The argument about creatine and the kidneys is almost always an argument about a blood test. Creatine metabolism produces creatinine, laboratories estimate kidney function from creatinine, so taking creatine makes the estimate look worse without anything having happened to the kidney. That confound is real, it is well documented, and our page on creatine's side effects sets it out.
This page is about what happened next. A smaller group of trials treated the confound as an obstacle to get around rather than a conclusion, and went and measured the thing itself.
Classifying the literature before counting it
A PubMed search for creatine supplementation crossed with kidney or renal, filtered to the randomised-controlled-trial publication type, returns 40 records. That number is not the size of the evidence base.
Reading the titles, the set includes a trial of low-dose ashwagandha on exercise endurance, one of Nigella sativa seed on memory, one of monomeric and oligomeric flavanols on kidney function in runners, a probiotic trial on half-marathon recovery, a sepsis trial of a Chinese herbal injection after the Wenchuan earthquake, and a feeding study in post-weaning piglets given zearalenone-contaminated diets. These match because creatinine is a routine safety chemistry in almost any trial, not because creatine was administered.
Classify by what was actually given and what was actually measured, and roughly a dozen records remain. That is a small literature — but it is a small literature that did the hard version of the experiment.
Two trials that measured filtration instead of estimating it
The reference way to establish glomerular filtration rate is to inject a tracer the kidney clears and measure how quickly it disappears. Two randomised, placebo-controlled trials did exactly that with chromium-51 EDTA.
| Postmenopausal women (2011) | Type 2 diabetes (2011) | |
|---|---|---|
| Duration | 12 weeks | 12 weeks, with exercise training |
| Intake | 20 g/day for 1 week, then 5 g/day | Creatine vs placebo |
| Creatine group, 51Cr-EDTA clearance | 86.16 → 87.25 mL/min/1.73 m2 | 90.4 → 96.1 mL/min/1.73 m2 |
| Placebo group | 85.15 → 87.18 | 97.9 → 96.4 |
| p | 0.81 | 0.58 |
The diabetes trial went further and confirmed that its subjects had actually absorbed what they were given: muscle phosphorylcreatine rose from 44 to 70 mmol/kg in the creatine group against 52 to 46 in placebo. That matters more than it sounds. A null safety result in people who did not take the supplement is worthless; this one was produced in people whose muscle chemistry shows the supplement arrived. Creatinine clearance, serum and urinary urea, electrolytes, proteinuria and albuminuria were all unchanged, and the trial was registered as NCT00992043.
A third trial approached the confound from the other side. It gave healthy males roughly 10 g a day for three months and tracked cystatin C — a protein produced at a steady rate by cells throughout the body, filtered by the kidney, and metabolically unconnected to creatine. Cystatin C fell in both groups across the 12 weeks (creatine 0.82 → 0.71 mg/L, placebo 0.88 → 0.75, p = 0.0001). The authors read the fall as an effect of the aerobic training both groups did, not of the supplement. Either way, it moved in the direction opposite to impairment.
The population everyone says was excluded
The standard caveat on pages like this one — including, until now, our own — is that people with existing kidney disease are systematically excluded from creatine trials, so the evidence does not reach them.
That is only partly true, and the part that is false is checkable.
- A placebo-controlled crossover trial gave 2 g a day to 20 male haemodialysis patients across two four-week periods separated by a four-week washout, measuring plasma guanidino compounds and inflammatory indices. Guanidinoacetate fell 15%, alpha-keto-delta-guanidinovaleric acid tripled and argininic acid doubled.
- A 1-year double-blind trial gave 5 g a day to haemodialysis patients and tracked body composition and the Malnutrition-Inflammation Score. After six months, 60% of the creatine group had increased fat-free mass against 36.8% on placebo, and 65% had increased skeletal muscle mass index against 15.8%. Total body water rose from 35.4 to 36.1 L, driven by intracellular water going from 20.2 to 21.0 L. The Malnutrition-Inflammation Score did not change.
Neither of those is a renal-safety trial, and saying so is the point. People on dialysis have already lost the function a safety study would be tracking, so these trials were designed around muscle mass and uraemic chemistry instead. The accurate statement is narrower than either the reassuring version or the cautious one: the population has been studied, for other questions.
That second trial is also a reminder of what creatine does to a body-composition reading — the fat-free-mass gain travelled with intracellular water, the same mechanism we traced through the compartment measurements on what creatine does to body water.
The one case that is not a blood-test artifact
In 2025, a paediatric nephrology journal published a case report of a 17-year-old male who developed acute kidney injury with cast nephropathy after a six-day high-dose creatine loading regimen. He presented with bilateral flank pain and enlarged kidneys on ultrasound, with adequate hydration and no rhabdomyolysis markers. Renal biopsy confirmed cast nephropathy without evidence of light chain disease.
A single case report cannot establish incidence, and it cannot establish causation. It is included here because of one detail that the usual framing of this subject cannot absorb.
Cystatin C and creatinine rose in parallel.
That is the opposite of the pattern the trials describe. The whole metabolic explanation for creatine's effect on kidney numbers is that it inflates one marker and leaves the other alone. When both move together, that explanation has nothing to say. It does not make the case representative — one adolescent on a six-day high-dose regimen is not the population taking 5 g a day — but it does make it the single data point in this subject that is not an argument about measurement.
Where this leaves the evidence
- Filtration measured directly with a tracer was unchanged in two 12-week randomised trials, one of them in people with type 2 diabetes.
- A marker creatine cannot inflate, cystatin C, moved away from impairment over three months at roughly 10 g a day.
- The dialysis population has been studied twice, but for muscle and uraemic chemistry, not for renal decline — so its safety question remains genuinely open.
- One biopsy-confirmed injury exists in the record, after high-dose loading in an adolescent, with both filtration markers rising together.
The reason this page is not simply the side-effects page again is the second column of that ledger. The reassuring evidence is stronger than the marker argument suggests, because someone went and measured filtration directly. And the cautionary evidence is more specific than the absence of it suggests, because one case exists in which the measurement explanation does not apply.
For the wider safety picture across all fifty-odd reported effects, see creatine's side effects; for what the loading protocols in these trials were actually built to test, see what the loading studies measured; and for how the compound sits against the research compounds this site covers, creatine versus peptides.
This is research journalism. Amounts are reported with their sources and dates. Nothing here is a recommendation, and nothing here is medical advice.
Sources
- Neves M Jr, Gualano B, Roschel H, et al. Effect of creatine supplementation on measured glomerular filtration rate in postmenopausal women. Appl Physiol Nutr Metab 2011;36(3):419–22 — PMID 21574777
- Gualano B, de Salles Painelli V, Roschel H, et al. Creatine supplementation does not impair kidney function in type 2 diabetic patients. Eur J Appl Physiol 2011;111(5):749–56 — PMID 20976468 · NCT00992043
- Gualano B, Ugrinowitsch C, Novaes RB, et al. Effects of creatine supplementation on renal function: a randomized, double-blind, placebo-controlled clinical trial. Eur J Appl Physiol 2008;103(1):33–40 — PMID 18188581
- Taes YE, Marescau B, De Vriese A, et al. Guanidino compounds after creatine supplementation in renal failure patients. Nephrol Dial Transplant 2008;23(4):1330–5 — PMID 18048424
- Marini ACB, Schincaglia RM, Candow DG, Pimentel GD. Effect of creatine supplementation on body composition and Malnutrition-Inflammation Score in hemodialysis patients. Nutrients 2024;16(5):615 — PMID 38474743 · PMC10934827
- Filler G, Maung E, Díaz González de Ferris ME, Chan NG, Sharma AP. Acute kidney injury with cast nephropathy following creatine loading in a 17-year-old. Pediatr Nephrol 2025;40(10):3089–92 — PMID 40304760