Caffeine: what the evidence actually shows

settled caffeine — well-studied; evidence and use are aligned

Reviewer and governance status are shown on each atomic claim below.

What it is

Caffeine is tracked as a health intervention in this evidence graph. No additional aliases are recorded.

Studied forms and doses

No form-specific catalog record is available. This does not mean all forms are equivalent.

cognition: 40–200 mg; form not reported.
energy / fatigue: 40–200 mg; form not reported.
strength: 3–6 mg/kg pre-exercise; form not reported.
cognition: not standardized; form not reported.

Mechanisms

Mechanistic plausibility is context, not proof that a clinical outcome improves.

adenosine-receptor antagonism — caffeine blocks adenosine receptors, reducing perceived fatigue

⚠ Safety in context

minorwith L-Theanine: Commonly paired; L-theanine is said to smooth caffeine’s jitter. Low-stakes.

Products and verification

No product label snapshot is mapped in this released catalog. Ingredient evidence does not establish any bottle’s identity, potency, purity, availability, or value.

Evidence by outcome

cognitiongrade Ahigh certainty— reliably improves alertness, reaction time, and vigilance

energy / fatiguegrade Bmoderate certainty— reduces perceived fatigue; tolerance develops with regular use

strengthgrade Bhigh certainty— reliable ergogenic effect on power and endurance performance

cognitiongrade Dvery-low certainty— Cochrane review, 13 RCTs, shift-work context: memory/attention improved, no verbal-skill benefit; high risk of bias flagged; mostly young people in simulated conditions, not real shift workers.

What it probably doesn't do

Caffeine for cognition — Grade D, Cochrane review, 13 RCTs, shift-work context: memory/attention improved, no verbal-skill benefit; high risk of bias flagged; mostly young people in simulated conditions, not real shift workers..

Grade history

2026-07-16B → BCorrection by Claude (gold-standard provenance sweep): live Postgres source drift replaced with the verified seed source set [pmid-29876876] details →
2026-07-16B → BCorrection by Claude (gold-standard provenance sweep): live Postgres source drift replaced with the verified seed source set [pmid-12204388] details →
2026-07-16A → ACorrection by Claude (gold-standard provenance sweep): live Postgres source drift replaced with the verified seed source set [pmid-27612937] details →

Re-review cadence: every 12 months (Grade A/B) · next scheduled by 2027-07-16 · methodology v1.

By outcome — the money pages

Is Caffeine hype?

sources: PMID:27612937 · PMID:12204388 · PMID:29876876 · PMID:20464765

Cite this page

Citation is encouraged. The proposed reuse license is CC-BY-4.0; status: pending-governance-ratification.

Plain-text citation

Substrate. Caffeine: what the evidence actually shows. https://evidencebased.info/interventions/caffeine. Updated 2026-07-16.

BibTeX

@misc{substrate_nterventionscaffeine,
  author = {Substrate editorial},
  title  = {Caffeine: what the evidence actually shows},
  year   = {2026},
  url    = {https://evidencebased.info/interventions/caffeine},
  note   = {evidencebased.info}
}