The short version

Plain English first. Everything below this section is the same material with the sources, the numbers and the caveats attached.

What SS-31 is

SS-31 is a four-residue peptide — D-Arg-Dmt-Lys-Phe-NH₂, where Dmt is 2,6-dimethyl-L-tyrosine, an amino acid that does not occur in nature. It concentrates in the inner mitochondrial membrane, and it has also been called MTP-131, Bendavia and elamipretide depending on who was developing it and when.

On 19 September 2025 the US FDA approved it as FORZINITY, made by Stealth BioTherapeutics. Read the indication carefully, because the wording is doing a lot of work:

Muscle strength, in Barth syndrome, in people over 30 kilograms, on an accelerated approval resting on an intermediate endpoint. Not cardiac function, not survival, not fatigue, not mitochondrial disease generally, and not ageing. Barth syndrome is a rare X-linked disorder with something on the order of 150 diagnosed patients in the United States.

Salt form and what the label on the vial means

The approved drug is the trihydrochloride, and its 40 mg dose is expressed as free base. Research-grade SS-31 is usually supplied as the acetate, which is a heavier salt. A vial labelled 10 mg is 10 mg of whatever salt was weighed out, and the peptide inside it is a smaller number. The certificate is the authority on peptide content, not the label — and this matters more here than for most compounds, because the clinical dose is expressed in free base and the comparison is the whole story of this page.

How it works

SS-31 accumulates in the inner mitochondrial membrane, where it associates with the membrane and improves the efficiency of electron transport. Two things about the standard explanation of that mechanism are wrong, and both have been corrected in the primary literature without the correction reaching the marketing copy.

It is not a general antioxidant

The name "SS-31" and the antioxidant story both come from a 2004 paper, and the story has not survived. The closely related analogue SS-20 has no radical-scavenging side chain at all and still works in preclinical models. As the mechanistic work puts it, "antioxidant chemistry is not likely to be the primary mechanism of SS peptides." Reduced reactive oxygen species is a consequence of better electron transport, not the cause of the benefit.

The practical implication: none of the reasoning people apply to CoQ10, vitamin E or any other antioxidant transfers here. Dose-response logic borrowed from that world does not apply.

It does not bind cardiolipin specifically

The FDA's own label describes elamipretide as "a mitochondrial cardiolipin binder." The mechanistic literature says something more precise: binding tracks the membrane's negative surface charge, not cardiolipin's identity. In yeast engineered to make no cardiolipin at all, binding was "nearly identical" to wild type — because phosphatidylglycerol rises to take its place and the surface charge is restored.

Cardiolipin is simply where most of the fixed negative charge in that membrane lives. It is the usual answer, not the required one. We are flagging the gap between the label's phrasing and the mechanism papers rather than picking whichever version sounds better.

Pharmacokinetics

Human pharmacokinetics for this compound are unusually well characterised, and they are split across two documents in a way worth knowing about.

Renal impairment raises exposure substantially: AUC increased 39% at creatinine clearance 60–89, 75% at 30–59, and 125% below 30. Anyone with reduced kidney function is getting a materially larger exposure from the same dose.

What the evidence shows

This is the only page in our library where the problem is not an absence of trials. Seven controlled trials were run. Not one met its prespecified primary endpoint.

MMPOWER-1 at p = 0.053 and MMPOWER-2 at p = 0.083 are the two anyone will reach for. Neither crossed. A near miss in a 30-person crossover is a reason to run a bigger trial, which is exactly what happened — and the bigger trial, at 218 patients, returned p = 0.69.

So what was it approved on?

Not a controlled trial. The approved label says so itself, in section 14, and this is the single most useful sentence anyone could read about this compound:

The efficacy finding comes from what happened after the randomised phase ended. Twelve patients were ever randomised in the Barth trial. Ten entered an open-label, single-arm extension, eight of whom reached week 168, and their knee extensor strength rose by a median of 63 newtons. That was supported by an externally controlled comparison against a natural history cohort in which, per the FDA's briefing document, eight subjects had evaluable data.

Open-label means every patient knew they were on the drug, and the endpoint is effort-dependent — you have to push against a machine. The FDA said so plainly: "The open-label, single-arm design, with all subjects aware that they were receiving elamipretide, can introduce performance bias." The agency accepted higher-than-usual uncertainty because Barth syndrome affects roughly 150 people in the country and there is no alternative treatment. That is a defensible regulatory judgement about a rare fatal disease. It is not evidence that the compound does anything for a healthy adult.

The regulator did not arrive here easily

FDA refused to file the original application in October 2021, issued a Complete Response Letter on 15 May 2025, and only approved on the second resubmission. It attached a required confirmatory randomised trial that does not report until March 2030. Accelerated approval means the question is still open.

Anti-ageing

One 39-person study gave a single intravenous dose to older adults with pre-documented mitochondrial impairment. The primary endpoint was met on percent change and missed on absolute change, the functional endpoint was not significant, and the effect was gone by day seven. A 2025 mouse study found improved muscle function with no change in epigenetic or transcriptomic age. There is no human evidence that this compound slows ageing.

Reconstitution

A 10 mg vial:

Our house volume is 1 mL, and the reason is the published 1 to 5 mg convention above. At 10 mg/mL one U-100 unit is exactly 0.1 mg, so 1 mg is 10 units, 2 mg is 20 and 5 mg is 50 — whole numbers across the whole range, and every one of them inside a single 0.5 mL barrel. At 2 mL a 5 mg draw becomes 100 units, which is an entire 1 mL barrel with nothing in reserve; at 3 mL it is 150 units and takes two draws. A larger volume does not make the dose smaller. It only makes the number harder to read. The calculator will do this for any volume you prefer.

Direct the stream down the inside wall of the vial rather than onto the cake, swirl rather than shake, and label the vial with the resulting concentration. The full method is in the reconstitution guide.

Reported protocols

The gap between the studied dose and the circulating dose is the largest in this catalogue, and unlike most such gaps this one cannot be explained away by absorption.

Why people converge downward anyway — and it is not efficacy

The honest explanation is tolerability. Injection-site reactions at 40 mg are dose-related and effectively universal (see below). People titrate down until the injections are bearable, and then the dose that is bearable becomes the dose that gets recommended. That is a real and understandable reason. It is not evidence that the lower dose does anything, and the two should not be confused.

Storage

Lyophilised, sealed, cold, dark and dry, this is a stable solid. The handling conventions in our storage guide apply. They are conventions carried across from peptides generally rather than a published stability study for this compound in a research presentation, and we say so.

Reconstituted in bacteriostatic water, the limiting figure is microbiological rather than chemical: the benzyl alcohol preservative is validated for 28 days. That is the number to work to for a multi-draw vial, and it is not a claim about when the peptide degrades. There is more on where the figure comes from in our FAQs.

Reported effects

Injection-site reactions, which are close to universal

From the approved label, in the Barth syndrome trial — twelve patients per arm, so read these as a strong signal rather than a precise population rate:

The same picture at scale, from MMPOWER-3 with 109 patients per arm: injection site erythema in 86.2% against 28.4% on placebo, any adverse event in 98.2% against 76.1%, and 7.3% discontinued for adverse events against 1.8% on placebo. This is not a compound people take without noticing.

Carcinogenicity — genuinely unanswered

The approved label states, verbatim: "Carcinogenicity studies have not been conducted with elamipretide." FDA approved it anyway, as it may for a rare fatal disease, and required the studies afterwards. The 26-week mouse study was due in February 2026; the two-year rat carcinogenicity study is not due until January 2030.

That is a reasonable trade for a patient with Barth syndrome and no alternative. It is a different calculation entirely for someone contemplating years of use in good health, and anyone in that position should be told the studies do not exist yet rather than left to assume they do.

Renal function

Elimination is entirely renal. Exposure rises 39%, 75% and 125% as creatinine clearance falls through the 60–89, 30–59 and below-30 bands. Reduced kidney function means a materially larger exposure from an unchanged dose.

Regulatory status

United States. Approved 19 September 2025 as FORZINITY, under accelerated approval, to improve muscle strength in Barth syndrome patients weighing at least 30 kg. Prescription only. A confirmatory randomised trial is required and reports in March 2030.

Elsewhere. Not approved in the European Union, the United Kingdom or Canada.

Canada. Health Canada's public advisory of 9 April 2026 names SS-31 among fourteen peptides sold online, states that peptides are generally regulated as prescription drugs in Canada, and states that research-use-only labelling "does not make these products legal or exempt from regulatory requirements." It is not a controlled substance under the Controlled Drugs and Substances Act. The advisory is linked in the references and is worth reading directly.

Sentient Biogenix supplies SS-31 for laboratory research use only. It is not for human or veterinary use.

Questions we actually get

If it failed every trial, how is it an approved drug?

Because accelerated approval for a rare disease with no alternative applies a different standard from the one most people picture. FDA weighed twelve randomised patients, an open-label extension of ten, and a natural-history comparison against the fact that Barth syndrome is fatal and untreated, and decided the uncertainty was acceptable while a confirmatory trial runs. That is a defensible decision about that disease. It is not a finding that transfers.

Is 1 to 5 mg a sensible starting dose?

It is the conventional one, and no trial has tested it. It is also 8 to 40 times below the dose the rare-disease programmes used — though that dose is not a validated target either, since MMPOWER-3 missed both of its primary endpoints at 40 mg. Because bioavailability is 92% and exposure is dose-proportional, the low dose gives proportionally low exposure — there is no mechanism by which it quietly matches the studied dose. Whether it does anything at all has not been tested.

Is it an antioxidant?

Not primarily. The analogue SS-20, which lacks the radical-scavenging side chain entirely, works in preclinical models — which is fairly decisive. Reduced oxidative stress is downstream of better electron transport rather than the mechanism itself.

Does it bind cardiolipin?

It binds negatively charged membrane surfaces, and cardiolipin is usually where that charge is. Mitochondria engineered to make no cardiolipin bind it about as well, because another anionic lipid rises to replace the charge. The FDA label's phrase "cardiolipin binder" is a simplification of what the mechanism papers actually found, and we would rather show you the disagreement than pick the tidier version.

Can I dose it weekly?

There is no pharmacokinetic basis for it. Plasma half-life is three to four hours, there is essentially no accumulation on daily dosing, and about all of a dose is out through the kidneys within two days. Every trial dosed daily.

References

1. FORZINITY (elamipretide) injection, prescribing information. US FDA, 2025. accessdata.fda.gov — the indication, "FORZINITY was not superior to placebo on these primary endpoints," the 40 mg dose, 92% bioavailability, dose proportionality 2 to 80 mg, renal impairment figures, injection-site reaction table, molecular formula and weight, and "Carcinogenicity studies have not been conducted with elamipretide."
2. FDA approval letter, NDA 215244, 19 September 2025. accessdata.fda.gov — accelerated approval, the required confirmatory trial due March 2030, and the two-year rat carcinogenicity study due January 2030.
3. FDA Integrated Review, NDA 215244. accessdata.fda.gov — the October 2021 refuse-to-file, the 15 May 2025 Complete Response Letter, the TAZPOWER p-values, and the composition of the evidence base.
4. FDA Briefing Document, elamipretide. fda.gov — the 3 to 4 hour plasma half-life, time to peak, renal elimination, the eight-subject externally controlled comparison, and the performance-bias caution. Note: the half-life figure is here and not in the approved label.
5. Reid Thompson W, et al. A phase 2/3 randomized clinical trial followed by an open-label extension to evaluate the effectiveness of elamipretide in Barth syndrome. Genetics in Medicine 2021. nature.com — TAZPOWER, twelve patients, p = 0.97 and p = 0.89.
6. Karaa A, et al. Randomized trial of elamipretide in primary mitochondrial myopathy (MMPOWER-3). Neurology 2023. digitalcommons.library.tmc.edu — 218 patients, p = 0.69 and p = 0.37, and the 7.3% discontinuation rate.
7. Karaa A, et al. A randomized crossover trial of elamipretide in primary mitochondrial myopathy (MMPOWER-2). 2020. pmc.ncbi.nlm.nih.gov — 19.8 m, p = 0.083.
8. Mitchell W, et al. The mitochondria-targeted peptide SS-31 binds lipid bilayers and modulates surface electrostatics as a key component of its mechanism of action. Journal of Biological Chemistry 2020. biorxiv.org — "antioxidant chemistry is not likely to be the primary mechanism of SS peptides"; binding driven by anionic surface charge density rather than cardiolipin identity; near-identical binding in cardiolipin-null yeast. Quoted from the open preprint because the journal blocks automated access.
9. Health Canada. Think twice before injecting peptides bought online: unauthorized products can seriously harm you. Public advisory, 9 April 2026. recalls-rappels.canada.ca

Every figure above was read from the source named beside it. Where a number is in the FDA briefing document rather than the approved label, this page says which. Where we could not open a primary publication — EMBRACE STEMI and ReCLAIM-2 in particular — the page states the outcome and not the fine detail, because we would rather print less than print something we took from a summary.