Can a GLP-1 Replace Your CPAP? Zepbound and Foundayo Cut Sleep Apnea Severity, but Most Patients Still Have the Disease After a Year

Two GLP-1 medications now have large randomized trials in obstructive sleep apnea. Zepbound (tirzepatide), a weekly injection, has been approved by the FDA since December 20, 2024 for moderate to severe obstructive sleep apnea in adults with obesity. Foundayo (orforglipron), a daily pill, completed two Phase 3 trials that were published in JAMA on November 9, 2026, and Eli Lilly plans to ask the FDA to approve it for sleep apnea before the end of the year.

Both drugs substantially reduced the number of breathing interruptions per hour of sleep, mainly by helping participants lose weight. Neither program showed that most patients could give up their breathing machine, however. Among participants taking Zepbound, 42% to 50% reached the level that sleep specialists consider remission or mild disease, and Lilly's corresponding figures for Foundayo are 20% to 32%, calculated with a method that generally flatters the results. That leaves half or more of the participants with moderate or severe sleep apnea after a full year of treatment.

The honest answer to "Can a GLP-1 replace my CPAP?" is therefore that it might for some people, but only after a repeat sleep study confirms that the problem has largely resolved. For most patients, the medication works alongside the machine rather than replacing it. This guide explains how sleep apnea is measured, what each trial found, what happens when people stop the drug, the side effects and the new eye warning, and how insurance coverage works. For a complete list of which GLP-1 drug is approved for which condition, see Which GLP-1 Drug Is FDA-Approved for What.

In obstructive sleep apnea, the soft tissue of the throat collapses or narrows repeatedly during sleep. Breathing either stops completely (an apnea) or becomes shallow (a hypopnea), oxygen levels in the blood fall, and the brain briefly rouses the sleeper to restart breathing. These repeated interruptions fragment sleep, and the condition is associated with reduced quality of life and with cardiovascular and metabolic disease.

Sleep specialists grade the condition with the apnea-hypopnea index (AHI), which is the average number of apneas and hypopneas per hour of sleep, measured during an overnight sleep study. Both drug programs used the same thresholds:

• Fewer than 5 events per hour: normal breathing, which the trials counted as remission.
• 5 to 14 events per hour: mild sleep apnea.
• 15 or more events per hour: moderate to severe sleep apnea, the population in which both drugs were tested.

The participants in these trials started with severe disease. Their average AHI was roughly 50 to 58 events per hour, which works out to almost one interruption every minute of sleep.

The trials also tracked hypoxic burden, a score calculated from the sleep study that combines how often blood oxygen drops, how long each drop lasts, and how deep it goes. Researchers follow it because two people with an identical AHI can experience very different oxygen deprivation.

The standard treatment is positive airway pressure (PAP), usually called CPAP, a bedside machine that pushes air through a mask to hold the airway open. It works well when people wear it consistently, but many patients find nightly use difficult, which is one reason a medication option draws so much interest. Each drug program therefore ran one trial in people who were not using PAP and a second trial in people who already used it.

On December 20, 2024, the FDA approved Zepbound for "moderate to severe obstructive sleep apnea (OSA) in adults with obesity," to be used with a reduced-calorie diet and more physical activity. It was the first medication ever approved for the condition, and the FDA stated that the improvement in breathing was likely due to the weight loss the drug produces.

The approval rested on SURMOUNT-OSA, a pair of 52-week randomized trials published in the New England Journal of Medicine that enrolled 469 adults with a body mass index (BMI) of 30 or more and an AHI of 15 or more. People with type 1 or type 2 diabetes were excluded. Participants received either tirzepatide at the highest weekly dose they could tolerate (10 or 15 mg) or a placebo injection, and everyone received counseling on diet and physical activity.

The table below summarizes the main results, counting every participant who started treatment, including those who later stopped taking the drug:

Hypoxic burden, a blood marker of inflammation called hsCRP, and systolic blood pressure also improved significantly more on tirzepatide than on placebo, and participants reported less sleep-related impairment during the day. In trial 2, participants paused their PAP machines for 7 days before each sleep study, so the AHI figures describe breathing without the machine rather than the combined effect of drug and device. Eli Lilly, which makes Zepbound, funded both trials.

Foundayo is a once-daily pill that the FDA approved for obesity in April 2026. Unlike Zepbound, it is a small molecule rather than a peptide, which is why it can be swallowed instead of injected. The ATTAIN-OSA program tested it in two 52-week trials at 66 sites in nine countries; the results were published in JAMA on November 9, 2026, and presented at the American Heart Association's Scientific Sessions.

The trials enrolled adults with moderate to severe sleep apnea and a BMI of 27 or more, a lower threshold than Zepbound's trials used. Trial 1 included 363 people who were not using PAP, with an average starting AHI of 50.7, and trial 2 included 349 people already using PAP, with an average starting AHI of 57.6. Participants took either the highest daily dose they could tolerate (9, 14.5, or 17.2 mg) or a placebo.

The JAMA paper reports results that count every participant who started treatment, including those who later stopped. In trial 1, the AHI fell by an average of 17.8 events per hour on orforglipron compared with 4.5 on placebo, a difference of 13.2. In trial 2, it fell by 17.3 compared with 5.3, a difference of 12.0. Six of the seven key secondary goals, including hypoxic burden and a sleep-impairment score, also favored orforglipron with statistical significance.

Lilly's press release reports results using a different statistical approach, called the efficacy estimand, which estimates what would have happened if every participant had continued the drug for the full year. By that measure, the AHI fell by 21.1 events per hour (43.8%) in the trial without PAP and by 18.3 (27.5%) in the trial with PAP, compared with 4.3 and 5.5 on placebo. At least a 50% reduction in AHI occurred in 47.2% and 33.7% of participants, and 31.5% and 19.6% reached remission (an AHI under 5) or mild disease (an AHI of 5 to 14 without excessive daytime sleepiness). Hypoxic burden fell by 55.9% and 45.8%, and body weight fell by 11.5% (28.0 pounds) and 11.2% (28.2 pounds).

Lilly plans to submit Foundayo to the FDA for moderate to severe sleep apnea in adults with obesity or overweight by the end of 2026. Until the FDA approves that use, Foundayo's label covers weight management, not sleep apnea. For how the pill compares with Novo Nordisk's Wegovy tablet, see Foundayo vs. the Wegovy Pill.

It is tempting to line up the two sets of numbers and declare a winner, but the trials were not designed for that comparison, and they differ in ways that affect the results.

• No head-to-head trial exists. No study has given both drugs to the same group of patients; each was compared only with a placebo.
• The results were counted differently. Lilly's Foundayo figures use the efficacy estimand, which usually looks more favorable because it assumes every participant continued the drug. The Zepbound figures in the table above count everyone who started treatment. Measured the same way, the average AHI reduction was larger with Zepbound (25.3 and 29.3 events per hour) than with Foundayo (17.8 and 17.3), although the trials still differed in other respects.
• The participants were different. Zepbound's trials required a BMI of 30 or more and excluded people with diabetes, while Foundayo's trials admitted people with a BMI as low as 27. The Foundayo group already using PAP also started with more severe disease.
• Weight loss was different. Participants lost about 18% to 20% of their body weight on Zepbound and about 11% on Foundayo. Because weight loss accounts for most of the improvement in breathing, that gap probably explains much of the difference.

A fair summary is that both drugs reduce sleep apnea severity compared with placebo, that the injection produced more weight loss in its trials, and that the pill gives people who dislike injections another option. Daniel Drucker, a GLP-1 researcher who was not involved in the Foundayo trials and has consulted for Lilly, told Scientific American that every structural change to a GLP-1 drug raises new questions about overall and cardiovascular safety.

Some people may eventually be able to, but most people in these trials could not, and the decision should not be made without a sleep specialist.

The trials defined a good outcome as remission (an AHI under 5) or mild disease (an AHI of 5 to 14) without excessive daytime sleepiness. The SURMOUNT-OSA authors note that at those levels, PAP therapy may no longer be recommended. On Zepbound, 42% to 50% of participants reached that point; on Foundayo, Lilly reports 20% to 32%. In other words, at least half of the participants taking either drug still had sleep apnea after a full year that would ordinarily call for continued treatment.

There are several other reasons for caution:

• Individual results vary widely. An average reduction of 50% combines people who improved dramatically with people whose breathing barely changed.
• The Zepbound trials measured breathing without the machine. Researchers asked PAP users to pause their machines for a week before each sleep study so they could isolate the drug's effect, which is not the same as advising patients to stop PAP.
• Only a sleep study can confirm improvement. Feeling more rested does not guarantee a normal AHI, so a repeat sleep study reviewed by a sleep physician is the reliable way to learn whether the machine is still needed.
• The benefit depends on continuing the medication, as the next section explains.

For many patients, the realistic expectation is that the drug makes sleep apnea milder and PAP easier to tolerate, while a minority may see their breathing return close to normal.

The sleep apnea trials lasted 52 weeks and did not examine what happens after people stop the medication. A 2026 review of the tirzepatide evidence identifies the durability of the benefit after discontinuation as one of the principal unanswered questions.

The best available clue comes from studies of body weight. In the SURMOUNT-4 trial, adults with obesity took tirzepatide for 36 weeks and lost an average of 20.9% of their body weight before being randomly assigned either to continue the drug or to switch to a placebo. Over the following year, the placebo group regained 14.0% of their body weight, while the group that continued tirzepatide lost an additional 5.5%. Because the improvement in sleep apnea closely follows weight loss, it is reasonable to expect breathing to deteriorate if the weight returns, although no trial has measured that directly.

For most people, therefore, a GLP-1 prescribed for sleep apnea should be viewed as a long-term treatment, much like the PAP machine it may partly replace.

Gastrointestinal side effects were the most common problems in both programs; in the Zepbound trials, they occurred most often while the dose was being increased and were mostly mild to moderate.

In SURMOUNT-OSA, diarrhea affected 22% to 26% of participants taking tirzepatide compared with 9% to 13% on placebo, nausea affected 22% to 25% compared with 5% to 10%, vomiting affected 9% to 18% compared with 1% to 4%, and constipation affected about 15% to 16% compared with 3% to 4%. Only 3.4% to 4.4% of participants stopped tirzepatide because of side effects, and two participants in trial 2 developed confirmed acute pancreatitis. In ATTAIN-OSA, gastrointestinal events affected 68% to 71% of participants taking orforglipron compared with about 37% on placebo, and according to Lilly, 12.2% stopped the drug because of side effects in the trial without PAP (compared with 3.8% on placebo) and 6.8% in the trial with PAP (compared with 0.6%).

On November 9, 2026, the FDA required every GLP-1 medication, including Zepbound and Foundayo, to add a warning about NAION (nonarteritic anterior ischemic optic neuropathy), a rare condition that can cause sudden, painless vision loss in one eye. The agency described the risk as "potential." In its pooled analysis of 126 clinical trials, probable NAION occurred at a rate of 6.4 events per 100,000 person-years on GLP-1 drugs compared with 3.8 in the comparison groups. The FDA also lists sleep apnea itself as a known risk factor for NAION, alongside diabetes, high blood pressure, and smoking. Anyone who notices sudden vision loss, blind spots, or blurry central vision should seek medical care immediately; for more detail, see GLP-1 Drugs and Your Eyes.

The Zepbound label also carries a boxed warning about thyroid C-cell tumors seen in rats, along with warnings about pancreatitis, gallbladder disease, low blood sugar when combined with certain diabetes medications, kidney injury, and the risk of inhaling stomach contents during anesthesia. Rapid weight loss can also reduce muscle mass, so resistance training and adequate protein intake matter; see Eating Enough Protein and Training Hard Enough on GLP-1s.

Zepbound's sleep apnea approval covers adults with obesity and moderate to severe obstructive sleep apnea; in the trials, that meant a sleep study showing an AHI of 15 or more and a BMI of 30 or more. The trials excluded people with central or mixed sleep apnea, in which the brain fails to send the signal to breathe, and the approval covers only the obstructive form.

Medicare. Federal law bars Medicare drug plans from covering drugs used only for weight loss. Sleep apnea is a medical condition, so a Part D plan can cover Zepbound when it is prescribed for that use. KFF Health News reported in August 2026 that people with conditions like moderate to severe sleep apnea are routed to their regular Part D plan rather than Medicare's $50-a-month GLP-1 Bridge pilot, and one physician said Part D copays can run $200 to $600 a month or more. An estimated 5.9 million Medicare enrollees are excluded from the Bridge discount for this kind of reason. See the Medicare GLP-1 Bridge patient guide for how that program works.

Private insurance. Coverage depends on the plan. Sharona Hoffman, a law and bioethics professor at Case Western Reserve University, told Ideastream in November 2026 that a sleep apnea diagnosis may qualify someone for coverage when obesity alone does not, and that appeals of denials often succeed.

Paying cash. Lilly's self-pay prices start at $299 a month for Zepbound and $149 a month for Foundayo, according to a November 6, 2026 Lilly announcement. Foundayo is sold for weight management, not yet for sleep apnea.

• What was my AHI on my most recent sleep study, and how far would it need to fall before I could reduce or stop using PAP?
• Should I continue using my CPAP machine while I start the medication? (In the Zepbound trials, participants who already used PAP kept using it, pausing only for the 7 days before each sleep study.)
• When should I schedule a repeat sleep study to measure my progress?
• Will the medication be prescribed specifically for sleep apnea, and how will that affect what my insurance or Medicare drug plan covers?
• How should I manage gastrointestinal side effects, and what can we do to protect my muscle mass while I lose weight?
• Because sleep apnea is itself a risk factor for NAION, which vision changes should prompt me to call immediately?
• If I eventually stop the medication, what is our plan for managing my sleep apnea?

GLP-1 medications have become a genuine treatment option for obstructive sleep apnea in people with excess weight. Zepbound is already approved for the condition, and Lilly expects to submit Foundayo to the FDA by the end of 2026. In year-long trials, both drugs reduced breathing interruptions substantially compared with placebo, mainly by producing weight loss.

They are not, however, a straightforward replacement for CPAP. Between half and four-fifths of participants still had sleep apnea that would ordinarily require treatment after a year, the benefit is likely to fade if the medication is stopped and the weight returns, and the drugs carry their own side effects. For most patients, the sensible approach is to use the medication alongside a sleep specialist's care and to let a repeat sleep study determine whether the machine can eventually be set aside.