What's the cutoff time for coffee before sleep?

A simple rule of thumb: stop caffeinated coffee 6 to 8 hours before bedtime. If you sleep at 11 p.m., that means no more coffee after 3-5 p.m. Polysomnography studies (Drake et al., Journal of Clinical Sleep Medicine 2013) show that a 400 mg dose taken 6 hours before bed still cuts total sleep by around one hour.

Pharmacokinetics set the boundary. With a median 5-hour half-life in healthy adults, a dose taken at 5 p.m. is only half cleared by 10 p.m. — roughly 65 mg of a double espresso still circulating when you try to fall asleep. Beyond onset, caffeine disrupts sleep architecture: it shortens slow-wave N3 phases responsible for physical recovery and trims REM phases linked to memory consolidation. A widely cited Wayne State University study (Drake et al., 2013) showed that a 400 mg dose taken 6 hours before bed cut sleep by about an hour and degraded its objective quality even when participants did not perceive the disruption.

The 'right' cutoff depends heavily on individual metabolism. A CYP1A2 fast metaboliser (4 h half-life) can often enjoy a 4 p.m. espresso and still sleep at 11 p.m. A slow metaboliser (6-8 h) needs to stop by 1-2 p.m. Pregnancy, oral contraceptives, certain antibiotics (ciprofloxacin) or antidepressants (fluvoxamine) lengthen the half-life and push the cutoff earlier. Active smokers, who clear caffeine twice as fast, tolerate late coffee better — but should expect a rebound when they quit smoking.

Three common mistakes. First: assuming 'I fall asleep fine after a 10 p.m. coffee' means 'coffee does not affect my sleep'. EEG recordings often show normal onset but shortened deep sleep. Second: ignoring hidden caffeine — a 50 g square of dark chocolate (40-60 mg), a green tea (25 mg), a Coke Zero (32 mg) each can push a careful cutoff over the line. Third: overestimating tolerance. Even heavy drinkers carry caffeine for hours; tolerance concerns the perceived lift, not the pharmacokinetics.

Three ways to keep the afternoon ritual. Swap to a Swiss Water or sugarcane EA decaf (1-7 mg per cup, no measurable impact on sleep). Try a late-afternoon green tea (20-30 mg, 3-4× less than espresso, with calming L-theanine). Or simply move your last cup two hours earlier for a week and check sleep quality with a tracker. This FAQ is informational; for any persistent sleep issue, please consult a doctor or a sleep clinic.

Coffee cutoff by profile

Evidence-based coffee cutoff times by individual profile

The concept of a universal coffee cutoff time — 'don't drink coffee after 2 PM' — is a population-average recommendation that fits no individual perfectly. Evidence-based personalisation requires knowing your caffeine half-life, your typical sleep time, and your desired caffeine concentration at bedtime. A common clinical target is having caffeine concentration below 25% of peak (two half-lives) at sleep onset. For a fast metaboliser with a 4-hour half-life targeting 11 PM sleep: an espresso consumed at 3 PM has one half-life elapsed by 7 PM (50% of peak) and two half-lives elapsed by 11 PM (25% of peak). This individual can drink coffee until 3 PM without measurably impairing sleep quality. A slow metaboliser with a 7-hour half-life targeting the same 11 PM sleep should consume their last espresso before 9 AM — before most people have even started work — to reach the same 25% threshold at bedtime.

Sleep architecture research provides the mechanistic explanation for caffeine's sleep disruption. Even when caffeine doesn't prevent sleep onset — many regular drinkers can fall asleep after afternoon coffee — it measurably reduces slow-wave sleep (SWS) and rapid eye movement (REM) sleep in the early part of the night. SWS is the physically restorative sleep phase where growth hormone is released and metabolic waste is cleared from the brain through glymphatic flow. Caffeine's reduction of SWS means that even apparently normal sleep after coffee is less physically restorative than caffeine-free sleep. This explains the puzzling phenomenon where people who 'sleep fine' after evening coffee still report feeling less rested than caffeine-free periods — they are experiencing quantity-intact but quality-reduced sleep.

Going deeper

Practical tools for personalising coffee cutoff times include: home genetic testing services that include CYP1A2 genotyping (numerous direct-to-consumer options available in Belgium and across the EU), wearable sleep trackers that measure sleep stage architecture and can reveal the signature SWS reduction from caffeine timing, and systematic self-experimentation where coffee cutoff times are adjusted by one hour per week and sleep quality metrics (morning grogginess, energy at midday, sleep onset latency) are tracked. The systematic self-experimentation approach requires no genetic testing or wearable device — only consistent logging and honest self-assessment. Most people who try it for four to six weeks discover their personal cutoff time with reasonable accuracy and report the experiment as one of the most practically useful coffee-health investigations they have undertaken.