REF-249 · Guest et al., 2021 · Journal of the International Society of Sports Nutrition, Vol. · Paper · PubMed

International Society of Sports Nutrition Position Stand: Caffeine and Exercise Performance (Updated)

Astrong evidenceMeasured on a mixed groupWritten before October 2026, being re-checked

Key Findings — Key Positions

Dosing: 3-6 mg/kg body mass consistently improves exercise performance. Minimal effective dose remains unclear but may be as low as 2 mg/kg. Very high doses (e.g. 9 mg/kg) carry a high incidence of side effects and do not seem needed for an ergogenic effect.

Timing: the most commonly used timing is 60 min pre-exercise; optimal timing likely depends on the source — caffeine chewing gum may need a shorter wait than capsules.

Forms: caffeinated chewing gum, mouth rinses, energy gels and chews improve performance, primarily in aerobic exercise. Energy drinks and caffeinated pre-workout supplements enhance both anaerobic and aerobic performance.

Performance effects:

DomainEffect
Aerobic enduranceMost consistent moderate-to-large benefit; magnitude differs between individuals
Muscular enduranceSmall-to-moderate benefit
Movement velocity, muscular strengthSmall-to-moderate benefit
Sprinting, jumping, throwingSmall-to-moderate benefit
Sport-specific aerobic and anaerobic actionsBenefit in many but not all studies
Cognition (attention, vigilance)Improved in most individuals
Under sleep deprivationCognitive and physical performance improved in some individuals

Caffeine improves physical performance in both trained and untrained individuals.

Special conditions: endurance exercise in the heat — 3-6 mg/kg well supported; at altitude — 4-6 mg/kg well supported.

Individual variation: differences in performance response and in adverse effects on sleep or anxiety may be attributed to genetic variation in caffeine metabolism and in physical/psychological response; habitual caffeine intake may also play a role. The full text names the CYP1A2 and ADORA2A variants (full text, checked 2026-09-27); the abstract itself does not grade genotypes as "better" or "worse" responders.

What's New vs Goldstein 2010 [REF-115]

Genetic variation as a source of individual response — not covered in 2010
Alternative delivery forms (gum, rinse, gel, chews) — now listed as effective
Heat and altitude dose ranges — new
Cognitive performance and sleep-deprivation evidence — expanded
Low-dose efficacy (possibly 2 mg/kg) — stated

Practical Application

ISSN 2021 position: caffeine at 3-6 mg/kg body mass, most commonly taken 60 min before exercise, consistently improves performance — most reliably aerobic endurance (moderate-to-large), plus small-to-moderate gains in strength, muscular endurance, sprint/jump/throw and cognition; gum acts faster than capsules; 2 mg/kg may already work and ~9 mg/kg brings side effects without extra benefit.
Not from this paper; the base's own practice: for a long endurance event an 80 kg athlete lands on 240-320 mg (3-4 mg/kg) 60 min pre-start, with a top-up later in the event — the base's practice dose and timing are in SYNTH-67 (2-3 mg/kg at 90-120 min for races over ~2 h), not a position-stand recommendation; the position stand has no dose-response for ultra-endurance.
Heat and altitude: the position stand supports 3-6 mg/kg in the heat and 4-6 mg/kg at altitude — no special reduction is required for hot conditions.
Test in training: response differs between individuals (genetics, habitual intake, sleep/anxiety side effects), so dress-rehearse the planned race dose in a B-race or key session before the A-race.
Not from this paper: the base advises against L-theanine co-supplementation for physical performance (Tuncer et al. 2026 — no synergy with caffeine; isolated theanine reduced isometric strength; not a base entry, not verified here).

Limitations

Position stand = synthesis, not new data. Genetic recommendations preliminary (few RCTs stratifying by genotype). Individual-variation advice based largely on observational data. No specific dose-response curve for ultra-endurance (>4 h). The abstract gives no figure for the share of non-responders and no withdrawal protocol.

How this entry was checked and changed
Why this level

(ISSN official, replaces Goldstein 2010 REF-115)

Content correction 2026-09-27
Removed numbers not in the abstract: "~10-15% non-responders", "withdraw caffeine 2-7 days before competition", "stop daily coffee 3-5 days before race" (the abstract says only that habitual intake may play a role in response variation).
Genetics paragraph claimed AA genotype = better response, AC/CC = reduced benefit -> not in the abstract; now stated as genetic variation in general, with CYP1A2/ADORA2A labelled full text.
Heat "(no additional risk)" and altitude "(due to altitude-specific fatigue)" rationales removed -> abstract gives the dose ranges only. Sleep deprivation "improved" -> "in some individuals".
Lead Practical bullet was the base's race-day protocol -> now the position stand's own dosing/timing/effects; race-day worked example and L-theanine note labelled "Not from this paper".
Lines labelled "full text; not checked" were read against the open-access full text (PMC7777221): 1 confirmed and relabelled.
2026-10-01: the entry carried its own house top-up (100-150 mg at 2-2.5 h) that disagreed with the one in SYNTH-67 and SYNTH-07 (2-3 mg/kg at 90-120 min); replaced with a pointer to SYNTH-67. Both were practice; the position stand gives neither.