Analysis finds sleep apnea improvements, not only weight loss, explain some tirzepatide benefits

A new analysis of the SURMOUNT-OSA trials reports that some of the heart and metabolic benefits seen with tirzepatide in people who have obstructive sleep apnea (OSA) and obesity were tied to improvements in the breathing disorder itself, not to weight loss alone ^[1][2]. The paper was published in Nature Medicine ^[2].

What the trials did

SURMOUNT-OSA was a master protocol with two 52-week, randomized, double-blind, placebo-controlled phase 3 studies. Study 1 enrolled people who were not using positive airway pressure (PAP) therapy; study 2 enrolled people who were using it ^[2].

Together, 469 people were randomized: 234 in study 1 (114 to tirzepatide, 120 to placebo) and 235 in study 2 (120 to tirzepatide, 115 to placebo) ^[2]. In both studies, tirzepatide was associated with greater relief of cardiometabolic risk factors than placebo, the authors report ^[1][2].

The change in the apnea-hypopnea index (AHI), a standard measure of sleep apnea severity, was the primary outcome of the original trials. This new paper reports prespecified results for HDL cholesterol, non-HDL cholesterol, LDL cholesterol, VLDL cholesterol and fasting insulin, along with a post hoc analysis of HOMA-IR, a measure of insulin resistance ^[2].

How the mediation analysis worked

A mediation analysis estimates how much of an observed change can be attributed to a specific factor. Here, the researchers looked at how much of each cardiometabolic change was attributable to reductions in body weight, in AHI, and in sleep apnea-specific hypoxic burden, which reflects drops in oxygen during sleep ^[1][2].

Changes in the sleep apnea measures had an independent mediation effect on high-sensitivity C-reactive protein (a marker of inflammation), HOMA-IR and triglycerides. For systolic blood pressure, weight loss alone and the combination of weight change plus sleep apnea changes both had a significant mediation effect; the estimated treatment difference at week 48 in study 1 was −7.9 mmHg ^[2]. For diastolic blood pressure, no significant mediation by weight or by sleep apnea measures was found ^[1][2].

The mediation analyses included 336 people for HOMA-IR, 341 for non-HDL cholesterol, 342 for HDL cholesterol, 341 for triglycerides, 342 for high-sensitivity C-reactive protein, and 364 each for systolic and diastolic blood pressure ^[1].

On the basis of the analysis, the authors conclude that treating both sleep-disordered breathing and obesity is likely required to optimize the treatment effect on cardiometabolic benefits in patients with moderate-to-severe OSA and obesity ^[1][2].

Several authors report consulting fees, advisory board roles or other income from Eli Lilly, which develops tirzepatide, and from other companies; four authors are employees and shareholders of Eli Lilly ^[1].

Why it matters for patients

For people who have both OSA and obesity, this analysis suggests the two conditions may contribute to heart and metabolic risk partly through separate routes. Weight loss appears to explain the blood pressure improvement seen in the trial, while better breathing measures appear to explain part of the improvement in inflammation, insulin resistance and triglycerides ^[1][2].

PAP has been first-line therapy for OSA but is not always well tolerated, and it has not shown consistent cardiovascular outcome benefits in trials. The authors note that one possible reason is that underlying risk factors such as obesity are not directly addressed by PAP ^[2].

The results do not show that tirzepatide prevents heart attacks or strokes. Whether these marker changes translate into fewer cardiovascular events has not been reported. Mediation analysis estimates proportions within a trial population and does not prove cause and effect, and the findings apply to people with moderate-to-severe OSA and obesity rather than to all patients ^[1][2].

These trials studied tirzepatide only. Other GLP-1 medications were not tested in SURMOUNT-OSA, and no comparisons between drugs are reported ^[1][2].

The two studies are registered with ClinicalTrials.gov under the number NCT05412004 ^[1][2].