Caffeine, alcohol and nicotine are often grouped together in sleep advice, but they interfere with sleep in quite different ways. Caffeine promotes wakefulness, nicotine is stimulating and can also create overnight withdrawal, while alcohol can make someone feel sleepy initially yet alter what happens after sleep begins. Treating all three simply as "substances to avoid" misses some useful distinctions.
Caffeine has the clearest dose-and-timing relationship. It blocks adenosine receptors, reducing one of the signals associated with accumulating sleep pressure. A 2023 systematic review and meta-analysis of 24 controlled studies found that caffeine reduced total sleep time by an average of about 45 minutes, reduced sleep efficiency by 7%, increased sleep-onset latency by around 9 minutes and increased wakefulness after sleep onset by approximately 12 minutes. Deep sleep was also modestly reduced.
That does not mean everyone needs to eliminate caffeine completely. Dose and timing matter substantially. A 2025 controlled study tested 100 mg and 400 mg of caffeine taken 4, 8 or 12 hours before bedtime. In the healthy young men studied, 100 mg taken four hours before bed did not significantly disrupt subsequent sleep, while 400 mg could impair sleep even when consumed 12 hours before bedtime. Individual sensitivity and caffeine metabolism also vary considerably. For one person an early-afternoon coffee may be inconsequential; for another it may still be interfering with sleep hours later.
Alcohol presents almost the opposite problem. Because it can initially make people drowsy, it is easy to mistake sedation for improved sleep. A 2025 systematic review and meta-analysis of 27 studies found that alcohol altered sleep architecture, particularly by delaying REM sleep and reducing the amount of REM obtained. Importantly, REM disruption was detectable even at relatively low doses of around two standard drinks and became progressively greater as the alcohol dose increased.
At high doses, alcohol can shorten the time taken to fall asleep. But that is not necessarily a benefit. The same analysis found that this apparent sedative effect occurred alongside greater disruption of subsequent REM sleep. Evidence for its effects on total sleep time, sleep efficiency and wakefulness after sleep onset was more uncertain, so the common claim that alcohol universally "makes you wake repeatedly in the second half of the night" is stronger than the newest pooled evidence allows. The more defensible conclusion is that alcohol can make sleep arrive more quickly at higher doses while simultaneously changing the architecture of the sleep that follows.
Nicotine introduces another mechanism. It is a stimulant, but regular users can also experience falling nicotine concentrations and withdrawal while asleep. Prospective evidence consistently associates smoking with sleep problems. A systematic review and meta-analysis found that smokers had approximately a 47% greater risk of developing sleep-related problems than non-smokers, although smoking behaviour is intertwined with many other health and lifestyle factors.
A particularly interesting study examined all three substances in people's normal lives rather than a sleep laboratory. Researchers followed 785 adults across 5,164 nights using wrist actigraphy and daily sleep diaries. Nicotine use within four hours of bedtime was associated with around six additional minutes awake after sleep onset and lower sleep efficiency. Alcohol was also associated with slightly lower sleep efficiency. In that particular study, evening caffeine consumption was not significantly associated with the measured sleep outcomes, illustrating why individual studies do not always reproduce the effects seen in controlled caffeine experiments.
There is also a practical trap connecting these substances with sleep. Poor sleep can increase the temptation to use caffeine for alertness the following day. More caffeine, particularly at higher doses or later in the day, can then interfere with the following night's sleep. Alcohol may be used in the opposite direction, as a way to become sleepy at night. Neither approach necessarily corrects the underlying sleep problem, and both can create behavioural cycles around sleep.
For sleep purposes, the most evidence-based approach is therefore more nuanced than "never consume any substances". Nicotine avoidance is strongly justified, and avoiding alcohol as a sleep aid makes particular sense because sedation should not be confused with normal sleep. With caffeine, experimentation with dose and timing is more scientifically defensible than insisting everyone must abstain completely. Someone struggling with unexplained insomnia may find a caffeine-free trial particularly informative, but many people may be able to retain modest morning caffeine without measurable sleep disruption.
This strategy is ultimately about removing potential interference. If sleep is already difficult, caffeine, alcohol and nicotine are three variables worth examining before assuming that the problem requires something additional. Sometimes improving sleep is not about adding another treatment. It is about identifying something already present that may be getting in the way.
