I have a friend who drinks an espresso at 8pm and “sleeps fine.” He also snorts and rolls over eight times a night, and his resting heart rate looks like he is late for a train. Fine is doing a lot of work in that sentence.
Caffeine blocks adenosine receptors. Adenosine is one of the reasons you get sleepy as the day wears on. The liver clears caffeine at a rate that is partly genetic (CYP1A2, if you like letters). Average half-life sits around five hours, with a wide spread. A 3pm mug can still be a quarter-strength mug in your blood at 10pm. You may fall asleep anyway. The sleep is often lighter. You remember it as “I am a light sleeper.”
Drake and colleagues gave people caffeine 0, 3, or 6 hours before bed and watched sleep fall apart even at the 6-hour mark. That paper is why I do not treat “I fall asleep fine” as the end of the argument.
The test that costs nothing
Pick a cutoff. 2pm is a useful default if you go to bed around 10 or 11. If you train at 6pm and you need something, use a small dose and accept that it is a trade.
Run it for fourteen days. Same bed time. No heroic other changes. Look at how long it takes to fall asleep, and whether you wake at 3am with a thought about a spreadsheet.
If nothing changes, you may be a fast metabolizer, or your problem is not caffeine. If it changes, congratulations, you found a cheaper intervention than a new mattress.
What I logged
I am a slow-ish metabolizer based on how a 4pm coffee feels at 10pm (wired, slightly sad). For two weeks I put the last caffeine at 13:30. Sleep latency on the diary dropped from “staring at the ceiling” to “boring.” Waking heart rate lost a few beats. I also became less fun at 4pm meetings. That is the trade. I took it.
When I put a double espresso back at 16:00 for a deadline week, the ceiling staring returned in two nights. I am not a mystery.
What I still do
Coffee with breakfast. Sometimes a second before noon. Tea in the afternoon if I am honest with myself. I do not treat energy drinks as a food group after 40. The crash is not mysterious. It is the bill.
Decaf in the evening is a ritual. The ritual is allowed. The 200mg “focus” packet at 4:30 is how you buy insomnia in convenient stick form.
People with panic disorder, uncontrolled reflux, or pregnancy (not our usual reader, still) need a clinician’s version of this, not mine. If your cardiologist said to cap caffeine, the cardiologist wins.
Common questions
Tea vs coffee? Dose is dose. A strong tea can match a weak coffee. Count milligrams, not folklore.
Does milk change it? Not enough to save a 7pm latte.
Naps? A short nap can help a deficit. A 90-minute nap at 4pm plus caffeine is how you ruin the night twice.
Pre-workout at 6pm? Then you accepted worse sleep for a session. Sometimes that is the right call. Do not pretend it is free.
Sources
- Drake C, et al. Caffeine effects on sleep taken 0, 3, or 6 hours before going to bed. J Clin Sleep Med. 2013.
- Institute of Medicine / FDA consumer materials on caffeine content ranges.
- Nehlig A. Interindividual differences in caffeine metabolism and effect. Pharmacol Rev. (CYP1A2 context)
- Roehrs T, Roth T. Caffeine: sleep and daytime sleepiness. Sleep Med Rev. 2008.
Science edit: Lena Park.
