Whether creatine causes hair loss and what the DHT study found
No controlled human experiment has established creatine as a cause of hair loss (Antonio et al., 2021; Lak et al., 2025). The newest trial followed scalp measurements for 12 weeks and did not find a creatine-specific decline, while the older paper that prompted the concern examined blood DHT rather than hair itself (Lak et al., 2025; van der Merwe et al., 2009).
The evidence therefore turns on a simple distinction: a hormone signal is not the same endpoint as a change in the scalp. The sections below compare the two key trials with broader reviews of creatine and pattern hair loss to show where the evidence is solid and where uncertainty remains.
Why creatine became linked with hair loss
Creatine is part of a rapid energy reserve used when muscle needs ATP faster than ordinary metabolism can supply it. For that reason, the research record is dominated by questions about exercise performance, lean tissue and safety rather than dermatology (Kreider et al., 2017). A claim about hair therefore sits outside the main set of outcomes that most creatine studies were designed to test. Cabdirect's sport nutrition and supplements overview gives the broader exercise-nutrition context in which supplement evidence is usually assessed.
The route from creatine to a hair-loss concern began with androgen biology. In men with androgenetic pattern loss, some scalp follicles inherit greater sensitivity to androgen signalling, and DHT is one of the hormones involved in that local process (Lolli et al., 2017; Ntshingila et al., 2023). Testosterone can be converted into DHT by the enzyme 5-alpha-reductase, and DHT can then interact with androgen receptors within susceptible follicular tissue.
That explanation is much less complete for female pattern loss, whose molecular basis remains less clearly resolved (Liu et al., 2025). If creatine caused a durable rise in DHT, the observation would justify asking whether susceptible follicles respond to it. It would not, by itself, show that hair becomes thinner or that shedding increases.
This difference in endpoints is the central issue. Blood hormones can shift without producing a visible scalp change because follicles vary in local enzyme activity, receptor behaviour and biological sensitivity. Evidence about causation therefore has to move beyond the bloodstream and ask whether the hair itself changes under controlled comparison.
What the 2009 DHT study actually found
The 2009 study used a double-blind crossover arrangement in college-aged male rugby players, with the two treatment periods separated by six weeks. Twenty players entered the experiment. Its Methods section says that four participants withdrew and their data were excluded, which implies 16 analysed participants, although the footnotes under Tables 1 and 2 inconsistently state that 20 completed each arm (van der Merwe et al., 2009).
The investigators followed testosterone and DHT in blood across a three-week supplementation period. Relative to the starting value for the creatine period, DHT was reported 56% higher after the first week and 40% higher at the end of week 3; the DHT-to-testosterone ratio was 36% higher at week 1 and 22% higher by the end (van der Merwe et al., 2009). Those figures are real study results, but they describe changes from that condition's own baseline rather than proof of a hair effect.
A later review added an important numerical detail. Before supplementation, DHT was 0.98 nmol/L in the creatine period and 1.26 nmol/L in the placebo period, so the creatine-period baseline was about 23% lower; Antonio et al. (2021) also noted that the reported DHT values and the ratio stayed within normal clinical limits. The baseline imbalance does not erase the observed rise, but it changes how the percentage increase should be read.
Nothing in that experiment measured baldness directly. The protocol included no hair counts, density measurements, shaft-width measurements, serial scalp photographs or clinical assessment of androgenetic alopecia. Its conclusion concerned androgen metabolism, and the authors called for further investigation rather than claiming that creatine had produced hair loss.
The finding also came from a short loading-and-maintenance schedule, so persistence cannot be inferred from it. A three-week hormone result does not show what happens after months or years, and it does not establish that a different supplementation pattern would produce the same response. These limits matter because the later public claim became much broader than the question the trial actually answered.
Replication provides another check. Antonio et al. (2021) reviewed creatine studies that measured androgens and did not find a consistent increase in total or free testosterone, while noting that human research available then had not demonstrated creatine-induced baldness. The 2009 DHT result therefore remained a single mechanistic observation rather than a repeatedly confirmed effect.
What the 2025 hair-follicle trial found
Direct scalp testing came much later. Lak and colleagues enrolled 45 men aged 18 to 40 who regularly performed resistance training; six left before the intervention, leaving 39 to be randomised, and 38 finished the protocol. One finisher did not complete the final hair assessment, so the hair analyses used data from 37 participants (Lak et al., 2025).
The design also avoided the loading phase used in 2009. Blood testing covered total testosterone, free testosterone and DHT, while the scalp assessment combined trichogram methods with digital imaging. The recorded endpoints included counts and density, follicular-unit measures, growth-phase distribution, terminal-versus-vellus proportions and cumulative hair thickness (Lak et al., 2025).
Across those scalp endpoints, the pattern of change in the creatine arm was not significantly different from the placebo arm (Lak et al., 2025). DHT and the DHT-to-testosterone ratio likewise showed no significant creatine-specific interaction over time. Total testosterone rose and free testosterone fell across the sample as time effects, but those shifts were not attributable to group assignment.
The authors also tested whether extreme statistical observations were driving the result. Removing those outliers did not materially change the interpretation. That makes the study more directly relevant to the hair question than the 2009 experiment, although its conclusions still apply only to the studied population and its 12-week follow-up.
How the two trials answer different questions
The two studies are easier to understand if they are treated as answering different layers of the question. The 2009 crossover work asked whether circulating androgens changed under its short supplementation protocol, and its own report contains conflicting participant counts between Methods and table footnotes (van der Merwe et al., 2009). The 2025 parallel trial asked a different question by adding objective scalp endpoints to hormone measurements (Lak et al., 2025).
One result is therefore mechanistic and the other is closer to the outcome people actually worry about. A rise in a blood marker can motivate a hypothesis, but count, density, thickness and follicular measurements test whether that hypothesis appears in the scalp. For the question of hair loss, the latter type of endpoint carries more direct relevance.
The protocols also differed in duration, dosing pattern and participant context. Treating them as one positive study and one negative study hides those design differences. A more accurate synthesis is that the older paper raised a hormonal question, while the later paper looked for the predicted downstream effect and did not detect it over 12 weeks.
Protocol, duration, population and endpoint all differ, so reducing the evidence to one positive and one negative study is misleading. A serum marker is upstream from the outcome people are worried about, whereas density, thickness and follicular measurements sit nearer to that outcome. A balanced reading keeps the 2009 signal in view but does not treat it as equivalent to observed hair loss.
Study | Design and participants | Measured | Hair-relevant result |
Op ’t Eijnde & Hespel, 2001 | Crossover study; 11 healthy young men; five days of creatine exposure | Testosterone response around a heavy resistance exercise session | No significant creatine-related change in the testosterone response (Op ’t Eijnde and Hespel, 2001). |
van der Merwe et al., 2009 | Crossover study; 20 entered; Methods imply 16 analysed, while table notes report n = 20; three weeks | Blood testosterone, DHT and the DHT-to-testosterone ratio | DHT increased from the creatine-period baseline; the protocol did not include a hair endpoint (van der Merwe et al., 2009). |
Lak et al., 2025 | Parallel placebo-controlled study; 45 enrolled, 39 randomised, 38 completed; 12 weeks | Androgen measures plus objective scalp counts, density, follicular units and thickness | No significant creatine-specific difference emerged for DHT or the recorded scalp outcomes (Lak et al., 2025). |
Blood DHT is an intermediate biological signal, not a direct measurement of hair loss. In the only controlled trial that tracked DHT alongside objective scalp outcomes, the creatine group did not show a worse 12-week pattern than placebo (Lak et al., 2025).
Why a DHT change is not the same as hair loss
Male-pattern androgenetic hair loss develops because susceptible follicles progressively miniaturise, not because one blood value crosses a simple threshold. Reviews describe inherited susceptibility together with local androgen activity, including DHT acting through androgen receptors within the follicle environment (Lolli et al., 2017; Ntshingila et al., 2023). Liu et al. (2025) present increased androgen signalling in susceptible follicles as a hypothesis for male disease and emphasise wide variation in onset, severity, distribution and genetic architecture.
That biology prevents a temporary serum DHT change from being translated into a predicted number of lost hairs. Different follicles and scalp regions do not share identical androgen sensitivity, and local enzyme and receptor activity add another layer beyond hormone concentration in blood. Visible pattern loss also emerges across repeated hair cycles, which operate on a longer timescale than a short biochemical measurement.
Where a measurement is taken also matters. A blood sample reflects circulating hormone at a particular moment, whereas androgenetic hair change develops within individual follicles and across repeated growth cycles. That is why a serum result can be biologically interesting without supplying a direct estimate of future density loss. The distinction is especially important when the original observation lasted only weeks but the clinical outcome being discussed develops gradually (Lolli et al., 2017; Ntshingila et al., 2023).
Shedding has many possible explanations as well. Pattern loss, telogen shedding, inflammatory scalp disease, medication effects, endocrine conditions and nutrient deficiencies can all alter hair through different routes, so temporal coincidence is weak evidence of cause. Cabdirect's article on iron deficiency and its health effects illustrates a separate nutrition-related mechanism that can include hair changes.
A related study digest on dietary iron and hair growth examines a different nutrition-related cause of hair changes in an animal model. Its findings cannot be applied directly to human hair loss, illustrating why the population and outcomes measured in each study must be considered separately.
What the wider creatine hormone evidence shows
Hormone studies outside the hair literature provide additional context. In a double-blind crossover experiment involving 11 healthy young men, short-term creatine did not significantly change the testosterone response to heavy resistance exercise (Op ’t Eijnde and Hespel, 2001). Antonio et al. (2021) identified 12 other creatine studies that measured testosterone apart from the 2009 DHT paper: two reported small increases in total testosterone, ten reported no change, and studies measuring free testosterone did not show increases.
That broader record does not prove the 2009 DHT observation was spurious. It shows that the observation has not become a stable pattern reproduced across creatine research. Confidence in a causal explanation would be stronger if separate studies repeatedly found the same androgen response and then linked it to objective deterioration in hair.
The same evidence rule applies elsewhere in the nutrition section: a plausible pathway and a measured health outcome are not interchangeable. Kreider et al. (2017) synthesised creatine evidence mainly around performance, metabolism and safety.
Antonio et al. (2021) addressed the hair-loss concern more directly, and Lak et al. (2025) then added controlled scalp measurements. That sequence moved the discussion from an intermediate hormonal possibility toward observation of the outcome people were actually concerned about.
The sequence of evidence therefore matters. First came a possible hormonal signal, then a broader literature that failed to show a consistent androgen pattern, and finally a controlled study that looked at hair itself. That progression is more informative than treating every paper as though it measured the same endpoint.
What is proven and what remains uncertain
The most defensible conclusion is deliberately narrow. Controlled human evidence has not shown that creatine causes hair loss, and the first trial built around objective scalp endpoints found no adverse difference from placebo across 12 weeks (Lak et al., 2025). The 2009 experiment did report a short-term DHT increase during its creatine period, but it did not measure hair and that hormone pattern has not been established as a consistent response in the wider literature (van der Merwe et al., 2009; Antonio et al., 2021).
Two opposite claims therefore go beyond the data. It is not justified to say that creatine is proven to cause baldness, but it is equally unjustified to say that every possible long-term effect has been ruled out. The direct trial was small, involved men aged 18 to 40 who performed resistance training, lasted 12 weeks and did not determine family history of hair loss (Lak et al., 2025).
The main remaining gaps concern duration and population overall. Women, adolescents, people followed for years, and participants selected for existing or strongly familial androgenetic alopecia have not been adequately tested in direct controlled hair research. Larger and longer studies could narrow those uncertainties, but the evidence currently available weighs against a demonstrated causal link under the conditions that have actually been studied.
Frequently asked questions
Does creatine cause hair loss?
Controlled human studies have not established creatine as a cause of hair loss. In the 2025 placebo-controlled trial, the recorded scalp measures changed similarly in the creatine and placebo groups over 12 weeks (Lak et al., 2025). Evidence remains limited for longer exposure, women and people selected specifically for high susceptibility to androgenetic pattern loss.
Why is creatine linked to DHT?
The concern largely comes from a 2009 crossover experiment in male rugby players. Blood DHT rose relative to the starting value for the creatine period, while testosterone did not change significantly (van der Merwe et al., 2009). The study did not record shedding, density, shaft thickness or other direct hair outcomes.
Did the 2025 creatine study measure actual hair changes?
Yes. Lak et al. (2025) combined trichogram assessment with digital imaging and recorded count, density, growth-phase distribution, follicular units, terminal-versus-vellus proportions and cumulative thickness. After 12 weeks, none of the reported scalp outcomes showed a significant creatine-specific difference compared with placebo.
Can a rise in DHT automatically cause baldness?
No. DHT is part of the biology of male androgenetic pattern loss, but follicular genetics and local androgen signalling also influence whether miniaturisation occurs (Ntshingila et al., 2023; Liu et al., 2025). A change in blood DHT therefore represents a mechanistic clue, not a direct measurement of visible hair loss.
What is still unknown about creatine and hair?
The largest gaps concern time and population. The direct trial followed resistance-trained men aged 18 to 40 for 12 weeks (Lak et al., 2025). Longer follow-up, larger samples, women and participants characterised for family history or established androgenetic alopecia would answer questions that the current evidence cannot resolve.
Sources
- Lak M, Forbes SC, Ashtary-Larky D, Dadkhahfar S, Robati RM, Nezakati F, Khajevandi M, Naseri S, Gerafiani A, Haghighat N, Antonio J, Tinsley GM. Does creatine cause hair loss? A 12-week randomized controlled trial. Journal of the International Society of Sports Nutrition. 2025;22(Suppl 1):2495229. doi: 10.1080/15502783.2025.2495229.
- van der Merwe J, Brooks NE, Myburgh KH. Three weeks of creatine monohydrate supplementation affects dihydrotestosterone to testosterone ratio in college-aged rugby players. Clinical Journal of Sport Medicine. 2009;19(5):399–404. doi: 10.1097/JSM.0b013e3181b8b52f.
- Antonio J, Candow DG, Forbes SC, Gualano B, Jagim AR, Kreider RB, Rawson ES, Smith-Ryan AE, VanDusseldorp TA, Willoughby DS, Ziegenfuss TN. Common questions and misconceptions about creatine supplementation: what does the scientific evidence really show? Journal of the International Society of Sports Nutrition. 2021;18(1):13. doi: 10.1186/s12970-021-00412-w.
- Kreider RB, Kalman DS, Antonio J, Ziegenfuss TN, Wildman R, Collins R, Candow DG, Kleiner SM, Almada AL, Lopez HL. International Society of Sports Nutrition position stand: safety and efficacy of creatine supplementation in exercise, sport, and medicine. Journal of the International Society of Sports Nutrition. 2017;14:18. doi: 10.1186/s12970-017-0173-z.
- Op ’t Eijnde B, Hespel P. Short-term creatine supplementation does not alter the hormonal response to resistance training. Medicine & Science in Sports & Exercise. 2001;33(3):449–453. doi: 10.1097/00005768-200103000-00018.
- Lolli F, Pallotti F, Rossi A, Fortuna MC, Caro G, Lenzi A, Sansone A, Lombardo F. Androgenetic alopecia: a review. Endocrine. 2017;57(1):9–17. doi: 10.1007/s12020-017-1280-y.
- Ntshingila S, Oputu O, Arowolo AT, Khumalo NP. Androgenetic alopecia: An update. JAAD International. 2023;13:150–158. doi: 10.1016/j.jdin.2023.07.005.
- Liu Y, Tosti A, Wang ECE, Heilmann-Heimbach S, Aguh C, Jimenez F, Lin SJ, Kwon O, Plikus MV. Androgenetic alopecia. Nature Reviews Disease Primers. 2025;11(1):73. doi: 10.1038/s41572-025-00656-9.
This article is for general information and is not medical advice. It was reviewed for accuracy by a qualified clinician; decisions about your health should be made with your own doctor.
Related
Sport Nutrition: Energy, Fuel and Athletic Performance
Iron deficiency and how it affects your health
Does Glycogen Hold 3 g of Water? What the 1941 Rat Study Found
Dietary Iron and Hair Growth - Cunningham’s 1932 Rat Study
What research shows about whether creatine is bad for your kidneys
Magnesium for anxiety and what the clinical trials actually found
What antioxidants are and why more is not always better