Chronic low-dose caffeine induces tolerance to its ergogenic effect
Protocol
n=18 low-habitual caffeine consumers (<75 mg/day) randomized to caffeine (titrated 1.5→3.0 mg/kg/day) or placebo for 28 days
Before supplementation: acute 3 mg/kg caffeine trial (precaf) vs placebo (testpla)
After 28 days: repeat acute 3 mg/kg trial (postcaf)
Task: 60 min cycling at 60% VO2peak → 30 min performance task
Key Results
Before chronic use: acute caffeine improved external work in both groups (caffeine group +38.4 kJ, ES=0.49, p=0.001; placebo group +21.4 kJ, ES=0.38, p=0.004)
After 28 days chronic caffeine: ergogenic benefit no longer apparent in caffeine group (precaf 383.3 → postcaf 358.0 kJ, ES=0.31, p=0.025 = decline); retained in placebo group
Circulating caffeine, hormones and substrate oxidation did NOT differ between groups
Conclusion (from source): chronic low-dose caffeine develops tolerance in low-habitual consumers → they should refrain from chronic supplementation to maximise acute benefits
Practical Application
Avoid chronic daily low-dose caffeine if the goal is to preserve acute race-day ergogenic response (relevant for athletes who don't habitually consume caffeine)
Race Day: 3-5 mg/kg 60 min before start = ergogenic effect (dose from REF-350/REF-249; NOT established by this study)
[not confirmed by the source] The recommendation "full caffeine withdrawal 7–10 days before a race restores sensitivity" — this work tests only the DEVELOPMENT of tolerance over 4 weeks in low-habitual consumers, NOT a recovery-via-withdrawal protocol in habitual consumers. Restoring sensitivity through withdrawal requires a separate source (evidence is mixed — see REF-350 "habituation evidence mixed").
Limitations
Small sample (n=18), low-habitual consumers only → NOT generalizable to habitual caffeine users
Low dose (1.5-3.0 mg/kg/day) — does not address tolerance from higher habitual intakes
Does not test any withdrawal/washout recovery protocol
Audit note (2026-07-06)
JOURNAL CORRECTED: was "PLOS One" — actually Journal of Sports Sciences (DOI 10.1080/02640414.2016.1241421, PMID 27762662).
CLAIM FLAGGED: the claim about "7-10 days of withdrawal to restore the effect" is marked [not confirmed by the source].
Level A → B: small sample, narrow population (low-habitual), the key conclusion is not confirmed by the source.
---