Evaluating Behavioral Weight Loss Programs for Sleep Apnea

Obesity is widely recognized as one of the leading risk factors for obstructive sleep apnea (OSA). Analyses from the Wisconsin Sleep Cohort Study suggest that 41% of all OSA cases, including 58% of moderate-to-severe cases, can be attributed to the presence of a body mass index (BMI) of 25 kg/m² or greater.¹ Importantly, obesity is the only major OSA risk factor that is reversible; however, longitudinal data from the Sleep Heart Health Study demonstrate that the impact of weight loss on the apnea-hypopnea index (AHI) is not simply the opposite of weight gain.² Thus, although a wide range of both longitudinal and interventional studies have established that the relationship between obesity and OSA is causative, the extent to which OSA improves with weight loss is unclear. The most recent American Academy of Sleep Medicine (AASM) guidelines for the management of OSA state that weight loss should be recommended for all overweight patients alongside a primary treatment. The AASM Task Force noted, however, that this recommendation was “a strategy that reflects uncertain clinical use,”³ and the guidelines do not specifically address the use of intensive behavioral weight loss programs as opposed to general weight loss education.

In an attempt to address this uncertainty, Araghi and colleagues⁴ in this issue of SLEEP report a systematic review and meta-analysis investigating the impact of diet and exercise programs on OSA severity (that is, nonsurgical and nonpharmaceutical weight loss approaches). Their resulting paper is an important contribution to the scientific literature, and the authors are to be commended on their thorough literature search, robust study design, and detailed statistical analyses. There was considerable variability in terms of the intervention tested amongst the studies included in their review. Most studies approached weight loss through either diet (standardized low calorie intake dependent on baseline weight, sometimes by use of liquid meal replacement) or exercise (aerobic and/or resistance training and/or breathing exercises), with very few studies incorporating both modalities. The majority of studies incorporated group sessions to encourage adherence.

Overall, summary results of Araghi et al.'s analyses are relatively disappointing. While behavioral weight loss interventions do appear to improve OSA severity, the magnitude of improvement is small. Araghi observed a high degree of heterogeneity across studies, and therefore went on to conduct two meta-regression analyses that are perhaps more informative in terms of guiding future research. It was found that OSA severity improved to a greater extent when the baseline AHI was high, and that the change in weight was predictive of the change in AHI. In Figure 8 of the paper by Araghi, it is apparent that there are three studies in which fairly large reductions in weight (> 15 kg) were obtained, and that these three studies were also those that resulted in the largest reductions in AHI (> 20 events/ hour). Future studies should therefore focus on modifying the content and/or delivery of behavioral interventions in order to achieve maximal weight loss.

It is also noteworthy in the Araghi report that the efficacy of behavioral interventions was much smaller in randomized controlled trials (RCTs) than in uncontrolled studies (pooled mean reductions in AHI of 6.0 and 12.3 events/h, associated with reductions in BMI of 2.3 kg/m² and 4.9 kg/m² in RCTs and uncontrolled studies, respectively). There are a number of potential reasons for uncontrolled studies to overestimate treatment effects, including a lack of intention-to-treat analysis, selection biases, regression to the mean, and greater susceptibility to publication bias. These results further emphasize the importance of conducting well-designed RCTs.

Another interesting finding is that the improvement in the oxygen desaturation index (ODI) tended to be greater than the improvement in the AHI. This may reflect the impact obesity has not only on the frequency of respiratory events, but also the severity of desaturation with each event.⁵ Given research suggesting desaturations may be more relevant to the cardiovascular risk attributable to OSA,⁶ future studies on weight loss should include a measure of oxygen saturation as an endpoint.

As noted by Araghi et al.,⁴ one major weakness of the literature to date on the role of behavioral weight loss interventions for OSA has been the relative lack of assessment of endpoints such as daytime sleepiness or cardiovascular outcomes. The danger of focusing solely on biomarkers such as the AHI or ODI is demonstrated by the recently published Look AHEAD study.⁷ In this landmark study of 5,145 diabetic patients followed for a median of 9.6 years, treatment with an intensive weight loss program led to improvements in glycated hemoglobin, blood pressure, and serum lipids, but had no impact on cardiovascular events.

Thus, the work by Araghi et al.⁴ suggests the potential for behavioral and other weight loss interventions in reducing the severity of OSA, but more must be done before these treatments can be routinely recommended. Firstly, work is needed to further optimize the weight loss strategies, including potentially tailoring the intervention to each patient, in order to maximize the weight loss achieved. Such strategies may need to include other approaches such as pharmaceutical and/or surgical treatments. Secondly, trials need to demonstrate improvements in outcomes of relevance to OSA patients. Thirdly, given the known weight regain that occurs following weight loss interventions, more research is needed on the durability of response and need for re-intervention. On the other hand, long term follow-up from two trials suggests a sustained effect on AHI despite weight regain of approximately 50%.^8,9 These tantalizing findings suggest weight loss may be more effective at slowing the progression of OSA than as a primary treatment. Finally, translational work will be needed to demonstrate both the generalizability of these interventions to typical overweight/ obese OSA patients, as well as the feasibility of implementing such therapies from a cost and infrastructure standpoint.

CITATION

Bakker JP; Patel SR. Evaluating behavioral weight loss programs for sleep apnea. SLEEP 2013;36(10):1419-1420.

DISCLOSURE STATEMENT

The authors have indicated no financial conflicts of interest.