Margen
You have two drinks before bed because your eyelids feel heavy and you assume you will fall asleep faster. A meta-analysis of 27 studies published in 2025 found a less flattering sum: that dose already disturbs REM sleep, while falling asleep sooner only showed up at much higher doses.
Feeling sedated is not the same as measuring sleep. Alcohol slows down the nervous system as its concentration in the blood rises, and you notice it: relaxation, clumsiness, heavy eyelids. None of those three things times how long you take to fall asleep.
It is signed by Carissa Gardiner with a team of sport and sleep physiologists at Australian universities, and it came out in April 2025 in Sleep Medicine Reviews, the journal that surveys the sleep literature. They gathered studies in healthy adults, all with sleep measured in the laboratory rather than self-reported, and split the results by the amount of alcohol. The paper defines a low dose, in these words, as “≤0.50 g/kg or approximately two standard drinks”. At that level they already saw two things: REM sleep took longer to arrive and lasted less.
A standard drink is 10 grams of pure alcohol in Australia, where the authors work, and 14 in the United States. A 150-millilitre glass of wine at 13% holds about 15 grams; a 330-millilitre beer at 5%, about 13. For someone weighing 70 kilos, the ceiling of the low dose, 0.50 grams per kilo, works out at 35 grams: two glasses of wine poured without stinginess already put you inside the range where REM shifts. The paper counts in ten-gram drinks; your glass weighs half as much again.
REM is involved in memory, learning and emotional regulation. You don't need to turn every night into an exam on sleep stages, but it helps to know what you are paying: you can feel drowsy while the sleep that arrives is rearranged. And the damage to REM grew as the dose rose.
A clear reduction in the time it takes you to fall asleep only appeared from what the paper calls a high dose: “≥0.85 g/kg or approximately five standard drinks”. At that point the descent into deep sleep, known as N3, also came earlier. The price was an even bigger disturbance of REM. That is why using alcohol as a sleeping pill is a bad trade: to get the effect you want you have to approach a quantity that spoils the night more and adds risks a sleep study does not even try to measure.
That is the trade: you fall asleep sooner in exchange for sleeping worse.
Two warnings before you use those figures. On total sleep time, on the efficiency of the night and on the number of awakenings after falling asleep the uncertainty was large, and there the meta-analysis concludes nothing; it did not measure next-day performance or perceived sleep quality either. And a standard drink is not a poured glass: it changes with the country, the strength and how much lands in the glass, and weight, sex, food, how fast you drink or medication all move the concentration you reach. These are research categories, not a personal calculator.
A 2013 review led by Irshaad Ebrahim concluded that alcohol reduced sleep onset latency at all doses and consolidated sleep during the first half of the night before fragmenting the second. The 2025 work, newer and quantitative, did not reproduce that advantage at low or moderate doses. The 2013 study was not badly done; when the data are pooled by dose and with another method, the small benefit stops showing up.
So what do you apply tomorrow? If you have one or two drinks for pleasure, don't file them mentally as minutes of sleep gained. Keep drinking separate from treating insomnia. Over several comparable nights, write down four things: what you drank and how much, what time you finish the last drink, what time you switch off the light, and how you function when you get up. Compare the nights with a drink and the nights without using your own numbers, not the feeling of having fallen asleep quickly. And don't use a bad night as an excuse to raise the next dose.
If trouble sleeping keeps coming back, work on your schedule, night-time light, late caffeine and whatever is keeping you awake. If it is persistent, or if you need alcohol to be able to go to bed, take it to a health professional. And if you pour the second drink tonight, pour it for the taste: the 27 studies did not find it a single minute of sleep gained.