Caffeine as an Ergogenic Aid: How to Read the ISSN Position Stand

The 2021 International Society of Sports Nutrition (ISSN) position stand puts caffeine's most consistent, moderate benefit in aerobic endurance, at doses of 3 to 6 mg/kg body mass taken roughly 60 minutes before exercise. For strength, sprinting, and team-sport actions the evidence is genuinely weaker and more scattered. And underneath every one of those group averages sits a wide spread, so the number that describes a study population is a poor forecast for you.

Key points#

• The strongest, most reproducible benefit is in aerobic endurance, and even there the average effect is statistically small (about a 2 to 3 percent gain).
• The best-supported dose is 3 to 6 mg/kg, taken about an hour before effort.
• Strength, power, repeated sprints, and team sports show smaller, less consistent effects.
• Individual response varies widely, partly by genetics (the CYP1A2 enzyme) and partly by habit.
• Higher doses (around 9 mg/kg) add side effects without adding performance.

A confidence map, not a recipe#

A position stand is easy to misread as a set of instructions. It is closer to a map that shades each region by how confident the authors can be. The 2021 document, led by Nanci Guest and published in the Journal of the International Society of Sports Nutrition, does not run a new experiment. It grades and pools the existing trials and meta-analyses, so its sentences carry very different weight depending on how much evidence sits beneath each one.

That distinction is the whole skill. Some statements rest on dozens of randomized trials combined into meta-analyses. Others rest on a few small studies pointing in opposite directions. Both can appear in the same paragraph, phrased with similar confidence, so you learn to watch the verbs and qualifiers rather than the headline. The useful question is never "what does it recommend?" but "how much is holding this sentence up?"

The claim that holds the most weight: endurance#

Aerobic endurance is the clearest win, and it is worth stating the actual numbers instead of the vibe. Pooled endurance data cited in the stand show mean power output rising by roughly 2.9 percent, with an average effect size near 0.22, and time-trial completion improving by about 2.3 percent, effect size near 0.28. In plain terms these are small-to-moderate effects. A 2 to 4 percent margin can still decide a close race, but nobody should expect a transformation.

Two things make this the trustworthy part of the document. The benefit shows up across different sports (cycling, running, swimming, cross-country skiing), so it is not a quirk of one lab or one protocol. And it survives being pooled across many trials, which is a far tougher standard than any single upbeat study clears on its own.

Dose and timing: a narrow, well-marked window#

The dosing guidance is among the better-supported claims. Benefits appear reliably in the 3 to 6 mg/kg range, with some signal as low as 2 mg/kg, and the most studied timing is about 60 minutes before effort, tracking the rise of caffeine in the blood. It helps to read these as descriptions of what the trials measured across groups, not as a personal prescription. The window is real, but it was drawn by averaging many people.

Where the signal fades#

Strength and power tell a messier story. The stand reports small effect sizes, roughly 0.16 to 0.38, across muscular endurance, movement velocity, and maximal strength, with the clearest gains among strength-sport athletes. Repeated-sprint performance is flagged in the document itself as inconsistent and needing more study.

Team sports scatter the most. Soccer trials showed more distance covered and better passing accuracy but no faster sprint time; basketball helped in some settings; volleyball improved on some measures and not others. When results fan out like that, the honest conclusion is that any real effect is small and easily buried under everything else that shapes a match.

Notice how the language itself shifts across these sections. Endurance earns "consistent" and "moderate." Sprinting and team play get "mixed," "inconsistent," and "requires further investigation." Those hedges are not filler. They are the document telling you where its own confidence runs out.

The caveat that outranks the average#

The single most important line in the whole stand is that these effect sizes are averages, and the spread around them is wide. A mean gain of 2.9 percent paired with a standard deviation of about 2.2 percent means some people respond strongly, some barely at all, and a few may do worse. An average is a weak predictor of a single result whenever the variance is that large.

Genetics is one named reason for the spread. More than 95 percent of caffeine is broken down by the liver enzyme CYP1A2, and a common variant (the -163A>C polymorphism) sorts people toward faster or slower metabolism. The stand cites work in which the endurance benefit appeared mainly in one genotype (AA), while some individuals with the CC genotype showed performance decrements at higher doses. Evidence for the ADORA2A adenosine-receptor gene is thinner and more preliminary, drawn from small samples, and it bears more on sleep and anxiety than on performance.

Habitual intake is genuinely unsettled. The stand calls the evidence equivocal: some studies suggest tolerance builds in low habitual consumers over weeks, others find the benefit persists regardless of daily intake. When a synthesis labels its own evidence equivocal, that is the honest place to sit with uncertainty rather than choose a side.

This is educational information rather than medical advice, and it describes what research populations showed, not what any one person should do.

More is not better#

The document is careful not to treat a bigger dose as a bigger benefit. It flags intakes around 9 mg/kg as carrying a high rate of side effects with no added performance value, and notes that the urine concentration monitored by anti-doping authorities corresponds to intakes several times higher than the amounts shown to help. The performance window and the tolerability window are two different windows, and the stand says so plainly.

Reading any evidence summary this way#

The practical takeaway is a reading habit, not a caffeine tip. The load-bearing claims here are narrow: aerobic endurance, a defined dose range, a defined timing window, and an average effect that is statistically small even when competitively meaningful. Everything past that carries visible hedges, and the between-person variability means the group number tells you little about the next individual. Grading a claim by how much evidence supports it and how large its caveats are is a skill that will serve you well beyond caffeine, in almost any health or performance claim worth taking seriously.