hraness

saved

Does creatine make you smarter?

by dynomightDynomightpublished

gist

Dynomight separates creatine’s established role in rapid cellular energy buffering from the much weaker case for cognitive enhancement. Supplementation reliably raises muscle creatine and improves short-burst strength, while brain levels appear to rise only modestly. A striking 2003 cognition result did not replicate; later trials and reviews point, at most, to a small uncertain benefit. The strongest conclusion is that large effects in healthy adults are unlikely, with stressed or deficient groups remaining plausible but unproven.

ideas

  • Creatine buffers and transports cellular energy. Phosphocreatine rapidly restores ATP during sudden demand and may shuttle energy from mitochondria to locally active structures.
  • The brain’s use differs from muscular sprinting. Whole-brain energy use stays relatively steady, so the proposed benefit is avoiding local neuronal energy deficits rather than supplying an organ-wide burst.
  • Replication changes the headline. The large cognitive effects reported in a 2003 vegetarian cohort shrank to small, statistically inconclusive estimates in a larger 2023 reproduction.
  • Meta-analysis cannot repair weak inputs automatically. Missing raw statistics, double-counted samples, and publication bias make pooled estimates look more settled than the underlying trials justify.
  • Separate an unproven benefit from a ruled-out large effect. The trials leave room for a small benefit while providing stronger evidence that healthy adults do not receive a large cognitive boost.

quotes

Overall, I think the RCTs provide very weak evidence in favor of a small benefit for healthy adults.

dynomight, weighing the randomized-trial evidence for a positive effect.

I also think they provide moderate evidence against a larger effect for healthy adults

dynomight, distinguishing uncertainty about small effects from evidence against large ones.

The RCTs suggest a benefit that is quite small, on the order of 1 to 3 IQ points.

dynomight, summarizing the scale implied by the inconclusive trials.

I don’t know. Maybe a little.

dynomight, giving the article’s final answer.