In brief
What breaks in Barth syndrome
Inside the mitochondrion there is an inner membrane folded into cristae. The respiratory chain complexes that generate energy sit in those folds. What holds them together is a particular lipid — cardiolipin.
Cardiolipin is built unusually: it has four tails instead of the usual two, so the molecule is wedge-shaped and a membrane made of it curves naturally. It is neither fuel nor a component of the engine, but fastening and framework at once.
Freshly made cardiolipin is not yet fit for use — it has to be finished, with some of its tails exchanged. That job belongs to the enzyme tafazzin, encoded by the TAZ gene. In Barth syndrome that gene is damaged.
The result: little mature cardiolipin, and an intermediate form — monolysocardiolipin — accumulating in its place. The ratio of the two, MLCL/CL, serves as a laboratory marker of the disease. The membrane loses its shape, the complexes drift apart, and less energy is produced.
The tissues that never stop working suffer most: heart muscle, skeletal muscle, bone marrow. Hence the pattern — cardiomyopathy, muscle weakness and fatigue, intermittent neutrophil deficiency, growth delay. The gene sits on the X chromosome and boys are affected. Until recently most died in the first years of life; today, with cardiological care, many reach adulthood.
The idea: not to mend the fastening but to brace the structure
Elamipretide is a short peptide of just four amino acids. It crosses membranes and settles on the cardiolipin of the inner mitochondrial membrane.
It is important to be clear about what it does not do. It does not correct the TAZ gene. It does not restore tafazzin. It does not turn defective cardiolipin into normal cardiolipin.
What it does is different: bound to the lipid, it stabilises the geometry of the membrane — the folds hold their shape and the respiratory chain complexes stay in working position. In effect, an external brace on a structure whose internal fastening has worn out. Indirect evidence that the intervention reaches its target: over long-term follow-up the MLCL/CL ratio shifted towards normal, and that shift tracked clinical measures.
What the TAZPOWER trial showed
Here begins the instructive part. The trial was competently designed: randomised, double-blind, crossover. Each of the 12 participants received 12 weeks of drug and 12 weeks of placebo, with a four-week washout between periods. With so few participants this is the best design available: each patient serves as their own control.
The primary endpoints were distance on the six-minute walk test and the Barth syndrome symptom scale.
Ten participants moved into the open-label extension, in which everyone received the drug and everyone knew it. There the measures began to rise:
| Open-label extension | Result |
|---|---|
| Walking distance, week 36 | *+95.9 m* (p=0.024) |
| Symptom scale, week 36 | *−2.1 points* (p=0.031) |
| Walking distance, week 168 | *+96.1 m* cumulative (p=0.003) |
| Fatigue | Below baseline at every time point |
| Left ventricular volumes | Significant trend towards improvement |
| MLCL/CL ratio | Shift towards normal, correlating with clinical outcomes |
Eight of the ten reached week 168. Tolerability was good, with injection-site reactions the most common finding.
The comparison with those who were not treated
An open-label extension alone is not enough: over three years a participant simply learns to walk the test better. So a separate study was done — 19 untreated patients were assembled from records and matched statistically to the 8 treated ones.
| Difference in favour of elamipretide | Week 64 | Week 76 |
|---|---|---|
| Walking distance | *+79.7 m* (p=0.0004) | *+91.0 m* (p=0.0005) |
| Muscle strength (dynamometry) | *+40.8 N* (p=0.0002) | *+56.7 N* (p=0.0005) |
One finding is harder to explain by expectation: left ventricular stroke volume rose in the treated and fell in the untreated. In the untreated the disease simply took its course — the heart weakened.
Why the drug was approved anyway
The regulator faced the choice familiar from every ultra-rare disease. A conventional large trial is physically impossible: there are a few hundred patients in the world, and twelve could be enrolled into the development programme. To demand full evidence is to guarantee that the disease is left untreated for good.
The decision: accelerated approval on an intermediate clinical endpoint, the increase in knee extensor strength. The indication was honestly narrowed to match: the drug is approved not 'for the treatment of Barth syndrome' but 'to improve muscle strength' in patients who have it. Continued approval is made contingent on a confirmatory trial.
The weaknesses of the evidence deserve to be stated plainly:
▸the open-label extension cannot separate drug effect from natural course, expectation, or practice at the test; ▸the untreated comparison was retrospective, drawn from records; statistical matching narrows the differences between groups but does not remove them; ▸twelve participants in the development programme, eight reaching the end — any chance event weighs disproportionately in a group that small; ▸employees of the manufacturer are among the authors of the key publications.
And the counter-argument worth holding alongside: every measure moved in the same direction, and cardiac function in treated and untreated patients diverged in opposite ways. That is not typical of a pure expectation effect.
A sobering check: the same molecule, a different disease
There is an experiment that clarifies much. The same elamipretide was tested in primary mitochondrial myopathy — a different group of diseases in which mitochondria also suffer, but cardiolipin has nothing to do with it.
MMPOWER-3 was everything the Barth syndrome trial could not be: 218 participants, randomised, double-blind, placebo-controlled, 24 weeks.
| MMPOWER-3, primary endpoints | Result |
|---|---|
| Walking distance | *−3.2 m* (95% CI −18.7 to 12.3), p=0.69 |
| Total fatigue score | −0.07 (95% CI −0.10 to 0.26), p=0.37 |
Not the faintest hint of an effect. Two conclusions follow. On one hand a warning: this is not a drug 'for mitochondria in general', and there is no ground for carrying hopes over to other diseases. On the other, an argument in favour of specificity: if the improvements in Barth syndrome were a pure artefact of open observation, the mechanism would not matter — and apparently it does.
What it costs
▸A daily subcutaneous injection — for life, at the same time each day; the most common adverse event is local, at the injection site. ▸Benzyl alcohol in the formulation. In low-birth-weight and preterm neonates this preservative has caused severe metabolic acidosis with neurotoxicity and fatal outcomes. The drug is therefore not used in neonates, and intravenous administration is not approved at all. ▸A weight limit: at least 30 kg, so the youngest children fall outside the indication. ▸Kidneys: in adults with severe renal impairment the dose is reduced.
Who it is not for
▸Neonates — because of the benzyl alcohol in the formulation. ▸Body weight under 30 kg — no dose has been established for such patients. ▸Serious hypersensitivity to the drug or its excipients — the only formal contraindication. ▸Other mitochondrial diseases — a large randomised trial found no effect. ▸As a replacement for cardiological care — the drug does not remove the need to treat heart failure or to monitor neutrophils.
What cannot be claimed yet
▸Effects on survival and on the need for heart transplantation were never studied. All endpoints were functional or laboratory. ▸The only randomised comparison against placebo was negative. Everything positive comes from unblinded observation and from a comparison with historical controls. ▸The confirmatory trial has not concluded — continued approval depends on it. ▸Children under 12 and lighter patients were not studied: the programme enrolled boys and men aged 12 to 35. ▸Long-term safety beyond three and a half years is unknown, while the drug is meant to be taken for decades.
Bottom line
▸The first drug in Barth syndrome aimed at cardiolipin, the structural lipid of the mitochondrial membrane. ▸It does not repair the gene but compensates for the lost membrane function by stabilising its geometry. ▸The randomised part of the trial failed — both primary endpoints were missed. ▸Improvements appeared in the open-label extension: +96 m of walking distance by week 168, and rising muscle strength. ▸Against untreated patients the difference in walking was 80–91 m; in the untreated the heart meanwhile weakened. ▸Approval is accelerated, on an intermediate endpoint, with the narrow wording 'to improve muscle strength'; confirmation is still to come.
Your own case — whether the diagnosis is genetically confirmed, what the MLCL/CL ratio and cardiac function show — can be reviewed at a consultation; the drug can be ordered here.
References
1. 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, a genetic disorder of mitochondrial cardiolipin metabolism. Genet Med. 2021;23(3):471–478. PMID 33077895
2. Thompson WR, et al. Long-term efficacy and safety of elamipretide in patients with Barth syndrome: 168-week open-label extension results of TAZPOWER. Genet Med. 2024;26(7):101138. PMID 38602181
3. Hornby B, et al. Natural history comparison study to assess the efficacy of elamipretide in patients with Barth syndrome. Orphanet J Rare Dis. 2022;17(1):336. PMID 36056411
4. Karaa A, et al. Efficacy and Safety of Elamipretide in Individuals With Primary Mitochondrial Myopathy: The MMPOWER-3 Randomized Clinical Trial. Neurology. 2023;101(3):e238–e252. PMID 37268435
5. FORZINITY (elamipretide) — US Prescribing Information, Stealth BioTherapeutics Inc.
Key facts
- Elamipretide (brand name Forzinity, Stealth BioTherapeutics) was approved by the FDA in September 2025 with the wording 'to improve muscle strength' in patients with Barth syndrome weighing at least 30 kg.
- Barth syndrome is an ultra-rare X-linked inherited disorder: a fault in the TAZ gene disrupts the maturation of cardiolipin, the lipid of the inner mitochondrial membrane.
- It presents with cardiomyopathy, skeletal muscle weakness, neutrophil deficiency and growth delay; boys are affected, and until recently most died in the first years of life.
- Cardiolipin is neither fuel nor a component: it is the structural fastening of the inner membrane, holding the respiratory chain complexes in working position.
- Elamipretide is a short peptide of four amino acids that binds cardiolipin and stabilises the membrane; it does not repair the gene and does not create normal cardiolipin.
- In the randomised crossover part of TAZPOWER, involving 12 patients, both primary endpoints — the six-minute walk test and the symptom scale — were missed.
- In the open-label extension, walking distance had improved by 95.9 m at week 36 (p=0.024) and by 96.1 m at week 168 (p=0.003); knee extensor strength improved too.
- Against 19 untreated patients, the difference in walking distance was 79.7 m at week 64 and 91.0 m at week 76 (both p<0.001), and in muscle strength 40.8 and 56.7 newtons.
- Approval was granted through the accelerated pathway on an increase in knee extensor strength — an intermediate clinical endpoint; confirmation of benefit is expected from a separate trial.
- In a large randomised trial in a different mitochondrial disease — primary mitochondrial myopathy, 218 participants — the same drug missed both primary endpoints.
