Alcohol disrupts REM sleep from about two drinks. It only speeds sleep onset at five.

A drink before bed is taken on the theory that it helps you fall asleep. The pooled evidence says the sedative half of that theory only shows up at high doses — around five standard drinks — while the disruption to rapid eye movement sleep begins at roughly two [s1]. The dose that produces the benefit is well past the dose that produces the cost.

That asymmetry is the finding, and it is the reason the nightcap question has a clearer answer than most in this field.

What the meta-analysis measured

A systematic review and meta-analysis published in Sleep Medicine Reviews pulled together 27 studies of alcohol consumed before bed by healthy adults, with the specific aim of separating the influence of dose and timing [s1].

The consistent architectural finding was on REM sleep: alcohol delayed the onset of REM sleep and reduced its duration [s1]. A dose-response relationship was identified, such that disruption to REM sleep occurred following a low dose of alcohol — defined as ≤0.50 g·kg⁻¹, or approximately two standard drinks — and progressively worsened as the dose increased [s1].

The sedative effect people are actually seeking behaved differently. Reductions in sleep onset latency, and in the time taken to reach deep sleep — non-REM stage three — were only observed following a high dose, defined as ≥0.85 g·kg⁻¹, or approximately five standard drinks [s1]. The review's own reading is that a high dose may shorten sleep onset latency but that this likely exacerbates the subsequent REM disruption [s1].

Everything else came back uncertain. The effect of alcohol on total sleep time, sleep efficiency and wake after sleep onset could not be determined, with large uncertainty around each estimate [s1]. That is a meaningful gap: three of the measures a person would most naturally use to judge whether they slept well are exactly the ones the pooled evidence cannot settle.

The breathing problem, which is separate and better quantified

Alcohol is a muscle relaxant, and a separate meta-analysis of 14 randomised controlled trials covering 422 participants — 71.9% of them male — measured what that does to breathing during sleep [s2].

The apnea-hypopnea index rose significantly after alcohol administration, by a weighted mean difference of 2.33 events per hour (95% CI 1.41 to 3.25), and mean oxyhemoglobin saturation fell by 0.60 percentage points (95% CI −0.72 to −0.49) [s2]. Respiratory events also lasted longer (WMD 0.86; 95% CI 0.18 to 1.55) and the lowest oxygen saturation reached during the night dropped further (WMD −1.25; 95% CI −2.00 to −0.50) [s2].

The effect was concentrated in the people with the most to lose. The increase in the apnea-hypopnea index was greater in snorers (WMD 4.20; 95% CI 1.19 to 6.50) and greater still in those with a diagnosis of obstructive sleep apnea (WMD 7.10; 95% CI 3.59 to 10.61) [s2]. The review's conclusion is that alcohol is a modifiable risk factor that can result in the development or worsening of obstructive sleep apnea [s2].

What happens over years, rather than over one night

The laboratory work describes single nights. A meta-analysis of 11 cohort studies asked the longer-run question: does drinking predict the later onset of a sleep disorder [s3]?

General drinking was significantly associated with incidence of sleep disorder, with a pooled odds ratio of 1.37 (95% CI 1.22 to 1.54), and no heterogeneity between studies (I² = 0.0%) [s3]. Heavy drinking, oddly, was not significantly associated — odds ratio 1.22 (95% CI 0.94 to 1.60) — but that estimate came with substantial heterogeneity (I² = 81.1%), meaning the constituent studies disagreed sharply with one another [s3].

The authors' summary is appropriately cautious: there was no evidence that alcohol consumption diminished sleep problems, some evidence that general drinking might increase them, and a need for further study [s3].

The limits

None of these three reviews establishes that a drink before bed causes chronic insomnia in a given person. The dose thresholds in the first review are expressed in grams per kilogram of bodyweight and translated into standard drinks by the authors as approximations, and individual alcohol metabolism varies — a limitation the authors flag explicitly, along with the possibility of differential effects by sex, which they call for future work on [s1]. The breathing meta-analysis is heavily male and covers single administrations rather than habitual drinking [s2]. The cohort meta-analysis is observational, and its inconsistent heavy-drinking result should temper confidence in the whole [s3].

The version that survives the evidence

The belief that alcohol helps you sleep is not baseless — it is dose-dependent, and at high doses alcohol does shorten the time to fall asleep and to reach deep sleep [s1]. But REM sleep is degraded at a dose several times lower than that [s1], breathing during sleep worsens measurably, particularly in snorers and people with sleep apnea [s2], and over years, drinking is associated with more sleep disorder rather than less [s3].

There is no dose in this literature that buys a faster sleep onset without paying for it somewhere else. This article describes what the studies found; it is not medical advice, and questions about alcohol use belong with a clinician.

Sources

• The effect of alcohol on subsequent sleep in healthy adults: A systematic review and meta-analysis — Sleep Medicine Reviews, 2024-11-19
• The impact of alcohol on breathing parameters during sleep: A systematic review and meta-analysis — Sleep Medicine Reviews, 2018-06-11
• Alcohol consumption and incidence of sleep disorder: A systematic review and meta-analysis of cohort studies — Drug and Alcohol Dependence, 2020-09-02