariate Association of Various Characteristics to Prevalent Cardiovascular (CVD) and Coronary Heart Disease (CHD)

N=6307 for all characteristics except AHI 3%/A (N=6131)

^ap≤0.05; ^bp≤0.01; ^cp≤0.001

There were 962 cases (15%) of CVD and 797 (13%) cases of CHD identified. Except for total cholesterol, all were either more prevalent or significantly higher or lower in participants with CVD or CHD. For both CVD and CHD, markedly higher prevalence rates were noted for sex (higher in men), hypertension, diabetes, depression, smoking (higher in ever smokers) and ability to engage in vigorous activity. Conversely, white race and good health status were much less common among those with CVD or CHD. Differences observed for the remaining characteristics were of lesser magnitude.

Figure 1 shows the prevalence rates of CVD or CHD as a function of OSA severity using the AHI3%A criteria. Both conditions were associated with increasing higher rates of disease as OSA became more severe.

Figure 1. Percentage of participants with either cardiovascular (CVD) or coronary heart (CHD) disease according to increasing severity of obstructive sleep apnea defined using a hypopnea definition characterized by a minimum 3% oxygen desaturation or an arousal (AHI3%A)

Table 2 shows the crude and adjusted odds ratios and their 95% confidence intervals for increasing complex models of the relationship between CVD and AHI3%A. The unadjusted model showed a strong, progressive association with increasingly severe OSA. However, as the models became increasingly complex, this relationship was attenuated and only approached statistical significance in the fully adjusted model (+Medical Conditions). Removal of hypertension and diabetes to create the Parsimonious model restored some of the association with a return of statistical significance.

Table 2. Adjusted Relative Odds (95% Confidence Interval) of Self-Reported Prevalent Cardiovascular Disease According to 3% or Arousal Apnea Hypopnea Index Severity Categories

^aDemographics model adds age, sex, race (White vs. American Indian)

^bAnthropometrics model adds BMI, Ankle Arm Index, diastolic blood pressure

^cSocial/Behavioral Factors model adds smoking, SF36 Physical Component Summary, SF36 General Health, SF36 Vigorous Activity

^dMedical Conditions model adds hypertension, diabetes, depression and HDL

^eParsimonious model includes factors in previous models, but removes hypertension and diabetes

^fN=6307

Presented in Table 3 are the models demonstrating the relationship between CHD and AHI3%A.

Table 3. Adjusted Relative Odds (95% Confidence Interval) of Self-Reported Prevalent Coronary Heart Disease According to 3% or Arousal Apnea Hypopnea Index Severity Categories^f

^aDemographics model adds age, sex, race (White vs. American Indian)

^bAnthropometrics model adds BMI, diastolic blood pressure

^cSocial/Behavioral Factors model adds smoking, SF36 Physical Component Summary, SF36 General Health, SF36 Vigorous Activity

^dMedical Conditions model adds hypertension, diabetes, triglycerides and HDL

^eParsimonious model includes factors in previous models, but removes hypertension and diabetes

^fN=6307

Similar to the findings for CVD, there was a progressively higher odds of having CHD as severity of OSA increased. The fully adjusted model (+Medical Conditions) was not significant, but the Parsimonious model approached statistical significance.

There were 3,326 participants who did not have OSA as defined by AHI4% criteria. Within this cohort, 2247 were classified as OSA using the AHI3%A criteria; 1966 (87.4%) were mild, 271 (12.0%) were moderate and 10 (0.4%) were severe. For this subgroup, Table 4 presents the increasingly complex models illustrating the relationship between the presence of CVD and increasing OSA severity.

Table 4. Adjusted Relative Odds (95% Confidence Interval) of Self-Reported Prevalent Cardiovascular Disease According to 3% or Arousal Apnea Hypopnea Index Severity Categories in Participants Without Obstructive Sleep Apnea According to 4% Desaturation Criteria^f

^aDemographics model adds age, sex, race (White vs. American Indian)

^bAnthropometrics model adds BMI, Ankle Arm Index, diastolic blood pressure

^cSocial/Behavioral Factors model adds smoking, SF36 Physical Component Summary, SF36 General Health, SF36 Vigorous Activity

^dMedical Conditions model adds hypertension, diabetes, depression and HDL

^eParsimonious model includes factors in previous models, but removes hypertension and diabetes

^fN=3326

Because of the relatively small number of cases with severe OSA, the moderate and severe cases were combined for these analyses. The unadjusted model showed a strong relationship with OSA severity; as model complexity increased, this finding was attenuated and only approached statistical significance in both the fully adjusted (+Medical Conditions) and Parsimonious models. As demonstrated in Table 5, similar findings were observed for CHD; the unadjusted model showed a strong association which was attenuated as the models became more complex; the fully adjusted (+medical conditions) and parsimonious models approached statistical significance.

Table 5. Adjusted Relative Odds (95% Confidence Interval) of Self-Reported Prevalent Coronary Heart Disease According to 3% or Arousal Apnea Hypopnea Index Severity Categories in Participants Without Obstructive Sleep Apnea According to 4% Desaturation Criterion^f

^aDemographics model adds age, sex, race (White vs. American Indian)

^bAnthropometrics model adds BMI, diastolic blood pressure

^cSocial/Behavioral Factors model adds smoking, SF36 Physical Component Summary, SF36 General Health, SF36 Vigorous Activity

^dMedical Conditions model adds hypertension, diabetes, triglycerides and HDL

^eParsimonious model includes factors in previous models, but removes hypertension and diabetes

^fN=3326

In sensitivity analyses, the natural log of AHI3%A was used as the index of OSA severity in lieu of a categorial representation. As shown in Table 6, in the entire cohort, for both CVD and CHD, a significant linear relationship with increasing severity of OSA was demonstrated in parsimonious models, but not the fully adjusted models. In the subgroup who did not have OSA as defined by AHI4% criteria but did have OSA using the AHI3%A criteria, linear relationships noted for both CVD and CHD in the fully adjusted and parsimonious models. For CHD in the fully adjusted model, the relationship was statistically significant and approached statistical significance in the others.

Table 6. Linear Adjusted Relative Odds (95% Confidence Interval) of Self-Reported Prevalent Cardiovascular and Coronary Heart Disease According to 3% or Arousal Apnea Hypopnea Index Severity

^aCohort restricted participants without OSA according to AHI4% criteria, N=3326

^bCovariates for CVD: age, sex, race, BMI, ankle-arm index, diastolic blood pressure

smoking, SF36 Physical Component Summary, SF36 General Health, SF36 Vigorous Activity

hypertension, diabetes, depression and HDL; Covariates for CHD: age, sex, race, BMI, diastolic blood pressure,smoking, SF36 Physical Component Summary, SF36 General Health, SF36 Vigorous Activity, hypertension, diabetes, triglycerides and HDL

^cExcludes diabetes and hypertension from fully adjusted model

Discussion

In this large community-based study, we demonstrated that OSA defined by apneas and hypopneas characterized by 3% desaturation events or arousals is associated with an increased likelihood of self-reported CVD and CHD after controlling for a number of relevant covariates. Importantly, in a subset of this cohort who did not have OSA as defined by apneas and hypopneas requiring a minimum 4% oxygen desaturation but did have OSA using the 3% desaturation or arousal criteria, we found that the association with both CVD and CHD remained, albeit weaker. Nevertheless, our analyses overall suggest that the regulatory requirement by the Centers for Medicare and Medicaid Services (CMS) in the United States of using a 4% desaturation definition to identify patients with OSA denies a substantial proportion of these individuals the opportunity to be treated for their OSA and thus reduce the risk of worsening or recurrence of their CVD or CHD.

Results from several large cohort studies including SHHS have found that OSA is associated with the presence of CVD and CHD, and that this association is stronger when the AHI as a metric of OSA severity increases (1, 2). These previous studies have used a definition of hypopnea that requires a minimum 4% oxygen desaturation (16-18). This definition has been adopted by CMS and other insurers to identify individuals as having OSA (5). However, the AASM recommends defining hypopneas with a minimum 3% desaturation or an arousal (4). This is based on evidence indicating that daytime sleepiness and other symptoms of OSA are associated with this less stringent definition of OSA (6). This distinction has important clinical implications because there are a large number of patients who do not meet the AHI4% criteria and but do meet the AHI3%A criteria (7, 19). In the former case, they are not considered to have OSA, but do have it in the latter.

To our knowledge, our study is the first to assess the association between OSA using the AHI3%A criteria and CVD and CHD. We found that as OSA severity increased, there was a greater likelihood of having CVD and CHD after adjusting for a number of relevant covariates. We acknowledge that in the fully adjusted model, this association only approached statistical significance. However, in our parsimonious model which removed the presence of hypertension and diabetes, likely mediators of this relationship, the association was strengthened. Sensitivity analyses using the natural log of AHI3%A validated the results we observed with categories of AHI severity. It has been well-established that hypertension and diabetes are independent risk factors for the development of CVD and CHD. However, a number of studies have demonstrated that OSA is a risk factor for the development of both hypertension and diabetes (1, 20). Therefore, both of the latter conditions lie on the causal path by which OSA may increase the risk for the development of CVD and CHD. Hence, we believe that inclusion of both these conditions in our fully adjusted model may be over-adjustment and that our parsimonious model best represents the association between OSA defined by AHI3%A and CVD or CHD.

We identified there was a large subset of our cohort that had OSA using the AHI3%A definition, but not the AHI4% definition of hypopnea. In this subset, we also observed an association between OSA severity and both CVD and CHD. This finding is analogous to the relationship we recently observed between OSA and the prevalence of hypertension (19). Similar to our findings with the full cohort, the fully adjusted model for both CVD and CHD was not statistically significant. However, it approached or became statistically significant in the parsimonious models. Although most of these cases were in the mild OSA category, 12.4% were moderate to severe where treatment is almost always recommended. Individuals with prevalent CVD or CHD and OSA are at risk for further complications of their CVD or CHD (21-23). However, if they do not meet the AHI4% definition of OSA, access to OSA treatment would be denied by CMS and some insurers.

Our findings with respect to CVD and CHD are consistent with recent analyses demonstrating that OSA defined by AHI3%A is associated with prevalent and incident hypertension in the SHHS cohort (19, 24). Similar findings also have been observed in other cohorts providing additional evidence that use of a hypopnea definition incorporating a minimum 3% oxygen desaturation or an arousal is important in the identification of individuals with OSA (25-27).

Although there is substantial evidence emerging that intermittent hypoxemia plays an important role in the cardiovascular consequences of OSA (28), the importance of arousals remains uncertain (29). Arousals involve an increase in the sympathetic activity and a decrease in the parasympathetic activity (28) and there is some evidence linking them in the development of hypertension (30). Data from our study would suggest that they may contribute to the development of CVD or CHD as well.

Some (31, 32), but not all (33) studies have suggested that the impact of OSA on the development of CVD or CHD is in part enhanced by the presence of sleepiness. However, in our initial assessment of potential covariates using a lasso regression, sleepiness did not emerge as a significant factor. Thus, our findings do not support the contention that sleepiness is an important factor impacting the relationship between OSA and CVD or CHD.

Our study does have a few limitations. Most important is that we identified prevalent CVD and CHD by self-report. While it is possible that some misclassification occurred, we do not think it was large. A large number of potential covariates were considered for inclusion in the models; we used a lasso regression to reduce the possibility of over-adjustment and collinearity. Furthermore, the possibility of residual confounding remains. Finally, this is a cross-sectional analysis, and causality cannot be assumed.

This study has several strengths. It uses a large, well characterized cohort with the availability of data from a number of potential covariates. Additionally, the cohort had a diverse racial/ethnic, age and sex distribution. Polysomnography was used to document the presence of OSA, and not more limited sleep apnea testing.

In summary, OSA as defined by apneas and hypopneas requiring a minimum 3% oxygen desaturation or arousal is associated with an increased likelihood of having CVD or CHD. Use of a more restrictive definition requiring a minimum 4% desaturation will misidentify a large number of individuals with OSA, and CVD or CHD. These individuals may be denied access to therapy which may prevent worsening of their underlying CVD or CHD. 

Acknowledgements

SHHS acknowledges the Atherosclerosis Risk in Communities Study, the Cardiovascular Health Study, the Framingham Heart Study, the Cornell/Mt. Sinai Worksite and Hypertension Studies, the Strong Heart Study, the Tucson Epidemiologic Study of Airways Obstructive Diseases (TESAOD), and the Tucson Health and Environment Study for allowing their cohort members to be part of the SHHS and for sharing such data for the purposes of this study. SHHS is particularly grateful to the members of these cohorts who agreed to participate in SHHS as well. SHHS further recognizes all the investigators and staff who have contributed to its success. A list of SHHS investigators, staff, and their participating institutions is available on the SHHS website (www.jhsph.edu/shhs).

The opinions expressed in this paper are those of the authors and do not necessarily reflect the views of the Indian Health Service.

This work was supported by National Heart, Lung and Blood Institute cooperative agreements U01HL53940 (University of Washington), U01HL53941 (Boston University), U01HL53938 (University of Arizona), U01HL53916 (University of California, Davis), U01HL53934 (University of Minnesota), U01HL53931 (New York University), U01HL53937 and U01HL64360 (Johns Hopkins University), U01HL63463 (Case Western Reserve University), U01HL63429 (Missouri Breaks Research).

SP was supported by Patient Centered Outcomes Research Institute (CER-2018C2-13262; PCS-1504-30430; DI-2018C2-13161; DI-2018C2-13161 COVID supplement, EADI-16493), NIH (HL126140, HL151254, AI135108, AG059202, HL158253) and American Academy of Sleep Medicine Foundation during the writing of this manuscript.

Authors’ Declarations: Dr. Budhiraja reports no conflicts of interest or grant funding. Dr. Quan reports research funding from the National Institutes of Health, serves as a consultant to Jazz Pharmaceuticals, Whispersom and is a committee chair and hypopnea taskforce member for the American Academy of Sleep Medicine. Dr. Javaheri serves as a consultant for Jazz Pharmaceuticals and Harmony Biosciences. Dr. Berry reports research funding from Philips Respironics, Res Med and the University of Florida Foundation. Dr. Parthasarathy reports grants from NIH/NHLBI as PI (HL138377, HL126140; IPA-014264-00001; HL095799) or site PI (HL128954; UG3HL140144), grants from Patient Centered Outcomes Research Institute as PI (IHS-1306-02505; EAIN-3394-UOA) or site-investigator (PCS-1504-30430), grants from US Department of Defense as co-investigator (W81XWH-14-1-0570), grants from NIH/NCI as co-investigator (R21CA184920) and NIH/NIMHD as co-investigator (MD011600), grants from Johrei Institute, personal fees from American Academy of Sleep Medicine, non-financial support from National Center for Sleep Disorders Research of the NIH (NHLBI), personal fees from UpToDate Inc., grants from Younes Sleep Technologies, Ltd., personal fees from Vapotherm, Inc., personal fees from Merck, Inc., grants from Philips-Respironics, Inc., personal fees from Philips-Respironics, Inc., personal fees from Bayer, Inc., personal fees from Nightbalance, Inc, personal fees from Merck, Inc, grants from American Academy of Sleep Medicine Foundation (169-SR-17); In addition, Dr. Parthasarathy has a patent UA 14-018 U.S.S.N. 61/884,654; PTAS 502570970 (Home breathing device) issued.

A preprint of this paper is available at: medRxiv, https://doi.org/10.1101/2020.09.22.20199745

Abbreviation List

AHI                             Apnea Hypopnea Index

AHI3%A                     Hypopneas with at least a 3% oxygen desaturation or an arousal

AHI4%                        Hypopneas with at least a 4% oxygen desaturation

AAI                              Ankle arm index

BP                               Blood pressure

BMI                             Body mass index

CHD                           Coronary heart disease

CVD                           Cardiovascular disease

CMS                           Centers for Medicare and Medicaid Services

ESS                            Epworth sleepiness scale

GenHlth                     General health rating subscale of SF36

Glmnet                       A statistical package used in R that fits a generalized linear model via penalized maximum likelihood

HDL                            High density lipoprotein

Lasso                         Least Absolute Shrinkage and Selection Operator

MICE                          Multiple imputation by chained equation

OSA                            Obstructive sleep apnea

PCS                            Physical component summary of the SF36

R                                 An open source programming language used for statistical computing and graphics

SHHS                         Sleep Heart Health Study

VigActiv                     Vigorous activity rating subscale of the SF36

References

1. Hou H, Zhao Y, Yu W, Dong H, Xue X, Ding J, Xing W, Wang W. Association of obstructive sleep apnea with hypertension: A systematic review and meta-analysis. J Glob Health. 2018 Jun;8(1):010405. [CrossRef] [PubMed]
2. Mashaqi S, Gozal D. The impact of obstructive sleep apnea and PAP therapy on all-cause and cardiovascular mortality based on age and gender - a literature review. Respir Investig. 2020 Jan;58(1):7-20. [CrossRef] [PubMed]
3. Berry RB, Budhiraja R, Gottlieb DJ, et al. Rules for scoring respiratory events in sleep: update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events. Deliberations of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. J Clin Sleep Med. 2012;8(5):597-619. [CrossRef] [PubMed]
4. Anonymous. CPAP for Obstructive Sleep Apnea. Last Updated: 2020. Accessed: 09/14, 2020. https://www.cms.gov/Medicare/Coverage/Coverage-with-Evidence-Development/CPAP.
5. Shamin-Uzzaman QA, Singh S, Chowdhuri S. Hypopnea definitions, determinants, and dilemmas: a focused review. Sci Sleep Pract. 2018;2(1):7. [CrossRef]
6. Raj R, Hirshkowitz M. Effect of the new Medicare guideline on patient qualification for positive airway pressure therapy. Sleep Med. 2003 Jan;4(1):29-33.  
7. [CrossRef] [PubMed]
8. Quan SF, Howard BV, Iber C, Kiley JP, Nieto FJ, O'Connor GT, Rapoport DM, Redline S, Robbins J, Samet JM, Wahl PW. The Sleep Heart Health Study: design, rationale, and methods. Sleep. 1997 Dec;20(12):1077-85. [PubMed]
9. Lind BK, Goodwin JL, Hill JG, Ali T, Redline S, Quan SF. Recruitment of healthy adults into a study of overnight sleep monitoring in the home: experience of the Sleep Heart Health Study. Sleep Breath. 2003;7(1):13-24. [CrossRef] [PubMed]
10. Iber C, Redline S, Kaplan Gilpin AM, et al. Polysomnography performed in the unattended home versus the attended laboratory setting--Sleep Heart Health Study methodology. Sleep. 2004;27(3):536-540. HTTPS://PUBMED.NCBI.NLM.NIH.GOV/ 15164911 [CrossRef] [PubMed]
11. Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14(6):540-545. [CrossRef] [PubMed]
12. Ware JE,Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care. 1992;30(6):473-483. [CrossRef] [PubMed]
13. O'Connor GT, Caffo B, Newman AB, et al. Prospective study of sleep-disordered breathing and hypertension: the Sleep Heart Health Study. Am J Respir Crit Care Med. 2009;179(12):1159-1164. [CrossRef] [PubMed]
14. Whitney CW, Gottlieb DJ, Redline S, et al. Reliability of scoring respiratory disturbance indices and sleep staging. Sleep. 1998;21(7):749-757. [CrossRef] [PubMed]
15. Rechtschaffen A, Kales A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subject. Washington, D.C.: US Government Printing Office, National Institute of Health Publication, 1968.
16. 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 Mar 19-25;365(9464):1046-53. [CrossRef] [PubMed]
17. Shahar E, Whitney CW, Redline S, et al. Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med. 2001;163(1):19-25. [CrossRef] [PubMed]
18. Young T, Finn L, Peppard PE, Szklo-Coxe M, Austin D, Nieto FJ, Stubbs R, Hla KM. Sleep disordered breathing and mortality: eighteen-year follow-up of the Wisconsin sleep cohort. Sleep. 2008 Aug;31(8):1071-8. [PubMed]
19. Budhiraja R, Javaheri S, Parthasarathy S, Berry RB, Quan SF. The Association Between Obstructive Sleep Apnea Characterized by a Minimum 3 Percent Oxygen Desaturation or Arousal Hypopnea Definition and Hypertension. J Clin Sleep Med. 2019 Sep 15;15(9):1261-1270 [CrossRef] [PubMed]
20. Qie R, Zhang D, Liu L, Ren Y, Zhao Y, Liu D, Liu F, Chen X, Cheng C, Guo C, Li Q, Zhou Q, Tian G, Han M, Huang S, Wu X, Qin P, Li J, Cao J, Zhang M, Huang J, Lu J, Li H, Wang J, Cheng R, Hu D. Obstructive sleep apnea and risk of type 2 diabetes mellitus: A systematic review and dose-response meta-analysis of cohort studies. J Diabetes. 2020 Jun;12(6):455-464. doi: 10.1111/1753-0407.13017. Epub 2020 Jan 22. [CrossRef] [PubMed]
21. Yudi MB, Clark DJ, Farouque O, Andrianopoulos N, Ajani AE, Brennan A, Lefkovits J, Freeman M, Hiew C, Selkrig LA, O'Brien J, Dart AM, Reid CM, Duffy SJ; Melbourne Interventional Group. Trends and predictors of recurrent acute coronary syndrome hospitalizations and unplanned revascularization after index acute myocardial infarction treated with percutaneous coronary intervention. Am Heart J. 2019 Jun;212:134-143. [CrossRef] [PubMed]
22. Lee CH, Khoo SM, Chan MY, Wong HB, Low AF, Phua QH, Richards AM, Tan HC, Yeo TC. Severe obstructive sleep apnea and outcomes following myocardial infarction. J Clin Sleep Med. 2011 Dec 15;7(6):616-21. [CrossRef] [PubMed]
23. Maia FC, Goulart AC, Drager LF, Staniak HL, Santos IS, Lotufo PA, Bensenor IM. Impact of High Risk for Obstructive Sleep Apnea on Survival after Acute Coronary Syndrome: Insights from the ERICO Registry. Arq Bras Cardiol. 2017 Jan;108(1):31-37. [CrossRef] [PubMed]
24. Budhiraja R, Javaheri S, Parthasarathy S, Berry RB, Quan SF. Incidence of hypertension in obstructive sleep apnea using hypopneas defined by a 3 percent oxygen desaturation or arousal but not by only 4 percent oxygen desaturation. J Clin Sleep Med. 2020 Jul 9. Epub ahead of print. PMID: 32643602.[CrossRef] [PubMed]
25. Bouloukaki I, Grote L, McNicholas WT, Hedner J, Verbraecken J, Parati G, Lombardi C, Basoglu OK, Pataka A, Marrone O, Steiropoulos P, Bonsignore MR, Schiza SE; European Sleep Apnoea Database Network. Mild obstructive sleep apnea increases hypertension risk, challenging traditional severity classification. J Clin Sleep Med. 2020 Jun 15;16(6):889-898. [CrossRef] [PubMed]
26. Hirotsu C, Haba-Rubio J, Andries D, Tobback N, Marques-Vidal P, Vollenweider P, Waeber G, Heinzer R. Effect of Three Hypopnea Scoring Criteria on OSA Prevalence and Associated Comorbidities in the General Population. J Clin Sleep Med. 2019 Feb 15;15(2):183-194. [CrossRef] [PubMed]
27. Heinzer R, Vat S, Marques-Vidal P, Marti-Soler H, Andries D, Tobback N, Mooser V, Preisig M, Malhotra A, Waeber G, Vollenweider P, Tafti M, Haba-Rubio J. Prevalence of sleep-disordered breathing in the general population: the HypnoLaus study. Lancet Respir Med. 2015 Apr;3(4):310-8. [CrossRef] [PubMed]
28. Sforza E, Roche F. Chronic intermittent hypoxia and obstructive sleep apnea: an experimental and clinical approach. Hypoxia (Auckl). 2016;499-108. [CrossRef] [PubMed]
29. Ryan S. Mechanisms of cardiovascular disease in obstructive sleep apnoea. J Thorac Dis. 2018;10(Suppl 34):S4201-S4211. [CrossRef] [PubMed]
30. Sulit L, Storfer-Isser A, Kirchner HL, Redline S. Differences in polysomnography predictors for hypertension and impaired glucose tolerance. Sleep. 2006;29(6):777-783. [CrossRef] [PubMed]
31. Kapur VK, Resnick HE, Gottlieb DJ, for the Sleep Heart Health Study Group. Sleep Disordered Breathing and Hypertension: Does Self-Reported Sleepiness Modify the Association? Sleep. 2008;31(8):1127-1132. [CrossRef] [PubMed]
32. Mazzotti DR, Keenan BT, Lim DC, Gottlieb DJ, Kim J, Pack AI. Symptom Subtypes of Obstructive Sleep Apnea Predict Incidence of Cardiovascular Outcomes. Am J Respir Crit Care Med. 2019;200(4):493-506. [CrossRef] [PubMed]
33. Ogilvie RP, Lakshminarayan K, Iber C, Patel SR, Lutsey PL. Joint effects of OSA and self-reported sleepiness on incident CHD and stroke. Sleep Med. 2018 Apr;44:32-37. [CrossRef] [PubMed]

Cite as: Quan SF, Budhiraja R, Javaheri S, Parthasarathy S, Berry RB. The Association Between Obstructive Sleep Apnea Defined by 3 Percent Oxygen Desaturation or Arousal Definition and Self-Reported Cardiovascular Disease in the Sleep Heart Health Study. Southwest J Pulm Crit Care. 2020;21(4):86-103. doi: https://doi.org/10.13175/swjpcc054-20 PDF 

Obstructive Sleep Apnea and Cardiovascular Disease: Back and Forward in Time Over the Last 25 Years

Stuart F. Quan, M.D.

Division of Sleep Medicine, Harvard Medical School and Brigham Womens Hospital, Boston, MA; Arizona Respiratory Center, University of Arizona College of Medicine, Tucson, AZ

Abstract

Over the past 25 years, there have been significant advances made in understanding the pathophysiology and cardiovascular consequences of obstructive sleep apnea (OSA). Substantial evidence now implicates OSA as an independent risk factor for the development of hypertension, coronary artery disease, congestive heart failure and stroke, as well as increased risk of death. Pathophysiologic mechanisms include release of inflammatory mediators, oxidative stress, metabolic dysfunction, hypercoagulability and endothelial dysfunction. Although non-randomized intervention studies suggest that treatment of OSA with continuous positive airway pressure may mitigate its impact of the development of cardiovascular disease, randomized clinical trials are lacking.

Introduction

The first report of the physiologic events occurring in obstructive sleep apnea (OSA) was published in 1965 by Gastaut and colleagues (1). However, literary and historical accounts of what most likely was OSA have existed since antiquity (2). Most well-known of these is the description of Burwell’s case of a man falling asleep while playing poker. This led to the widespread use of the “Pickwickian Syndrome” because of the similarity of the patient’s symptoms to the literary imagery of “Joe, the fat boy”, a character in Charles Dicken’s Posthumous Papers of the Pickwick Club (3). In the 50 years since Gastaut’s original description of OSA, there has been an exponential growth in the recognition that it is a clinical condition with a high prevalence, subtle to disabling clinical symptoms, and more recently, substantial cardiovascular morbidity and mortality. However, much of the progress in our understanding of OSA with respect to cardiovascular disease has occurred in the past 25 years. Thus, it is appropriate to look back to where we were 25 years ago, our current state of knowledge and identify avenues for future research. Before doing so, however, one must examine whether there is sufficient evidence suggest that a biologic association between OSA and cardiovascular disease is plausible.           

Obstructive Sleep Apnea and Cardiovascular Disease: Biologic Plausibility

Obstructive sleep apnea is characterized by repetitive episodes of occlusion or near occlusion of the upper airway at the level of the pharynx despite increasing inspiratory efforts (4). These episodes are terminated by brief arousals from sleep resulting in sleep fragmentation. There are a number of physiologic consequences to what essentially are repetitive involuntary Müeller maneuvers (5,6). With cessation or near cessation of airflow, transient oxygen desaturation and hypercarbia occur (4,6). Inspiratory efforts against an occluded airway result in large intrathoracic pressure swings (5,6). As a result, there is increased sympathetic nervous system activity (7), fluctuations in parasympathetic tone (8), large cyclical changes in heart rate, arterial and pulmonary vasoconstriction and hypertension, and increases in cardiac preload and afterload (9-11). Given the chronicity of OSA, it is not difficulty project that such physiologic changes might lead to daytime cardiovascular dysfunction. Indeed, daytime hypertension has been induced in a canine model of OSA (12) as well as in rodent models of intermittent hypoxia (13). Thus, there are potential biologic mechanisms to explain why OSA might be an independent risk factor for cardiovascular disease.

Obstructive Sleep Apnea and Cardiovascular Disease: Circa 1970s-1980s

In the 1980’s, studies in Europe found cross-sectional associations between snoring as a surrogate for OSA and both hypertension and cardiovascular disease (14,15). Slightly earlier, evidence also was emerging of an association between obstructive sleep apnea and hypertension. These studies noted a greater than 50% prevalence of hypertension among patients with OSA (16,17) and conversely, a 20-30% prevalence of OSA was observed in patients with hypertension (18,19). Furthermore, several case control and cohort studies indicated that the risk of stroke or heart disease was 2.1 to 10.3 fold greater for those with snoring (20-23). Cardiac arrhythmias were frequently noted to occur in OSA patients as well (24).

Later during this time period, there were 3 retrospective studies analyzing the relationship between OSA and mortality. From the Henry Ford Sleep Disorders Center, the vital status of 385 male patients with OSA studied with polysomnography between 1978 and 1986 was determined. Mortality was significantly greater in those with an apnea index (not apnea hypopnea index [AHI]) greater than 20 events/hour. Uvulopalatopharyngoplasty did not attenuate the mortality rate, but better survival was reported in those who had a tracheostomy or were prescribed nasal continuous positive airway pressure (CPAP) (25). At the same time, the Stanford Sleep Disorders Center analyzed their experience in 198 patients who had either a tracheostomy (n=71) or were treated conservatively with a recommendation for weight loss (n=127).  There were 14 deaths of which 8 were from stroke or myocardial infarction over a 5 year follow-up period.  All occurred in the conservatively treated group (26). In contrast, another series from the University of Florida failed to find any mortality differences between 91 treated and untreated OSA patients and 35 patients with symptoms consistent with OSA, but negative findings on polysomnography over a 7-98 month follow-up (27).

Given the known acute effects of repetitive obstruction of the upper airway on systemic and pulmonary blood pressure, and heart rate, as well as the recurrent episodes of hypoxemia (11), linkages between these physiologic findings and the development of cardiovascular disease were proposed (28). One such pathophysiologic pathway highlighted important roles for hypoxemia and increased sympathetic activity, both of which are currently considered important mechanistic factors for the development of cardiovascular disease (28).

Obstructive Sleep Apnea And Cardiovascular Disease 2012: Epidemiology

Starting in the mid 1990’s, important observations were made that solidified linkages between OSA and cardiovascular disease. In the Wisconsin Sleep Cohort, the first large prospective population-based cohort study to use polysomnography to confirm the presence of OSA, Peppard et al. (29) demonstrated that OSA was an independent risk factor for the development of hypertension. Furthermore, the risk progressively increased with greater levels of OSA severity (OR: 1.42, 2.03, 2.89 [AHI: <5, 5-<15, >15 /hour vs. referent=0]) (29). These findings were subsequently confirmed in the Sleep Heart Health Study (OR: 1.13, 1.54, 2.119 [AHI: 5-<15, 15-<30, >30 /hour vs. referent=<5], although they were significantly attenuated when body mass index (BMI) was included in the analytic models (30). In addition, the development of hypertension was found to be associated with the presence of nocturnal hypoxemia. A more recent prospective study from a clinical cohort showed similar findings (31). Moreover, a number of studies have shown that blood pressure will decrease after treatment of OSA with continuous positive airway pressure (32). Although the magnitude of improvement in large clinical trials is only 2-3 mm Hg, such changes are large from an epidemiologic and public health perspective, and may be greater in individual patients and those with resistant hypertension (33). Given this evidence, the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure has concluded that OSA is an identifiable cause for the development of hypertension (34).

Evidence now indicates that OSA is an independent risk factor for the development of coronary heart disease (CHD). In a Spanish clinical cohort of only men, there was an increased risk of incident CHD over a 12 year follow-up period (35). Subsequently, this observation was confirmed in the Sleep Heart Health Study (36). In this latter study, however, increased risk (Hazard Ratio: 1.75 for AHI > 30 /hour vs. AHI < 5) was only observed in men less than 70 years of age. It is possible that the absence of an impact of OSA on risk of CHD in older men is related to a “healthy survivor” effect (37). Thus, whether OSA increases the likelihood of developing CHD in older men and women remains unclear.

In addition to findings linking OSA to greater risk of incident CHD, OSA appears to enhance the likelihood of new events in those with prevalent CHD. In a study of 407 consecutive patients with CHD, those with an oxygen desaturation index greater than 5 per hour had a 70% relative increase and a 10.7% absolute increase in the composite endpoint (cerebrovascular event, myocardial infarction or death) (38). In another study, 89 patients who had undergone percutaneous coronary intervention had polysomnography with OSA found in 51. In comparison to those who did not have OSA, 23.5% vs. 5.3% had a major adverse coronary event (cardiac death, reinfarction, and target vessel revascularization) in the ensuing follow-up period (mean=227 days) (39).

Sleep apnea, in particular central sleep apnea, frequently is observed in association with congestive heart failure (CHF). However, recent data from the Sleep Heart Health Study suggest that severe OSA is a risk factor for incident CHF in men (Hazard Ratio: 1.71 for AHI > 30 /hour vs. AHI < 5), but not women (36).

Recent observations indicate that stroke incidence also may be increased in those with OSA. In the Wisconsin Sleep Cohort over a 4 year follow-up, a significantly increased odds ratio for incident stroke after age and sex adjustment of 4.48 was observed in those with an AHI > 20 /hour (40). This was attenuated to 3.08 after controlling for body mass index. In the Sleep Heart Health Study, a significant increase in incident stroke risk also was found in those with an AHI > 20 /hour, but only in men (41).

As initially reported over 25 years ago (24), more recent observations have confirmed an association between OSA and cardiac arrhythmias. In the Sleep Heart Health Study, those with severe OSA were more likely to have both ventricular and atrial ectopy (42). Furthermore, those with severe OSA were 4.5 and 1.8 times more likely to have episodes of atrial fibrillation, and complex ventricular ectopy or non-sustained ventricular tachycardia.42 Additional analyses reveal that the relative risk of having an episodes of atrial fibrillation is 17 times greater after an apnea or hypopnea episode and there is 1 excess episode of paroxysmal atrial fibrillation or nonsustained ventricular tachycardia for every 1000 hours of sleep or 40000 respiratory disturbances (43). Obstructive sleep apnea with attendant hypoxemia also may be a risk factor for recurrence of atrial fibrillation after cardioversion (44).

Finally, there is now relatively conclusive evidence from 3 longitudinal cohort studies demonstrating that OSA contributes to excess mortality. In the Busselton Health Study of 380 individuals followed for a mean of 13.4 years, a respiratory disturbance index of greater than 15 /hour yielded a hazard ratio of 6.24 for excess mortality (45). Subsequently, in an 18 year follow-up of 1496 participants in the Wisconsin Sleep Cohort, the adjusted hazard ratio for excess all cause mortality related to severe OSA was 3.0 in comparison to no OSA (46). Furthermore, the hazard ratio related to cardiovascular disease mortality was 5.2 (46). More recently, the Sleep Heart Health Study reported a hazard ratio of 1.46 for all cause mortality over an 8.2 year average follow-up. Similar to the Wisconsin Sleep Cohort, it appeared that cardiovascular deaths accounted for much of the excess risk (47). In this latter study, indices of nocturnal hypoxemia also were associated with excess all cause mortality. Although sudden cardiac death in the general population usually occurs during the 6 am to 12 noon time frame, in those with OSA, it is shifted to night-time hours, 12 midnight to 6 am providing further evidence of the adverse impact of OSA on the heart (48).

Obstructive Sleep Apnea and Cardiovascular Disease: Mechanistic Observations

Since the initial observations of the physiologic events that might be operative in the pathogenesis of cardiovascular sequelae of OSA over 25 years ago, substantial progress has been made towards understanding how OSA is a risk factor for cardiovascular disease. These findings include OSA induced changes in cardiac structure and function, abnormalities in metabolic function, and increases in inflammation, coagulability and sympathetic nervous system activity. These latter issues then interact to enhance atherogenesis and cardiac dysfunction.

Obstructive sleep apnea is associated with increases in left ventricular mass. In the Sleep Heart Health Study, those with an AHI > 30 /hour in comparison to those with an AHI < 5 were more likely to have left ventricular hypertrophy on echocardiography, and estimates of left ventricular mass were higher (49). These associations were even stronger when indices of nocturnal hypoxemia were used instead of the AHI. This further highlights the potential role of hypoxemia in the pathogenesis of cardiovascular disease attributable to OSA. Given the presence of left ventricular hypertrophy related to OSA, it is not surprising that diastolic dysfunction is more common among individuals with OSA (50). However, use of CPAP may improve left ventricular function (50). This raises the possibility that early intervention to treat OSA may reduce cardiac morbidity and mortality.    

A number of studies have demonstrated that OSA is associated with metabolic abnormalities. For example, in the Sleep Heart Health Study, levels of cholesterol and triglycerides increased as a function of increasing OSA severity (51). These findings may be related to OSA induced intermittent hypoxia (52). Furthermore, the prevalence of metabolic syndrome is higher among persons with OSA in comparison to those without OSA (53). This finding appears to be driven primarily by the higher frequency of hypertension in persons with OSA. Whether these findings are causally related to OSA remains to be determined. However, OSA might potentially increase the risk of CHD by promoting dyslipidemia.

The prevalence of OSA among persons with type 2 diabetes mellitus is high with one study observing that 86% of obese type 2 diabetics had an AHI >5 /hour, indicative of mild OSA (54). Therefore, it is not surprising that considerable evidence now implicates OSA as a determinant of glucose regulation. Both cross-sectional and prospective studies have demonstrated that OSA is a risk factor for glucose dysregulation and in some studies incident diabetes mellitus (55). Furthermore, some studies have demonstrated that treatment of OSA with CPAP results in improved glucose control (55). Data suggest that OSA induced hypoxemia may be a causative mechanism (56). The close association between OSA and type 2 diabetes mellitus raises the distinct possibility that a positive interaction exists to increase the risk of CHD in persons with both conditions.

It is generally accepted that obesity is a risk factor for the development of OSA (57). However, a few studies have reported that reverse causality may be present such that OSA promotes weight gain (58,59). Recent data from the Sleep Heart Health Study support this hypothesis. Over an approximate 5 year follow-up, weight gain was greater among those with an AHI > 15 /hour in comparison to those with an AHI < 5 /hour (60). Thus, promotion of weight gain may be another mechanism by which OSA increases cardiovascular risk.

Especially in those with severe OSA, recurring apneas and hypopneas result in repetitive episodes of hypoxia and reoxygenation. This produces oxidative stress leading to an increase in the flux of free radicals, induction of endothelin expression, suppression of nitric oxide generation, local vasoconstriction and changes in vascular permeability (61). All of these effects have the potential of enhancing the development of cardiovascular disease.

Substantial data now is available demonstrating that OSA is associated with release of a number of inflammatory mediators such as IL6, sIL6R, IL-8, TNFα, CRP and NF-Kappa β (62). In addition, there is evidence for elevated levels of pro-thrombotic factors such as PAI-1, P-selectin, fibrinogen and VEGF in persons with OSA (62). With the recent findings of the importance of inflammation and thrombosis in the pathogenesis of cardiovascular disease, these observations may be important causal mechanistic links that lead OSA to cardiovascular disease.

Using observations from 25 years ago in combination with current data, a plausible pathogenic pathway from OSA to cardiovascular can be summarized as follows (Figure 1). Recurrent episodes of apnea and hypopnea lead to intermittent hypoxia, increased sympathetic activity, hypercapnia, sleep fragmentation with arousals and large swings in intrathoracic pressure. These physiologic perturbations result in increased oxidative stress, release of inflammatory mediators, metabolic dysfunction, weight gain, hypercoagulability, glucose dysregulation and endothelial dysfunction All of these mechanisms can lead to the development of hypertension, diabetes mellitus, CHD, CHF, stroke and increased risk of death.

Figure 1. Proposed mechanisms leading from physiologic alterations occurring during obstructive sleep apnea and hypopnea to the development of cardiovascular disease.

Obstructive Sleep Apnea and Cardiovascular Disease: Knowledge Gaps

Although significant advances have been made in our understanding of the relationship between OSA and CVD in the past 25 years, there are a number of important areas which require further investigation. With respect to OSA and hypertension, it remains unclear whether treatment of OSA reduces the risk of developing hypertension. Most studies to date have used non-randomized cohorts. In the most recently published randomized controlled trial, CPAP treatment did not decrease the incidence of hypertension in nonsleepy subjects over a median 4 year follow-up (63). However, post-hoc analyses did suggest an effect in subjects who were compliant with CPAP for more than 4 hours per night (63). Furthermore, if CPAP or other treatments for OSA are beneficial in reducing CVD risk, are there subsets of the population for whom it is more advantageous?

As to the impact of OSA on CVD and stroke, there also have not been any published large scale clinical trials demonstrating an impact of treatment on changing the incidence of disease. Similar to hypertension, published studies are from non-randomized cohorts. However, there are several large clinical trials such as Randomized Intervention with CPAP in Coronary Artery Disease and Sleep Apnoea (RICCADSA), Sleep Apnea Cardiovascular Endpoints Study (SAVE), and Heart Biomarker Evaluation in Apnea Treatment study (HeartBEAT). In the RICCADSA study, 400 CAD participants will be randomized to one of 4 groups: 1) non-sleepy with OSA treated with CPAP, 2) non-sleepy with OSA and no CPAP treatment, 3) sleepy with OSA treated with CPAP, 4) CAD but no OSA. The participants will be followed for 3 years for CVD morbidity and mortality (64).  In the multinational SAVE trial, participants with OSA at high risk for CVD will be randomized to CPAP or conventional medical therapy, and followed for 3-5 years (65). Because of the length of follow-up required and expense, and some would argue the ethical dilemmas in performing a long-term interventional trial, studies such as the recently completed HeartBEAT have attempted to assess intermediate outcomes. In the HeartBEAT trial, 270 subjects with CHD or at high risk for CHD were randomized to healthy lifestyle instruction, CPAP or nocturnal oxygen with a primary endpoint of 24 hour blood pressure (66). Whether findings from trials using intermediate endpoints will be predictive of an impact on “hard” endpoints such as incident myocardial infarction or stroke remains to be determined.

Conclusions

Substantial progress has been made in the past 25-30 years in our understanding of the relationship between OSA and CVD. Accumulating evidence implicates OSA as an independent risk factor for hypertension, CHD and stroke. However, the risk may not be the same for all segments of the population. A variety of mechanisms may be operative. Non-randomized trials suggest that treatment appears to mitigate the risk in some clinical populations, but it is unclear whether treatment is beneficial in patients without symptoms. Large scale randomized clinical trials are needed to clearly demonstrate that current treatment modalities for OSA can mitigate CVD risk and to delineate which populations will accrue the most benefit.

References

1. Gastaut H, Tassinari CA, Duron B. Polygraphic study of diurnal and nocturnal (hypnic and respiratory) episodal manifestations of Pickwick syndrome. Rev Neurol (Paris) 1965; 112(6):568-579.
2. Kryger MH. Sleep apnea. From the needles of Dionysius to continuous positive airway pressure. Arch Intern Med 1983; 143(12):2301-2303.
3. Bickelmann AG, Burwell CS, Robin ED, Whaley RD. Extreme obesity associated with alveolar hypoventilation; a Pickwickian syndrome. Am J Med 1956; 21(5):811-818.
4. White DP. Sleep apnea. Proc Am Thorac Soc 2006; 3(1):124-128.
5. Podszus T, Greenberg H, Scharf SM. Influence of Sleep State and Sleep-Disordered Brathing on Cardiovascular Function. In: Sleep and Breathing. Saunders NA and Sullivan CE, eds. New York, NY: Marcel Dekker, 1994; 257-310
6. Schaub CD, Schneider H, O'Donnell CP. Mechanisms of acute and chronic blood pressure elevationin animal models of obstructive sleep apnea. In: Sleep apnea: implications in cardiovascular and cerebrovascular disease. Bradley TD and Floras JS, eds. New York: Dekker, 2000; 554
7. Somers VK, Dyken ME, Clary MP, Abboud FM. Sympathetic neural mechanisms in obstructive sleep apnea. J Clin Invest 1995; 96(4):1897-1904.
8. Leung RS. Sleep-disordered breathing: autonomic mechanisms and arrhythmias. Prog Cardiovasc Dis 2009; 51(4):324-338.
9. Parker JD, Brooks D, Kozar LF, et al. Acute and chronic effects of airway obstruction on canine left ventricular performance. Am J Respir Crit Care Med 1999; 160(6):1888-1896.
10. Lorenzo-Filho G, Bradley TD. Cardiac function in sleep apnea. In: Sleep Apnea--Pathogeneis, Diagnosis, and Treatment. Pack AI, ed. New York, NY: Marcel Dekker, 2002; 377-410
11. Schroeder JS, Motta J, Guilleminault C. Hemodynamic studies in sleep apnea. In: Sleep Apnea Syndromes. Guilleminault C and Dement WC, eds. New York: Alan R. Liss, 1978; 177-196
12. Brooks D, Horner RL, Kozar LF, Render-Teixeira CL, Phillipson EA. Obstructive sleep apnea as a cause of systemic hypertension. Evidence from a canine model. J Clin Invest 1997; 99(1):106-109.
13. Bosc LV, Resta T, Walker B, Kanagy NL. Mechanisms of intermittent hypoxia induced hypertension. J Cell Mol Med 2010; 14(1-2):3-17.
14. Lugaresi E, Cirignotta F, Coccagna G, Piana C. Some epidemiological data on snoring and cardiocirculatory disturbances. Sleep 1980; 3(3-4):221-224.
15. Koskenvuo M, Kaprio J, Partinen M, Langinvainio H, Sarna S, Heikkila K. Snoring as a risk factor for hypertension and angina pectoris. Lancet 1985; 1(8434):893-896.
16. Guilleminault C, Tilkian A, Dement WC. The sleep apnea syndromes. Annu Rev Med 1976; 27:465-484.
17. Kales A, Cadieux RJ, Bixler EO, et al. Severe obstructive sleep apnea--I: Onset, clinical course, and characteristics. J Chronic Dis 1985; 38(5):419-425.
18. Lavie P, Ben-Yosef R, Rubin AE. Prevalence of sleep apnea syndrome among patients with essential hypertension. Am Heart J 1984; 108(2):373-376.
19. Fletcher EC, DeBehnke RD, Lovoi MS, Gorin AB. Undiagnosed sleep apnea in patients with essential hypertension. Ann Intern Med 1985; 103(2):190-195.
20. Spriggs DA, French JM, Murdy JM, Bates D, James OF. Historical risk factors for stroke: a case control study. Age Ageing 1990; 19(5):280-287.
21. Koskenvuo M, Kaprio J, Telakivi T, Partinen M, Heikkila K, Sarna S. Snoring as a risk factor for ischaemic heart disease and stroke in men. Br Med J (Clin Res Ed) 1987; 294(6563):16-19.
22. Palomäki H, Partinen M, Juvela S, Kaste M. Snoring as a risk factor for sleep-related brain infarction. Stroke 1989; 20(10):1311-1315.
23. Partinen M, Palomaki H. Snoring and cerebral infarction. Lancet 1985; 2(8468):1325-1326.
24. Boudoulas H, Schmidt HS, Clark RW, Geleris P, Schaal SF, Lewis RP. Anthropometric characteristics, cardiac abnormalities and adrenergic activity in patients with primary disorders of sleep. J Med 1983; 14(3):223-238.
25. He J, Kryger MH, Zorick FJ, Conway W, Roth T. Mortality and apnea index in obstructive sleep apnea. Experience in 385 male patients. Chest 1988; 94(1):9-14.
26. Partinen M, Jamieson A, Guilleminault C. Long-term outcome for obstructive sleep apnea syndrome patients. Mortality. Chest 1988; 94(6):1200-1204.
27. Gonzalez-Rothi RJ, Foresman GE, Block AJ. Do patients with sleep apnea die in their sleep? Chest 1988; 94(3):531-538.
28. Tilkian AG, Guilleminault C, Schroeder JS, Lehrman KL, Simmons FB, Dement WC. Hemodynamics in sleep-induced apnea. Studies during wakefulness and sleep. Ann Intern Med 1976; 85(6):714-719.
29. Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med 2000; 342(19):1378-84.
30. O'Connor GT, Caffo B, Newman AB, et al. Prospective study of sleep-disordered breathing and hypertension: the Sleep Heart Health Study. Am J Respir Crit Care Med 2009; 179(12):1159-1164.
31. Marin JM, Agusti A, Villar I, et al. Association between treated and untreated obstructive sleep apnea and risk of hypertension. JAMA 2012; 307(20):2169-2176.
32. Bazzano LA, Khan Z, Reynolds K, He J. Effect of nocturnal nasal continuous positive airway pressure on blood pressure in obstructive sleep apnea. Hypertension 2007; 50(2):417-423.
33. Lozano L, Tovar JL, Sampol G, et al. Continuous positive airway pressure treatment in sleep apnea patients with resistant hypertension: a randomized, controlled trial. J Hypertens 2010; 28(10):2161-2168.
34. Chobanian AV, Bakris GL, Black HR, et al. Seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension 2003; 42(6):1206-1252.
35. 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.
36. Gottlieb DJ, Yenokyan G, Newman AB, et al. Prospective study of obstructive sleep apnea and incident coronary heart disease and heart failure: the sleep heart health study. Circulation 2010; 122(4):352-360.
37. Murphy TE, Han L, Allore HG, Peduzzi PN, Gill TM, Lin H. Treatment of death in the analysis of longitudinal studies of gerontological outcomes. J Gerontol A Biol Sci Med Sci 2011; 66(1):109-114.
38. Mooe T, Franklin KA, Holmstrom K, Rabben T, Wiklund U. Sleep-disordered breathing and coronary artery disease: long-term prognosis. Am J Respir Crit Care Med 2001; 164(10 Pt 1):1910-1913.
39. Yumino D, Tsurumi Y, Takagi A, Suzuki K, Kasanuki H. Impact of obstructive sleep apnea on clinical and angiographic outcomes following percutaneous coronary intervention in patients with acute coronary syndrome. Am J Cardiol 2007; 99(1):26-30.
40. Arzt M, Young T, Finn L, Skatrud JB, Bradley TD. Association of sleep-disordered breathing and the occurrence of stroke. Am J Respir Crit Care Med 2005; 172(11):1447-1451.
41. Redline S, Yenokyan G, Gottlieb DJ, et al. Obstructive Sleep Apnea Hypopnea and Incident Stroke: The Sleep Heart Health Study. Am J Respir Crit Care Med 2010;
42. Mehra R, Benjamin EJ, Shahar E, et al. Association of nocturnal arrhythmias with sleep-disordered breathing: The Sleep Heart Health Study. Am J Respir Crit Care Med 2006; 173(8):910-6.
43. Monahan K, Storfer-Isser A, Mehra R, et al. Triggering of nocturnal arrhythmias by sleep-disordered breathing events. J Am Coll Cardiol 2009; 54(19):1797-1804.
44. Kanagala R, Murali NS, Friedman PA, et al. Obstructive sleep apnea and the recurrence of atrial fibrillation. Circulation 2003; 107(20):2589-2594.
45. Marshall NS, Wong KK, Liu PY, Cullen SR, Knuiman MW, Grunstein RR. Sleep apnea as an independent risk factor for all-cause mortality: the Busselton Health Study. Sleep 2008; 31(8):1079-1085.
46. Young T, Finn L, Peppard PE, et al. Sleep disordered breathing and mortality: eighteen-year follow-up of the Wisconsin sleep cohort. Sleep 2008; 31(8):1071-1078.
47. Punjabi NM, Caffo BS, Goodwin JL, et al. Sleep-disordered breathing and mortality: a prospective cohort study. PLoS Med 2009; 6(8):e1000132.
48. Gami AS, Howard DE, Olson EJ, Somers VK. Day-night pattern of sudden death in obstructive sleep apnea. N Engl J Med 2005; 352(12):1206-1214.
49. Chami HA, Devereux RB, Gottdiener JS, et al. Left ventricular morphology and systolic function in sleep-disordered breathing: the Sleep Heart Health Study. Circulation 2008; 117(20):2599-2607.
50. Arias MA, Garcia-Rio F, Alonso-Fernandez A, Mediano O, Martinez I, Villamor J. Obstructive sleep apnea syndrome affects left ventricular diastolic function: effects of nasal continuous positive airway pressure in men. Circulation 2005; 112(3):375-383.
51. Newman AB, Nieto FJ, Guidry U, et al. Relation of sleep-disordered breathing to cardiovascular disease risk factors: the Sleep Heart Health Study. Am J Epidemiol 2001; 154(1):50-59.
52. Adedayo AM, Olafiranye O, Smith D, et al. Obstructive sleep apnea and dyslipidemia: evidence and underlying mechanism. Sleep Breath 2012;
53. Parish JM, Adam T, Facchiano L. Relationship of metabolic syndrome and obstructive sleep apnea. J Clin Sleep Med 2007; 3(5):467-472.
54. Foster GD, Sanders MH, Millman R, et al. Obstructive sleep apnea among obese patients with type 2 diabetes. Diabetes Care 2009; 32(6):1017-1019.
55. Clarenbach CF, West SD, Kohler M. Is obstructive sleep apnea a risk factor for diabetes? Discov Med 2011; 12(62):17-24.
56. Punjabi NM, Shahar E, Redline S, et al. Sleep-disordered breathing, glucose intolerance, and insulin resistance: the Sleep Heart Health Study. Am J Epidemiol 2004; 160(6):521-530.
57. Young T, Peppard PE, Taheri S. Excess weight and sleep-disordered breathing. J Appl Physiol 2005; 99(4):1592-1599.
58. Traviss KA, Barr SI, Fleming JA, Ryan CF. Lifestyle-related weight gain in obese men with newly diagnosed obstructive sleep apnea. J Am Diet Assoc 2002; 102(5):703-706.
59. Phillips BG, Hisel TM, Kato M, et al. Recent weight gain in patients with newly diagnosed obstructive sleep apnea. J Hypertens 1999; 17(9):1297-1300.
60. Brown MA, Goodwin JL, Silva GE, et al. The Impact of Sleep-Disordered Breathing on Body Mass Index (BMI): The Sleep Heart Health Study (SHHS). Southwest J Pulm Crit Care 2011; 3:159-168.
61. Prabhakar NR. Oxygen sensing during intermittent hypoxia: cellular and molecular mechanisms. J Appl Physiol 2001; 90(5):1986-1994.
62. Kent BD, Ryan S, McNicholas WT. Obstructive sleep apnea and inflammation: relationship to cardiovascular co-morbidity. Respir Physiol Neurobiol 2011; 178(3):475-481.
63. Barbe F, Duran-Cantolla J, Sanchez-de-la-Torre M, et al. Effect of continuous positive airway pressure on the incidence of hypertension and cardiovascular events in nonsleepy patients with obstructive sleep apnea: a randomized controlled trial. JAMA 2012; 307(20):2161-2168.
64. Peker Y, Glantz H, Thunstrom E, Kallryd A, Herlitz J, Ejdeback J. Rationale and design of the Randomized Intervention with CPAP in Coronary Artery Disease and Sleep Apnoea--RICCADSA trial. Scand Cardiovasc J 2009; 43(1):24-31.
65. Anonymous . Continuous Positive Airway Pressure Treatment of Obstructive Sleep Apnea to Prevent Cardiovascular Disease (SAVE). Last Updated: 2011. Accessed: October 1, 2012. http://www.clinicaltrials.gov/ct2/show/NCT00738179.
66. Anonymous . Heart Biomarker Evaluation in Apnea Treatment (HeartBEAT). Last Updated: 2011. Accessed: October 1, 2012. http://www.clinicaltrials.gov/ct2/show/NCT01086800.

Contact Information:

Stuart F. Quan, M.D.

Division of Sleep Medicine

Harvard Medical School

401 Park Dr., 2^nd Floor East

Boston, MA 02215

Voice: 617-998-8842

Fax: 617-998-8823

Email: Stuart_Quan@hms.harvard.edu

Conflicts of Interest: The author does not have any conflict of interests pertinent to the subject matter of this review.

Reference as: Quan SF. Obstructive sleep apnea and cardiovascular disease: back and forward in time over the last 25 years. Southwest J Pulm Crit Care 2012;5:206-17. PDF

The Impact of Sleep-Disordered Breathing on Body Mass Index (BMI): The Sleep Heart Health Study (SHHS)

Mark A. Brown, M.D. ¹

James L. Goodwin, Ph.D.²

Graciela E. Silva, Ph.D, MPH.³

Ajay Behari, M.D.⁴

Anne B. Newman, M.D., M.P.H^5,6

Naresh M. Punjabi, M.D., Ph.D.⁷

Helaine E. Resnick, Ph.D., M.P.H.⁸

John A. Robbins, M.D., M.S.H.⁹

Stuart F. Quan, M.D.^2,10

¹Department of Psychiatry, Kaiser Permanente, Portland, OR (markbrownmd@gmail.com); 

²Sleep and Arizona Respiratory Centers, University of Arizona College of Medicine, Tucson, AZ(jamieg@arc.arizona.edu);

³College of Nursing & Health Innovation, Arizona State University, Tempe, AZ (Graciela.Silva@asu.edu); 

⁴Pulmonary and Critical Care Associates of Baltimore, Baltimore, MD (ajaybehari@yahoo.com);

⁵Graduate School of Public Health, Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA

⁶Division of Geriatric Medicine, University of Pittsburgh School of Medicine, Pittsburgh, PA (NewmanA@edc.pitt.edu);

⁷Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD (npunjabi@jhmi.edu);

⁸American Association of Homes and Services for the Aging, Washington, DC (heresnick@gmail.com); 

⁹Center for HealthCare Policy and Research, University of California, Davis, Sacramento, CA (jarobbins@ucdavis.edu);

¹⁰Division of Sleep Medicine, Harvard Medical School, Boston, MA (squan@arc.arizona.edu)

Address for correspondence and reprint requests: Stuart F. Quan, M.D., Division of Sleep Medicine, Harvard Medical School, 401 Park Dr., 2^nd Floor East, Boston, MA 02215, Tel (617) 998-8842, Fax (617) 998-8823, Email: squan@arc.arizona.edu

Conflict of Interest Statement: None of the authors have conflicts of interest pertinent to the subject matter of this manuscript. 

Reference as: Brown MA, Goodwin JL, Silva GE, Behari A, Newman AB, Punjabi NM, Resnick HE, Robbins JA, Quan SF. The impact of sleep-disordered breathing on body mass index (BMI): the sleep heart health study (SHHS). Southwest J Pulm Crit Care 2011;3:159-68. (Click here for PDF version of the manuscript)

Abstract

Introduction: It is well known that obesity is a risk factor for sleep-disordered breathing (SDB). However, whether SDB predicts increase in BMI is not well defined. Data from the Sleep Heart Health Study (SHHS) were analyzed to determine whether SDB predicts longitudinal increase in BMI, adjusted for confounding factors.

Methods: A full-montage unattended home polysomnogram (PSG) and body anthropometric measurements were obtained approximately five years apart in 3001 participants. Apnea-hypopnea index (AHI) was categorized using clinical thresholds: < 5 (normal), ≥ 5 to <15 (mild sleep apnea), and ³ 15 (moderate to severe sleep apnea). Linear regression was used to examine the association between the three AHI groups and increased BMI. The model included age, gender, race, baseline BMI, and change in AHI as covariates.

Results: Mean (SD) age was 62.2 years (10.14), 55.2% were female and 76.1% were Caucasian. Five-year increase in BMI was modest with a mean (SD) change of 0.53 (2.62) kg/m² (p=0.071). A multivariate regression model showed that subjects with a baseline AHI between 5-15 had a mean increase in BMI of 0.22 kg/m² (p=0.055) and those with baseline AHI ≥ 15 had a BMI increase of 0.51 kg/m² (p<0.001) compared to those with baseline AHI of <5.

Conclusion: Our findings suggest that there is a positive association between severity of SDB and subsequent increased BMI over approximately 5 years. This observation may help explain why persons with SDB have difficulty losing weight.

Key Words: Sleep Apnea, Weight Gain, Obesity

Abbreviation List: PSG-polysomnogram, SDB-sleep disordered breathing, AHI-apnea hypopnea index, SHHS-Sleep Heart Health Study, BMI-body mass index, SD-standard deviation, SEM-standard error of the mean, ANOVA-analysis of variance

Introduction

There is overwhelming epidemiological and clinical data indicating that obesity is a risk factor for sleep disordered breathing (SDB).^1-8 The association between obesity and SDB is substantial, with high body mass index (BMI) contributing to moderate to severe SDB in 58% of affected persons.⁹ The effect of obesity is greater in men than women^1,10-12 although it decreases with increasing age.^6,7 In addition, weight loss has been demonstrated to decrease the severity of SDB.^10,13,14 Longitudinal data from population studies including the Sleep Heart Health Study (SHHS),¹⁰ the Wisconsin Sleep Cohort,¹⁵ and the Cleveland Family Study⁶ have initially focused on the impact of increased weight on SDB severity. However, examination of the opposite causal pathway has yet to be prospectively addressed.

Anecdotally, patients with SDB appear to have more difficulty losing weight than obese patients without SDB. They also report marked weight gain prior to confirmation of their diagnosis. Two small studies support these empiric observations.^5,16  Given this limited information on the impact of SDB on BMI, data from SHHS was analyzed to examine the impact of SDB on BMI after controlling for change in AHI and severity of SDB.

Methods

Study Design and Population. The SHHS is a multi-center, community-based prospective cohort study of the natural history and cardiovascular consequences of SDB. Details of the study design, sampling, and procedures have been reported.¹⁷ Briefly, between November 1995 and January 1998 participants were recruited from several ongoing prospective cohort studies--the Framingham Offspring and Omni Studies, the Atherosclerosis Risk in Communities Study, the Cardiovascular Health Study, the Strong Heart Study, and the cohort studies of respiratory disease in Tucson and of hypertension in New York. Participants were eligible if they were ≥ 40 years of age and were not being treated for sleep apnea with positive pressure therapy, an oral appliance, oxygen, or a tracheostomy. Habitual snorers < 65 years were over sampled to increase the prevalence of obstructive sleep apnea. Subjects were required to provide written consent and the protocol was approved by the institutional review boards of each of the eight investigative sites. 

Data Collection. A total of 6,441 subjects completed the baseline polysomnogram (PSG), and 4,586 consented to have a second evaluation approximately five years later. This analysis focuses on the 3,040 participants who had PSG and BMI data at both time points. Data from all 215 participants who had a follow-up PSG from the New York center were excluded because they did not meet quality standards for the follow-up examination. The remaining participants died, were too ill to participate, refused to participate, were lost to follow-up or had incomplete covariate data such as weight. This latter group had a higher percentage of Whites (85%) compared with the study group (75.5%) (p-value <0.001). There also were statistically significant differences in baseline BMI, baseline AHI, and age between the study group compared with the excluded group, however, these differences were very small and were not clinically significant. There was no gender difference between the two groups.

Weight was measured on the night of the PSG examination with the participant in light clothes on a calibrated portable scale. Height was obtained at the baseline home visit if not already measured within + 3 months of the parent study. BMI was calculated as weight in kilograms divided by the square of height in meters. Baseline height was used for baseline and follow-up BMI calculations. Age, sex, and ethnicity were self-reported. 

The PSG was conducted using a portable monitor (PS-2 System; Compumedics Limited, Abbotsford, Victoria, Australia), using methods previously described.¹⁸ Apnea was present if there was an absence or near absence of airflow or thoracoabdominal movement (at least < 25% of baseline) for > 10 seconds. Hypopnea was defined as a decrease in the amplitude of the airflow or thoracoabdominal movement below 70% of baseline for > 10 seconds. The apnea-hypopnea index (AHI) was calculated as the number of apnea and hypopnea events, each associated with at least a 4% decrease in oxygen saturation, divided by total sleep time in hours.

Results

Participant characteristics are provided in Table 1. 

Table 1: Characteristics of participants of the Sleep Heart Health Study cohort with complete baseline and follow-up polysomnography and weight measurements as a function of sleep apnea severity.

As expected, women were over-represented in the baseline AHI < 5 group (64.9%) and men were over-represented in the AHI ³ 15 group (63.5%) (p<0.001). Baseline BMI increased as baseline AHI severity increased.  Overall unadjusted five-year increase in BMI was modest with a mean (SD) BMI change of 0.53 (2.61) kg/m². The unadjusted five-year increase in BMI was 0.63 (2.54) kg/m² for those with baseline AHI < 5, 0.43 (2.48) kg/m² among those with AHI ≥ 5 to < 15 and 0.37 (3.05) kg/m² for the AHI group ≥ 15. These values were not statistically different from each other.

A multivariate regression model was constructed predicting five-year change in BMI by baseline AHI category adjusted for age, gender, race, baseline BMI, and AHI change (Table 2).

Table 2: Adjusted β coefficients of BMI change according to AHI and continuous variables in the Sleep Heart Health Study*.

Compared to baseline AHI group of < 5, those with AHI between ≥ 5 to < 15 had a mean adjusted increase in BMI of 0.21 that approached statistical significance (p=0.055). However, those with AHI ≥ 15 had a statistically significant adjusted BMI increase of 0.51 (p<0.001).  Younger age, lower baseline BMI and greater AHI change also were associated with a larger BMI increase. There was a trend for women to have a greater increase in BMI, but no effect of race was observed. However, the model only accounted for 7% of the total variance. Adjusted means by baseline AHI group are displayed graphically in Figure 1.

Figure 1: Estimated Adjusted Means of BMI increase according to AHI in the Sleep Heart Health Study. Data are adjusted for baseline age (continuous), race (categorical), gender (categorical), baseline BMI (continuous), change in AHI (continuous). Covariates fixed at: baseline BMI = 28.7, baseline age 62.1, change in RDI = 2.7. Bars represent 95% confidence intervals.

Discussion

Our findings indicate that there is a positive association between severity of SDB and five-year increase in BMI. The finding was demonstrated after controlling for key covariates including age, gender, race, baseline BMI, and AHI change. This observation may help explain the difficulty patients with SDB have in trying to lose weight.

Two previous small studies have demonstrated a positive association between newly diagnosed SDB and weight gain. A retrospective study by Phillips et al. compared one-year weight histories of 53 men and women patients who were recently diagnosed with SDB with 24 control subjects matched for gender, age, BMI and percent body fat.⁵ Subjects in that study were somewhat younger than the SHHS cohort with an age difference of approximately 10 years. The SDB among subjects in the previous study tended to be moderate to severe with mean ± SEM AHI 33 ± 5 /h for men and 37 ± 10 /h for women. Mean ± SEM of BMI at time of diagnosis was somewhat higher than in the SHHS with 35 ± 1 kg/m² for men and 44 ± 2 kg/m² for women.  Men and women patients with SDB had reported a recent weight gain of 7.4 ± 1.5 kg compared with a weight loss of 0.5 ± 1.7 kg (p=0.001) in obese controls without SDB. However, given the design of this study it is not possible to determine whether weight gain contributed to the onset of SDB or was a result of SDB. The study was also limited by reliance on self-report of weight gain history.

Another study by Traviss et al. prospectively evaluated 49 obese patients with newly diagnosed SDB.¹⁶ Mean ± SD of AHI at diagnosis was severe at 45 ± 27 /h.  BMI at diagnosis was elevated at 36.5 ± 6.2 kg/m². Of the 49 subjects, 43 could estimate the duration of their symptoms with 84% reporting weight gain since becoming symptomatic.  Weight gain was relatively large, with a reported 17 ± 15 kg over 5.3 ± 4.8 years. However, this study was limited by the lack of a control group and reliance on self-report of weight history.  These two small studies, in addition to our findings, suggest that there is an association between SDB and increased BMI.

Interestingly, unadjusted BMI change in our study was quite modest and not statistically different as a function of SDB severity. However, BMI change over time is a complex phenomenon influenced by several variables. A large (29,799 subjects) prospective study examining 5-year change in weight in a multi-ethnic cohort of men and women explored several of these relationships.¹⁹ In that study, younger men and to a greater degree, younger women were at greater risk for weight gain compared to older adults. This is consistent with our initial findings. In addition, there was a trend for women in the higher baseline BMI categories of ‘overweight’ (BMI >25– 30 kg/m²) and ‘obese’ (BMI >30 kg/m²) in the aforementioned cohort to gain more weight than men in the higher baseline BMI categories. In order to more precisely examine the effect of AHI on weight change, we controlled for these confounders in our final multivariate model thus resulting in the finding of an increase in BMI as a function of SDB severity in this study. 

Several mechanisms could explain why SDB contributes to increased BMI. First, persons with SDB may have a reduction in the quantity and quality of their sleep. Recent data indicate that insufficient sleep may be a risk factor for obesity.²⁰ Experimental sleep restriction increases ghrelin and reduces leptin production favoring appetite enhancement,²¹ a finding that also has been observed in a large population cohort.²² Second, those with SDB may eat a diet that favors weight gain. In support of this hypothesis, sleep restriction has been shown to increase craving for calorie dense food with high carbohydrate content. The Apnea Positive Pressure Long-Term Efficacy Study (APPLES) demonstrated that those with severe SDB consumed a diet higher in cholesterol, protein, total fat and total saturated fatty acids, even after adjusting for BMI, age, and daytime sleepiness.²³ Third, a cardinal symptom of SDB is excessive daytime sleepiness. Thus, it is possible that persons with SDB engage in less physical activity because they are too fatigued to exercise. Data from APPLES indicate that recreational physical activity is less in those with SDB. However, this finding appears to be principally explained by concomitant obesity. 

Weight loss frequently results in an improvement and sometimes resolution in SDB. This is most evident in those who undergo bariatric surgical procedures.^24,25 Persons with SDB are frequently counseled to treat their SDB by losing weight through diet and exercise,²⁶ an approach that is usually unsuccessful.²⁵ Failure to primarily address SDB in conjunction with a weight reduction program may diminish the latter’s success. However, evidence to date indicates that treatment of SDB does not consistently result in weight loss. In a sample of clinical patients with SDB, treatment with CPAP did not result in weight loss. Moreover, in female patients, there was actually an increase in weight.²⁷ In addition, consistent weight reduction was not observed in a small number of patients with severe OSA who underwent tracheostomy.²⁸ Thus, it appears that weight gain engendered by the presence of OSA is not easily reversed despite therapy. Prospective studies will be required to determine whether primary treatment for OSA enhances weight loss programs in those with OSA.

Although this analysis demonstrated a positive association of severity of SDB on five-year increase in BMI, there are several caveats that deserve consideration. The BMI of participants tended to be lower than that seen in clinical SDB populations and a relatively small number of subjects had large changes in BMI. As previously noted, the mean BMI increase was, at best, quite modest. When converted for illustrative purposes to weight using an average height of 167 cm of the participants, those with an AHI between ≥ 5 to < 15 had a mean adjusted increase in BMI of 0.21 kg/m² equal to 0.59 kg or 1.30 lbs. Similarly, those with AHI ≥ 15 had an adjusted BMI increase of 0.51 kg/m² equal to 1.42 kg or 3.13 lbs. Thus, the magnitude of the changes we observed may not be applicable to clinical populations where patients with SDB may have a higher BMI. In addition, it is not known when the participants developed SDB, thus definitive inference of causality cannot be made. However, following a large undiagnosed cohort over an extended period of time to determine incidence of SDB onset and subsequent change in weight would be exceedingly difficult and costly.  Additionally, the model only accounted for a small amount of the total variance in five-year BMI increase, suggesting that there are likely other unmeasured variables influencing the amount of BMI increase over time in this cohort. Finally, while not statistically significant, the unadjusted mean change in BMI was slightly less in the high RDI group in comparison to the lower RDI groups. This observation underscores the biological complexity of the interactions among weight change, SDB, age, gender and other factors.

In conclusion, our findings suggest that although weight gain is a risk factor for developing or worsening SDB, SDB may, in a reciprocal fashion, lead to increased weight gain. This may help explain why patients with SDB find it difficult to lose weight.

Acknowledgements

This work was supported by National Heart, Lung and Blood Institute cooperative agreements U01HL53940 (University of Washington), U01HL53941 (Boston University), U01HL53938 (University of Arizona), U01HL53916 (University of California, Davis), U01HL53934 (University of Minnesota), U01HL53931 (New York University), U01HL53937 and U01HL64360  (Johns Hopkins University), U01HL63463 (Case Western Reserve University), and U01HL63429 (Missouri Breaks Research).

Sleep Heart Health Study (SHHS) acknowledges the Atherosclerosis Risk in Communities Study (ARIC), the Cardiovascular Health Study (CHS), the Framingham Heart Study (FHS), the Cornell/Mt. Sinai Worksite and Hypertension Studies, the Strong Heart Study (SHS), the Tucson Epidemiologic Study of Airways Obstructive Diseases (TES) and the Tucson Health and Environment Study (H&E) for allowing their cohort members to be part of the SHHS and for permitting data acquired by them to be used in the study.  SHHS is particularly grateful to the members of these cohorts who agreed to participate in SHHS as well. SHHS further recognizes all of the investigators and staff who have contributed to its success. A list of SHHS investigators, staff and their participating institutions is available on the SHHS website, www.jhucct.com/shhs.

The opinions expressed in the paper are those of the author(s) and do not necessarily reflect the views of the Indian Health Service.

These data have been presented in part at the Annual Meeting of the Associated Professional Sleep Societies, June 11, 2009, Seattle, WA.

References

1. Bixler EO, Vgontzas AN, Lin HM, Ten Have T, Rein J, Vela-Bueno A, et al. Prevalence of sleep-disordered breathing in women: Effects of gender. Am J Respir Crit Care Med 2001;163:608-13.
2. Bixler EO, Vgontzas AN, Ten Have T, Tyson K, Kales A. Effects of age on sleep apnea in men: I. prevalence and severity. Am J Respir Crit Care Med 1998;157:144-8.
3. Carmelli D, Swan GE, Bliwise DL. Relationship of 30-year changes in obesity to sleep-disordered breathing in the western collaborative group study. Obes Res 2000;8:632-7.
4. Duran J, Esnaola S, Rubio R, Iztueta A. Obstructive sleep apnea-hypopnea and related clinical features in a population-based sample of subjects aged 30 to 70 yr. Am J Respir Crit Care Med 2001;163:685-9.
5. Phillips BG, Hisel TM, Kato M, Pesek CA, Dyken ME, Narkiewicz K, et al. Recent weight gain in patients with newly diagnosed obstructive sleep apnea. J Hypertens 1999;17:1297-300.
6. Tishler PV, Larkin EK, Schluchter MD, Redline S. Incidence of sleep-disordered breathing in an urban adult population: The relative importance of risk factors in the development of sleep-disordered breathing. JAMA 2003;289:2230-7.
7. Young T, Shahar E, Nieto FJ, Redline S, Newman AB, Gottlieb DJ, et al. Predictors of sleep-disordered breathing in community-dwelling adults: The sleep heart health study. Arch Intern Med 2002;162:893-900.
8. Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328:1230-5.
9. Young T, Peppard PE, Taheri S. Excess weight and sleep-disordered breathing. J Appl Physiol 2005;99:1592-9.
10. Newman AB, Foster G, Givelber R, Nieto FJ, Redline S, Young T. Progression and regression of sleep-disordered breathing with changes in weight: The sleep heart health study. Arch Intern Med 2005;165:2408-13.
11. Ip MS, Lam B, Tang LC, Lauder IJ, Ip TY, Lam WK. A community study of sleep-disordered breathing in middle-aged chinese women in hong kong: Prevalence and gender differences. Chest 2004;125:127-34.
12. Millman RP, Carlisle CC, McGarvey ST, Eveloff SE, Levinson PD. Body fat distribution and sleep apnea severity in women. Chest 1995;107:362-6.
13. Barvaux VA, Aubert G, Rodenstein DO. Weight loss as a treatment for obstructive sleep apnoea. Sleep Med Rev 2000;4:435-52.
14. Grunstein RR, Stenlof K, Hedner JA, Peltonen M, Karason K, Sjostrom L. Two year reduction in sleep apnea symptoms and associated diabetes incidence after weight loss in severe obesity. Sleep 2007;30:703-10.
15. Peppard PE, Young T, Palta M, Dempsey J, Skatrud J. Longitudinal study of moderate weight change and sleep-disordered breathing. JAMA 2000;284:3015-21.
16. Traviss KA, Barr SI, Fleming JA, Ryan CF. Lifestyle-related weight gain in obese men with newly diagnosed obstructive sleep apnea. J Am Diet Assoc 2002;102:703-6.
17. Quan SF, Howard BV, Iber C, Kiley JP, Nieto FJ, O'Connor GT, et al. The sleep heart health study: Design, rationale, and methods. Sleep 1997;20:1077-85.
18. Redline S, Sanders MH, Lind BK, Quan SF, Iber C, Gottlieb DJ, et al. Methods for obtaining and analyzing unattended polysomnography data for a multicenter study. sleep heart health research group. Sleep1998;21:759-67.
19. Ball K, Crawford D, Ireland P, Hodge A. Patterns and demographic predictors of 5-year weight change in a multi-ethnic cohort of men and women in australia. Public Health Nutr 2003;6:269-81.
20. Chaput JP, Despres JP, Bouchard C, Tremblay A. The association between sleep duration and weight gain in adults: A 6-year prospective study from the Quebec Family Study. Sleep 2008;31:517-23.
21. Spiegel K, Tasali E, Penev P, Van Cauter E. Brief communication: Sleep curtailment in healthy young men is associated with decreased leptin levels, elevated ghrelin levels, and increased hunger and appetite. Ann Intern Med 2004;141:846-50.
22. Taheri S, Lin L, Austin D, Young T, Mignot E. Short sleep duration is associated with reduced leptin, elevated ghrelin, and increased body mass index. PLoS Med. 2004;1:e62.
23. Vasquez MM, Goodwin JL, Drescher AA, Smith TW, Quan SF. Associations of dietary intake and physical activity with sleep disordered breathing in the apnea positive pressure long-term efficacy study (APPLES). J Clin Sleep Med 2008;4:411-8.
24. Rasheid S, Banasiak M, Gallagher SF, Lipska A, Kaba S, Ventimiglia D, et al. Gastric bypass is an effective treatment for obstructive sleep apnea in patients with clinically significant obesity. Obes Surg 2003;13:58-61.
25. Fritscher LG, Canani S, Mottin CC, Fritscher CC, Berleze D, Chapman K, et al. Bariatric surgery in the treatment of obstructive sleep apnea in morbidly obese patients. Respiration 2007;74:647-52.
26. Veasey SC, Guilleminault C, Strohl KP, Sanders MH, Ballard RD, Magalang UJ. Medical therapy for obstructive sleep apnea: A review by the medical therapy for obstructive sleep apnea task force of the standards of practice committee of the american academy of sleep medicine. Sleep 2006;29:1036-44.
27. Redenius R, Murphy C, O'Neill E, Al-Hamwi M, Zallek SN. Does CPAP lead to change in BMI? J Clin Sleep Med 2008;4:205-9.
28. Haapaniemi JJ, Laurikainen EA, Halme P, Antila J. Long-term results of tracheostomy for severe obstructive sleep apnea syndrome. ORL J Otorhinolaryngol Relat Spec 2001;63:131-6.