Does Caffeine Actually Improve Endurance? The 1978 Study That Started It All

Short answer: yes, but the honest number is smaller than the famous one. At 3 to 6 mg of caffeine per kilogram of bodyweight, taken about an hour before, endurance time-trial performance improves by roughly 2 to 3 percent on average. The 19.5 percent figure everyone quotes comes from one nine-person study in 1978 that used a very different test.

Both numbers are real. The gap between them is the most useful thing in this article.

The 1978 study: nine cyclists, two rides, one cup of coffee

David Costill and his colleagues took nine competitive cyclists, seven men and two women, and had each of them ride a bike ergometer to exhaustion at 80 percent of VO2 max. Twice.

One trial came an hour after drinking decaffeinated coffee. The other came an hour after drinking coffee containing 330 mg of caffeine. Same riders, same bike, same intensity. The only variable was what was in the cup.

	Decaf trial	Caffeine trial
Time to exhaustion	75.5 min	90.2 min
Fat oxidized	57 g (0.75 g/min)	118 g (1.31 g/min)
Carbohydrate oxidized	about 240 g	about 240 g
Rated effort	harder	significantly easier

Just under fifteen extra minutes. A 19.5 percent improvement out of the same set of legs.

And the detail that makes the study famous is not the time at all. It is that the riders rated the caffeine trial as significantly easier, on a standard perceived exertion scale, while doing fifteen minutes more work.

Why did it feel easier?

Costill's own explanation leaned on fat. His riders burned more than double the fat per minute on caffeine, 1.31 grams against 0.75, while burning the same amount of carbohydrate. The idea was that caffeine frees up fatty acids, the body leans on them harder, and glycogen lasts longer. He also noted a likely effect on nerve impulse transmission.

Nearly fifty years later, the second half of that sentence has held up better than the first.

The leading explanation today is adenosine blockade. Adenosine accumulates in your brain as you burn through energy and binds to receptors that tell you that you are tired. Caffeine is shaped closely enough to park in those spots without triggering them. The message never lands. The work does not get easier. Your read on it does.

The glycogen-sparing story is now considered a smaller piece of the puzzle than it looked in 1978, because the performance benefit shows up in events far too short for fuel availability to be the limiting factor. Costill measured something real. The interpretation moved.

So why is the modern number only 2 to 3 percent?

This is where most articles quietly skip a step, so here it is plainly.

A 2018 systematic review and meta-analysis in Sports Medicine pooled 46 randomized placebo-controlled studies on caffeine and endurance. At moderate doses of 3 to 6 mg per kilogram it found mean power output improved by 3.03 percent and time-trial completion time improved by 2.22 percent.

That is a long way from 19.5 percent, and the reason is the test, not the coffee.

Costill used a time-to-exhaustion protocol: hold a fixed intensity until you cannot anymore. Those tests are famously sensitive. How long you can hold a pace is partly a question of how long you are willing to suffer, so anything that changes your perception of effort moves the number a lot. Modern research mostly uses time trials instead: cover a set distance as fast as you can. That is closer to racing, more reproducible, and produces smaller, more honest effect sizes.

So the 1978 study is not wrong. It answered a slightly different question, in a way that magnifies exactly the thing caffeine is best at.

A 2 to 3 percent improvement is still a lot. In a field race that is the difference between a podium and fourth.

What the research does not support

• It does not work for everyone. In that same meta-analysis, two studies reported slower time-trial performance on caffeine and five reported lower mean power output. Individual response varies, and it is worth testing in training rather than on race day.
• Nine people is nine people. The 1978 study is a landmark, not a large trial. It matters because it replicates, not because it was big.
• More is not better. Past roughly 6 mg/kg the performance benefit flattens out while the jitters, the elevated heart rate and the wrecked sleep do not.
• It will not fix being undertrained or underfed. A few percent on top of a good engine is a few percent. A few percent on top of nothing is nothing.

The practical protocol

1. Dose 3 to 6 mg per kilogram of bodyweight. For a 180 pound athlete that is roughly 245 to 490 mg. The 330 mg Costill used sits right in the middle.
2. Take it about 60 minutes out. Caffeine is close to fully absorbed inside 45 minutes and blood levels peak somewhere around 30 to 60 minutes in, which is why the research protocols dose it an hour before. Drinking it on the drive to the gym puts your peak in the middle of the session, which is a perfectly good place for it.
3. Coffee counts. The 1978 study used actual coffee, not a pill. So have coffee.
4. Test it in training first. See the non-responder note above.

If you want coffee built for that job, PR Blend carries naturally about 50 percent more caffeine than standard coffee and is 100 percent organic. The rest of the Caffeine and Kilos lineup is roasted to order and shipped fresh.

Frequently asked questions

How much caffeine should I take before a workout?

Three to six milligrams per kilogram of bodyweight, roughly 60 minutes before. Going higher does not reliably buy more performance and does cost you sleep.

Does coffee work as well as a caffeine pill?

The landmark 1978 study used brewed coffee and found a 19.5 percent improvement in time to exhaustion. Coffee works.

How long before exercise should I drink coffee?

About an hour. Absorption is largely complete within 45 minutes and blood levels peak around 30 to 60 minutes after drinking.

Does caffeine make exercise feel easier?

That is one of the most consistent findings in the literature. In the 1978 study the riders rated the caffeine trial as significantly easier while lasting fifteen minutes longer. The current explanation is that caffeine blocks adenosine receptors, which is how your brain tracks fatigue.

Why do some people say caffeine does nothing for them?

Because for some people it does very little. In the 46-study meta-analysis a handful of studies found slower times or lower power output on caffeine. Habitual intake, genetics and sleep debt all play a part.

Sources

1. Costill DL, Dalsky GP, Fink WJ. Effects of caffeine ingestion on metabolism and exercise performance. Medicine and Science in Sports. 1978;10(3):155-158.
2. Southward K, Rutherfurd-Markwick KJ, Ali A. The Effect of Acute Caffeine Ingestion on Endurance Performance: A Systematic Review and Meta-Analysis. Sports Medicine. 2018;48(8):1913-1928.
3. Guest NS, VanDusseldorp TA, Nelson MT, et al. International Society of Sports Nutrition position stand: caffeine and exercise performance. Journal of the International Society of Sports Nutrition. 2021;18(1):1.

This article covers exercise performance research. It is not medical advice, and caffeine is not appropriate for everyone. Talk to your doctor about your own situation.