Alcohol and sleep — why two glasses destroy six hours of rest

Mariana is 44. A marketing director, two children, late dinner, wine in the glass. Doctor, I sleep. I'm out in five minutes. But I wake up at four in the morning, sweating, heart racing, and I toss until six. Oh, I thought it was perimenopause.

I asked for the report from the ring she has been wearing for nine months. We crossed it with the food diary. On every night with two or more glasses of wine, three patterns repeated: resting heart rate 8 to 12 beats above baseline, deep sleep cut by 35%, REM sleep cut by 40%. On nights without alcohol, the same numbers returned to normal. It was not perimenopause destroying her sleep. It was dinner.

Mariana is the common case. She is not an alcoholic. She is a functional adult doing what she was taught to do to “relax” — and paying a bill no one explained to her.

The illusion of the bedtime glass

Western culture wrapped alcohol in an elegant package: wine with dinner, a drink to wind down, “one glass is good for the heart.” Cardiovascular literature has already demolished the second argument — the “benefit” disappeared in analyses with proper controls. But the sleep argument is the most entrenched, because it has the appearance of truth. You drink wine. You're out fast. Therefore, you slept.

The problem is that falling asleep is not sleeping. Falling asleep is only the first stage. What restores body and brain are the deep stages — slow-wave sleep (N3) and REM. And it is precisely in those two that alcohol cuts deep.

The classical review by Ebrahim and colleagues, published in 2013 in Alcoholism: Clinical and Experimental Research, described the pattern clearly: in the first half of the night, alcohol acts as a sedative — it reduces sleep latency and increases deep sleep in the first two hours. In the second half, the effect inverts. As alcohol is metabolized, an excitatory rebound sets in — sleep fragmentation, REM suppression, increased heart rate, awakenings.

The most complete meta-analysis to date — Gardiner and colleagues, in Sleep Medicine Reviews (2025), pooling 27 studies — confirmed and quantified the effect. Doses as low as 0.5 g of alcohol per kilogram of body weight (approximately two standard drinks, or two 150 mL glasses of wine for a 65 kg / 143 lb woman) already significantly reduce REM duration and delay its onset. Higher doses worsen the effect in a dose-dependent way.

It is not “if you drink a lot.” It begins at two.

What wearables show in high resolution

What polysomnography showed in the lab, rings and watches now show at scale. The Pietilä and colleagues study, published in 2018 in JMIR Mental Health with more than 4,000 Finnish employees monitored for months, observed a simple and forceful finding: alcohol, even at moderate doses and in healthy people, depressed parasympathetic tone in the first three hours of sleep in a dose-dependent way. Heart rate climbed, heart rate variability (HRV) fell, and the autonomic quality of rest collapsed. Regular physical activity and youth did not protect.

That is why Mariana's Oura, Whoop, and Apple Watch all showed the same pattern. They are not “wrong.” They are too right.

In patients who have used wearables for more than six months, the signature is almost pathognomonic. A night with two glasses of wine: HRV falls 20 to 35%, resting heart rate climbs 7 to 15 bpm, deep sleep falls 25 to 50%, REM falls 20 to 40%. The first half of the night looks better. The second half — the one that matters for REM — is where the damage shows up. And the early awakening between 3 and 5 a.m., with a sense of racing heart, anxiety, heat, is the most common clinical manifestation.

Why this matters far beyond fatigue

Cutting 30 to 50% of deep sleep is not a cosmetic problem. Slow-wave sleep is when the brain's glymphatic system does its housekeeping — removing metabolic waste, including beta-amyloid, from neural tissue. It is when growth hormone is released in pulse. It is when declarative memories consolidate.

REM is when the brain processes emotional load, integrates learning, regulates mood. Reducing REM chronically is associated with greater emotional reactivity the next day, worse cognitive performance, and — in longitudinal studies — higher risk of cognitive decline and depression.

Add that alcohol relaxes pharyngeal musculature and worsens obstructive sleep apnea, especially the female and mild form, often undiagnosed. Add that it increases nighttime awakenings to urinate (diuretic effect from inhibition of vasopressin). Add that it dehydrates, increases night sweats, and — in perimenopausal women — amplifies hot flashes.

It is not that wine “slightly disturbs” sleep. It is that it transforms a biologically restorative night into a biologically unproductive one.

Tolerance is not protection

One of the most common defenses in the office is: I sleep well even when I drink, I always have. Subjective tolerance is real — the brain habituates to the sensation. But the architecture of sleep does not tolerate. Studies of chronic drinkers show that REM suppression and fragmentation persist; what is lost is the perception of them. The wearable keeps showing the same numbers. The patient simply stopped noticing.

And there is a gender layer that deserves attention. Women metabolize alcohol more slowly — lower total body water, lower gastric alcohol dehydrogenase activity. The same amount of alcohol reaches higher blood levels in a woman than in a man of the same weight. Perimenopausal women metabolize even more slowly. The “two glasses” that for an 80 kg / 176 lb man corresponds to 0.4 g/kg, for a 60 kg / 132 lb woman is easily 0.7 g/kg — clear REM-suppression territory.

What to do (without moralism)

The literature is consistent enough for some practical positions, without requiring full abstinence if that is not the goal:

• A window of at least 3 to 4 hours between the last drink and bedtime. Alcohol consumed at 7 p.m. has much less impact than alcohol at 10 p.m. It allows most of it to be metabolized before primary REM sleep.
• A limit of one standard drink for women, two for men, on nights you drink. Not as a magic formula, but as the point above which evidence shows measurable impact.
• At least 4 nights a week without any alcohol. The evidence does not support “moderate daily consumption” as neutral for sleep.
• If you use a wearable, observe the pattern for two weeks. Compare HRV, resting heart rate, and deep sleep between nights with and without alcohol. The clarity of your own data convinces more than any article.
• Substitutes for the “wind-down” ritual: warm tea (chamomile, mulungu), a hot bath 90 minutes before bed, low-light reading, stretching, 4-7-8 breathing. Not as immediate as the glass — but they don't collect the early-morning bill.

The reading the Continuum makes

In the Continuum Plenya panel, wearable data enters alongside blood, symptoms, and history. When sleep is fragmented and HRV is low, the first question — before ordering polysomnography, before considering hormone replacement, before medicating anxiety — is alcohol, caffeine, dinner timing, evening light. These are variables modifiable in weeks, with enormous impact over months. It is low-cost, high-return medicine — visible only when someone looks at the numbers in the right context.

Mariana stopped wine at dinner for 30 days. HRV climbed 28%. Deep sleep returned to 1h40 per night. Waking at 4 a.m. disappeared in the second week. She did not return to total abstinence — she now drinks twice a month, on occasions. I didn't know one could sleep like this, she said at follow-up.

Bad sleep at 44 is not perimenopause. It is not stress. It is not “life.” It is, almost always, a sum of operable variables — and the dinner glass is the first in line.