The WattPlan Podcast · Episode 44 · 10 min
Caffeine for cycling: how much, when, and who it does nothing for
Next episodes: What should you eat before a sportive or race? · How much sleep does a cyclist need? · How much should you eat on a long ride?
Read the transcript
James. This is the WattPlan Podcast and I'm James. Family, day job, WattPlan, the odd podium when nobody's looking, and more sports science papers on my phone than photos of my kids. Today it's the one legal performance aid nearly every rider already uses. Caffeine. How much, how long before the start, and why does it do nothing at all for some people? Lyra, our AI research voice, has read the studies, and every one of them is linked in the show notes.
Lyra. Three to six milligrams per kilo of body weight, about an hour before the effort you care about. That's the range the trials keep landing on, and for most riders it's worth roughly two to four percent on a time trial. Two things to know before you go and buy gels. Some riders get nothing from it, a few get slower, and taken late in the day it costs you sleep.
James. Two to four percent. On a 25 mile time trial that's over a minute. For a gel?
Lyra. For a measured dose, yes. The International Society of Sports Nutrition pulled the evidence together in a position stand in its own journal in 2021. Doses of three to six milligrams per kilo consistently improved performance, the smallest dose that worked was about two, and going up to nine or so brought a lot of side effects and nothing extra. So for a 70 kilo rider, about 140 milligrams at the very bottom and somewhere between 210 and 420 in the working range. [1]
James. And a strong coffee is, what, 150?
Lyra. That's the trouble with coffee. The same stand reports that similar coffees from similar shops varied eight to nine fold in how much caffeine they held. So you cannot dose from a cup for an event. A gel, a chew, a gum or a tablet tells you on the packet what's in it, and coffee doesn't.
James. Right. And the hour before. Is that a real number or just when people happen to take it?
Lyra. A real one, and it's about how the body handles it. Caffeine in the blood peaks somewhere between half an hour and two hours after you take it, depending on the form. Capsules are slow, about an hour and a half to two hours. Chewing gum is quick, under an hour and a half, because it goes in through the mouth. And once it's in, the half-life is four to six hours on average, with a huge spread between people, from an hour and a half to ten.
James. So if I take it an hour before a four hour sportive, it's still there at the finish.
Lyra. Still there, but fading. Which is why the second dose matters on a long day. The stand reports a trial where riders took caffeinated gum late in a 30 kilometre race and rode the last 10 kilometres about four percent faster. So the pattern for a long event is a dose an hour before the start, then a smaller one before the hardest bit. A gel about an hour out, or gum right at the foot of it, because gum is quick and the stand found it helped when chewed just before the effort, not an hour or two earlier. [1]
James. I always forget to eat on the climb, never mind the gel timing.
Lyra. Then let the reminders nag you. WattPlan's in-ride reminders tell you when to eat on a longer ride, and the feed before the big climb is the one to make the caffeinated one. And if you want the rest of the day's eating, the race day nutrition episode covers it.
James. Okay, here's the one I really want answered. I drink three or four coffees a day. Surely by now it does nothing for me?
Lyra. That's the most common belief in the sport and the best trial we have says it's wrong. The position stand reports a study of 40 trained cyclists split into light, moderate and heavy coffee drinkers, and given a proper race dose before a 30 minute time trial. The heavy drinkers gained as much as the light ones. Habitual intake made no difference at all. [1]
James. So the whole give-up-coffee-for-a-week-before-the-race thing?
Lyra. There's a nuance. In one trial, light users who took caffeine every single day for a month lost the acute effect. In another, three weeks of daily use left the effect smaller but still there. So a race-sized dose every day blunts it. Whether a daily coffee does the same, the trials can't say, only that the heavy coffee drinkers in that study still got the full effect from the race dose. Keep the coffee, add a measured dose on the day, and skip the week of misery. That's how WattPlan would call it too, for what it's worth.
James. Right, that's a relief. Now the bit that worries me. You said some riders get slower.
Lyra. More than 95 percent of caffeine is cleared by one liver enzyme, and a common variant in its gene splits people into fast and slow metabolisers. Guest and colleagues tested that on the bike, in Medicine and Science in Sports and Exercise in 2018. 101 competitive men rode a 10 kilometre time trial with no caffeine, a small dose and a moderate dose, double blind. Across everyone the moderate dose was about three percent faster. The fast metabolisers were nearly seven percent faster. The slow ones were nearly fourteen percent slower. [3]
James. Fourteen percent slower. From a gel.
Lyra. From a moderate dose, in the slow group, yes. And a third group, in the middle, got nothing at either dose. So the two to four percent we started with is an average of people who flew, people who stood still and a few who went backwards.
James. But I don't know which one I am. Nobody's tested my genes.
Lyra. Then test the one thing you can, which is you. Same loop or the same turbo session, same time of day, with your intended dose and without, and do each a few times, because one ride's power moves day to day and expectation is part of what you feel. If you're jittery and the numbers are down every time, that isn't nerves, that's your answer.
James. And the evening thing. Most of my hard sessions are a six o'clock turbo after work. If I have a pre-workout before that, how bad is it really?
Lyra. Bad enough to measure. Gardiner and colleagues pooled 24 sleep studies in Sleep Medicine Reviews in 2023. Caffeine took about 45 minutes off total sleep, and took a chunk out of deep sleep, the kind you recover in. And they worked out the timing. A cup of coffee should be at least about nine hours before bed. A standard pre-workout serving, at least about thirteen. [5]
James. Thirteen hours. So a pre-workout at six is a pre-workout aimed at a seven in the morning bedtime.
Lyra. That's exactly it. For a weekday evening session, ride it without, or have a coffee by mid afternoon and leave it there. Save the full dose for the weekend rides and the event. You'll lose far more from a short night than you'll gain from a sharper Tuesday turbo.
James. Fine. Let's make this useful for everyone. If you're new to this and just want to get round your first sportive?
Lyra. If you're new to this, your normal coffee on the morning is plenty, and the thing to practise is eating and drinking on the bike, not caffeine. Get the fuelling right first.
James. And if you train with power?
Lyra. If you train with power, you've got a test rig. Do the with-and-without session, keep the numbers, and if caffeine helps you, start at three milligrams per kilo from something that states its content, about an hour before the start. Go up only if that dose does nothing and brings no side effects.
James. And if you race?
Lyra. If you race, the two-dose pattern is yours. One before the start, then a smaller one, about two milligrams per kilo, before the decisive climb or the last third, a gel an hour out or gum as it starts, and keep the whole day, coffee included, inside six milligrams per kilo. And protect your sleep the week of the event, because that's where the evening doses do their damage.
James. What does the research still not know?
Lyra. Quite a lot, honestly. Nearly all of this is men. The genotype trial had no women in it, the coffee-drinkers trial had no women in it, and the one small study on the other caffeine gene had twelve. Nobody has tested the two-dose sportive pattern over four or five hours in amateur riders, so it's built from hour-long time trials and one late-race gum trial. And part of what any of us feel from caffeine is expectation, which the article gets into. The gains in the trials were against a placebo, so they sit on top of that.
James. And the medical bit, because someone's going to ask.
Lyra. If you take any medication or have a heart condition, the caffeine question is one for your GP, not a podcast. Ask the Coach in WattPlan and it will tell you the same, then help you with the training side.
James. So, for me, keep the coffee, test the gel on a loop I know, and nothing after three on a turbo night.
Lyra. That's the whole thing. The dose that helps you is the one you've measured on yourself, taken when it will peak, and never the night before a hard day.
Lyra. The full article and every study we mentioned are on the WattPlan blog, linked in the show notes. And if you've enjoyed this, follow the WattPlan Podcast wherever you're listening, so you don't miss the next one.
James. And if you want a plan built around the time you've actually got, get WattPlan on the App Store or Google Play. Thanks for listening.
Voices produced with ElevenLabs. James's voice is a clone of his own, Lyra is an AI voice, and James reviews every episode.
Caffeine for cycling works at 3 to 6 mg per kilo, taken about an hour before you need it. The International Society of Sports Nutrition's 2021 position stand puts the endurance gain at about 2 to 4%, and a 2018 meta-analysis of 44 placebo-controlled trials found time trials 2.26% faster. Some riders get nothing, and a few get slower.
How much caffeine should a cyclist take?
The position stand, written by Guest and colleagues and published on 2 January 2021, is the pooled view. Doses of 3 to 6 mg/kg consistently improve performance, the smallest dose that works may be as low as 2 mg/kg, and very high doses of about 9 mg/kg bring a high rate of side effects with no extra benefit. For a 70 kg rider that is 140 mg at the bottom and 210 to 420 mg in the working range. Side effects, which the stand lists as a racing heart, palpitations, anxiety, headaches and poorer sleep, rise roughly in line with the dose, so the smaller end of the range is where to start.
Southward and colleagues pooled the time trial evidence in Sports Medicine in 2018. In the corrected version of their meta-analysis, 44 placebo-controlled trials at 3 to 6 mg/kg gave a mean power 2.92% higher, give or take 2.18, and time trials 2.26% faster, give or take 2.60. Two trials found slower time trials and five found lower power, so the average hides a spread. On a 25 mile time trial ridden in an hour, 2.26% is about 80 seconds.
The stand reports that the caffeine content of similar coffees from similar shops varied eight- to nine-fold, so for an event you cannot dose from a cup. A gel, a chew, a gum or a tablet states its content.
When should you take caffeine before a ride?
About 60 minutes before the effort is the most common timing in the trials, and the stand explains why. Blood caffeine peaks between 30 and 120 minutes after you take it, depending on the form, and the half-life averages 4 to 6 hours, with a range of about 1.5 to 10 hours between people. Capsules peak at 84 to 120 minutes, chewing gum at 44 to 80, and coffee in about half an hour.
For a four-hour sportive, a dose an hour before the start is still working through the second hour, and the stand reports a trial in which 3 to 4 mg/kg of caffeinated gum taken late in a 30 km race made the final 10 km about 4% faster. So the second, smaller dose goes in before the hardest part of the route. A gel goes in about an hour before it. Gum is absorbed through the mouth, and the stand reports it helped when chewed immediately before the effort but not one or two hours earlier, so gum goes in at the foot of the climb. On rides of 60 minutes or more, WattPlan's in-ride reminders prompt a feed every 45 minutes, and the feed before the big climb is the one to make caffeinated. Our race day nutrition guide covers the rest of the day's eating.
What the trials found
| Study | Who was studied | What they tested | Key number |
|---|---|---|---|
| Guest 2021, position stand | Pooled evidence | Dose, timing and forms | 3 to 6 mg/kg about 60 min before, endurance gain of about 2 to 4% |
| Southward 2018 | 44 placebo-controlled trials | Time trials at 3 to 6 mg/kg | 2.26% faster (SD 2.60), power 2.92% higher (SD 2.18) |
| Guest 2018 | 101 competitive men, 10 km cycling time trial, double-blind | 0, 2 and 4 mg/kg by CYP1A2 genotype | 4 mg/kg 3% faster overall, 6.8% faster in AA carriers, 13.7% slower in CC carriers |
| Gonçalves 2017 | 40 trained male cyclists, double-blind crossover | 6 mg/kg against placebo by habitual intake | 29.92 min against 30.81 min, habitual intake made no difference |
| Gardiner 2023 | 24 sleep studies | Caffeine before bed | 45 min less sleep, sleep efficiency down 7%, coffee at least 8.8 h before bed |
| Hurst 2026 | 17 studies | A placebo described as caffeine | Effect size 0.49 (95% CI 0.28 to 0.69), low-certainty evidence |
Do you need to give up coffee before a race?
The evidence says no. Gonçalves and colleagues at the University of São Paulo tested 40 endurance-trained male cyclists, split into thirds by their habitual intake, 58, 143 and 351 mg a day. With 6 mg/kg of caffeine their 30-minute time trial took 29.92 minutes against 30.81 with placebo and 31.14 with nothing, and the heavy coffee drinkers gained as much as the light ones. The number who improved by more than the test's own variation was similar in all three groups.
In one trial, 18 light users who took caffeine daily for four weeks lost the acute effect, while in another, 20 days of daily use left the effect smaller but still there. And 12 habituated cyclists who went four days without still improved with 3 mg/kg. Our view at WattPlan is that keeping your normal coffee and adding a measured dose on the day is the right trade, and that cutting caffeine for days before an event buys little and feels worse than it is worth.
Why does caffeine do nothing for some riders?
More than 95% of caffeine is cleared by one liver enzyme, CYP1A2, and a common variant in its gene splits people into fast and slow metabolisers. Guest and colleagues' 2018 trial in Medicine and Science in Sports and Exercise is the largest test on the bike. 101 competitive men rode a 10 km time trial with 0, 2 and 4 mg/kg, double-blind. Across everyone, 4 mg/kg was 3% faster. The fast AA carriers were 4.8% faster at 2 mg/kg and 6.8% faster at 4 mg/kg. The slow CC carriers were 13.7% slower at 4 mg/kg, and the AC group got nothing at either dose.
Few riders know their genotype, so your own repeated test on a known loop, with and without, is the practical one, and a rider who feels jittery and slower every time is not imagining it.
Part of what everyone feels is expectation, and Philip Hurst and colleagues published a meta-analysis in the Journal of the International Society of Sports Nutrition on 20 October 2026 of 17 studies in which people were given a placebo and told it was caffeine. Performance improved by a small-to-moderate margin, an effect size of 0.49 with a 95% confidence interval of 0.28 to 0.69, but 9 of the 17 studies were at high risk of bias and the authors rated the certainty of the evidence as low. The gains in the trials above were measured against a placebo, so they sit on top of whatever belief adds, and on the day, belief is on your side too.
Will caffeine before an evening turbo session wreck your sleep?
Gardiner and colleagues pooled 24 studies in Sleep Medicine Reviews in 2023. Caffeine cut total sleep by 45 minutes, lowered sleep efficiency by 7%, added 9 minutes to the time taken to fall asleep and 12 minutes of waking in the night, and took 11.4 minutes out of deep sleep. A cup of coffee, 107 mg in 250 ml, should be at least 8.8 hours before bed, and a standard pre-workout serving of 217.5 mg at least 13.2 hours before.
For an evening session, ride it without caffeine or stop at a mid-afternoon coffee, and save the full dose for weekend rides and event days. Our sleep guide explains what the lost deep sleep costs the next day's training.
What to do before your next event
- Test your dose in training first, on a known loop or a repeatable turbo session, with and without, two or three times each, because one ride's power moves day to day and expectation is part of what you feel. Trust only a gap that keeps showing up.
- Start at 3 mg/kg, about 210 mg for a 70 kg rider, from a product that states its caffeine content, taken about 60 minutes before the start. Go up only if that dose gives you nothing and no side effects.
- On a long sportive, add a second dose of about 2 mg/kg, 140 mg for a 70 kg rider, before the hardest climb, a gel about an hour out or gum at the foot of it, and keep the whole day, coffee included, inside 6 mg/kg.
- Keep drinking your usual coffee in the weeks before. There is no need to abstain, and no evidence that abstaining helps.
- On a weekday evening session, skip the dose. If you take any medication or have a heart condition, the caffeine question is one for your GP, and the WattPlan Coach will say the same if you ask it.
What we don't know yet
Almost all of this evidence comes from men. The genotype trial and the habituation trial had no women at all, the pilot study the position stand cites on the other caffeine gene, ADORA2A, had 12 women. Nobody has tested the two-dose sportive pattern in amateur riders over four or five hours, so it is assembled from hour-long time trials and one late-race gum trial. And how much of your 2% is belief remains an open question.
Sources
- International society of sports nutrition position stand: caffeine and exercise performance, Journal of the International Society of Sports Nutrition
- Correction to: The Effect of Acute Caffeine Ingestion on Endurance Performance: A Systematic Review and Meta-Analysis, Sports Medicine
- Caffeine, CYP1A2 Genotype, and Endurance Performance in Athletes, Medicine and Science in Sports and Exercise
- Dispelling the myth that habitual caffeine consumption influences the performance response to acute caffeine supplementation, Journal of Applied Physiology
- The effect of caffeine on subsequent sleep: A systematic review and meta-analysis, Sleep Medicine Reviews
- Placebo effects of caffeine on sport performance: a systematic review and meta-analysis, Journal of the International Society of Sports Nutrition