Zepbound and sleep apnea, the real numbers

Obstructive sleep apnea is a condition where the soft tissue at the back of the throat repeatedly collapses or narrows during sleep, partially or fully blocking airflow. Each blockage forces the brain to briefly rouse the body enough to reopen the airway, which happens without full waking awareness but still fragments sleep, drops blood oxygen, and stresses the cardiovascular system, often dozens of times an hour, night after night. Left untreated, OSA is linked to hypertension, coronary heart disease, stroke, heart failure, atrial fibrillation, and type 2 diabetes — it's a disease with real downstream consequences, not just a snoring nuisance.

The standard way to measure OSA severity is the apnea-hypopnea index, or AHI. It's calculated during a sleep study — either an overnight polysomnogram in a lab or a home sleep apnea test — and represents the average number of apneas (complete pauses in breathing) and hypopneas (partial reductions in airflow, usually paired with an oxygen drop or an arousal) per hour of sleep. A single AHI number is the primary way clinicians classify severity and track whether a treatment is working.

The clinical cutoffs are fairly standardized: an AHI under 5 events per hour is considered normal, 5 to 14 is mild OSA, 15 to 29 is moderate, and 30 or more is classified as severe. Someone with an AHI of 45, for example, is experiencing a breathing disruption roughly every 80 seconds of sleep, on average — which gives some sense of just how disruptive severe OSA is to actual sleep architecture, independent of whether the person remembers waking up.

Zepbound's OSA indication specifically targets the moderate-to-severe range — an AHI of 15 or higher, confirmed by a sleep study, in a patient who also has obesity. That's a deliberately narrower population than "everyone who snores" or "everyone who feels tired," and it's worth understanding that distinction before assuming the approval applies broadly to any sleep complaint.

The scale of undiagnosed OSA is part of why this approval got so much attention. Lilly's own trial materials cite an estimated 80 million adults in the US living with OSA, with more than 20 million falling into the moderate-to-severe category the indication targets — and, strikingly, roughly 85% of OSA cases going undiagnosed and therefore untreated altogether. A meaningful share of people with moderate-to-severe disease don't know they have it, which is part of why a documented sleep study, not a guess based on symptoms, is the starting point for this specific indication rather than an afterthought.

SURMOUNT-OSA was actually two parallel phase 3 studies run under one master protocol (NCT05412004), both randomized, double-blind, and placebo-controlled, both running 52 weeks, both enrolling adults with obesity and a confirmed moderate-to-severe OSA diagnosis. What separated the two studies was PAP status at enrollment — and that split wasn't incidental, it was the whole design logic, because the two populations needed to be measured differently.

Study 1 enrolled 234 participants who were not using PAP therapy — either unwilling or unable to tolerate it, which describes a large share of real-world OSA patients, since PAP adherence is notoriously difficult to sustain long-term. Of those, 114 were randomized to tirzepatide and 120 to placebo. This study answers a fairly direct question: in people getting no other OSA treatment at all, does tirzepatide alone move the needle on AHI.

Study 2 enrolled 235 participants who were already using PAP therapy for at least three months before screening and who continued using it for the full 52 weeks of the trial. Of those, 120 were randomized to tirzepatide and 115 to placebo. This study asks a different question: in people who are already on standard-of-care treatment, does adding tirzepatide produce further improvement on top of what PAP is already providing.

Across both studies, tirzepatide was titrated to each participant's maximum tolerated dose of either 10 mg or 15 mg once weekly, starting at 2.5 mg and increasing by 2.5 mg every four weeks — the same titration approach used in Zepbound's weight-management trials. The primary endpoint in both studies was the same: change in AHI from baseline to week 52, measured against placebo. Total enrollment across the two studies was 469 participants, at sites across the US, Australia, Brazil, China, Czechia, Germany, Japan, Mexico, and Taiwan.

These are the efficacy-estimand figures Lilly reported from the trial — the numbers reflecting drug effect prior to any treatment discontinuation. Baseline AHI is the mean across each study's full population before treatment began; the "approx. AHI at 52 weeks" row is calculated by subtracting the reported mean reduction from that baseline, rounded to the nearest whole number, since Lilly reported the change rather than a single endpoint mean directly.

A note on precision, because it matters here: these are group means, not a promise about any individual's outcome, and they're rounded from Lilly's reported figures. "Approx. AHI at 52 weeks" is a derived estimate (baseline minus mean reduction), not a separately reported endpoint value — the underlying published data reports the change, not a re-stated endpoint mean, so treat that row as a reasonable approximation rather than a trial-reported statistic. What is directly reported and more precise: in Study 1, 62.3% of tirzepatide participants achieved at least a 50% reduction in AHI, versus 19.2% on placebo; in Study 2, that was 74.3% versus 22.9%. "Disease resolution" in the trial meant either an AHI under 5, or an AHI of 5–14 combined with a low score on the Epworth Sleepiness Scale, a standard daytime-sleepiness questionnaire.

Both studies also tracked secondary cardiometabolic markers — systolic blood pressure, a measure of overnight low-oxygen burden, and high-sensitivity C-reactive protein, an inflammation marker — and tirzepatide showed significant improvement over placebo on all of them. That matters because OSA's health risk isn't just about daytime fatigue; it's tied to the same cardiovascular and metabolic pathways that weight-related disease already stresses.

The leading physiological explanation connecting tirzepatide's weight loss to its AHI improvement is mechanical: excess fat tissue accumulates not just around the waist but around the neck, tongue base, soft palate, and the walls of the upper airway itself. That extra tissue narrows the airway's resting diameter and makes it more prone to collapsing inward when the surrounding muscles relax during sleep, which is exactly the physical event that produces an apnea or hypopnea. Reduce the surrounding fat mass, and there is measurably more room in the airway before it collapses.

Neck circumference is, in fact, one of the more consistent predictors of OSA severity independent of overall body weight — two people with similar BMIs can have meaningfully different AHI scores depending on how their fat is distributed, with more central, upper-body fat tracking more closely with airway obstruction. That's consistent with why substantial weight loss, whether from bariatric surgery, intensive lifestyle intervention, or a GLP-1/GIP medication like tirzepatide, has shown up as an effective lever against OSA severity across different treatment modalities over the years, not just in this specific trial.

That said, researchers running SURMOUNT-OSA noted that the AHI improvement observed appeared somewhat larger than some earlier weight-loss studies (including bariatric surgery data) would have predicted from weight change alone, which has raised the open question of whether tirzepatide's effect on OSA runs partly through some additional pathway beyond simple mass reduction — possibly related to fluid shifts, reduced systemic inflammation, or effects on respiratory drive and neuromuscular tone during sleep. That mechanism hasn't been fully isolated or confirmed, and it's an active area of ongoing research rather than a settled explanation.

This is the single most important practical caveat in the entire trial, and it's easy to miss if you only see the headline percentages: everyone in Study 2 kept using PAP therapy for the full 52 weeks. The trial did not test whether tirzepatide works as well as PAP, or whether it's safe to swap one for the other — it tested whether adding tirzepatide on top of existing PAP use produces further improvement. It did. But that's a fundamentally different question from "can this replace my machine."

Study 1 shows that tirzepatide alone, with no PAP at all, produced a substantial AHI reduction in people who weren't using PAP to begin with — which is meaningful, especially given how many people with OSA never successfully adopt or sustain PAP therapy in the first place. But even in that group, the average AHI at 52 weeks in the tirzepatide arm remained well above the "normal" threshold of under 5 events per hour for the group as a whole, even though a substantial share of individuals reached resolution-level improvement. Group averages and individual outcomes aren't the same thing, and standard-of-care PAP therapy remains the established first-line treatment for moderate-to-severe OSA.

The honest, current-evidence framing is that Zepbound looks like a genuinely useful adjunct or complementary treatment for moderate-to-severe OSA in people with obesity — something that can meaningfully lower AHI on its own or add further benefit on top of PAP — rather than an established substitute for PAP therapy. Any decision to reduce, pause, or stop PAP or CPAP use should be made together with a sleep medicine physician, generally anchored to a repeat sleep study confirming where your AHI actually stands, not to how rested you happen to feel on a given morning. Subjective improvement is a reasonable thing to mention to your doctor; it is not, on its own, a substitute for measured data when the disease in question carries real cardiovascular risk.

The OSA indication is distinct from Zepbound's original weight-management approval, and eligibility isn't automatic just because someone is already prescribed Zepbound for weight loss. The label covers adults with obesity who have a confirmed diagnosis of moderate-to-severe obstructive sleep apnea — meaning a sleep study showing an AHI of 15 events per hour or higher, not a self-reported sense of poor sleep or loud snoring reported by a partner.

In practice, that means getting this indication generally starts with a referral to a sleep medicine specialist and a formal sleep study — either an in-lab polysomnogram or, increasingly common, a home sleep apnea test for straightforward cases. Without a documented AHI in the moderate-to-severe range, a prescriber would be treating under the general weight-management indication rather than the OSA-specific one, which can matter for how the diagnosis is coded, how insurance evaluates medical necessity, and how outcomes get tracked going forward.

People already on PAP therapy are not excluded from this indication — quite the opposite, since Study 2 specifically demonstrated benefit in exactly that population. Someone with moderate-to-severe OSA and obesity who is struggling with PAP tolerance, or who has coverage or access barriers to a CPAP machine, or who is already using PAP but still has significant residual disease burden, all represent scenarios where a sleep physician might reasonably discuss whether the OSA-specific Zepbound indication is appropriate, alongside — not instead of — their existing treatment plan.

Because this is a distinct FDA indication, it can also affect insurance coverage discussions in a useful way. Some health plans that exclude weight-management drugs from their formulary handle OSA-indicated prescriptions differently, since the documented diagnosis and medical necessity are different from a weight-loss-only claim. That doesn't guarantee coverage — plan design still varies enormously — but it's a distinction worth raising directly with your prescriber and insurer if cost has been a barrier, rather than assuming the answer is automatically the same as it would be for the weight-management indication alone.