Does Alcohol Help You Sleep?

Last checked 3 November 2026 · 11 min read

It shortens the time it takes to fall asleep, at doses high enough to matter, and it damages the night at every dose. Those two facts are not in tension; they describe two different halves of the same night. Alcohol consolidates the first half and fragments the second, and the second half is the one that decides whether you wake up rested. The most useful finding in the literature is also the least quoted: at around two drinks, the sleep benefit largely disappears while the damage to REM sleep is already there. So the common version — a glass of wine to get to sleep — is the version that pays the price without collecting the goods.

Two summaries of the same literature disagree

Start with a contradiction, because it is the most honest place to start. The standard review of alcohol and sleep, covering decades of laboratory nights, concludes that at all doses alcohol shortens sleep onset, consolidates the first half of the night and increases disruption in the second. That is the version most people have heard.

A newer meta-analysis, pooling 27 studies and stratifying by dose, found something narrower. Reductions in the time to fall asleep — and in the time to reach deep sleep — appeared only at the higher dose, around 0.85 grams of alcohol per kilo of body weight, or roughly five standard drinks. At the lower dose, around two drinks, the onset benefit was not there. The REM damage, by contrast, appeared at that lower dose and got worse as the dose rose.

The table scrolls sideways on a phone.

Which one is right? Probably both, in different senses. The older review counts every study that ever measured the effect, including small ones and ones without a placebo night. The newer one pools them with dose as a variable and finds that the low-dose effect is too inconsistent to survive the pooling. What is not in dispute is the direction of the REM finding, or the shape of the night: better at the start, worse at the end.

The first half is bought, the second half is charged

Almost every laboratory study finds the same architecture. In the first three or four hours, alcohol increases slow-wave sleep — the deep, high-amplitude stage — and reduces the number of awakenings. Sleep looks, on the recording, unusually solid. Then, as the alcohol clears, the pattern inverts: more time awake, more brief awakenings, and a first REM period that arrives late and a total REM time that is shorter than it should be.

This is why the morning report and the night's recording disagree so often. People describe falling asleep fast and sleeping deeply, and they are describing the first half accurately. The part they do not experience is the part that matters, because the second half is where the night is repaired. Deep sleep in the first half is not a substitute for the REM that was pushed out of it; the stages are not interchangeable currency.

The size of the effect is not trivial. In the largest laboratory comparison, 93 healthy adults drank to a breath alcohol of about 0.11 per cent — a heavy evening by any standard — or a matched placebo, and then slept under electrodes from eleven to seven. Alcohol shortened sleep onset and then reduced sleep efficiency, reduced REM sleep, and increased wakefulness across the whole night, while increasing slow-wave sleep. Every one of those changes is in the direction of a worse night, even though the first hour felt easier.

Why the second half breaks

The mechanism is not simply that alcohol wears off. If it were, you would return to your normal night, not to a worse one. What seems to happen is that alcohol borrows from the system that regulates sleep pressure.

The best-supported account comes from animal work, and it is worth being explicit that it is animal work. Alcohol increases extracellular adenosine, the molecule that accumulates while you are awake and that caffeine blocks. Higher adenosine inhibits the wake-promoting neurons of the basal forebrain, which is a clean explanation for the sedation: the sleepiness is real, and it is arriving through the same channel as a long day. But the same research shows that this interferes with sleep homeostasis — with the accounting system that is supposed to balance sleep and wakefulness. The debt is not paid down; it is deferred, and the repayment shows up as fragmented sleep later in the night and as a hangover that feels like a debt.

People describe the timing precisely without knowing the mechanism. One says a couple of drinks guarantees a wake-up at four in the morning. Another says the alcohol makes him pass out, and then he sobers up an hour or two later and cannot get back to sleep. A third describes the same thing in the language of rebound: the drink is calming while it is present, and the system overshoots when it leaves. That overshoot is the step that wakes you, and unlike a door slam it is coming from inside the room.

The morning you remember is wrong

The 93-person study measured this directly, and the result is the most practically useful one on the page. After alcohol, people rated themselves sleepier at bedtime than after the placebo. The next morning, they rated their sleep quality lower. Both of those are subjective, and they point in opposite directions at opposite ends of the night.

That is the trap. The feeling that alcohol helps is measured at the moment of falling asleep, when the sedative effect is at its peak. The feeling that it did not help is measured in the morning, when the second half has already happened. If you only ever notice the first measurement, you will keep concluding that it works.

There is a second reason the morning rating is the more trustworthy one. It is the only one that includes the part of the night you were not conscious for. A person cannot report the awakenings they do not remember, and alcohol makes brief awakenings both more frequent and less memorable. The wrist monitor and the electrodes do not have that problem, which is why they disagree with the diary.

Women pay more for the same night

In that same study, the disruption was not evenly distributed. Total sleep time, sleep efficiency, the number of night-time awakenings and the time spent awake after falling asleep were all more disrupted in women than in men, at a dose adjusted for body weight. The study was not designed to explain why; the plausible reasons are differences in body water, in the enzymes that clear alcohol, and in how the dose was calculated, and none of them was tested here.

The practical point is not that one group should drink and another should not. It is that "a couple of drinks" is not a fixed dose. The same two glasses are a different experiment on two different bodies, and the person who reports that alcohol ruins their sleep while their partner sleeps fine may both be telling the truth.

When it stops working

The pattern that worries clinicians is not the bad night after a party. It is the slow conversion of alcohol from an occasional tool into a requirement, and the people who have been through it describe the sequence with unusual consistency.

One started at seventeen, using it for panic attacks and sleep, was at a six-pack a night by twenty-one, and at eight or nine beers a night by forty-three. His description of the end state is the important part: eventually it stops working, or you need more of it, and even then you wake after five hours and need a couple more to get back down. Another, eighteen and already unable to sleep without seven or eight drinks, was asking whether once or twice a month was acceptable — a question that is reasonable in the moment and almost always the first sentence of the other story.

The pharmacology explains the escalation. Sedation is not sleep, and the brain adapts to a sedative that arrives every night. The dose that used to end the evening stops ending it, and the second-half disruption does not adapt at all — it gets worse as the dose rises. So the thing you are using to get to sleep is also the thing making the back half of the night worse, and the only way to get more of the first effect is to accept more of the second.

What to do instead

If alcohol is currently part of how you get to sleep, the useful experiment is not abstinence forever. It is separation: find out how much of your night is alcohol and how much is everything else.

One more thing belongs here, because it is the question people actually ask: is six hours after drinking better than three hours without it? On the evidence, probably yes for the acute night and no for the pattern. A single night of sedated sleep beats a single night of none. A habit of sedated sleep is a different calculation, and the second half of the night is where the bill arrives.

Where this fits

Alcohol is one of the things that reliably produces the specific pattern of waking in the small hours — the night that starts well and ends early, with the second half doing the damage. If the wake-up is at the end of the night rather than the middle, waking before the alarm is closer. If you are using a drink the way other people use a tablet, why melatonin does not work covers the difference between a sedative and a clock signal, which is the same distinction in a different bottle.

And if the reason you drink before bed is that the room is too quiet to fall asleep in, the answer may be a sound rather than a substance: whether white noise helps sets out what a continuous noise can and cannot do.