Issue Navigator

Volume 12 No. 12
Earn CME
Accepted Papers

Journal Club


Nancy Collop, MD, FAASM1; Tracy L. Stierer, MD2; Shirin Shafazand, MD3
1Emory Sleep Center, Emory University, Atlanta, GA; 2Johns Hopkins Center for Sleep at Howard County General Hospital, Johns Hopkins University, Baltimore, MD; 3Miller School of Medicine, Miami University, Miami, FL

Summary and commentary on McEvoy et al. CPAP for prevention of cardiovascular events in obstructive sleep apnea. N Engl J Med 2016;375:919–931.



In individuals with moderate to severe obstructive sleep apnea (OSA) and cardiovascular disease, does the treatment of OSA with continuous positive airway pressure therapy (CPAP), in addition to medical management of cardiovascular disease (usual care), reduce the incidence of the primary composite end point (death from cardiovascular causes, myocardial infarction, stroke, or hospitalization for unstable angina, heart failure, or transient ischemic attack)?



Multi-center, randomized, parallel-group, open-label trial, with blinded end-point assessment; identifier: NCT00738179; Australian New Zealand Clinical Trials Registry number: ACTRN12608000409370.


Randomization was conducted by a centralized computer system. The randomization sequence was concealed. Randomization was performed with the use of a minimization procedure and stratification was according to study site, type of cardiovascular disease (cardiac, cerebrovascular, or both), and severity of daytime sleepiness (Epworth Sleepiness Scale score < 11 vs. ≥ 11).


The investigators and participants were not blinded to study arm assignment. The primary and secondary outcome measurements were blinded.

Follow-up period

The end of the study was when a statistically pre-determined number of efficacy endpoint events had been observed or every patient had participated for at least 1.5 years. The mean follow-up was 3.7 years.


Participants were recruited from 89 clinical centers in 7 countries.


INCLUSION CRITERIA: (1) Age between 45 and 75 years, (2) Diagnosis of coronary artery disease or cerebrovascular disease, (3) Diagnosis of moderate to severe OSA defined as an ODI (≥ 4% oxygen desaturation) ≥ 12 per hour.

EXCLUSION CRITERIA: (1) Medical or behavioral comorbidities that made it difficult to participate in study, (2) Any planned coronary or carotid revascularization procedure within the next 6 months, (3) Severe pulmonary disease, (4) New York Heart Association (NYHA) class III-IV, (5) Other household member enrolled in SAVE trial or using CPAP, (6) Prior use of CPAP for OSA, (7) Increased risk of a sleep-related accident and/or excessive daytime sleepiness (Epworth sleepiness scale > 15), (8) Severe nocturnal desaturation documented on the ApneaLink device, defined as > 10% overnight recording time with arterial oxygen saturation of < 80%, (9) Cheyne-Stokes Respiration.


OSA was diagnosed using ApneaLink portable device. Subjects meeting eligibility criteria completed a 1 week run in period with sub therapeutic CPAP (sham CPAP) to determine their level of treatment adherence. Participants who had a minimum average use of 3 hours per night of CPAP use during that week, were then randomized to CPAP and usual care vs. usual care alone. Subjects who were assigned to CPAP therapy underwent a 1-week home trial of auto-titrating CPAP after which the appropriate fixed level of CPAP (i.e. 90th percentile of pressures determined on auto CPAP) needed to treat OSA was determined. This fixed pressure was used until study end. Best fit mask was provided. Clinic visits took place at baseline, 1, 3, 6, 12 months, and every 12 months thereafter, until the end of the study. Participants were contacted by telephone at 6 months between annual clinic visits.

Outcome Measures and Sample Size Calculations

The primary endpoint was defined as a composite endpoint of death from any cardiovascular cause, myocardial infarction (including silent myocardial infarction), stroke, or hospitalization for heart failure, acute coronary syndrome (including unstable angina), or transient ischemic attack. Secondary endpoints were other combinations of composite endpoints, individual components of the primary composite endpoint, health related quality of life, mood, OSA symptom and sleepiness, new onset atrial fibrillation, new onset diabetes, and all-cause mortality.

Pre-specified safety endpoints were evaluated at each participant contact point including serious adverse events, self-reported accidents causing personal injury that occurred while the participant was driving or while at work, and any accidents or near-miss accidents that occurred as a result of the participant falling asleep. Participants were also asked about the number of traffic accidents due to any cause and the number of days off from work due to poor health.

This was a multi-center, international study with parallel group design. An initial sample size of 5,000 was estimated to provide at least 80% power (alpha 0.05) to detect a 20% relative risk reduction in the primary endpoint in the intervention group compared with the control group, allowing for 10% loss to follow-up in each group, and for 80% of participants who remained in the CPAP to adhere approximately 3–4 hours per night to treatment and 20% to have zero adherence to treatment. In mid-2012, recruitment challenges prompted a review of the trial experience, and CPAP adherence. Additionally, an updated meta-regression of cardiovascular events in moderate to severe OSA, suggested an increased incidence of cardiovascular events in patients with OSA than what was originally estimated. A new power calculation indicated a sample size of 2,500 would be able to detect a 25% relative risk reduction (alpha 0.05, power 90%) in the primary composite CV endpoint. The study was revised to reflect the reduced sample size. This was an intention to treat analysis.

Patient follow-up

2,717 patients were randomized from December 2008 through November 2013; 1,359 received CPAP plus usual care and 1,358 were randomized to usual care alone; 21 participants from one site were excluded due to protocol irregularities. 2,687 participants were included in the primary analysis (1,346 in CPAP group and 1,341 in usual care group). The mean duration of follow-up was 3.7 years with > 90% complete follow-up.

Main Results

Over 80% of participants were male, mean age 61 years, and 63% were Asian. The mean body-mass index of the participants was 29; the mean oxygen desaturation index (ODI), 28 events per hour; and the mean Epworth Sleepiness Scale score, 7.4. There were no significant differences in baseline characteristics (including AHI, ODI, distribution of cardiac or cerebrovascular disease) between the usual care group and the CPAP group.

During the 1-week run-in phase, sub therapeutic CPAP device was used for a mean of 5.2 hours per night. For the CPAP group, the mean (± SD) duration of adherence to CPAP therapy in the first month was 4.4 ± 2.2 hours per night, which decreased to 3.5 ± 2.4 hours per night by 12 months and remained relatively stable to study end (mean adherence during follow-up, 3.3 ± 2.3 hours). Of the 1,346 patients in the CPAP group, 566 (42%) had good adherence to treatment (≥ 4 hours per night) during follow-up.

The incidence of the primary end point did not differ significantly between the CPAP group and the usual care group (17.0% in the CPAP group and 15.4% in the usual-care group, hazard ratio with CPAP, 1.10; 95% confidence interval [CI], 0.91 to 1.32; p = 0.34). No significant difference on any individual or composite cardiovascular outcomes, or all-cause mortality was observed. No significant difference was noted in the primary outcome, across gender, location, severity of OSA, age, sleepiness, presence of diabetes, and type of cardiovascular disease.

Propensity-score matching was performed to compare 561 patients who were adherent to CPAP therapy with 561 patients in the usual-care group. There was no statistically significant difference in the primary outcome. However, the propensity score-matched analyses showed that the patients who were adherent to CPAP therapy had a lower risk of stroke than those in the usual-care group (hazard ratio, 0.56; 95% CI, 0.32 to 1.00; p = 0.05), as well as a lower risk of the composite end point of cerebral events (hazard ratio, 0.52; 95% CI, 0.30 to 0.90; p = 0.02), but these results were not adjusted for multiple testing. A post hoc CPAP dose-response analysis of the primary and secondary cardiovascular end points showed no significant association.

There was a statistically significant reduction from baseline in depression and anxiety score, daytime sleepiness, number of days lost from work, and an improvement in health related quality of life in subjects on CPAP therapy.


In predominately male subjects with moderate to severe obstructive sleep apnea and cardiovascular disease, randomized to CPAP therapy in addition to usual care or usual care alone, there was no statistically significant difference in a composite primary outcome of death from any cardiovascular cause, myocardial infarction (including silent myocardial infarction), stroke, or hospitalization for heart failure, acute coronary syndrome (including unstable angina), or transient ischemic attack. However, there were improvements with CPAP therapy in daytime sleepiness and quality of life measures.

Sources of funding: Supported by project grants (1006501 [2011–2015] and 1060078 [2014–2016]) from the National Health and Medical Research Council (NHMRC) of Australia and by Respironics Sleep and Respiratory Research Foundation and Philips Respironics. Supplementary trial funding was provided by Fisher & Paykel Healthcare, the Australasian Sleep Trials Network (enabling grant 343020 from the NHMRC), the Spanish Respiratory Society (grant 105-2011 to Drs. Barbe and Mediano), and Fondo de Investigaciones Sanitarias (grant 13/02053 to Drs. Barbe and Mediano). In-kind donations were provided by Respironics for CPAP equipment and by ResMed for sleep apnea diagnostic devices. For correspondence: Dr. McEvoy at the Adelaide Institute for Sleep Health, Flinders University and Respiratory and Sleep Services, Southern Adelaide Local Health Network, Repatriation General Hospital, Daw Park, Adelaide SA 5041, Australia,; or to Dr. Luo at the First Affiliated Hospital of Guangzhou Medical University, State Key Laboratory of Respiratory Disease, Guangzhou, China,


The SAVE trial concluded that use of CPAP in patients with a prior cardiovascular event did not reduce the likelihood of a second event compared to “usual” therapy.1 We certainly applaud the authors for doing this study and providing these valuable results. However, as with any study, it has its pitfalls.

First, we would like to point out the unusual inclusion and exclusion criteria. Where did the “ODI ≥ 12 per hour” come from? The supplemental data shows that this criteria was based on validation of the ApneaLink home sleep apnea test, but “twelve” is not one that is usually utilized as a cutoff. Regarding the exclusion criteria, the authors effectively excluded many patient populations that might have benefited greatly from CPAP use with regards to cardiac events and death.2 For example, there are studies showing patients with advanced heart failure (NYHA Class III and IV) have improved survival with use of CPAP.3 Similarly, patients with pulmonary hypertension and heart failure have been shown to have modest hemodynamic improvement with use of CPAP.4 Although the authors chose to exclude those with a high Epworth (> 15) due to safety concerns, there certainly are data that those with excessive sleepiness have a dramatic reduction in risk for CV death compared to those who are not sleepy.5 Finally, severe oxygen desaturation in patients has been shown to be an important predictor of a poor prognosis following acute myocardial infarction.6

An interesting note in demographic make-up of the study was that 63% of participants were Asian and had lower BMI scores. This population with low BMI does not represent the epidemiological distribution of patients with OSA and cardiovascular disease, at least in the USA. We also agree with the accompanying New England Journal of Medicine editorial which denoted concern, consistent with many sleep specialists, that an average CPAP use of 3.3 hours may have not been enough to make a difference.7 The study allowed for 80% of participants to adhere to 3–4 hours per night treatment and 20% to have zero adherence to treatment; in fact, only 42% had greater than 4 hours adherence per night! It is not clear if the study monitors employed any type of individual encouragement or help other than providing “the best mask fit” to improve CPAP usage even though there are methods that can improve adherence.8,9,10 It is abundantly clear that more hours with CPAP results in improved measures such as lower blood pressure, reduced sleepiness and increased quality of life.11,12 One would assume that such improvements would similarly translate to fewer cardiovascular events.

While we think that the New England Journal of Medicine editorial is a worthy commentary on the study, we also believe in being realistic. Having taken care of thousands of sleep apnea sufferers during our careers, we know that some patients are just not going to adequately adhere to their prescribed CPAP therapy—no matter how much we urge, cajole, beg or insist. Despite all we know about the risks of OSA, and despite all the improvements that have been developed and are available in the current CPAP units, most studies like the SAVE trial have shown suboptimal CPAP usage. In fact, a recent literature review suggested that this has not really improved much over time.13

Why is this the case? CPAP use is not much different than use of any chronically-required medication—some will use their medications and some won't, and those who don't have higher mortality rates.14 Specifically, it has been shown that those who did not use their beta blocker after a myocardial infarction were 2.6 times more likely to die than those who were adherent.15 Likewise, adherence to candesartan in heart failure proved life saving.16 Although there was a secondary analysis using propensity score matching in this trial examining those with better CPAP use, such an analysis is not equivalent to a controlled study. Most of us can understand that taking a regular medication is challenging—due to cost, side effects, ambivalence about having a disorder or busy lives/schedules. We have a unique challenge in treating OSA patients —because their disorder occurs while they are asleep—there is often a lack of awareness and easy denial. Those that clearly note an immediate and dramatic improvement, do use CPAP more consistently. There are others who are extremely disciplined and use it because they are told and acknowledge it may improve future health. However, we would say that the majority of our patients may note some initial benefit—better sleep, less sleepiness, less nocturia, fewer headaches—but over time, these issues become less important and usage drops. Also, many patients will admit to better sleep with their PAP device but usage will fall over time due to other issues that come up, getting out of the habit, or not noticing as much improvement as they initially did.

The bottom line is that we need to continue to try and move beyond CPAP with innovative treatments for OSA. Until we do, there will always be the question: was CPAP used enough to make a difference?


This was not an industry supported study. Dr. Collop is Editor-in-Chief of the Journal of Clinical Sleep Medicine. Dr. Stierer is Associate Editor for the Journal of Clinical Sleep Medicine. Dr. Shafazand is Associate Editor for the Journal of Clinical Sleep Medicine. She has received grant support from DOD and NIDDR. There was no investigational or off label use.


Collop N, Stierer TL, Shafazand S. SAVE me from CPAP. J Clin Sleep Med 2016;12(12):1701–1704.



McEvoy R, Antic N, Heeler E, et al. CPAP for prevention of cardiovascular events in obstructive sleep apnea. N Engl J Med. 2016;375:919–31. [PubMed]


Javaheri S, Javaheri S, Javaheri A. Sleep apnea, heart failure, and pulmonary hypertension. Curr Heart Fail Rep. 2013;10:315–320. [PubMed Central][PubMed]


Kasai T, Narui K, Dohi T, et al. Prognosis of patients with heart failure and obstructive sleep apnea treated with continuous positive airway pressure. Chest. 2008;133:690–6. [PubMed]


Sajkov D, Wang T, Saunders NA, Bune AJ, McEvoy RD. Continuous positive airway pressure treatment improves pulmonary hemodynamics in patients with obstructive sleep apnea. Am J Respir Crit Care Med. 2002;165:152–8. [PubMed]


Empana JP, Dauviliers Y, Dartigues JF, et al. Excessive daytime sleepiness is an independent risk indicator for cardiovascular mortality in community dwelling elderly: the three city study. Stroke. 2009;40:1219–24. [PubMed]


Xie J, Sert Kuniyoshi FH, Covassin N, et al. Nocturnal hypoxemia due to obstructive sleep apnea is an independent predictor of poor prognosis after myocardial infarction. J Am Heart Assoc. 2016;5:e003162. [PubMed Central][PubMed]


Mokhlesi B, Ayas NT. Cardiovascular events in obstructive sleep apnea - can CPAP therapy SAVE lives? N Engl J Med. 2016;375:994–6. [PubMed]


Richards D, Bartlett DJ, Wong K, Malouff J, Grunstein RR. Increased adherence to CPAP with a group cognitive behavioral treatment intervention: a randomized trial. Sleep. 2007;30:635–40. [PubMed]


Lettieri CJ, Walter RJ. Impact of group education on continuous positive airway pressure adherence. J Clin Sleep Med. 2013;9:537–41. [PubMed Central][PubMed]


Parthasarathy S, Wendel C, Haynes PL, Atwood C, Kuna S. A pilot study of CPAP adherence promotion by peer buddies with sleep apnea. J Clin Sleep Med. 2013;9:543–50. [PubMed Central][PubMed]


Weaver TE, Maislin G, Dinges DF, et al. Relationship between hours of CPAP use and achieving normal levels of sleepiness and daily functioning. Sleep. 2007;30:711–9. [PubMed Central][PubMed]


Yang MC, Huang YC, Lan CC, Wu YK, Huang KF. Beneficial effects of long-term CPAP treatment on sleep quality and blood pressure in adherent subjects with obstructive sleep apnea. Respir Care. 2015;60:1810–8. [PubMed]


Rotenberg B, Murariu D, Pang K. Trends in CPAP adherence over 20 years of data collection: a flattened curve. J Otolaryngol Head Neck Surg. 2016;45:43. [PubMed Central][PubMed]


Simpson SH, Eurich DT, Majumdar SR, et al. A meta-analysis of the association between adherence to drug therapy and mortality. BMJ. 2006;333:15. [PubMed Central][PubMed]


Horwitz RI, Viscoli CM, Berkman L, et al. Treatment adherence and risk of death after a myocardial infarction. Lancet. 1990;336:542–5. [PubMed]


Granger BB, Swedberg K, Ekman I, et al. Adherence to candesartan and placebo and outcomes in chronic heart failure in the CHARM programme: double-blind, randomised, controlled clinical trial. Lancet. 2005;366:2005–11. [PubMed]