Elamipretide: a peptide tested against physical function
In this investigation
Elamipretide binds cardiolipin in mitochondrial membranes. Its human trials help distinguish a proposed target from a useful outcome.
Early symptom signals had a failed statistical gate
MMPOWER-2 randomized 30 participants from an earlier mitochondrial muscle-disease trial into four-week elamipretide/placebo periods, separated by washout. It therefore reused participants rather than providing wholly independent replication.
The primary walking difference was 19.8 metres, 95% CI −2.8 to 42.5, P=.0833. Reported fatigue improved by 1.7 points, but the statistical hierarchy required primary success before confirmatory secondary testing. That fatigue finding remains exploratory. Timed mobility and accelerometer activity did not improve significantly.
One participant withdrew because of injection-site pain. Stealth BioTherapeutics funded the trial; investigators disclosed sponsor grants, travel and consulting payments. Original crossover trial
The larger myopathy trial missed both primary outcomes
MMPOWER-3 randomized 218 people with genetically confirmed primary mitochondrial myopathy; 205 completed 24 weeks.
Walking distance differed by −3.2 metres, CI −18.7 to 12.3, P=.69. Full Results report null fatigue (P=.81). A favorable nuclear-DNA subgroup and later genotype analysis reused this trial; they do not overturn the overall result or constitute replication.
Adverse events were reported in 98.2% versus 76.1%; adverse-event discontinuations were 8 versus 2. Injection-site reactions predominated. Stealth funded the study, participated in its design, and employed or consulted with several authors. Original phase-three trial, same-trial genotype analysis
The later nuclear-DNA-focused NuPOWER record is completed with 102 enrolled and no posted results at this review. Completion is not a positive outcome.
Acute energy production did not establish endurance
A blinded experiment treated 39 adults aged 60–85, selected for low mitochondrial function, after 41 were randomized and two excluded before treatment. The single intravenous exposure differs from chronic subcutaneous trials.
The protocol-defined absolute change in ATP (cellular energy) production capacity had P=.055; a percentage-change analysis had P=.045. The planned hand-muscle force-time comparison was null. Other post-hoc functional analyses suggested improvement, but were exploratory; resting mitochondrial coupling and the seven-day between-group ATP comparison were null.
This distinguishes an acute physiological signal from proven durable training benefit. Stealth, NIH and university sources funded the work; three authors disclosed paid sponsor consulting. Original older-adult trial, original protocol
Barth syndrome: extensions and accelerated approval
TAZPOWER randomized 12 males. Walking, fatigue and strength did not significantly outperform placebo after twelve weeks. Ten entered the uncontrolled extension; eight reached week 168. Strength and walking improvements concern that same uncontrolled cohort. Safety was the extension’s primary outcome; functional measures were secondary. Two stopped for injection-site reactions. Stealth funded and helped analyze the trial; a coauthor was its employee. Original trial
FDA granted accelerated approval in September 2025 to Forzinity for muscle strength in Barth syndrome at ≥30 kg, based on knee-extensor strength as an intermediate endpoint. Confirmatory testing remains required. Serious hypersensitivity is a labeled risk. FDA label
The approval authority accepted greater uncertainty for this rare disease; clinical and statistical reviewers had recommended against approval. FDA assessment
What this establishes for me
These trials do not establish improved lifting, sprinting or recovery for me. Their populations and measured outcomes limit what the proposed target proves.
The peptide-precision investigation develops the broader framework.