Three health questions dominate the search traffic around creatine that is not about lifting: diabetes, blood pressure and cancer. Each has a confident answer circulating, and in each case the confident answer is larger than the evidence under it. This page counts what the trial record actually contains, names what is missing, and takes no position on what anyone should do — those decisions belong with a clinician, as this site's editorial standards commit it to saying.
Type 2 diabetes: one trial, 25 people
The clinical case rests on a single randomised, double-blind, placebo-controlled trial: Gualano and colleagues, University of São Paulo, published in Medicine & Science in Sports & Exercise in 2011, registered as NCT00992043.
Twelve weeks. 5 g a day of creatine or placebo, both arms enrolled in an exercise training programme. Twenty-five participants analysed, thirteen on creatine and twelve on placebo. The primary endpoint was HbA1c, and it moved:
| Creatine (n = 13) | Placebo (n = 12) | |
|---|---|---|
| HbA1c before | 7.4 ± 0.7% | 7.5 ± 0.6% |
| HbA1c after | 6.4 ± 0.4% | 7.6 ± 0.7% |
The between-group difference was −1.1 percentage points (95% CI −1.9 to −0.4, p = 0.004), with lower glucose at 0, 30 and 60 minutes of a meal tolerance test and increased GLUT-4 translocation to the sarcolemma offered as the mechanism.
That is a real result and it is one trial of twenty-five people with a surrogate endpoint. A 2026 commentary in Nutrition and Health, reviewing this exact field, calls it a translational gap — mechanistic and preclinical plausibility, clinical evidence that remains "limited and inconsistent", and a need for work that moves "beyond short-term surrogate endpoints" to longer follow-up, real clinical outcomes, and integration with the drugs people are actually taking.
One detail this site can add, because it already published the other half of it. The same São Paulo group's kidney-safety report on this trial is the source of the tracer-clearance data on creatine and kidneys — chromium-51 EDTA clearance unchanged, muscle phosphorylcreatine confirming the supplement had actually arrived. It is the same 12-week study, in the same patients, reported for a different endpoint. Anyone counting "trials of creatine in type 2 diabetes" by counting papers will count it twice.
Blood pressure: the population that has not been studied
Here the interesting fact is an absence, and it is stated by the people best placed to know it.
The International Society of Sports Nutrition's 2025 expert review, an eighteen-author narrative review addressing sixteen recurring claims about creatine, works through the blood-pressure question and reaches this: relatively high doses (10–20 g/day) over 5 to 31 days in healthy young men and women did not affect blood pressure; a systematic review in heart-failure, ischaemic-heart-disease and myocardial-infarction patients found no impact at up to 20 g/day for six weeks; and a two-year trial in postmenopausal women, run for bone outcomes, measured pressure at one site and found no group difference.
| Two-year trial, completers | Creatine (n = 60) | Placebo (n = 52) | p |
|---|---|---|---|
| Systolic, mmHg | 121 ± 15 → 124 ± 11 | 116 ± 13 → 121 ± 15 | 0.34 |
| Diastolic, mmHg | 77 ± 8 → 76 ± 7 | 75 ± 8 → 76 ± 9 | 0.65 |
And then the sentence that is the actual state of the evidence: "Studies of CrM in people who were hypertensive at baseline are missing from the literature." The nearest substitute the panel could name is a study in spontaneously hypertensive rats.
So the honest statement is narrow. Nothing in the measured record shows creatine raising blood pressure in people whose blood pressure was normal, over periods from five days to two years. The group whose question this actually is has not been enrolled. An untested population is an untested population, and this page will not convert that into reassurance.
Cancer: two papers, neither about supplement users
The claim comes from two identifiable places, published decades apart.
The heterocyclic-amine review. Creatine and creatinine in meat and fish react during prolonged high-temperature cooking to form amino-imidazo-azaarenes, some of which are mutagenic in bacterial assays; creatinine can also be nitrosylated in the stomach to compounds including N-nitrososarcosine. The review that laid this out concluded those compounds "may impose only a minor health risk" — and the practical conclusion that follows from it is about blackened, barbecued and heavily processed meat, not about a scoop of powder. The distinction survives in the ISSN panel's own summary, which pairs "no support for increased cancer risk at 3–5 g/day" with advice to limit overcooked meat.
The metastasis study. A murine model reported that cancer cells upregulate glycine amidinotransferase to make their own creatine, and that dietary creatine at 5% of feed by wet weight increased liver metastases while leaving the primary colonic tumour unchanged, or mildly suppressed. The caveats belong with the headline and they are large: that dose scales to roughly 28 g a day in a human, five to nine times ordinary use; the metastases were hepatic, and liver inflammation on creatine appears to be a mouse-specific effect not seen in rats or humans; and tumour cells are energetically promiscuous, expressing high levels of creatine kinase isoforms whether or not creatine is supplied.
Meanwhile the randomised trials that do exist in cancer are asking the opposite question. Six are indexed, and they test whether creatine protects muscle through treatment. The most recent is unambiguous and null: a 2025 double-blind randomised trial in 30 men with prostate cancer on androgen deprivation therapy found that creatine added to resistance exercise did not improve lean mass, strength or physical function beyond resistance exercise alone. Another examined short-term muscular performance in breast cancer survivors; a third, from 2026, was a feasibility study.
That pattern — a scare built on a rodent dose, and a human literature quietly testing something else — is the same one this site traced through the DHT study behind creatine and hair loss and the four claims on creatine and the brain. The mechanism is always real; the leap to the person holding the tub is where the evidence stops.
What the three questions have in common
Each is a claim about a clinical population resting on evidence from somebody else. Diabetes has one small trial of people who have the condition. Blood pressure has good data from people who do not have the condition. Cancer has a mouse study at a dose nobody takes, plus human trials asking a different question.
That is not a reason to dismiss the questions — they are searched precisely because they matter to the person asking. It is a reason to state the boundary each time rather than let a mechanism or a rodent stand in for a result.
Limits of this page
The counts are title-and-abstract classifications of PubMed on 2026-09-20 and a different phrasing returns a slightly different set. Two of the three sections rest partly on a narrative expert review rather than on a systematic one, and that review's author list includes researchers who have published much of the primary literature it summarises. Nothing here is medical advice, a recommendation, or a reason to start, stop or change any treatment or supplement; anyone with diabetes, raised blood pressure, a cancer diagnosis, or any other condition should put the question to their own clinician, who can see the rest of the picture.
Sources and dates
- Gualano B, de Salles Painelli V, Roschel H, et al. Creatine in type 2 diabetes: a randomized, double-blind, placebo-controlled trial. Med Sci Sports Exerc 2011;43(5):770–8 — PMID 20881878 · NCT00992043
- Machado M. Creatine supplementation in type 2 diabetes: a critical appraisal of the evidence gap. Nutr Health 2026;32(4):1135–7 — PMID 41810769
- Antonio J, Brown AF, Candow DG, et al. Part II. Common questions and misconceptions about creatine supplementation: what does the scientific evidence really show? J Int Soc Sports Nutr 2025;22(1):2441760 — PMID 39720835 (sections 8, 10)
- Fairman CM, Kendall KL, Newton RU, et al. Creatine supplementation does not add to resistance training effects in prostate cancer patients under androgen deprivation therapy: a double-blind randomized trial. J Sci Med Sport 2025;28(2):118–24 — PMID 39366880
- PubMed, searched 2026-09-20:
"creatine supplementation"[tiab] AND cancer[tiab] AND "randomized controlled trial"[pt]— 6 records.