When to Stop Drinking Coffee, According to Its Half-Life

Caffeine has a half-life of about five to six hours in a typical adult. That single number explains most of what people get wrong about coffee and sleep.

A half-life means a constant proportion leaves per unit time, not a constant amount. So the decay is slow at the tail, and the tail is longer than intuition suggests.

Following one coffee through an evening

A 200 mg coffee at 4pm, with a 5.5 hour half-life:

 4:00pm   200 mg
 9:30pm   100 mg    still a full espresso
 3:00am    50 mg
 8:30am    25 mg    before you have had today's

At bedtime you have not “had a coffee earlier”. You have the equivalent of a fresh espresso circulating.

The uncomfortable implication for daily drinkers is that you probably never reach zero. There is residue from yesterday sitting underneath today’s dose, permanently.

The study that should change your cut-off

The most useful piece of research here gave participants 400 mg of caffeine at three points before bed: at bedtime, three hours before, and six hours before.

All three reduced total sleep time significantly against placebo. The dose taken a full six hours before bed cut sleep by more than an hour.

That is the headline. The finding that actually matters is what participants reported about themselves.

Their subjective assessments of sleep disruption did not match the objective recordings. The people who had taken caffeine six hours before bed largely did not notice any effect, while the instruments showed it clearly.

So “caffeine doesn’t affect my sleep” is not evidence that caffeine is not affecting your sleep. It is precisely what the affected group said.

Why you can sleep and still feel terrible

Caffeine does not only shorten sleep. It changes its composition.

The damage concentrates in deep slow-wave sleep, the stage that does most of the physical restoration and memory consolidation. Total time in bed can look entirely normal while the quality underneath is worse.

This produces the pattern people find most confusing: eight hours in bed, no memory of waking, and feeling unrefreshed. The sleep happened. The useful part of it was reduced.

It also creates a tidy feedback loop. Poor sleep leads to more caffeine tomorrow, which leads to worse sleep tomorrow night, and the coffee that is supposed to be solving the tiredness is manufacturing it.

What caffeine is actually doing

It does not create energy. It blocks the perception of tiredness, which is a different thing.

Adenosine accumulates in the brain through the day and binds to receptors that promote sleepiness. Caffeine is shaped closely enough to adenosine to occupy those receptors without activating them. The tiredness signal is muted while the underlying sleep pressure keeps building normally.

When the caffeine clears, all of that accumulated adenosine is still there and binds at once. That is the afternoon crash. It is not a deficiency of caffeine; it is the bill arriving.

Tolerance, and what your morning coffee is really for

With regular use the brain grows more adenosine receptors, so a given dose blocks a smaller share of them. This is why the effect fades after a few weeks of daily drinking.

The consequence is more interesting than it sounds. In a habitual drinker, overnight abstinence produces mild withdrawal: headache, fog, low alertness. The morning coffee relieves that withdrawal and returns you to baseline. It does not lift you above it.

Studies comparing habitual consumers on caffeine against people who never drink it find much smaller differences in alertness than either group expects. The genuine lift is available mainly to people who have not built tolerance, and it comes back after a break of a week or two, which is roughly how long receptor levels take to normalise.

Why your friend can drink espresso at 10pm

Metabolism varies enormously, and most of the variation is one enzyme.

• CYP1A2 genotype. Fast metabolisers clear caffeine in as little as two to three hours. Slow metabolisers take eight or more. This is a single common variant, and it splits the population substantially.
• Smoking roughly halves the half-life by inducing the same enzyme. People who quit often find their usual intake suddenly feels excessive, and it is not imagination.
• Pregnancy can extend the half-life to 10 to 15 hours by the third trimester, which is much of why the guideline halves to 200 mg.
• Oral contraceptives can roughly double it.
• Liver function and some medications, including certain antibiotics and antidepressants, slow it further.

If an afternoon coffee reliably wrecks your night while a colleague drinks it at nine and sleeps fine, you are probably not being precious. You are probably a slow metaboliser and they are probably a fast one.

A practical cut-off

Work backwards from your bedtime using your own dose rather than a rule of thumb.

As a rough guide, to have under 50 mg left at bedtime from a single 200 mg coffee at a 5.5 hour half-life, you need about eleven hours. For a 10pm bedtime that means stopping around 11am, which strikes most people as absurd.

A more liveable target is to be under about 50 mg at bedtime in total, which usually means a hard stop somewhere between noon and 2pm for a moderate drinker, and earlier if you drink a lot or metabolise slowly.

Other adjustments worth making

Delay the first cup. Cortisol is naturally high on waking, so caffeine then adds relatively little. Waiting an hour or two produces a noticeably stronger effect from the same dose.

Watch the hidden sources. Green tea, cola, energy drinks, dark chocolate, pre-workout supplements and several over-the-counter painkillers all contain caffeine. A large energy drink can carry more than two coffees.

Know the ceiling. Most health authorities put the limit at 400 mg a day for healthy adults, about four cups of filter coffee. Pregnancy guidance is 200 mg. Adolescents are generally advised to stay under 100 mg.

Taper rather than quit. Withdrawal headaches peak at 24 to 48 hours and can last a week. Cutting by a quarter every few days avoids most of it.

Is caffeine bad for you?

For most healthy adults, moderate intake looks neutral to mildly positive, and coffee consumption is associated with slightly lower all-cause mortality in large observational studies.

The clearest downside is not cardiovascular. It is sleep: real, dose-dependent, and routinely unnoticed by the person experiencing it.

Which makes the timing question more important than the quantity question, and it is the one almost nobody asks.