The Sleep Apnea Cardiovascular Endpoints (SAVE) trial made waves with its findings and unexpected conclusion that CPAP for obstructive sleep apnea does not have certain cardiovascular benefits. A sleep physician addresses the nuances that appear to have been overlooked in the media and puts the results in perspective.

The Sleep Apnea Cardiovascular Endpoints (SAVE) trial presented at the European Society of Cardiology Conference and in The New England Journal of Medicine1 made waves with its unexpected finding that CPAP for obstructive sleep apnea (OSA) does not have certain cardiovascular benefits. While the SAVE trial provides new and interesting data, the conclusion appears contrary to conventional “wisdom” based on generally accepted principles and a large body of literature. The findings have been watered down through various media channels, and the ripples from these waves could have far-reaching effects on the practice patterns of healthcare providers and their patients (diagnosed or not) suffering from OSA.

The intent of this article is to address nuances that appear to have been overlooked in other media channels and put the results in perspective for patients and physicians who may question the benefits of CPAP for patients suffering from OSA. The goal is to plug the leak, SAVE our ship (the sleep medicine industry), steady the course, and minimize any unintended consequences for its crew (healthcare providers), and its precious cargo (patients with OSA).

Study Summary

The SAVE trial studied 2,717 adults (mostly men ages 45 to 75 years with minimal sleepiness) from 7 countries (though only 4 patients were from the United States) who had moderate to severe OSA and coronary or cerebrovascular disease. Half the patients received CPAP plus usual care or usual care alone. The primary composite endpoint was death from cardiovascular causes, myocardial infarction, stroke, or hospitalization for unstable angina, heart failure, or transient ischemic attack. Secondary endpoints included other cardiovascular outcomes, quality of life, snoring, daytime sleepiness, and mood. The mean duration of adherence to CPAP therapy was 3.3 hours per night.

Using a limited-channel home sleep testing (HST) device, the diagnosis of moderate to severe OSA was defined as an oxygen desaturation index or ODI (drop in oxygen saturation ? 4% from baseline per hour of recording) of at least 12. Patients were excluded from the study for severe daytime sleepiness, severe hypoxemia, or Cheyne–Stokes respiration.

After a mean follow-up of 3.7 years, no significant effect on any individual or other composite cardiovascular endpoint was observed between the CPAP group and the usual care group. The authors concluded that CPAP plus usual care, as compared with usual care alone, did not prevent cardiovascular events in patients with moderate to severe OSA and established cardiovascular disease. However, in the CPAP group there was significantly reduced snoring and daytime sleepiness, improved health-related quality of life and mood, and a suggestion that the stroke risk was lower.

Limitations of SAVE

The authors acknowledged the many known adverse effects of OSA and hypoxemia on the cardiovascular system, the benefits of CPAP, and recognized that the short duration of CPAP usage may have been insufficient to provide beneficial effects on cardiovascular outcomes. They referenced other similar studies that reported better cardiovascular outcomes among patients who were adherent to CPAP than patients who did not receive CPAP or were not adherent.

Before definitive conclusions are drawn from any study and widespread practice parameters are changed, it is important to understand not only the study’s limitations but also the patient populations to which the results apply. Here are some limitations of SAVE.

Duration of therapy. A major limitation (recognized by the authors) is the mean CPAP duration of 3.3 hours per night. In this writer’s opinion, even the so called “rule” of 4 hours/night for at least 70% of the nights (primarily used to justify payment or occupational fitness and apparently adopted without objective basis) is not sufficient PAP adherence. The goal should be all night every night.

A subsequent review of the SAVE trial addressed several limitations including the mean duration of CPAP use and the specific portion of the night that CPAP was used as related to REM sleep (when respiratory events and desaturations may be longer, more frequent, and more severe).2 These and other limitations have also been recently addressed by others.3

Limited diversity of patients. Most of the patients were Asian and only 1 US site with 4 patients included. Both risk factors and response to treatment for OSA, as well as frequency and types of comorbid conditions, may vary among ethnic groups.

There is substantial evidence that OSA is a heterogeneous condition even within the same degree of severity and clinical presentation. Multiple subgroups exist with distinctly different phenotypes, variability of clinical symptomology such as the frequency and severity of excessive daytime sleepiness, type of disturbed sleep or insomnia, and specific physiological issues such as age, airway collapsibility, arousal threshold, and stability of respiratory control.4

Effects of O2 desaturation. Oxidative stress due to OSA results in serious cardiovascular damage long before cardiac symptoms appear. If OSA is not diagnosed and treated, damage is progressive and often irreversible.5 The endpoints to determine CPAP effectiveness in SAVE and the conclusions are not a reason to withhold CPAP in any patient with OSA with or without cardiovascular disease.

Technical limitations. HST and/or data downloads from PAP machines were used to determine adherence, frequency of respiratory events, and ODI. These methods do not accurately determine true sleep time and staging. Parameters based on HST recording time or PAP machine runtime can result in significant over and/or underestimation of “true” parameters, and the lower the sleep efficiency the worse the inaccuracy. Furthermore, in this writer’s opinion, the use of automated algorithms to analyze HST signals is not sufficient and data should be scored manually.

Known Cardiovascular Benefits of CPAP

An analysis of 7 prospective cohort studies with a total of 1,087 patients showed that the use of CPAP in patients with OSA is associated with a significant reduction in atrial fibrillation (AF) recurrence independent of medical therapy, irrespective of pulmonary vein isolation, and suggested that CPAP has the potential to become another major treatment pathway for AF reduction in patients with OSA.6


Ed Michaelson, MD


In a cohort of more than 3 million US veterans, both untreated and treated OSA was linked with higher mortality risk, compared with OSA negative patients; however, untreated OSA was associated with a significantly higher risk of both incident coronary heart disease and chronic kidney disease.7

Using the primary endpoints of the first event of repeat revascularization, myocardial infarction, stroke, or cardiovascular mortality, a randomized controlled trial (RICCADSA) conducted in Sweden concluded that routine treatment of non-sleepy OSA patients with CAD did not significantly reduce long-term adverse cardiovascular outcomes. However, there was a significant reduction after adjustment for baseline comorbidities and compliance with the treatment.8

Conclusions, Questions, and Future Research

The heterogeneous nature of OSA patient populations and subgroups is important for both design and interpretation of study results. Future studies should address questions such as how to define OSA severity (eg, what combination of AHI and oxygen desaturation); what constitutes adequate therapy or improvement from CPAP and based on what parameters (eg, what combination of AHI, oxygen saturation, performance, and subjective improvement) and what can be reasonably concluded from such endpoints; what type of interface is used and whether it may have compromised therapy effectiveness9;and whether there is a CPAP dose-response based on adherence or other parameters to achieve cardiovascular or other benefits. Conclusions should be interpreted in context and with emphasis of limitations since misleading conclusions can diminish the value of information.

It is likely that awareness of the SAVE trial decreases as the provider specialty moves from sleep medicine to cardiology to internal medicine and primary care and, in some cases, due to media attention, patients may be more aware of SAVE than their provider. So how should sleep physicians address these issues among their colleagues and with their patients in situations where cardiologists may be less inclined to refer patients to a sleep specialist or patients are looking for a reason not to accept or continue with CPAP? Here are a few suggestions.

  • Explain that the study has had a lot of attention in the media but that important details have been omitted and address some of the limitations described in this article.
  • Tell them about the proven adverse effects of untreated OSA on comorbid conditions such as hypertension, diabetes, obesity, and AF, which if untreated will get progressively worse, leading to earlier onset of heart failure, stroke, and death.
  • Emphasize the other known benefits of CPAP, which include better more restful sleep, reduced daytime sleepiness, improved mental and physical performance, and a better quality of life.


Is the SAVE Study Influencing Your Practice?

Ninety-six healthcare providers took Sleep Review’s “SAVE Trial-Provider Survey” between Dec 7 and Dec 22, 2016. Providers identified themselves as physician (45%), RPSGT (28%), RT (15%), and other (12%, which included dentists, nurse practitioners, and others). Half of the physicians (50%) identified as sleep medicine-certified, 21% as pulmonary, 10% as neurology, 8% as internal medicine, and only a few as cardiology, family practice, ear-nose-throat, critical care, and behavioral sleep medicine. Ninety-one percent of respondents routinely screen patients for sleep apnea, and 95% of respondents treat patients suffering from sleep apnea.

Relating to the SAVE trial, the findings were as follows.

Awareness of SAVE

  • 32% had in-depth awareness of the SAVE trial (such as reading the published paper) prior to receiving the Sleep Review survey
  • 26% had cursory awareness of SAVE prior to receiving the survey
  • 42% had no awareness of SAVE prior to receiving the survey

Patient Discussion Plans

  • 93% of respondents now say they are discussing or plan to discuss the SAVE trial with patients and felt the following issues should be addressed (respondents could select more than one).
    -Emphasize other (non-cardiovascular) benefits of CPAP (69%)
    -Explain limitations of SAVE trial (69%)
    -Address it via other means (15%)
  • Most respondents focused their comments on the SAVE trial’s mean CPAP usage of 3.3 hours per night. Here’s a sampling of comments:
    -“The SAVE trial concludes that using CPAP for 3.3 hours per night does not improve CV outcomes. This is like saying that taking half a dose of hypertension medicine does not address CV disease, which is plausible, but says nothing about the efficacy of a full dose of the medicine. The SAVE trial does nothing to advance the understanding of the relationship between CV disease and CPAP if the patient actually complies with their therapy. The takeaway should be the importance of CPAP compliance, not calling into question the effectiveness of a full dose of CPAP, when only half a dose was applied to this patient population.”
    -“Use of less than 4 hours of CPAP per night unlikely to provide risk reduction of stroke, but that routine use during entire sleep period may reduce that risk.”
    -“Post hoc adherence analysis suggests there may be a benefit to treating OSA for cerebrovascular events. Also, SAVE was secondary prevention of cardiovascular events and PAP may provide primary prevention.”


Edward D. Michaelson, MD FACP FCCP FAASM, is based in Ft Lauderdale, Fla, and a member of Sleep Review’s Editorial Advisory Board.

1. McEvoy RD, Antic NA, Heeley E, Luo Y, Ou Q, et al. CPAP for Prevention of Cardiovascular Events in Obstructive Sleep Apnea. N Engl J Med. 2016 Sep 8;375(10):919-31.
2. Mokhlesi B, Ayas NT. Cardiovascular Events in Obstructive Sleep Apnea—Can CPAP Therapy SAVE Lives? N Engl J Med. 2016 Sep 8;375(10):994-6.
3. Collop N, Stierer TL, Shafazand S. SAVE Me From CPAP. J Clin Sleep Med. 2016 Nov 14.
4. Ryan CM, Kendzerska T, Wilton K, Lyons OD. The Different Clinical Faces of Obstructive Sleep Apnea (OSA), OSA in Older Adults as a Distinctly Different Physiological Phenotype, and the Impact of OSA on Cardiovascular Events after Coronary Artery Bypass Surgery. Am J Respir Crit Care Med. 2015 Nov 1;192(9):1127-9.
5. Chan MP, Antonio N, Chan AQ. Sleep Apnea and Cardiac Remodeling, a Call for Routine Echocardiogram. SLEEP Suppl April 2014. A357.
6. Shukla A, Aizer A, Holmes D, Chinitz L. Effect of Obstructive Sleep Apnea Treatment on Atrial Fibrillation Recurrence. JACC Clinical Electrophysiology. March 2015. 1(1-2):41-51.
7. Molnar MZ, Mucsi I, Novak M, Szabo Z, Freire AX, et al. Association of incident obstructive sleep apnoea with outcomes in a large cohort of US veterans. Thorax. 2015 Sep;70(9):888-95.
8. Peker Y, Glantz H, Eulenburg C, Wegscheider K, Herlitz J, et al. Effect of Positive Airway Pressure on Cardiovascular Outcomes in Coronary Artery Disease Patients with Nonsleepy Obstructive Sleep Apnea. The RICCADSA Randomized Controlled Trial. Am J Respir Crit Care Med. 2016 Sep 1;194(5):613-20.
9. Andrade RGS, Madeiro F, Piccin VS, Moriya HT, Schorr F, et al. Impact of Acute Changes in CPAP Flow Route in Sleep Apnea Treatment. Chest. 2016;150(6):1194-1201.