Can a Pill Replace CPAP for Sleep Apnea?
For an adult who can't live with a CPAP mask, the answer is starting to be yes. AD109, a once-nightly pill, is the first drug in decades of trying that credibly opens the sleeping airway. In Phase 3 trials it cut breathing interruptions by roughly 44 to 47 percent from baseline versus about 7 to 18 percent on placebo, and the FDA is set to decide on approval by February 28, 2027. It has real side effects, and no trial has yet shown it lowers heart attacks or strokes, so it's a partial fix aimed at people the mask has failed, not a replacement for a CPAP that's working.
AD109 is a once-nightly pill that acts on the nerves controlling the muscles at the back of your throat, keeping them from going slack while you sleep. It's the first oral drug shown in large randomized trials to meaningfully open the sleeping airway in obstructive sleep apnea (OSA). The question isn't whether it works. It's who it works well enough for, and whether cutting breathing events roughly in half is enough to matter for the diseases OSA actually causes.
What the Phase 3 trials found
In SynAIRgy, 646 adults with mild-to-severe OSA who couldn't or wouldn't use a CPAP mask took AD109 or placebo nightly for 26 weeks. The pill cut their apnea-hypopnea index, the count of breathing pauses and shallow breaths per hour, by about 44% from baseline. Placebo cut it by about 18%. LunAIRo, a second trial of 660 adults, showed a 47% reduction against 7% on placebo. The median starting AHI was about 20 events per hour, so a 44% cut left many people with residual mild apnea rather than fully controlled disease. The FDA has accepted the application; a decision is expected by February 28, 2027. If it clears, AD109 will be the first oral drug ever approved for this condition.
Why it works where earlier drugs failed
Sleep apnea drugs have been tried and failed for decades. AD109 works because it combines two old drugs used in a new way. Atomoxetine, better known for ADHD, raises noradrenaline and turns up the nerve signal that keeps your tongue and pharynx toned. Aroxybutynin, a close cousin of the oxybutynin used for overactive bladder, blocks muscarinic inhibition of those same motor neurons during sleep. Together they leave the airway more open. That's a different approach from CPAP, which splints the airway open with pressurized air, and from tirzepatide, which reduces apneas largely by driving weight loss in adults with obesity.
Where CPAP still wins
For the airway obstruction itself, CPAP is still more effective. Across trials it cuts the AHI by roughly 30 events per hour and drops most adherent users well into the normal range. A 44% cut on AD109, starting from a median of about 20 events per hour, leaves many people with residual mild apnea. In SynAIRgy, patient-reported fatigue only improved significantly in the subgroup who came in with excessive daytime sleepiness at baseline. AD109 isn't trying to unseat a mask that already works. It's aiming at everyone the mask has failed.
| Treatment | AHI reduction | Main drawback | Long-term outcome evidence |
|---|---|---|---|
| CPAP (mask) | Cuts about 30 events per hour; near-normal in adherent users | 20 to 30 percent refuse or quit the mask | Lower all-cause and cardiovascular mortality in observational data; three large randomized trials did not show fewer cardiovascular events overall, largely because of poor adherence |
| AD109 pill (FDA decision Feb 2027) | Cuts breathing events roughly in half from a median around 20 per hour | About 1 in 5 quit for dry mouth, nausea, insomnia, or urinary hesitancy | None yet on heart attacks, strokes, or death |
| Mandibular advancement device | Moderate cut; less than CPAP | Dental fit and jaw comfort limit long-term use | Similar blood pressure benefit to CPAP; equivalent to CPAP for cardiovascular death in one observational cohort |
| Tirzepatide (for OSA with obesity) | Cuts around 20 to 24 events per hour over a year in adults with obesity | Weekly injection; requires substantial weight loss | None yet for hard cardiovascular endpoints in OSA specifically |
The tolerability catch
About 21% of AD109 users quit the trial because of side effects, versus about 3% on placebo. The list is what you'd expect from these two drug classes: dry mouth, nausea, insomnia, and urinary hesitancy. Heart rate rose a mild 5 to 9 beats per minute. No serious drug-related events and no deaths were reported. But the people still taking the pill by the end were self-selected for tolerating it. The formulation the FDA is reviewing uses a slightly lower dose of aroxybutynin (2.3 mg) than the 2.5 mg used in the pivotal trial, which may help tolerability. It may not.
What the trials don't answer yet
OSA matters because untreated it raises the risk of heart attack, stroke, and death, and the reason to treat it is to lower those risks. No AD109 trial has yet shown it does. CPAP's own mortality benefit is strongest in observational data and depends on wearing the mask at least four hours a night. In three large randomized trials (SAVE, RICCADSA, and ISAACC), CPAP didn't reduce cardiovascular events on an intention-to-treat basis, largely because adherence averaged just over three hours a night. In an on-treatment analysis pooling those trials, adherent users had about a third fewer recurrent cardiovascular events. That's the gap a pill is meant to fill: a treatment people will actually take every night. Whether taking AD109 translates into fewer heart attacks and strokes is still unproven, and the trial that would answer it hasn't been run.
Who this actually helps
If you have OSA and CPAP is working for you, nothing in these trials argues for switching. Adherent CPAP users still get a bigger reduction in apneas than AD109 delivers. If you have mild-to-moderate OSA and have given up on the mask, or never tried it because you knew you wouldn't stick with it, AD109 will be, once approved, the first drug worth asking your sleep specialist about. Knowing your baseline AHI from a sleep study, and repeating it after a few months on the pill, is the way to tell whether it's doing for you what it did in the trials.
For a first-generation pill in a field where every previous drug has been a disappointment, AD109 is a step. It's proof that pharmacology can open the sleeping airway. Whether it will also cut the cardiovascular toll of untreated apnea, which is the thing that ultimately matters, is the trial that hasn't been run yet.