Caffeine has a half-life of roughly five hours in a healthy adult. Drink a cup at 3pm and about half of it is still in your bloodstream at 8pm.
In a controlled trial, 400 mg taken six hours before bed cut measured sleep by more than an hour.
So the 3pm cup isn't really a decision about this afternoon. It's a decision about tonight.
That's the finding. Here's where it comes from, what it does and doesn't mean, and — the part nobody writes about — what people actually lose when they give caffeine up.
Where your 3pm cup is at 10pm
Half-life is the time your body takes to clear half a dose. The StatPearls caffeine monograph, hosted by the US National Library of Medicine and last updated in May 2024, puts caffeine's half-life in the average adult at approximately 5 hours.
Run the arithmetic on a cup at 15:00:
- 20:00 — about 50% of it, still circulating
- 01:00 — about 25%
- 06:00 — about 12.5%, which is roughly when you're deciding whether you need a cup to wake up
You can see the loop from here.
Five hours is only the average, though. The same monograph records that the half-life is reduced by up to 50% in smokers, prolonged to up to 15 hours in pregnancy — particularly the final trimester — and lengthened by liver disease and by drugs that inhibit cytochrome P450 enzymes. Two people can drink the identical cappuccino at the identical moment and be in genuinely different situations by midnight.
Which is also why quitting smoking makes coffee feel suddenly much stronger, and why pregnancy guidance is stricter than general guidance: the World Health Organization recommends that pregnant women with a high daily intake — more than 300 mg a day — lower it, to reduce the risk of pregnancy loss and low birth weight.
The study that made 3pm the cut-off
The specific experiment worth knowing is Drake, Roehrs, Shambroom and Roth, published in the Journal of Clinical Sleep Medicine in 2013.
Twelve healthy normal sleepers took 400 mg of caffeine — in pill form, so the dose was exact — at either 0, 3 or 6 hours before their usual bedtime, with a placebo on the other nights. They slept at home, and sleep was measured both by diary and by a validated portable sleep monitor.
The results:
- Measured objectively, caffeine reduced total sleep time by 1.1 to 1.2 hours at all three timings — including the six-hours-before dose.
- By self-report, the six-hour dose cut total sleep time by 41 minutes, though this only approached statistical significance (p = 0.08).
- Caffeine taken three hours before bed lengthened the time to fall properly asleep by 17.2 minutes.
The authors' own conclusion is the quotable part: the size of the reduction in total sleep time supports the standard sleep-hygiene advice to avoid substantial caffeine for a minimum of six hours before bedtime.
Two caveats we're not going to hide. Twelve people is a small study. And 400 mg is a large single dose — the US Food and Drug Administration describes it as roughly two to three 12-fluid-ounce cups of coffee, which is closer to an afternoon's total than one flat white.
The direction holds up elsewhere, though. A systematic review by Clark and Landolt in Sleep Medicine Reviews found caffeine typically prolonged the time taken to fall asleep, reduced total sleep time and sleep efficiency, and worsened perceived sleep quality.
"Safe" and "won't touch your sleep" are two different sentences
This is where most caffeine articles quietly cheat, so it's worth being precise about it.
The European Food Safety Authority's 2015 Scientific Opinion on the safety of caffeine concluded that:
- Single doses up to 200 mg (about 3 mg per kg of body weight for a 70 kg adult) do not give rise to safety concerns.
- Habitual intake up to 400 mg per day does not give rise to safety concerns for non-pregnant adults.
- Habitual intake up to 200 mg per day by pregnant women does not give rise to safety concerns for the foetus.
- For children and adolescents, the available information is insufficient to derive a safe intake at all.
Not one word of that is a statement about your sleep.
A dose can be entirely safe and still cost you forty minutes in bed. So if you're drinking coffee well inside EFSA's limits and still awake at 3am, you aren't doing it wrong. You're running a stimulant with a five-hour half-life through your evening, and it's doing exactly what it says on the tin.
What you actually lose when you quit
Here's the thing the sleep research doesn't capture, and the reason most people who quit caffeine start again.
Nobody misses the adenosine blockade.
They miss the ten minutes at the kitchen counter. The grinder. The smell. The pause before the day starts. The second cup that's really just an excuse to sit down.
Caffeine is a molecule. The morning cup is a ritual. Those are not the same thing.
And that distinction has actually been tested. In "Rituals Enhance Consumption", published in Psychological Science in 2013, Vohs, Wang, Gino and Norton ran four experiments showing that performing a short ritual before eating or drinking made people rate what followed as more flavourful and more valuable, and made them savour it more — an effect that held for chocolate, lemonade and even carrots.
The ritual wasn't decoration around the experience. It was part of the experience.
Which explains the pattern almost everyone reports. Giving up caffeine is easy for about nine days. What breaks people is the fourth morning of standing at the counter with a mug of rooibos, performing a ritual that no longer has anything in it.
Keep the ritual. Drop the molecule.
The practical version of everything above isn't "cut down". It's: change what's in the cup, and leave every other part of it exactly where it is.
Which is why Mesqi is a grind rather than a powder you stir into hot water. It goes in the moka pot, French press, AeroPress, pour-over or espresso machine you already own, at one heaped teaspoon per 250 ml. Roasted mesquite pod, one ingredient, no caffeine — because mesquite is a legume, and legumes don't make caffeine.
Same grinder. Same pot. Same ten minutes.
The 3pm cup stays. The 3am wake-up doesn't. And it costs R7.11 a cup, which is a different argument, but not an irrelevant one.
If you're using coffee to keep you going — where are you going?
The questions we actually get asked
How long does caffeine stay in your system?
Caffeine's half-life in an average healthy adult is about five hours, so half a dose remains after five hours and a quarter after ten. In pregnancy the half-life can extend to 15 hours; in smokers it can be up to 50% shorter.
Is 3pm too late for coffee?
For most people, yes, if you go to bed before 21:00. The standard advice is to avoid substantial caffeine for at least six hours before bed — and in Drake et al. (2013), 400 mg six hours before bedtime still reduced objectively measured sleep by 1.1 to 1.2 hours.
Does caffeine affect sleep if I fall asleep fine?
Falling asleep is only one measure. Clark and Landolt's systematic review found caffeine also reduces total sleep time and sleep efficiency and worsens perceived sleep quality — all of which you can have while still dropping off in five minutes.
How much caffeine is safe per day?
EFSA found habitual intake up to 400 mg a day raises no safety concerns for non-pregnant adults, and up to 200 mg for pregnant women; the FDA cites the same 400 mg figure. Worth repeating: safety limits say nothing at all about the effect on your sleep.
How much caffeine is safe during pregnancy?
The World Health Organization recommends that pregnant women taking in more than 300 mg a day reduce it; EFSA's figure is 200 mg. Caffeine also clears far more slowly in pregnancy. Ask your doctor or midwife about your own situation.
What can I drink instead of coffee in the afternoon?
Anything caffeine free that keeps the ritual intact — rooibos, herbal tea, or a roasted-pod brew such as mesquite. Given what the research says about ritual, pick something you brew rather than something you stir.
Is caffeine-free coffee the same as decaf?
No. Decaf is a coffee bean with most of the caffeine removed, usually leaving a small residue. A caffeine-free alternative such as roasted mesquite never contained any in the first place.
Sources
- Evans, J., Richards, J.R. & Battisti, A.S. Caffeine. StatPearls [Internet], StatPearls Publishing; last updated 29 May 2024. Half-life, and factors that change it.
- Drake, C., Roehrs, T., Shambroom, J. & Roth, T. (2013). Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed. Journal of Clinical Sleep Medicine, 9(11), 1195–1200.
- Clark, I. & Landolt, H.P. (2017). Coffee, caffeine, and sleep: a systematic review of epidemiological studies and randomized controlled trials. Sleep Medicine Reviews, 31, 70–78.
- EFSA Panel on Dietetic Products, Nutrition and Allergies (2015). Scientific Opinion on the safety of caffeine. EFSA Journal, 13(5):4102.
- US Food and Drug Administration. Spilling the Beans: How Much Caffeine is Too Much?
- World Health Organization, e-Library of Evidence for Nutrition Actions. Restricting caffeine intake during pregnancy.
- Vohs, K.D., Wang, Y., Gino, F. & Norton, M.I. (2013). Rituals Enhance Consumption. Psychological Science, 24(9), 1714–1721.
This article is general information about caffeine and sleep, not medical advice. Mesqi is a food, not a medicine. If caffeine, sleep or pregnancy is a live question for you, speak to a doctor.