Uncovering the effect of CPAP on cardiovascular outcomes in obstructive sleep apnea
SLEEP · 09.09.2024 · 47 (9)
DOI: 10.1093/sleep/zsae153 ↗Kurz gefasst
Editorial in der Zeitschrift SLEEP, online erschienen am 09.07.2024. Verfasst von Bhajan Singh und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Pulmonary Physiology & Sleep Medicine, Sir Charles Gairdner Hospital, Nedlands, WA, Australia. Hauptthemen laut MEDLINE: Sleep Apnea, Obstructive, Continuous Positive Airway Pressure, Cardiovascular Diseases. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 26 Literaturangaben.
Abstract (Originalsprache)
This issue of SLEEP features a timely and important Pro–Con debate on how to assess the effectiveness of continuous positive airways pressure (CPAP) in reducing major adverse cardiovascular events (MACE) in patients with obstructive sleep apnea (OSA) [1, 2]. The debate is adversarial by design but serves to highlight a lack of consensus on this important question. Proponents and opponents of this debate provide compelling arguments for and against the value of studies that estimate the effect of CPAP on MACE in observational cohorts by using CPAP adherence to define intervention and control groups. Prior to discussing the arguments raised, let us reflect on areas of general agreement. It is generally accepted that patients with OSA are at increased MACE risk [3–5], OSA is associated with physiologic insults known to cause vascular injury [6–8] and CPAP prevents such insults during sleep [9]. Many long-term observational studies have found that CPAP reduced MACE risk [9] but these studies could have overestimated the effect of CPAP because of limitations in design and failure to consider potentially important confounding variables. In contrast, randomized controlled trials (RCTs) found that CPAP did not reduce MACE risk [9]. There is wide agreement that RCTs are the optimal study design to assess the effectiveness of any therapy because the random allocation of participants to an intervention should minimize bias. Evidence from nonrandomized studies is regarded as lower quality and is less likely to influence treatment decisions and clinical guidelines [10]. However, limitations in RCTs examining the role of CPAP in OSA could have underestimated the effect of CPAP on MACE risk [11]. As a result, we remain uncertain about the effect of CPAP on MACE risk and whether it is possible to conduct an unbiased RCT to answer this question. Debates such as these are essential to direct future research capable of providing evidence that can be widely accepted. The proponents highlight some of the important limitations of published RCTs. Given the low adherence to CPAP in RCTs, the proponents argue that RCTs are answering the question of whether assigning CPAP rather than whether using CPAP improves MACE risk. While intention-to-treat analyses (e.g. assigning CPAP) preserve the random allocation of intervention [12], the proponents argue that we should be primarily interested in per-protocol effects (e.g. using CPAP). A post hoc individual-level meta-analysis of recent CPAP trials found that CPAP adherence was associated with reduced MACE risk [13]. The proponents also highlight participant selection biases of these RCTs that have largely enrolled less sleepy participants atypical of sleep clinic patients. They question the ability of RCTs to overcome these limitations because of the difficulty of withholding effective CPAP therapy for extended periods in patients with sleep disturbance, excessive sleepiness, reduced productivity, and impaired quality of life. They advocate an alternate approach of comparing CPAP adherent with non-adherent patients in observational cohorts balanced for a rich set of MACE risk factors (including exercise and diet) and other health behaviors (e.g. adherence with vaccinations and cancer screening guidelines). Substantial detail about how to achieve this is provided. In contrast, the opponents of this debate contend that it is impossible to overcome the bias associated with adherence using other measures of healthy behavior. They highlight that adherence to placebo is consistently associated with lower MACE risk in drug trials and contend that this healthy adherence effect is likely with CPAP therapy and would lead to overestimating the benefit of CPAP on MACE risk. While this raises important questions about the validity of using CPAP adherence in observational cohorts to examine MACE risk, a compelling explanation of how to conduct an unbiased RCT has not been provided. Although the debate has not led to a consensus, it has highlighted important future directions. Before embarking on observational studies comparing CPAP adherent versus non-adherent, we need to better understand the relationship between CPAP adherence, medication adherence, and other healthy behaviors. We need to know whether CPAP non-adherent patients are also likely to be non-adherent to prescribed medication and other health behaviors, and have a higher MACE risk than CPAP adherent patients due to characteristics unrelated to CPAP. In addition, to make valid causal inferences, observational studies need meticulous designs to overcome bias using advanced techniques such as those described by the proponents rather than relying on matching alone. Other strategies to improve observational studies in the context of CPAP and MACE risk have been proposed and include the use of instrumental variables, differences in differences and regression discontinuity designs [14]. In addition, approaches to emulate RCTs have been described [15–17], but are yet to be assessed in the context of CPAP and MACE risk. Before embarking on further RCTs, we should address the design flaws of published studies. Study populations should be more representative of patients attending sleep clinics [18]. This requires the enrollment of patients with OSA symptoms. To do this, RCTs may need to include measures to alleviate OSA symptoms that do not modify upper airway patency or MACE risk. For example, it may be necessary to treat participants with stimulant medication that is effective, safe, and devoid of significant cardiovascular effects to alleviate excessive sleepiness. Newer stimulants such as pitolisant appear promising in this regard [19]. We may also need to find measures to alleviate other troublesome OSA symptoms such as sleep disruption and morning headaches. Study populations should also consider including younger patients without preexisting cardiovascular disease to examine the role of CPAP in the primary rather than secondary prevention of MACE. The relationship between OSA and MACE is weaker in the elderly [20, 21], and it is possible that CPAP has a more important role in preventing the development of vascular damage rather than its progression. Trials focusing on improving CPAP adherence through patient and provider education, and reducing OSA through any available therapy, not only CPAP, should also be considered. Finally, studies assessing the effects of active comparators that are part of the OSA clinical care pathway, such as weight loss therapies may also mitigate some limitations of published RCTs. There are additional considerations for future observational and randomized studies. We should consider selecting patients with OSA based on physiologic derangements associated with vascular damage such as impaired oxygenation (e.g. oxygen desaturation indices, hypoxic burden [22]) or exaggerated autonomic responses (e.g. OSA-specific pulse-rate response [23]) rather than the apnea–hypopnea index. It may also be necessary to explore the role of CPAP on specific MACE outcomes such as coronary events, heart failure, or stroke as the effect of CPAP on individual outcomes may be inconsistent [24]. Ultimately, all approaches assessing the effect of CPAP on MACE risk have strengths, weaknesses, and potential biases. Rather than relying on a single approach, we could consider strategies that support the triangulation of evidence [25]. This allows us to consider evidence generated from multiple angles for a more holistic understanding of the question under investigation. This may be achieved by combining information from well-designed RCTs and observational studies using robust causal inference approaches. While the hierarchy of evidence-based medicine is conventionally represented as a “pyramid,” improvements in trial design and causal inference methods suggest that evidence generation may be more nuanced than originally proposed [26]. To get to the truth, we may need to consider collecting and evaluating every piece of robust evidence. We commend the contributors of this Pro–Con debate for helping to frame the challenges ahead. None.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Department of Pulmonary Physiology & Sleep Medicine, Sir Charles Gairdner Hospital, Nedlands, WA, Australia.
- Faculty of Human Sciences, University of Western Australia, Crawley, WA, Australia.
- West Australian Sleep Disorders Research Institute, QEII Medical Centre, Nedlands, WA, Australia.
- Division of Medical Informatics, Department of Internal Medicine, University of Kansas Medical Center, Kansas City, KS, USA.
- Division of Pulmonary Critical Care and Sleep Medicine, Department of Internal Medicine, University of Kansas Medical Center, Kansas City, KS, USA.
ORCID-Profile: Diego Robles Mazzotti
Zitieren
APA
Singh, B., & Mazzotti, D. R. (2024). Uncovering the effect of CPAP on cardiovascular outcomes in obstructive sleep apnea. SLEEP, 47(9). https://doi.org/10.1093/sleep/zsae153
Vancouver
Singh B, Mazzotti DR. Uncovering the effect of CPAP on cardiovascular outcomes in obstructive sleep apnea. SLEEP. 2024;47(9). doi:10.1093/sleep/zsae153
BibTeX
@article{Singh2024,
author = {Singh, Bhajan and Mazzotti, Diego Robles},
title = {Uncovering the effect of CPAP on cardiovascular outcomes in obstructive sleep apnea},
journal = {SLEEP},
year = {2024},
volume = {47},
number = {9},
doi = {10.1093/sleep/zsae153}
}
EUVIMED