Coffee before bed doesn’t affect you? What sleep studies found

Key takeaways

• Falling asleep quickly is only one part of sleep; caffeine can also shorten it, lighten it and add wake time later in the night.
• In a 2013 trial, 400 mg taken 6 hours before bed cut sleep by more than an hour on a sleep monitor, while people’s diaries showed a smaller drop that did not reach significance.
• In a 2025 trial, 400 mg cut deep sleep at every timing tested, up to 12 hours before bed, while 100 mg taken 4 or more hours before bed made no significant difference.
• Regular drinkers grow tolerant to some caffeine effects, but whether that protects sleep remains unclear.
• Comparing two weeks of your own mornings is a better guide than how you feel after one night.

Falling asleep after coffee is not the same as sleeping well

When people say coffee before bed doesn’t affect them, they usually mean they can still fall asleep. That can be true. In a 2023 meta-analysis of 24 studies, caffeine made falling asleep take about 9 minutes longer on average, a delay that is easy to miss.

But falling asleep is only the start of the night. The same analysis found caffeine cut total sleep by about 45 minutes and lowered sleep efficiency (the share of time in bed actually spent asleep) by 7%. It shifted sleep toward lighter stages and reduced deep sleep by about 11 minutes. Those changes happen while you are unconscious, so you cannot watch them happen; caffeine and deep sleep covers that part in detail.

What sleep monitors caught that coffee drinkers didn’t report

In a 2013 trial, 12 healthy sleepers took 400 mg of caffeine at bedtime, 3 hours before bed or 6 hours before bed, or a placebo, while wearing a home sleep monitor and keeping a sleep diary. At every timing, the monitor recorded about 1.1 to 1.2 hours less sleep than with placebo.

The diaries told a softer story. At 6 hours before bed, people’s own records showed about 41 minutes less sleep, a difference that did not reach statistical significance, and only the monitor clearly detected a change at that timing. People did notice something, just less than the monitor measured.

A 2025 double-blind trial found the same pattern. In 23 healthy young men, 400 mg taken 4 or 8 hours before bed added about 26 to 29 minutes of time awake after first falling asleep on the recordings, yet the men’s own estimates of time awake did not change significantly. They did rate their sleep quality lower after 400 mg taken 4 hours before bed, but not at 8 or 12 hours. The authors concluded that people may not be able to tell accurately how much caffeine is affecting their sleep.

Daily life looks similar. In a 2023 randomized trial of 100 adults wearing wrist trackers, people slept about 36 fewer minutes on days they were assigned to drink coffee, with the amount and timing left up to them.

Does tolerance protect regular coffee drinkers’ sleep?

Tolerance is real for some effects. In a controlled study, people taking 300 mg three times a day for 18 days became fully tolerant to caffeine’s subjective effects, and another trial found frequent users develop substantial tolerance to its anxiety-raising effect. More on that in caffeine tolerance.

Sleep is a different question. The authors of the 2025 trial wrote that tolerance to caffeine-induced sleep disturbance remains unclear in habitual consumers. Their participants already used caffeine regularly (under 300 mg a day) and still lost deep sleep after a 400 mg dose.

One 10-day study of 20 men who were heavy regular caffeine users, taking 150 mg three times during the day, found no significant change in total sleep or sleep stages, but REM sleep started later, and the men found it harder to wake up and felt more tired on waking. Its unusual setup, with an evening nap and recordings that began about 5 hours after the men’s usual bedtime, means it cannot settle the question either. So "I’m used to it" may hold for the buzz while your sleep still pays a price, or it may not. Nobody has shown which.

Why coffee before bed can seem to help you sleep

Some people go further and say coffee before bed helps them sleep. None of the trials or reviews cited here found caffeine improving sleep. A systematic review of trials and population studies found that caffeine usually made people slower to fall asleep, cut their sleep time and efficiency, and left them rating their sleep as worse, with the size of the effect depending on dose and timing.

A few things can still make an evening coffee feel harmless. A small or early dose may be too little to matter by bedtime: in the 2025 trial, 100 mg taken 4 or more hours before bed made no measurable difference. The FDA notes wide variation in how sensitive people are to caffeine and how fast they clear it. And if you are very short on sleep, you may drop off despite caffeine, which says more about your tiredness than about how deep the rest of the night is.

If the ritual of a warm evening drink is what you value, decaf keeps it with far less caffeine: the FDA says an 8 oz cup typically has 2 to 15 mg. See caffeine in decaf coffee.

Who is more likely to sleep worse after evening coffee

Both trials above that tested timing were small, and a systematic review noted that most caffeine sleep studies involved male adults in Western countries. The evidence is weakest for some of the people who may be most affected.

• Older and middle-aged adults: a systematic review found older adults’ sleep may be more sensitive to caffeine, and in a lab study 400 mg disrupted sleep more in middle-aged than in young adults.
• People who clear caffeine slowly: half-life ranges from about 2 to 8 hours in healthy adults, and it was longer with estrogen-containing birth-control pills (about 7.9 vs 5.4 hours in one small study) and in pregnancy (8.3 hours on average in one study).
• People on certain medicines: the antidepressant fluvoxamine raised caffeine’s half-life from about 5 to 31 hours in one study. Ask your doctor or pharmacist about yours.
• Anyone having a bigger dose than they think: lab tests of coffee-shop drinks found 58 to 259 mg per serving, and one drink from the same shop ranged from 259 to 564 mg over six days.

A two-week test of whether evening coffee affects your sleep

Because feelings are an unreliable gauge, a simple comparison beats guessing. It will not measure deep sleep, but it can show changes you feel the next day.

• Week one: drink as usual and note the time and amount of your last caffeine each day.
• Week two: keep the same daily total but move all of it before a cutoff; the caffeine cutoff calculator suggests one for your bedtime.
• Each morning, before checking any app, rate how rested you feel and how often you remember waking.
• Compare the weeks. If mornings feel better in the early-cutoff week, the evening coffee was costing you something.
• If sleep is poor either way, caffeine may not be the main issue; goodsleep.today has tools for working on sleep itself.

Common questions