EVIDENCE VS HYPE · CAFFEINE & SLEEP
The hype: As long as your last coffee is six hours before bedtime, caffeine cannot affect your sleep.
Many people use a fixed afternoon cut-off and assume that falling asleep normally proves it works. Others avoid every afternoon coffee because they expect a ruined night. Research supports a more useful middle ground: caffeine can disturb sleep hours after consumption, but the dose and individual response matter. Six hours is neither a universal guarantee nor evidence that every small coffee has the same effect as a large supplement dose.
The short answer
Yes, caffeine can disrupt sleep when taken six hours before bed, particularly at a substantial dose. Controlled trials and systematic reviews support that conclusion. Newer research also finds effects from a 400 mg dose taken even earlier. A small trial did not detect sleep changes from 100 mg four hours before bed, but that does not establish a safe late dose for everybody.
Why people believe it
The useful rule “avoid caffeine late in the day” becomes misleading when it is treated as an exact countdown. People differ in how quickly they clear caffeine and how sensitive their sleep is to it. A cup of coffee is also an unreliable unit: size, bean, preparation and serving strength vary.
Sleep has several dimensions. Time taken to fall asleep is only one. A person can fall asleep readily and still spend less time asleep, wake more often or experience altered sleep stages. A single question such as “did I feel tired?” cannot capture all of those outcomes.
What the evidence says
The six-hour experiment directly tested the claim
Drake and colleagues' 2013 placebo-controlled experiment compared 400 mg of caffeine at bedtime, three hours before bed and six hours before bed. Both reported sleep and a portable objective monitor were assessed. All three caffeine timings disturbed sleep relative to placebo. This directly challenges a six-hour guarantee, but the tested dose was substantial; it is not equivalent to every espresso or cup of tea. Read the trial ↗
The wider literature shows consistent disruption
A 2023 systematic review of 24 studies found average reductions of 45 minutes in total sleep time and seven percentage points in sleep efficiency, with longer sleep onset and more wakefulness after sleep began. The researchers modelled different cut-offs for coffee and higher-dose pre-workout products. Those estimates are population-level modelling, not precise personal deadlines. Differences in protocols mean the pooled average cannot predict exactly what one drink will do to you. Read the review ↗
A newer trial separates a typical dose from a high dose
A randomised double-blind crossover trial published in 2025 studied 23 young men with moderate habitual caffeine intake. Each completed placebo and 100 mg or 400 mg conditions at 12, eight and four hours before bed. The 100 mg conditions did not produce statistically significant objective or subjective sleep changes. The 400 mg dose affected sleep initiation and architecture within 12 hours of bedtime, with greater fragmentation within eight hours. Subjective sleep quality did not always reveal the objective disruption. Read the dose-and-timing trial ↗
An updated review supports the overall direction
A 2025 review of 22 crossover trials involving 956 participants examined polysomnography measurements. It estimated about 35 minutes less total sleep and roughly five percentage points lower efficiency, alongside longer sleep latency. Its subgroup patterns should not be treated as proven differences when statistical tests did not establish those differences. The exact average differs from the earlier review, but both support a real sleep cost across studied conditions. Read the updated review ↗
The catch: “not significant” does not mean “impossible”
The newer trial's 100 mg result is useful, but it comes from 23 men, not a representative sample of every age, sex, medical condition or sensitivity. A small study may miss a modest effect. It also cannot establish what happens after several servings across a day, because total exposure differs from a single scheduled dose.
Most trials assess relatively short periods. They tell us more about an acute night's sleep than about years of habits, insomnia treatment or every real-world caffeine pattern. Habitual use may change perceived stimulation without ensuring that sleep is unaffected. Conversely, someone who is sensitive should not feel obliged to tolerate a dose simply because it was tolerated in a study.
What this means for you
Start by recording all caffeine sources and serving amounts, not just the last coffee. If sleep is poor, move the last serving earlier and reduce the amount. Compare several ordinary nights with broadly similar bedtimes and routines, rather than drawing a conclusion from one unusually stressful night. Decaffeinated alternatives can preserve a familiar ritual without the same stimulant exposure.
Pay attention to sleep duration and waking, as well as whether you can fall asleep. A wearable can provide context, but its sleep-stage estimates are not a diagnostic test. Persistent insomnia deserves appropriate assessment; caffeine reduction can be helpful without being a complete treatment.
EFSA notes that even 100 mg close to bedtime may affect some adults. Its general daily safety assessment of up to 400 mg for healthy adults does not mean that 400 mg is sleep-neutral. The pregnancy limit is 200 mg daily, and medicines or health conditions may require different advice. Avoid concentrated powders and stacking high-caffeine products. Read the safety guidance ↗
A late performance boost may be a poor bargain if it costs recovery sleep. Our exercise-performance investigation explains the separate benefits and why they do not erase this trade-off.
The Evidence vs Hype verdict
Six hours is a useful warning point, not a protective barrier. Match your cut-off to dose and response, and do not assume that feeling able to sleep proves caffeine had no effect.
Research & Sources
- Drake et al. (2013): caffeine zero, three or six hours before bedtime.
- Gardiner et al. (2023): systematic review and meta-analysis of caffeine and sleep.
- Gardiner et al. (2025): randomised dose-and-timing crossover trial.
- 2025 systematic review: caffeine and polysomnography outcomes.
- EFSA: caffeine safety and sleep considerations.
Evidence reviewed: October 2026. Sources were checked against guidance and indexed research records; abstract-only assessment is not represented as full-text review. No commercial product was tested or endorsed. Educational information, not individual medical advice. Editorial method · Affiliate disclosure.