Weaver TE. Do we finally know whether CPAP improves quality of life in non-sleepy patients? J Clin Sleep Med. 2019;15(9):1193–1194.
The importance of the outcome of quality of life in disease management is underscored by Engleman in a Shattuck Lecture who stated, “The centerpiece and unifying ingredient of outcomes management is the tracking and measurement of function and well-being or quality of life.”1 Further, in considering measurement of quality to improve care, Morgenthaler and colleagues delineated that measurement of health care quality has utility in identifying opportunities for improvement, assessing the effectiveness of interventions to improve quality and helps to prioritize areas for improvement.2 It is with this intent that Wallström and colleagues in this issue of Journal of Clinical Sleep Medicine, analyzed data from the RICCADSA trial to address the effect of continuous positive airway pressure (CPAP) treatment on quality of life (QOL) in nonsleepy patients with obstructive sleep apnea (OSA) who received interventions for acute myocardial infarction.3
It has been solidly demonstrated that in sleepy patients with OSA, CPAP treatment improves QOL, even in those with milder disease.4–6 However, the question remains if CPAP therapy improves quality of life in those who do not manifest sleepiness. About 27% of patients with OSA do not experience sleepiness and many of these are those with coronary artery disease (CAD).7 Thus, the documentation of the effect of CPAP in those who are not sleepy, especially those with the comorbidity of CAD, is of keen interest to the field. Using the SF36 to measure quality of life, Wallström and coworkers compared participants with OSA, an Epworth Sleepiness Scale score < 10, and had undergone treatment for coronary artery disease who received CPAP therapy against those with the same characteristics, but had not received CPAP.3 Although half of the SF36 within group subscale scores improved with CPAP, there were no statistically significant differences between groups.
The data from Wallström et al’s secondary analysis is plagued by the same limitations common to previous studies answering the question of whether CPAP makes a difference in nonsleepy patients with OSA and cardiovascular disease. Whether participants in this study would be classified as having CAD having received treatment for their arterial occlusion, could come into question. Nevertheless, even with the intervention, they remain a nonsleepy OSA phenotype where demonstrating efficacy or treatment effectiveness, as is the case with the comparison to usual care, remains challenging when symptoms are not present. Indeed, the mean values for both experimental groups at baseline tended to be what might be considered within the normal range for the SF36. This brings to the forefront the primary obstacle to determining whether CPAP makes a difference in nonsleepy individuals—how do you make individuals receiving CPAP treatment who have no impairment more normal? That is, if the score value is normal for a metric, you cannot illustrate utility, in the terms described by Morganthaler,2 if there is limited bandwidth to improve the score.
The second threat to the ability to detect change is sufficient statistical power, related to the potential range to normalize QOL scores. With 102 and 104 participants per group (CPAP versus no CPAP) in the secondary analysis of the RICCADSA trial, adequate power was attained for the SF36 Mental Component Summary Scale (89% power), but not for the Physical Component Summary Scale (36% power).3 In contrast, the SAVE study, almost 3,000 participants with cardiovascular disease and a mean Epworth Sleepiness Scale value < 8, showed adequate power finding significant differences with CPAP treatment compared to usual care for both the SF36 Physical and Mental Component Summary Scales.8 Where previous research, for the most part, failed to illustrate benefit of CPAP treatment in those without sleepiness, this study succeeded principally due to the power generated by the large sample size.6
Finally, optimal treatment exposure is a necessary requisite to a successful clinical trial. This has been the nemesis of randomized controlled trials testing the efficacy/effectiveness of CPAP either against a placebo or usual care. Indeed, the average mean nightly use of active treatment in such studies has been less than 5 hours, which is considerably below the 7 hours needed to produce a normal value in daily functioning.9,10 Only 47.2% of participants in the Wallström et al study had nightly use of at least 5 hours with an overall sample mean nightly use of 4.5 ± 2.3 h/night on all nights.3 Use greater than 4 hours did not alter the findings of this analysis, although CPAP use was weakly related to improvement in Mental Component Summary score, but negatively correlated to the Physical Component Summary score. Of course, it is acknowledged that low nightly use often occurring in clinical trials reflects what is typically seen in clinical practice and has spurred the identification of interventions to promote higher use. However, including all levels of use provides a more real-world portrayal of the impact of CPAP therapy.
We will never be able to conclusively determine whether CPAP improves QOL in nonsleepy patients with CAD unless we have adequately powered studies with > 5 hours use. Large sample sizes are needed to compensate for low use or have proportionately greater numbers of participants with higher levels of nightly use. Obtaining adequate power and sample size will no doubt necessitate costly multisite studies, and likely be international. Execution of such clinical trials will necessitate funding from international government agencies in the form of cooperative agreements or substantial philanthropic entities. In addition to treatment effectiveness comparisons, randomized controlled trials will provide the bottom line regarding the efficacy of CPAP to improve QOL, but only with the utilization of appropriate placebos.
Dr. Weaver is a consultant for Jazz Pharmaceuticals, receives royalty fees for use of the FOSQ from: Philips Respironics, Nyxoah, Bayer AG, ResMed, ResMed Germany, Jazz Pharmaceuticals, Cook Medical, Merck Col Inc.