creatine monohydrate · graded evidence · pulled 2026-08-31

Creatine (monohydrate)

The one item in this entire catalog where the evidence base is genuinely enormous and genuinely boring. Saturation kinetics measured in the 1990s, hundreds of randomized trials, and 2026's mega-review closing the safety question — plus a surprising number of nulls where people expect benefits.

The short version
  • Loading 20 g/day for 6 days raises muscle creatine ~20%; 3 g/day reaches the same place in 28 days; 2 g/day then holds it. Loading is a speed choice, not a requirement. DATA
  • The 2026 umbrella analysis of 684 RCTs, >12,800 participants: no increase in renal, liver, GI or musculoskeletal adverse events — even at high doses. DATA
  • Adolescent data shows no consistent short-term safety signals in systematic review. DATA
  • Bone: three independent meta-analyses, all null. Creatine does not build bone. DATA

Saturation — the study every dosing scheme comes from

ProtocolDurationMuscle creatineReading
20 g/day (4 × 5 g)6 days≈ +20%the classic load — fast saturation
3 g/day28 dayssame level, graduallyloading is optional
2 g/day after loadingmaintenancelevel held2 g replaces what is lost
stop entirely~30 daysback to baselinefull washout in a month
Hultman 1996 — the original quantification of human muscle-creatine loading. Individual response varies (some subjects saturate poorly), but the protocol arithmetic is settled.

Effect size — what saturation buys

The ISSN position-statement arithmetic: ~1–1.5 kg mass, ~8% strength DATA
The International Society of Sports Nutrition's 2023 position stand, synthesizing the RCT base: typical gains of about 1–1.5 kg lean mass and ~8% on strength over 8–12 weeks of training — consistently the largest effect of any legal supplement, and modest in absolute terms. This is review-level consensus, not a single trial number.

Safety — the question that took 30 years to close

684 RCTs, >12,800 participants: no organ signal DATA
The 2026 umbrella/dose-response review of the entire randomized base (Gonzalez et al.): no increased risk of renal, hepatic, gastrointestinal or musculoskeletal adverse events versus placebo — including at high-dose protocols. Healthy kidneys handle the load; the decades of "creatine ruins your kidneys" had no trial behind them.
Adolescents: no consistent short-term signals DATA
2026 systematic review of creatine in adolescent athletes: no consistent short-term safety concerns identified across the studied protocols. Long-term data in this age group remains thinner than the adult base — that is the honest boundary, not a signal.

The nulls people don't expect

Bone: three independent meta-analyses, three nulls DATA
Bone mineral density under creatine supplementation, meta-analyzed independently three times: n=763, p=0.430; n=1,093, p=0.557; n=608 — null. Whatever creatine does, skeletal mineral accrual is not on the list. The hypothesis that it helps growing bone has now been tested and discarded.
Cognition claims outperform the evidence DATA
Where sleep-deprivation and cognitive benefits were tested systematically, effects are inconsistent and small — unlike the muscle endpoint, the brain claims do not survive aggregation. The strongest supplement in the stack is strong exactly where it was first measured, and nowhere else.

What does not exist

No long-term (>5 y) adolescent cohort · no interaction studies with AAS DATA
Verified by search (2026-08-31): no multi-year randomized follow-up in adolescents (the safety case there is short-term and review-level), and — as with everything else in this catalog — no study of creatine in combination with anabolic steroids. Its clean record stands on its own; the stacks people actually run have never been studied as stacks.