Doxepin for Insomnia

Doxepin has been prescribed in the United States since the 1960s as a tricyclic antidepressant. In 2010 the FDA approved the same molecule at a small fraction of that dose under a separate application, NDA 022036, and the label for those tablets reads nothing like the capsule label. The sleep product is indicated for the treatment of insomnia characterized by difficulty with sleep maintenance.

That is the whole indication. No depression, no anxiety, no general sedation.

The capsule is a different clinical animal. Its label indicates it for major depressive disorder in adults and nothing else, with a starting dosage of 25 mg three times daily or 75 mg once daily and a target range between 75 mg and 150 mg per day. It opens with a boxed warning about suicidal thoughts and behaviors in younger adults, carries dosage reductions for CYP2D6 and CYP2C19 poor metabolizers, and instructs prescribers to taper rather than stop.

The sleep label states the gap plainly: doxepin is an antidepressant at doses 10 to 100 fold higher than in the sleep product. Pharmacology gives a reason. Doxepin has high binding affinity to the H1 receptor, with a Ki below 1 nM, so a few milligrams is enough to occupy that target while leaving little in circulation to hit the muscarinic and adrenergic receptors that a 100 mg dose reaches. The label is careful to say the mechanism in sleep maintenance is unclear and the effect could be mediated through H1 antagonism.

If your problem is lying awake for an hour at bedtime, this is probably the wrong drug. The trials behind the approval used objective and subjective time spent awake after sleep onset as the primary efficacy measures, which is the 3 a.m. problem, not the midnight one. The American Academy of Sleep Medicine reached the same split. Its 2017 guideline suggests clinicians use doxepin as a treatment for sleep maintenance insomnia in adults, a weak recommendation under GRADE, and doxepin appears nowhere in its sleep onset recommendations.

A systematic review of nine randomized placebo controlled trials found that low dose doxepin had a small to medium effect size against placebo for sleep maintenance and sleep duration but not for sleep initiation. The one caveat worth knowing is that the 35 night adult trial did show improved latency to persistent sleep on night 1 for both 3 mg and 6 mg, but that head start did not hold at night 15 or night 29. Treat any first night speed as a bonus rather than the reason to start.

The approval rests on six randomized double-blind studies of up to three months that enrolled 1,423 subjects aged 18 to 93, with chronic insomnia in 858 and transient insomnia in 565. Doses studied were 1 mg, 3 mg and 6 mg against placebo, in sleep laboratories and at home. That number matters later: 10 mg was never in this program, so nobody has trial data on it for sleep.

The 35 night adult trial randomized 75 to 3 mg, 73 to 6 mg and 73 to placebo. Both doses significantly improved wake time after sleep onset at night 1, night 15 and night 29, and improved total sleep time. The longest trial ran 12 weeks in 240 adults over 65 on 1 mg or 3 mg, where 3 mg improved wake time after sleep onset, total sleep time and sleep efficiency on night 1 with the gains sustained at night 85.

Outside the sponsor's own program, a network meta-analysis of 24 trials in 5,917 older adults found low dose doxepin to be the effective treatment for increasing objective sleep efficiency, at 6.08 percent, and concluded it looked like the preferred pharmacotherapy for total sleep time and sleep efficiency in that age group. The discontinuation picture is unusually quiet as well. The label reports no evidence of rebound insomnia at 3 mg or 6 mg and no indication of a withdrawal syndrome on the Tyrer's Symptom Checklist.

This is the substitution that gets suggested when the price gap shows up, and it does not hold. The capsule label lists strengths of 10 mg, 25 mg, 50 mg, 75 mg and 100 mg, so the smallest capsule already sits above the sleep label's instruction that the total daily dose should not exceed 6 mg. A gelatin capsule cannot be halved into a reliable 5 mg, and the low dose tablets themselves are not scored, so there is no clean way to land on 6 mg from either direction.

The oral concentrate is the other workaround, and it has the same problem. It is supplied at 10 mg per mL with a dropper calibrated at 5 mg, 10 mg, 15 mg, 20 mg and 25 mg, which means the smallest measurable dose is close to double the elderly starting dose. Beyond the arithmetic, the evidence simply stops. Every efficacy and next day safety result quoted on this page comes from 1 mg, 3 mg or 6 mg arms. A 10 mg nightly dose is an extrapolation beyond anything the label offers, and the anticholinergic reactions that show up on the antidepressant label start becoming relevant as the dose climbs.

Section 9 of the sleep label is four words long where it counts: doxepin is not a controlled substance. The label adds that doxepin is not associated with abuse potential in animals or in humans and that no symptoms indicative of a withdrawal syndrome were observed after chronic administration.

Whatever else you weigh, that is a genuinely different regulatory object from the standard prescription sleep aid.

Compare zolpidem, which is classified as a Schedule IV controlled substance by federal regulation and now leads with a boxed warning that complex sleep behaviors including sleep-walking and sleep-driving may occur and some of these events may result in serious injuries, including death. The scheduling has practical teeth. Federal rules say no Schedule III or IV prescription may be filled or refilled more than six months after it was issued, and none may be refilled more than five times. Doxepin is not bound by either clock, which is part of why it fits a remote prescribing workflow more cleanly.

None of that makes it consequence-free, and the honest reading of the label refuses the easy version of this argument. The doxepin sleep label carries its own warning that complex behaviors such as sleep-driving have been reported with hypnotics and discontinuation should be strongly considered for anyone who reports an episode. It is also contraindicated with MAOIs used within the past two weeks, and in untreated narrow angle glaucoma or severe urinary retention, and ordinarily not recommended in severe sleep apnea.

The dosing instructions are short and unusually specific. The label sets an initial dose of 6 mg once daily for adults and 3 mg once daily for the elderly, taken within 30 minutes of bedtime, with a total daily dose that should not exceed 6 mg. If you have hepatic impairment, treatment starts at 3 mg with close monitoring for adverse daytime effects. There is no titration schedule to climb, which makes this one of the simpler nightly regimens to follow correctly.

The food rule is the part most likely to get ignored and the part with the clearest pharmacokinetic reason behind it. A high fat meal increased AUC by 41 percent and Cmax by 15 percent and delayed time to peak concentration by roughly three hours, which is exactly the wrong shape for a drug you want out of your system by morning. Taken fasted, median time to peak concentration was 3.5 hours after a 6 mg dose. A late dinner does not just blunt the onset, it pushes the tail of the curve into your alarm.

The tolerability numbers are mild by hypnotic standards. Across three long-term placebo-controlled trials the label reports somnolence or sedation in 9 percent on 6 mg and 6 percent on 3 mg against 4 percent on placebo, with nausea at 2 percent and upper respiratory infection at 2 to 4 percent. Dropping out for an adverse reaction was almost indistinguishable from placebo, at 0.7 percent on 6 mg against 0.6 percent on placebo. Notably absent from that table are dry mouth and constipation.

On next-day function the sources disagree, and this page will not smooth that over. The systematic review concluded there was no significant next day residual effect with low dose doxepin, and both pivotal trials reported no significant next-day residual effects and no spontaneous reports of memory impairment or complex sleep behaviors. The label is more guarded. It records that in a one night study of 565 healthy adults, 6 mg showed modest negative changes on the symbol copying test and the sleepiness visual analog scale, and that over 35 days small decreases in the digit symbol substitution and symbol copying tests occurred in the 6 mg group.

Only 1 mg and 3 mg in the elderly trial were comparable to placebo across all three measures. The practical version: at 3 mg the morning-after evidence is clean, at 6 mg it is good but not spotless, and the label still tells you to avoid operating a motor vehicle or heavy machinery at night after taking it and to expect possible impairment the following day. Combinations matter more than the drug alone. The label says not to consume alcohol with it, and warns of additive effects with CNS depressants and sedating antihistamines, which quietly rules out stacking it on a diphenhydramine sleep aid.

The geriatric argument for this drug is not a marketing line, it is written into the criteria that usually rule tricyclics out. The 2023 American Geriatrics Society Beers Criteria lists doxepin above 6 mg per day among antidepressants with strong anticholinergic activity to avoid in older adults, and states in the same entry that the safety profile of low dose doxepin at 6 mg per day or less is comparable to that of placebo. The dose threshold is the criterion. Almost nothing else on that list gets a carve-out written into it.

The adverse reaction tables back that distinction up. The capsule label's most common reactions at 5 percent or more are somnolence, dry mouth, dizziness, constipation and fatigue, the classic anticholinergic set, and it carries a dedicated instruction to reduce the dose if dry mouth, blurred vision or constipation develop. Neither reaction shows up in the sleep label's table at all, which is the clearest single sign that the two doses behave like different drugs in the body. Both pivotal trials recorded no spontaneous reports of anticholinergic effects, weight gain or increased appetite, and the sleep label notes that 362 subjects aged 65 or older and 86 aged 75 or older received it in controlled studies with no overall differences in safety or effectiveness.

Beers still lists plain doxepin under its tertiary tricyclic entry for orthostatic hypotension and syncope, so the conversation with a prescriber does not simply end at 6 mg.

Every figure here comes from the CMS National Average Drug Acquisition Cost file, which is what retail pharmacies pay wholesalers, sampled from invoices. On the file effective 2026-08-19, the 6 mg doxepin tablet is listed at $1.44465 per tablet and the 3 mg tablet at $1.36437, both flagged as generic. Thirty nights of the 6 mg dose is therefore about $43 of acquisition cost before anyone has dispensed it.

The antidepressant capsules are on a different planet. The same file lists the 10 mg capsule at $0.07759 each, the 25 mg at $0.13679, the 150 mg at $0.52776, and the 10 mg per mL oral concentrate at $0.77890 per mL. That puts a month of 10 mg capsules near $2.33 at acquisition, close to one eighteenth of the low dose tablet. The gap is real and it is the entire reason the substitution question keeps coming up, which is why the previous section spends so long on why the swap does not track.

Two caveats before you budget from those numbers. Acquisition cost excludes the dispensing fee, the pharmacy's margin, and any telehealth visit fee, so your counter price will be higher, sometimes substantially. And the low dose product is no longer brand-only: generic 3 mg and 6 mg doxepin tablets carrying the sleep maintenance indication are marketed in the US, which is what those generic-flagged acquisition rows represent. If a price quote you are shown assumes brand pricing, it is worth asking which version is being dispensed.

Three questions separate a considered prescription from a reflex one. First, which product: 3 mg or 6 mg tablet, or a capsule. If the answer is a 10 mg capsule for sleep, you want to hear why, because it is outside the labeled dose for that indication. Second, whether the complaint being treated is sleep maintenance, since a prescriber reaching for this drug for sleep onset is working against the guideline. Third, whether they screened the label contraindications: MAOI use within two weeks, untreated narrow angle glaucoma, severe urinary retention, and severe sleep apnea.

The label also sets a review date most prescriptions never get. It says that failure of insomnia to remit after 7 to 10 days of treatment may indicate a primary psychiatric or medical illness that should be evaluated, so a service that writes 90 nights with no check-in is not following it. Whether any of this applies to you is a call for the clinician writing the prescription, and this page is information rather than medical advice. If you are 65 or older, confirm the plan starts at 3 mg, which is what the label instructs.