HMB — β-hydroxy β-methylbutyrate — is one of the few sports supplements whose evidence base genuinely cannot be summarised in a single verdict. Read the trials in trained lifters and you would conclude it does almost nothing. Read the trials in older adults confined to a hospital bed and you would conclude it does something worth having. Both readings are correct, and averaging them into one “does HMB work?” answer is the single most common mistake made about this compound. The split is not noise. It is the finding, and it follows directly from what HMB actually does inside a muscle cell.
What HMB is, and where it comes from
HMB is not an exotic molecule. Your body already makes it: leucine, the branched-chain amino acid that acts as the trigger for muscle protein synthesis, is partly catabolised inside the muscle, first to the keto acid KIC and then, in a small fraction of cases, onward to HMB. Only a modest share of leucine takes that route — conventionally put at around five percent — which is why you cannot eat your way to supplemental doses. Getting the ~3 g used in trials from food would mean an implausible quantity of protein. That gap between endogenous production and the studied dose is the entire commercial rationale for the supplement.[14]
It has been sold in two chemical forms for most of its history: the calcium salt (HMB-Ca), which is what the great majority of trials used, and a free acid form (HMB-FA), marketed on the strength of better absorption. We come back to whether that distinction matters below — the short answer is less than the marketing implies.
The mechanism: a brake first, a throttle second
The mechanistic story told on supplement labels is usually the leucine story: HMB activates mTORC1, mTORC1 drives muscle protein synthesis, therefore HMB builds muscle. That is not wrong, but it is the less important half, and the human data make the ranking fairly clear.
The cleanest human experiment fed roughly 3 g of free-acid HMB to healthy young men while tracking both sides of protein turnover with stable-isotope tracers. HMB did stimulate muscle protein synthesis — by about 70%, versus about 110% for an equivalent dose of leucine itself, so it is weaker than its own parent amino acid on the anabolic side. But it also suppressed muscle protein breakdown by roughly 57%, insulin-independently, which leucine was not shown to do in the same way.[2] A follow-up study using the calcium form found the same dual pattern: a rise in synthesis with mTORC1 substrate phosphorylation, and a simultaneous fall in breakdown.[3] The professional summary of the field describes exactly this — a dual mechanism of enhanced synthesis and suppressed breakdown, with HMB's mTORC1 activation apparently running independently of the leucine-sensing pathway.[14]
The original 1996 human work pointed the same direction before anyone was talking about mTOR: HMB supplementation blunted the exercise-induced rise in urinary 3-methylhistidine, a marker of muscle proteolysis, and reduced creatine kinase.[1] HMB looked like a brake on breakdown from the very first study.
Everything downstream follows from that. A brake only produces a visible result where something is actually being lost. In a well-fed young lifter in a good training programme, net protein balance is already being driven positive and there is very little catabolism left to prevent. In someone immobilised, ill, elderly, or in a substantial energy deficit, there is a great deal. That is the population split, and it is mechanistically predictable rather than mysterious.
The population split: five different answers
Bed rest and disuse: the most convincing signal
The most persuasive single trial is not an athletic one. Twenty-four healthy older adults were confined to complete bed rest for ten days while taking either 3 g/day of calcium HMB or placebo. The control group lost about 2 kg of total lean body mass over those ten days; the HMB group essentially did not.[10] That is a large, clinically meaningful protection of lean tissue over a short window, in exactly the catabolic setting the mechanism predicts. It was a small study and its authors said plainly that it needed replication in a larger trial, but as a proof of concept it is the strongest thing HMB has. The professional position stand reaches the same conclusion in more cautious language: HMB may be effective in countering disuse atrophy during illness or injury.[14]
Sarcopenia and older adults: real but genuinely mixed
Here the literature disagrees with itself in an instructive way. A meta-analysis of seven randomised trials in adults aged 65 and over found greater muscle-mass gain with HMB than control, with no change in fat mass — the authors concluded HMB contributes to preserving muscle mass in older adults and might be useful specifically against bed rest–induced atrophy.[11] A 2025 meta-analysis restricted to patients meeting formal clinical sarcopenia criteria found benefits for skeletal muscle mass index and handgrip strength — but explicitly no benefit for physical performance measured as gait speed, and told readers to interpret the whole thing with “utmost caution” given only five eligible trials.[13]
And then the counterweight: a meta-analysis of ten RCTs asking a narrower and more practical question — does adding HMB to an exercise programme in adults aged 50–80 beat the exercise alone? — concluded that it had no or only fairly low impact on body composition, strength, or physical performance.[12] That is the honest shape of the older-adult evidence: HMB looks most useful where muscle is being lost and least useful as an add-on to a training programme that is already working. Gaining muscle and not losing it are different problems, and HMB is better evidenced for the second.
Untrained beginners: small, and probably real
The cleanest quantification of the training split comes from a meta-analysis of nine studies (394 subjects) that deliberately subcategorised results by training experience. In previously untrained men, HMB produced small but clear gains: about 9.9% for lower-body strength and 6.6% for average strength.[5] Effects on fat and fat-free mass were trivial even there, which is worth sitting with — the strength benefit in beginners was not obviously a body-composition benefit.
Trained athletes: weak to null
In the same meta-analysis, all strength outcomes in trained lifters were trivial, and the authors' conclusion was blunt: benefits in trained lifters are trivial and the effect on body composition is inconsequential.[5] Individual trials in this population agree. Six weeks of 3 g/day in resistance-trained male athletes, in either standard or time-release capsules, produced significant strength and lean-mass gains from the training itself — with no influence of HMB on any of it, and no effect on markers of protein turnover or muscle damage either.[6] Even the supplement field's own position stand concedes that performance benefits in trained athletes are “mixed.”[14]
| Population | What the human evidence shows |
|---|---|
| Older adults on bed rest / disuse | Positive — lean mass largely preserved vs placebo in a small RCT |
| Clinically defined sarcopenia | Mixed-positive — muscle mass and grip strength improved; gait speed not |
| Older adults already exercising | Little to none beyond the exercise itself |
| Untrained beginners lifting weights | Small strength benefit; body-composition effect trivial |
| Trained athletes | Weak to null — all strength outcomes trivial in meta-analysis |
| People losing weight on a GLP-1 | No trials — rationale is extrapolated only |
Calcium HMB vs HMB free acid
The pharmacokinetic difference is real and well characterised. In a crossover study, the free-acid gel form reached peak plasma HMB in about 38 minutes versus roughly 128 minutes for the calcium salt, roughly doubled peak concentration and area under the curve, and improved plasma clearance — from which the authors inferred that free acid “could improve HMB availability and efficacy.”[4] Note the conditional. That study measured blood levels, not muscle outcomes.
The follow-up that actually tested the implication is the more useful one. When calcium HMB was given to healthy young men with the same tracer methodology used earlier for the free acid, it produced a comparable stimulation of protein synthesis and comparable suppression of breakdown— explicitly “despite proposed differences in bioavailability.”[3] The position stand's own wording has softened accordingly: HMB-FA appears to raise blood HMB more than HMB-Ca, “though recent results are mixed.”[14] The honest read is that better kinetics has not clearly translated into better outcomes, and the overwhelming majority of the positive clinical evidence — including the bed-rest trial and the older-adult meta-analyses — was generated with the cheaper calcium salt. Paying a premium for free acid is buying a plasma curve, not a demonstrated result.
The GLP-1 claim, stated honestly
HMB is now being marketed for preserving lean mass during weight loss on semaglutide and tirzepatide, and the reasoning is not stupid. A meaningful fraction of the weight lost on these drugs is lean tissue rather than fat — we cover that in detail in GLP-1 medications and muscle — the deficit is large, intake is suppressed, and HMB's best-supported action is precisely a brake on protein breakdown in catabolic states. On paper it is the most mechanistically coherent supplement pitch in the entire GLP-1 space.
What would settle it is unremarkable and eminently doable: a randomised, placebo-controlled trial in people on a GLP-1 receptor agonist, with DXA or MRI lean-mass outcomes plus function (grip strength, chair-stand, gait speed), running long enough to matter — twelve weeks at minimum, ideally through the titration period when weight loss is fastest — and, crucially, layered on top of adequate protein and resistance training rather than compared against nothing. The last point is what makes or breaks the commercial case. HMB has already failed to add much to exercise in older adults who were training properly.[12]
Until that trial exists, the evidenced approach to protecting muscle on a GLP-1 is the boring one, and it is the same advice given by reviews written specifically for this drug class: adequate dietary protein paired with resistance training.[17] That pairing is not a hypothesis. HMB, in this population, still is.
Dose, timing, safety, and what HMB legally is
Dose. Effectively every trial worth citing used 3 g/day, and that number is not the result of dose-optimisation research — it is a convention inherited from the original 1996 experiments, which compared 0, 1.5, and 3.0 g/day and found 3 g at least as good as 1.5 g.[1] It became the standard because it was the top of the first dose ladder, and almost nobody has revisited it since. The position stand now expresses the same thing as roughly 38 mg per kg of body weight per day, which lands near 3 g for a typical adult and is the more defensible way to scale it.[14] Doses are usually split — the bed-rest trial gave 1.5 g twice daily.[10]
Timing. There is a case, though not a strong one, for taking a dose near training: the position stand suggests intake close to an exercise bout may help synthesis and blunt the inflammatory response.[14] Given that the benefit accrues from sustained suppression of breakdown, total daily intake and consistency almost certainly matter more than the clock — and if you are taking it for disuse or illness rather than training, timing is moot.
Safety. This is HMB's least controversial dimension. A pooled analysis of nine human studies at 3 g/day, spanning young and old, exercising and sedentary, found no adverse effect on any surrogate marker of organ or tissue function, and incidentally reported small reductions in total and LDL cholesterol and systolic blood pressure.[15] The 2025 position stand concludes that chronic HMB-Ca and HMB-FA intake is safe for at least a year, with no negative effect on glucose tolerance or insulin sensitivity.[14] Two honest caveats: much of that safety literature was generated by parties with a commercial interest in HMB, and “no signal in short trials” is not the same as long-term safety data in frail, polypharmacy patients — the very group most likely to be offered it.
Regulatory status. HMB is a dietary supplement. It is not an FDA-approved drug for sarcopenia, cachexia, disuse atrophy, GLP-1–associated muscle loss, or anything else, and no manufacturer may lawfully claim it treats or prevents a disease. Purity, dose accuracy, and label honesty are the buyer's problem, as with any supplement.
HMB vs creatine: not a close contest
If you are choosing one, the comparison is lopsided and it is not close. Creatine monohydrate has hundreds of human trials, replicated meta-analytic effects on strength, power, and lean mass, a mechanism that is fully worked out, and a cost of cents per day — see our creatine monohydrate monograph for the full grading. HMB has, in the same population where creatine shines, essentially nothing: trivial strength outcomes in trained lifters and inconsequential body-composition effects.[5]
The two are not really competitors, though, because they answer different questions. Creatine is a training amplifier — it helps you do more work, which is why it needs a training stimulus to act on. HMB is closer to a tissue-loss brake — which is why it shows up best when the stimulus is absent and the losses are real. For a healthy person who lifts, creatine is the obvious first choice and HMB is largely redundant. For someone facing ten days in a hospital bed, the ranking is less clear-cut, and HMB has the more specific evidence. The same logic separates HMB from the injectable muscle-growth peptides, most of which cannot yet show either kind of result in humans.
The honest bottom line
HMB is a legitimate compound that has been sold on the wrong claim for thirty years. It was marketed to bodybuilders, which is the population where it works least well, on the basis of an anabolic mechanism that is its weaker one. Its actual strength is anti-catabolic and its actual constituency is people losing muscle: the immobilised, the ill, the sarcopenic. Reasonably evidenced: preserving lean mass during bed rest and disuse, and improving muscle mass and grip strength in clinical sarcopenia. Small: strength in untrained beginners. Weak to null: anything in trained athletes. Unproven and currently untested: the GLP-1 lean-mass pitch, which is a good hypothesis wearing the costume of a finding. At roughly 3 g/day of the cheap calcium salt, with a clean short-term safety record, HMB is a defensible thing to take if you are in a catabolic situation — and a largely pointless one if you are a healthy person who lifts and already eats enough protein.
This article is research information, not medical advice. HMB is sold as a dietary supplement and is not approved by the FDA to treat or prevent any condition, including sarcopenia, cachexia, disuse atrophy, or muscle loss during GLP-1 treatment. Nothing here is a recommendation to start, stop, or substitute any supplement or medication. If you are losing muscle because of illness, immobility, aging, or rapid weight loss on a GLP-1, that is a clinical situation — discuss it with a licensed clinician or registered dietitian, who can assess your protein intake, kidney and liver function, and other medications first.