The Most-Prescribed Sleep Drug in America Was Never Approved to Treat Sleep
The drug he’s about to be handed has never been FDA-evaluated for the thing it’s about to be prescribed for. That fact alone doesn’t settle whether it’s the right choice — but it’s the fact everyone in the room has to reckon with before agreeing to it.
J.O., a 39-year-old man, has coached high school football for eleven years, the last six as head coach of a program he took over after a difficult stretch and has slowly rebuilt into a genuine contender — until this season, which has gone the other direction, three close losses in a row and a fanbase that has started to notice. His in-season sleep has never been especially good, a pattern he’s lived with for most of his coaching career, but this year it has genuinely fallen apart: real difficulty falling asleep most nights, replaying game film in his head, and a new pattern of waking at 2 or 3 a.m. and lying there until his alarm rather than falling back asleep.
He attributes it honestly, without much prompting, to the losing season itself and to a contract renewal decision that his athletic director has left unresolved longer than he expected, both real and identifiable stressors rather than anything vaguer. He has no psychiatric history, no other medications, and no substance use beyond a beer most Friday nights after games, a routine he’s kept for years without it becoming anything more. He mentioned, almost as an aside near the end of the visit, that a friend takes “trazodone, one of those old antidepressants” for sleep and asked, half-joking, whether that might work for a coach who clearly isn’t depressed, just exhausted and anxious about his job.
His physician was already considering trazodone before he brought it up — it is, by prescription volume, the most commonly used sleep medication in the country — but his question prompted the team to actually name, out loud, something usually left unsaid in these conversations: trazodone has never gone through an FDA approval process for insomnia at any dose, in a country where it is nonetheless the single most-prescribed drug for exactly that purpose.
In clinic, mid-season
It’s worth being precise about what “off-label” actually means here, because it sounds worse than it is once you see the mechanism. Trazodone at low, sleep-targeted doses (25–100 mg) acts predominantly through 5-HT2A and H1 antagonism, producing sedation at doses well below its antidepressant range (150–600 mg) — it isn’t being repurposed at its labeled dose, it’s being used at a genuinely different, lower dose for a genuinely different effect.
That said, “used this way for decades” is real-world experience, not the same evidentiary standard as an FDA-reviewed indication — the randomized trial evidence specifically for low-dose trazodone in primary insomnia is thinner than its prescription volume would suggest, and I think that gap deserves to be said plainly rather than assumed away by how familiar the drug is.
I’d actually lean on the prescription-volume point more than the pharmacologist is willing to — not as proof of efficacy, but as a genuinely large, if uncontrolled, safety record. Decades of widespread low-dose use in exactly this population have not surfaced a signal remotely resembling the complex sleep behavior FDA later attached to Z-drugs, and for a patient with no cardiac history and no other medications, the realistic risk profile — daytime sedation, rare priapism, orthostatic effects — is well characterized even without a formal FDA insomnia indication.
I take the point that trial evidence is thinner than the prescribing volume implies, but “thinner trial evidence” isn’t the same claim as “no real evidence,” and for a young, healthy, treatment-naive patient with a clear situational trigger, I don’t think the evidentiary gap should block a drug this well-characterized in practice.
I’d frame the actual choice slightly differently than either of you. His insomnia sounds genuinely situational — a hard season, a real professional anxiety — and both trazodone and zolpidem treat the symptom without touching the cause. Given that, and given he has no factor pushing him toward a controlled substance specifically, I’d rather start with the drug that isn’t a scheduled hypnotic at all, precisely because his trigger is likely to resolve on its own timeline.
That’s not really a disagreement with either of you on the evidence — it’s a vote for trazodone on different grounds: not because its off-label safety record is reassuring, though it is, but because a non-scheduled option makes the most sense for a problem this likely to be temporary.
Agreed: trazodone 50 mg nightly, with a planned reassessment after the season ends to see whether his sleep resolves alongside the situational trigger.
Not agreed, and worth keeping visible:
The randomized evidence specifically supporting low-dose trazodone for primary insomnia remains thinner than its prescribing volume would suggest, regardless of how this particular case turns out.
Decades of large-scale, low-dose real-world use without a safety signal like Z-drugs’ is itself meaningful evidence, even without a formal FDA indication behind it.
The sleep physician’s framing — that the real driver of today’s choice was the drug’s non-scheduled status given a likely-temporary trigger — was accepted by both other voices as the practical tiebreaker, without either backing off their own read of the underlying evidence question.