The One Number You Need to Know: 5.7 Hours
The average caffeine half-life in a healthy adult is 5.7 hours (range: 1.5–9.5 hours). A half-life is the time for your body to eliminate 50% of a substance. This means:
- A 200mg coffee at 2:00pm → ~100mg still active at 7:42pm
- That same dose → ~50mg at 1:24am
- Full clearance (>95%) → around 2:00pm the *next day*
Caffeine Content by Drink
| Drink | Serving | Caffeine (mg) | Half-life clearance (90%) |
|---|---|---|---|
| Espresso (single) | 30ml | 60–75mg | ~19 hours |
| Filter coffee | 250ml | 80–150mg | ~19–24 hours |
| Cold brew | 250ml | 150–200mg | ~24–29 hours |
| Red Bull (250ml) | 250ml | 80mg | ~19 hours |
| Green tea | 250ml | 25–45mg | ~14–17 hours |
| Matcha latte | 250ml | 60–80mg | ~18–20 hours |
| Dark chocolate (50g) | 50g | 20–30mg | ~13–16 hours |
How Caffeine Works: Blocking Adenosine
Caffeine produces wakefulness by a specific mechanism: it is structurally similar to adenosine, a neurotransmitter that accumulates in the brain during waking hours and progressively builds sleep pressure (the feeling of needing to sleep). Caffeine molecules occupy adenosine receptors without activating them — effectively blocking the signal.
When caffeine wears off and the blockade clears, all that accumulated adenosine is still there, binding simultaneously. This is the cause of the classic caffeine crash: a sudden wave of fatigue as the delayed sleep pressure hits at once.
The deeper implication is that caffeine does not create energy — it borrows it. The adenosine debt must still be repaid, usually through sleep.
Why Your Clearance Speed Differs from the Average
The enzyme CYP1A2 in your liver is responsible for ~95% of caffeine metabolism. Genetic variations in this enzyme explain 35–50% of the variation in individual clearance rates:
| Population group | Approximate half-life | Why |
|---|---|---|
| Fast metabolisers (CYP1A2 fast variant) | 2–4 hours | Genetic CYP1A2 efficiency |
| Average healthy adults | 5–7 hours | Standard enzymatic activity |
| Slow metabolisers (CYP1A2 slow variant) | 7–10 hours | Reduced enzyme activity |
| Pregnant women (3rd trimester) | 15–18 hours | Oestrogen inhibits CYP1A2 |
| Women on oral contraceptives | 7–11 hours | Pill hormones slow metabolism |
| Smokers | 3–5 hours | Tobacco induces CYP1A2 |
| Newborns | 65–130 hours | Immature liver enzymes |
Caffeine and Deep Sleep: The Hidden Cost
Even caffeine that produces no subjective alertness effect still reduces deep sleep quality. Research using polysomnography shows that caffeine consumed 6 hours before bedtime reduces slow-wave sleep (N3) — the most physically restorative sleep stage — by 15–25%, without subjects reporting any difficulty falling asleep.
This is the insidious problem with afternoon caffeine: you fall asleep on time, feel like you slept 8 hours, but your recovery and cognitive performance the following day are measurably impaired because N3 was compressed.
Time-to-Clear Reference Table
Use this table to find when 90% of your caffeine dose will be eliminated, assuming a 5.7-hour half-life:
| Caffeine dose | Consumed at | 50% cleared (half-life) | 90% cleared | 95% cleared |
|---|---|---|---|---|
| 80mg (1 espresso) | 8:00am | 1:42pm | 7:04pm | 9:46pm |
| 80mg | 12:00pm | 5:42pm | 11:04pm | 1:46am |
| 160mg (2 espressos) | 2:00pm | 7:42pm | 1:04am | 3:46am |
| 200mg (filter coffee) | 2:00pm | 7:42pm | 1:04am | 3:46am |
| 200mg | 4:00pm | 9:42pm | 3:04am | 5:46am |
| 400mg (high intake) | 2:00pm | 7:42pm | 1:04am | 3:46am |
Practical Cutoff Guidelines
Based on a 5.7-hour half-life and a target of less than 25mg active at bedtime:
| Bedtime | Maximum caffeine cutoff (200mg dose) | Maximum cutoff (80mg dose) |
|---|---|---|
| 10:00pm | 2:00pm (8 hours before) | 4:00pm (6 hours before) |
| 11:00pm | 3:00pm | 5:00pm |
| 12:00am | 4:00pm | 6:00pm |
| 6:00am (shift worker) | 10:00pm previous night | 12:00am |
Common Myths Debunked
"Drinking water flushes caffeine faster" — False. Caffeine clearance is enzymatic, not renal. Hydration does not accelerate CYP1A2 activity. Drinking water helps with the diuretic effects of caffeine but does not speed up half-life.
"Eating food slows caffeine absorption, so it's safer to drink coffee with a meal" — Partially true. Food slows *absorption* rate (peak plasma level is delayed by 30–45 minutes), but the *total amount* metabolised and the half-life remain unchanged. You still process the same total dose.
"Decaf has no caffeine" — False. Decaffeinated coffee typically contains 2–15mg of caffeine per 250ml serving. For caffeine-sensitive individuals or those in late pregnancy, this can still be meaningful.