Caffeine keeps you awake by blocking adenosine, the tiredness signal your brain builds up all day 1. Your liver clears it through one enzyme, CYP1A2, and how fast that runs is mostly genetic 2. An average non-smoker halves a dose in about five hours; a fast metaboliser does it in two, a slow one in ten 3.
Half-life is not clearance. After five hours half your coffee is still on board; being properly clear takes three to five half-lives. The sleep effect outlasts even that, because the blocked receptors stay quiet after the caffeine has gone 4.
Afternoon coffee costs measured sleep, even when you don't feel it. A morning 200 mg still altered the overnight EEG thirteen hours later, long after the blood had cleared 4. Across trials: less deep sleep, slower onset, and habitual drinkers underestimating what the polysomnograph records 5.
You may be a slow metaboliser without knowing. Any of these puts you there:
- The pill. Combined oral contraceptives roughly double your caffeine half-life 7.
- Pregnancy. By the third trimester you clear it about three times slower 8.
- Just quit smoking. Smoking sped clearance up by half; it renormalises over four to six days, which is why coffee suddenly wires you 3.
- One cup makes you jittery. About a third of people carry an ADORA2A variant that makes caffeine more anxiety-provoking whatever the clearance 9.
Set the cut-off by your category, then hold it.
The loop is self-feeding. The afternoon coffee shortens tonight's sleep; the short sleep makes tomorrow harder; you drink more coffee to cover it. Fatigue and fix share a source you can't tell apart, and years of it become calling yourself a bad sleeper 5.
Give it two weeks. The first sober nights read differently on a tracker within days: faster to fall asleep, more deep sleep, fewer 3 am wakings 4. The felt change lags a week, because habitual drinkers have lowered what they expect from a night. By week two the afternoon coffee is optional.
The fine print — when to skip it, and what people get wrong
Some cases need a smaller dose too. Pregnancy: cap at 200 mg/day, morning-only 10. Fluvoxamine or ciprofloxacin can stretch the half-life past thirty hours, so cut the dose 11. Uncontrolled hypertension or arrhythmia: ask your clinician 12.
"I sleep fine on late coffee": subjective onset is a weak detector; the EEG shows the loss 6. "I'm tolerant now": tolerance builds for the alertness, barely for the sleep damage 13. Decaf isn't zero, at 2–15 mg a cup.
- 1Bjorness TE, Greene RW (2009). Adenosine and sleep. Current Neuropharmacology. link
- 2Thorn CF, Aklillu E, McDonagh EM, Klein TE, Altman RB (2012). PharmGKB summary: caffeine pathway. Pharmacogenetics and Genomics. link
- 3Carrillo JA, Benitez J (2000). Clinically significant pharmacokinetic interactions between dietary caffeine and medications. Clinical Pharmacokinetics. link
- 4Landolt HP, Werth E, Borbely AA, Dijk DJ (1995). Caffeine intake (200 mg) in the morning affects human sleep and EEG power spectra at night. Brain Research. link
- 5Clark I, Landolt HP (2017). Coffee, caffeine, and sleep: a systematic review of epidemiological studies and randomized controlled trials. Sleep Medicine Reviews. link
- 6Drake 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. link
- 7Abernethy DR, Todd EL (1985). Impairment of caffeine clearance by chronic use of low-dose oestrogen-containing oral contraceptives. European Journal of Clinical Pharmacology. link
- 8Knutti R, Rothweiler H, Schlatter C (1981). Effect of pregnancy on the pharmacokinetics of caffeine. European Journal of Clinical Pharmacology. link
- 9Childs E, Hohoff C, Deckert J, Xu K, Badner J, de Wit H (2008). Association between ADORA2A and DRD2 polymorphisms and caffeine-induced anxiety. Neuropsychopharmacology. link
- 10EFSA Panel on Dietetic Products, Nutrition and Allergies (2015). Scientific opinion on the safety of caffeine. EFSA Journal. link
- 11Jeppesen U, Loft S, Poulsen HE, Brosen K (1996). A fluvoxamine-caffeine interaction study. Pharmacogenetics. link
- 12Robertson D, Frolich JC, Carr RK, Watson JT, Hollifield JW, Shand DG, Oates JA (1978). Effects of caffeine on plasma renin activity, catecholamines and blood pressure. New England Journal of Medicine. link
- 13Weibel J, Lin YS, Landolt HP, Garbazza C, Kolodyazhniy V, Kistler J et al. (2020). Caffeine-dependent changes of sleep-wake regulation: evidence for adaptation after repeated intake. Progress in Neuro-Psychopharmacology and Biological Psychiatry. link
დაკავშირებული სახელმძღვანელოში (6)
- — An afternoon coffee masks sleep debt and steals the night's sleep you'd use to repay it.
- — Like a late coffee, an evening drink quietly degrades the sleep you don't blame it for.
- — The whole reason for a 2 PM cutoff is caffeine's long half-life — it's still in you at bedtime.
- — Timing matters both ways — an afternoon nap and a too-late coffee compete for the same window; coffee before a short nap can sharpen the wake-up.
- — Caffeine is the workhorse nootropic, and getting the timing right is the difference between sharper days and wrecked sleep.
- — How fast you clear caffeine is largely genetic — the same metaboliser difference these tests read.
Caffeine Half-Life
Anchor dimension. Drake 2013 demonstrated 400 mg caffeine 6 hours pre-bed costs roughly an hour of polysomnography-measured total sleep time; Landolt 1995 showed morning caffeine measurably alters overnight EEG slow-wave power. Moving caffeine to morning-only produces same-night improvement in sleep onset, slow-wave sleep, and total sleep time, particularly in slow metabolisers.
A mild but persistent lifestyle shift: saying no to the afternoon coffee daily, navigating a 1–3 day withdrawal window when changing established habits, and learning to recognise hidden caffeine sources (decaf, tea, pre-workouts). No willpower-intensive routine, but a real daily choice.
Multiple consistent polysomnography RCTs (Drake 2013, Landolt 1995) anchor the sleep-architecture finding; decades of clinical pharmacology document the half-life variation and its modifiers (Carrillo and Benitez 2000, Knutti 1981, Abernethy and Todd 1985); ADORA2A-anxiety link replicated independently (Alsene 2003, Childs 2008). Not a 5 only because no large pragmatic RCT of cut-off timing as a stand-alone intervention exists.
The dominant felt benefit. Habitual caffeine drinkers compensate for caffeine-caused sleep loss with more caffeine; breaking that loop produces a clear, sustained increase in baseline daytime energy within two weeks, often making the afternoon dose unnecessary (Clark and Landolt 2017, Weibel 2020).
Restoration of slow-wave sleep and total sleep time within days of moving caffeine to morning-only produces measurable next-day wellness — fewer headaches, steadier appetite, less afternoon crash (Drake 2013, Landolt 1995). Effect size modest in light users; substantial in slow metabolisers.
Improvement is secondary — morning caffeine still provides direct focus lift; the timing intervention adds the focus benefit of better-rested sleep on top, which is real but smaller than the morning dose itself.
Two pathways. ADORA2A risk-allele carriers (Alsene 2003, Childs 2008) show measurable anxiety reduction within hours of removing caffeine. Better sleep produces steadier baseline mood for everyone (Clark and Landolt 2017). Effect modest for non-anxious users with adequate sleep already.
Indirect, via the long-term skin and facial effects of restored sleep quality once mis-timed evening caffeine is removed. Effect is small relative to direct dermatological interventions but real over months to years (Clark and Landolt 2017).
Indirect via chronic sleep restriction, which is itself associated with mortality and disease endpoints. The half-life-timing intervention's longevity effect is contingent on it actually moving sleep, and the effect chain is several links long.