Caffeine: how much, how late, and what it does to sleep

Caffeine: how much, how late, and what it does to sleep

“I can have coffee at 6pm and still fall asleep fine” is one of the most common things people say about their own caffeine tolerance, and it’s often true, in the narrow sense that they do fall asleep. What the research shows is that falling asleep and sleeping well are different things, and caffeine’s effects on the second one persist for far longer than most people realise.

Why caffeine lingers longer than it feels like it does

Caffeine works by blocking adenosine receptors in the brain, adenosine being a chemical that builds up through the day and makes you feel sleepy. Blocking it is what produces the alert, awake feeling caffeine is taken for. The catch is caffeine’s half-life, the time it takes your body to clear half of it, averages around 5 to 6 hours in healthy adults, though this varies considerably between individuals. That means a coffee at 3pm can still have a meaningful amount of caffeine active in your system at 9pm, even if you don’t consciously feel “wired” anymore.

This is the crux of why people underestimate caffeine’s evening impact: feeling ready for bed and having cleared the caffeine from your system are not the same thing. Subjective sleepiness can return well before the drug is actually gone.

What the landmark study actually found

The most-cited research on this is a 2013 randomised, double-blind, placebo-controlled trial by Christopher Drake and colleagues at the Henry Ford Sleep Disorders and Research Center. Twelve healthy adults took 400mg of caffeine (roughly the amount in two strong coffees) at one of three points: right at bedtime, 3 hours before, or 6 hours before, and their sleep was measured objectively.

Even at the 6-hour mark, the results were striking: total sleep time dropped by more than an hour, sleep latency (time to fall asleep) increased by roughly 24 minutes on average, and slow-wave sleep, the deepest, most physically restorative stage, was significantly reduced. The researchers’ conclusion was blunt: caffeine taken 6 hours before bed still meaningfully disrupts sleep, providing direct empirical support for the general sleep-hygiene advice to avoid substantial caffeine intake for at least that long before bed.

Dose matters as much as timing

A more recent 2025 randomised crossover trial, published in the journal Sleep, tested this more precisely by varying both dose and timing together. Twenty-three adults completed multiple conditions: placebo, and 100mg or 400mg of caffeine consumed 12, 8, or 4 hours before bedtime, with objective sleep measured via in-home polysomnography.

The results added useful nuance to the picture: the 100mg dose (roughly a single coffee) showed no significant effect on objective or subjective sleep compared with placebo across the timings tested. The 400mg dose, however, produced significant disruption, and did so even at the 12-hour mark, the longest gap tested. In other words, a single normal coffee appears to carry meaningfully less risk than the classic study’s 400mg dose, but a genuinely high-caffeine day (two or more strong coffees, energy drinks, certain pre-workout supplements) can still measurably disrupt sleep architecture even taken very early in the day.

Why your own experience might be misleading you

Caffeine metabolism varies substantially between individuals, driven largely by genetic variation in the CYP1A2 gene (which controls how quickly your liver breaks caffeine down) and the ADORA2A gene (which affects how sensitive your adenosine receptors are to caffeine’s blocking effect in the first place). Roughly half the population are considered “fast metabolisers” and clear caffeine notably quicker than the other half. Certain medications and conditions can push half-life even further in the other direction, some CYP1A2-inhibiting drugs have been shown to extend caffeine’s half-life from around 5 hours to over 30 hours in specific cases.

This genetic variability is part of why anecdotal claims of “caffeine doesn’t affect my sleep” deserve some scepticism, even from the person making them. Research has found that people can genuinely feel they slept fine after evening caffeine, while objective measurements of their sleep, reduced slow-wave sleep and more fragmented sleep architecture, tell a different story. Subjective sleep quality and objectively measured sleep quality aren’t always the same thing, particularly with caffeine specifically.

What this means practically

Given caffeine’s roughly 5-6 hour average half-life and the trial evidence that even a 6-hour gap didn’t fully protect sleep at higher doses, a genuinely conservative cutoff sits somewhere around 8-10 hours before bed for anyone particularly sensitive to caffeine’s sleep effects, earlier than the “no coffee after 2pm for a normal bedtime” rule of thumb many people already follow, but broadly consistent with it. If you drink a single, moderate coffee (around 100mg), the 2025 trial suggests the objective sleep risk is considerably lower even fairly close to bedtime. If your day includes multiple coffees, energy drinks, or other stacked caffeine sources reaching 400mg or more, treating that total as needing a longer buffer, potentially most of the day, is better supported by the evidence.

The honest takeaway

Caffeine’s reputation as something you either “can” or “can’t” have in the evening oversimplifies a genuinely dose- and person-dependent picture. A single normal coffee appears to carry limited objective sleep risk even reasonably close to bedtime; a high-dose day can measurably disrupt sleep architecture even 12 hours out. And because individual metabolism varies so much, feeling unaffected isn’t reliable proof that your sleep isn’t being quietly disrupted underneath.

FAQs

How long does caffeine actually stay in your system?

Caffeine’s average half-life in healthy adults is roughly 5 to 6 hours, meaning about half of what you consumed is still active in your body that long after drinking it, with meaningful individual variation either side of that average.

Is it really true that coffee 6 hours before bed can disrupt sleep?

Yes, according to a landmark randomised trial: 400mg of caffeine taken 6 hours before bed reduced objectively measured total sleep time by over an hour and reduced slow-wave (deep) sleep, even though participants weren’t necessarily consciously aware of feeling “wired.”

Does a single normal coffee affect sleep the same way as multiple coffees?

No. A 2025 randomised trial found 100mg (roughly one coffee) showed no significant objective sleep disruption across the timings tested, while 400mg significantly disrupted sleep even 12 hours before bed. Total daily dose matters as much as timing.

Why do some people say caffeine doesn’t affect their sleep at all?

Genetic variation (particularly in the CYP1A2 and ADORA2A genes) means caffeine metabolism and sensitivity differ substantially between people. Some people also report feeling fine subjectively while objective measurements show their sleep architecture was still disrupted.

What’s a reasonable caffeine cutoff time before bed?

Given caffeine’s roughly 5-6 hour half-life and evidence that higher doses disrupt sleep even 6-12 hours out, a cutoff of around 8-10 hours before bed is a reasonably conservative, evidence-consistent guideline for anyone sensitive to caffeine’s sleep effects.

Does the type of caffeinated drink matter, or just the total amount?

The research generally focuses on total caffeine dose rather than the specific drink, so coffee, tea, energy drinks and supplements should generally be added together when estimating your total intake and how it might affect a given evening’s sleep.