Caffeine and sleep: why dose, timing and sensitivity matter

Start with the amount as well as the clock

“Coffee in the afternoon” leaves the amount and proximity to bedtime unclear. Drink sizes and caffeine content vary, so cups are not a consistent dose. The FDA also notes differences in personal sensitivity and the speed at which caffeine leaves the body.

For someone sleeping after a night shift, “afternoon” may describe a different relationship to bedtime than it does for a daytime worker. Describe the actual pattern instead of applying a rule based only on a clock hour. This guide concerns real consumption; a cup appearing in a dream is a separate subject.

What the recent dose-and-timing trial tested

Gardiner and colleagues analysed 23 men in a blinded, randomised crossover trial, published online in 2024 and in a 2025 journal issue. Eligibility required healthy men aged 18–40, excluding sleep conditions and recent shift work. Participants compared placebo with 100 mg or 400 mg, taken 12, eight or four hours before bed, with 48-hour washouts.

They also took their habitual caffeine amount each morning; the test dose was additional. Home sleep assessment used partial polysomnography and diaries.

Read the results without turning them into a guarantee

With 400 mg four hours before bed, recorded total sleep fell about 51 minutes versus placebo. At eight and 12 hours, deep-sleep duration fell but total-sleep reductions were not significant. None of the 100 mg comparisons was significant. A non-significant result does not prove no effect or define a safe cutoff.

The selected sample, single doses, one recorded night per condition and incompletely controlled home conditions limit generalisation. The main-paper methods, results and limitations were inspected; supplements were not independently audited.

An earlier experiment adds a different comparison

Drake and colleagues' blinded crossover study involved 12 healthy sleepers: six women and six men, aged 19–48. They compared 400 mg at bedtime, three or six hours before bed with placebo, using diaries and a portable headband monitor. Some habitual caffeine use continued under the study rules.

Recorded total sleep fell at all three timings; the six-hour diary estimate was not significant. The small sample and device limit generalisation. This account uses indexed original methods, results and discussion after direct PMC access failed.

Falling asleep and feeling fine do not measure the whole night

Falling asleep, staying asleep and feeling restored are different outcomes. In the earlier experiment, the diary–monitor disagreement shows why a reassuring impression cannot settle every measured effect.

Conversely, a difficult night is not automatically caused by caffeine. An original example is someone who drinks their usual afternoon coffee before an unusually noisy night. The drink and the noise are both context; the diary alone cannot isolate either cause. Record what happened without forcing one explanation, and keep a persistent sleep concern open to broader assessment.

A product note can be more useful than counting cups

Caffeine may come from coffee, tea, chocolate, supplements or non-prescription medicines. Decaffeinated coffee and tea can still contain some. Check product information where available; an unknown amount can remain unknown.

Here is an invented label-reading example: a can says “80 mg per serving” and contains two servings. Drinking the whole can means 160 mg from that product. That arithmetic identifies exposure; it does not set a target or judge whether the amount suits you. If a café cannot provide an amount, “large brewed coffee, amount unknown, around 14:00” remains a useful description. Avoid silently treating every cup as the same dose.

Make any change manageable rather than turning sleep into an experiment

For regular use, FDA guidance favours gradual reduction because withdrawal can be unpleasant. Discuss changes to a caffeine-containing medicine with a clinician or pharmacist.

An optional sleep note can record what changed, approximately when, and whether the practical difficulty changed too. It need not become a daily optimisation project. Keep uncertainty visible: “I moved the drink earlier and had several easier nights” describes an observation, while leaving other influences possible. Continuing difficulty deserves attention even if your record does not reveal a simple pattern.

Individual advice matters in circumstances the trials did not cover

The FDA recommends professional advice about caffeine when pregnancy, breastfeeding, medicines or medical conditions are relevant, and about children's or teenagers' consumption. Adult experiments cannot settle those individual questions. A general daily-intake figure is not a promise of undisturbed sleep.

If sleep difficulty is affecting daily activities, NHLBI recommends a medical conversation. Describe the products, approximate timing, actual sleep problem and daytime effects. You do not need to prove caffeine caused it. The sleep-diary guide can help prepare a short account; the substance-and-dream guide addresses the separate question of what research measures about dream reports.

Sources & their limits

These sources support the statements indicated above. Their authors do not endorse this website or provide an assessment of your experience.

1. Gardiner et al. (2025 issue; online 2024) — Dose and timing effects of caffeine on subsequent sleep

   Publisher main-paper methods, results and limitations inspected; supplements not audited.

   Cited in this guide · 3 places

   • What the recent dose-and-timing trial tested
   • Read the results without turning them into a guarantee
   • Did the 100 mg result prove that dose is safe before bed?
2. Drake et al. (2013) — Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed

   Indexed original methods, results and discussion inspected; direct access failed.

   Cited in this guide · 4 places

   • An earlier experiment adds a different comparison
   • Falling asleep and feeling fine do not measure the whole night
   • Is there one time everyone should stop caffeine?
   • Can caffeine affect sleep if I still fall asleep easily?
3. FDA — Spilling the Beans: How Much Caffeine Is Too Much?

   Official advice on products, sensitivity, reduction and individual circumstances.

   Cited in this guide · 6 places

   • Start with the amount as well as the clock
   • A product note can be more useful than counting cups
   • Make any change manageable rather than turning sleep into an experiment
   • Individual advice matters in circumstances the trials did not cover
   • Is there one time everyone should stop caffeine?
   • Does decaffeinated mean caffeine-free?
4. NHLBI — Insomnia: Diagnosis

   Advice to discuss sleep difficulties affecting daily activity and describe relevant habits and medicines.

   Cited in this guide · 2 places

   • Make any change manageable rather than turning sleep into an experiment
   • Individual advice matters in circumstances the trials did not cover

Put the context beside an experience

Continue through the science library