To our knowledge, this is the first study to evaluate the effect of liraglutide in people with obesity, COPD, and OSA. We examined the effect of liraglutide (3.0 mg) for 40 weeks on measures related to HRQoL, OSA, and sleep. To obtain specific results, we stratified the participants in subgroups depending on the presence and severity of OSA and whether they were treated for OSA with CPAP.
In obese individuals with COPD and mild OSA (subgroup 3), 40 weeks of liraglutide compared to placebo reduced AHI by 9.25 points. In fact, this reduction downgraded the condition from mild OSA to almost a normal state. There was no significant change in ODI in this subgroup. Among obese individuals with COPD and moderate-to-severe OSA who started treatment for OSA with CPAP (subgroup 5), there was a significant decline in ODI of 12.4 in the liraglutide group compared to placebo. There was no significant change in AHI in this subgroup.
CPAP is the standard treatment for moderate-to-severe OSA. However, an oral appliance (OA) is a useful alternative or additional treatment option for mild, moderate, and severe OSA [20]. A mandibular advancement device is a sort of OA, where the lower jaw is advanced and attached to the teeth [21]. Additionally, it is well known that weight loss improves OSA-related parameters especially in moderate-to-severe OSA, so weight loss is recommended as an initial treatment. Intensive lifestyle intervention over 10 years improved OSA severity due to change in body weight and AHI compared to baseline [22]. Furthermore, a study of obese individuals with moderate-to-severe OSA reported that liraglutide reduced AHI by 6.1 events per hour compared to placebo [13]. This finding is in line with our results on reduced AHI, though we found the reduction in AHI in subgroup 3 (obese individuals with COPD and mild OSA). In a bariatric surgery cohort, AHI declined by 9.1 events per hour due to weight loss after the surgery [23]. In all three studies, the participants in question were not treated with CPAP. Regarding ODI, in another obese population, a weight loss program combined with CPAP or OA in obese people with severe OSA reduced ODI [24]. This is also in line with our results from subgroup 5 (obese individuals with COPD and moderate-to-severe OSA who started treatment for OSA with CPAP), where there was a reduction in ODI in the liraglutide group compared to placebo. However, the mechanisms underlying the effects of weight loss (in the previous study) [24] and the presence and impact of COPD (in our study) differ.
The co-existence of OSA and COPD depends on the COPD phenotype: increased lung volume, low BMI, and emphysema protect against OSA, while increased BMI with bronchitis promote OSA. The latter phenotype describes our population. OSA occurs in 65.9% of people with moderate-to-severe COPD, so it is a common comorbidity and also a challenge in this population [4, 25].
We did not find a significant change in ESS overall between the liraglutide and placebo groups. A similar result occurred in the abovementioned bariatric surgery study, in which weight reduction achieved by bariatric surgery did not affect ESS [23]. Another study showed that ESS did not change in people with the triad of OSA, COPD, and obesity [26]. This population is comparable to our population in terms of the presence of obesity, COPD, and OSA.
In general, quality of life is lower in COPD individuals than non-COPD individuals [27]. Quality of life is also reduced in overweight and obese people, with lower PCS, GH, and PF [28]. Our participants had a mean PCS of 40.1 ± 5.6 and a mean MCS of 42.9 ± 9.2, which are higher than both the PCS (26.1) and MCS (27.9) in COPD patients referred to pulmonary rehabilitation [29] and higher than the PCS (35.7) and lower than the MCS (48.3) in normal-weight COPD patients with pain complaints [30]. Regarding the second study, interference of pain with physical activities may explain the lower PMC [30] in the second study, while decreased MCS is common in obese people with chronic conditions [15] such as in our study.
OSA also impacts HRQoL. In a cohort study with different severities of OSA, severe OSA, higher age, and higher BMI were associated with lower SF-36 PCS scores [31]. In people with COPD, obesity is common and the number of comorbidities increases with the class of obesity. This combination is associated with worse COPD-related outcomes such as impaired quality of life, and weight loss may impact COPD-related outcomes in obese people with COPD [32]. We found significant improvements in two of the eight health domains of SF-36v2: GH and RP. Although these changes were significant, the findings were at the border of clinical importance. We found a 13.1-point improvement in ΔGH and a 9.1-point improvement in ΔRP (from baseline to the end of the medication period) in liraglutide group compared to placebo group. Studies in people with COPD have suggested clinically important differences are 15 for GH and 12.5 for RP [33]. However, some individuals might experience clinically relevant improvement.
COPD and obesity are inflammatory diseases and both predispose individuals to OSA. Intermittent hypoxemia is comparable to ischemia-reperfusion injury, leading to oxidative stress and contributing to reactive oxygen species and inflammatory mediator production, which triggers upper airway and systemic inflammation [17]. Our cardiorespiratory recordings revealed unknown/undiagnosed cases of OSA in 32 of the participants (84%). Undiagnosed OSA has been linked to neurocognitive impairment and cardiovascular morbidity, and it is a well-known public health concern [17]. It is likely that the presence of OSA affects HRQoL in this obese population with COPD and any accompanying treatment will improve HRQoL in this group.
Strengths and limitations
The strengths of our study are the randomized controlled design combined with both objective (cardiorespiratory monitoring) and subjective (SF-36v2 and ESS) measures.
We included and randomized 40 participants, but ended up with 30 completing participants, due to comorbidities diagnosed during the study, invalid cardiorespiratory data, and dropouts. Our reduced sample size limits the strength of our study (power calculation indicated the need for 32 participants). Some participants were diagnosed with OSA, but either they declined OSA treatment (CPAP) or discontinued treatment due to side effects (subgroup 4). Due to the differences in receiving treatment in the various subgroups, the results should be considered with caution. It is conceivable that, eventually, continued treatment would affect the results further.