A Prescription Instead of a Mask
Tirzepatide's approval for moderate-to-severe OSA raises a genuinely new question: when PAP therapy is effective but a patient won't reliably use it, does starting a slower-acting drug instead of the fast-acting standard trade an immediate safety risk for a treatment he might actually keep taking?
R.F. has worked from a home office for six years, an arrangement that let a slow creep of weight and a slower creep of daytime exhaustion go largely unnoticed by anyone but his wife, who finally pushed him toward a sleep study after a close call merging onto the highway last month — he doesn't remember drifting, only the horn. He is 47, has a BMI of 38, and the study came back unambiguous: an AHI of 42, severe obstructive sleep apnea, oxygen desaturations into the low 80s overnight. In the same visit where his sleep physician explained the diagnosis, R.F. said plainly that he'd tried a CPAP mask once, years ago at a friend's urging, hated it within a week, and would rather not be prescribed something he already knows he won't use.
That statement changes the shape of the conversation more than the AHI number does. Positive airway pressure remains the fastest-acting, most immediately effective therapy for severe OSA, correcting the airway obstruction the same night it's used — but only the night it's actually used, and PAP's real-world effectiveness is limited less by its physiology than by adherence rates that hover well below what trials report under close supervision. Tirzepatide's 2024 approval for moderate-to-severe OSA rests on the SURMOUNT-OSA program reported by Malhotra and colleagues, and for once the population question is easy: Study 1 enrolled adults with obesity and an AHI of 15 or above who were not on PAP, which is R.F. at a BMI of 38 and an AHI of 42, and its mean AHI reduction of roughly 27 events per hour was measured in exactly that group. What the trial cannot supply is speed — its mechanism is sustained weight loss unloading the upper airway, and the separation from placebo did not appear until around week 20. That is the number worth setting against his near-miss last month. The gap he actually falls into is not efficacy but adherence: the device's effectiveness figures come from closely-monitored trial cohorts, and unsupervised everyday use runs well below them — a therapy that works every night in a study and one that works only on the nights someone will wear it are not, in practice, the same treatment. His own single week on a mask years ago is the closest thing this case has to a real-world adherence datum, and it points the wrong way for the option that would work tonight.
New OSA diagnosis, treatment planning
He had a drowsy-driving near-miss last month, and his AHI is 42 with desaturations into the low 80s. PAP corrects that the same night it's worn. Tirzepatide's benefit builds over months as weight comes off. Given that he has an actual, dated safety event on record, I don't think we can treat a months-long ramp-up as an acceptable substitute for something that works tonight.
I hear the urgency, but he's told us directly he won't use PAP — he tried it once, hated it, stopped within a week. A prescription he doesn't use provides exactly zero benefit, no matter how fast it theoretically works. SURMOUNT-OSA showed real, substantial AHI reduction with tirzepatide, through a mechanism that actually addresses why his airway collapses in the first place. Starting a therapy he's realistically going to take is not a lesser option; a merely-good therapy he uses beats an ideal one he doesn't.
Recommending PAP again on the strength of "it works faster" without accounting for his stated, specific history with it isn't a safety-first plan — it's a plan that predictably fails the same way it failed before.
I don't think this actually has to be one or the other. Start PAP today — a different mask style than his prior attempt, with proper fitting and follow-up this time, since a single bad experience years ago shouldn't be the last word on it — while tirzepatide starts in parallel to treat the weight driving the disease. If his AHI comes down substantially as weight loss progresses, PAP pressure or duration can be reassessed later. That covers the immediate safety risk the pulmonologist is right to weight heavily, without asking him to commit to a device he's already told us he doesn't want.
Agreed: PAP restarted today with a new mask interface and a two-week adherence check, and tirzepatide started in parallel, with AHI and weight reassessed at three months.
Not fully agreed: the endocrinologist remains doubtful he will sustain PAP use even with a better-fitted mask, and would have preferred tirzepatide alone with PAP offered but not pushed; the pulmonologist views the combined start as the only responsible plan given his documented safety event. Both agreed that if his PAP adherence data at two weeks shows he is not using it, that finding — not a repeat of the original argument — should drive whatever happens next.