Why elamipretide's blinded trial and extension tell different stories

Elamipretide's Barth syndrome research can seem contradictory at first. The initial randomized trial did not meet its main goals, yet a longer follow-up reported improvements, and FDA later granted accelerated approval to the prescription product Forzinity. How can those statements all be true?

They describe different comparisons, different observation periods and, for the approval, a different endpoint. The longer findings don't erase the blinded result, and the blinded result doesn't make the extension worthless. Each answers part of the clinical question while leaving a different amount of uncertainty.

For someone living with Barth syndrome, the distinction is practical. It explains what benefit the approval addresses, why ongoing evidence matters and why a treatment discussion should be more specific than whether “the trial worked.” Our SS-31 guide introduces elamipretide's mechanism and wider clinical evidence.

Start with the disease and the outcomes

Barth syndrome is a rare genetic disorder involving abnormal cardiolipin biology in mitochondria. Elamipretide interacts with cardiolipin, providing a rationale for investigating its effects on muscle and other affected tissues. The clinical question, however, is whether treatment improves something meaningful for patients.

The original TAZPOWER study selected two main outcomes: distance walked during six minutes and fatigue recorded with a Barth syndrome symptom assessment. Researchers also measured other outcomes, including knee-extensor strength. These are related aspects of function, but they are not interchangeable. Original TAZPOWER report

Walking integrates several systems and the person's effort during a task. A fatigue questionnaire captures the patient's experience. Knee-extensor strength measures force from muscles that straighten the knee. A change in one does not guarantee a corresponding change in the others.

That is why the original endpoint hierarchy matters. The primary outcomes were the main questions the blinded comparison was designed to answer. Other measurements can be valuable, but their interpretation needs to preserve their role in the study.

What happened during the randomized period?

Twelve participants entered a randomized, double-blind, placebo-controlled crossover study. They received elamipretide or placebo for twelve weeks, had a four-week washout, and then received the opposite treatment for another twelve weeks. Neither primary endpoint was met. TAZPOWER primary publication

In a crossover design, participants experience both treatments, allowing comparisons within the same people. This can be useful in a rare disease where recruiting a large parallel-group trial is difficult. The washout separates treatment periods, while blinding helps reduce the effects of knowing which intervention is being received.

The design still has challenges. Period effects, changes over time and any lingering influence from earlier treatment need consideration. A small study also has limited precision. Those limitations do not justify treating its negative primary findings as positive, but they help explain why researchers continued investigating.

The clear factual result is that elamipretide was not superior to placebo on the main walking and fatigue endpoints during the randomized comparison. That result belongs in any account of the program.

What changed in the extension?

Ten participants continued into the open-label extension. Everyone began elamipretide, and treatment was known. Eight participants reached the 36-week point described in the first publication and the 168-week visit reported in the later extension paper. Original report, 168-week extension

Longer follow-up provides information unavailable from twelve-week treatment periods. It can show whether changes continue, whether treatment remains tolerable and what happens to patients who stay on therapy. These are important questions in a chronic disease.

At the same time, the comparison changes. The extension primarily follows participants against their earlier measurements rather than comparing them with a concurrent placebo group. If function improves, the observed change includes any treatment effect together with other changes occurring during follow-up.

This is the central tradeoff: more time and repeated observation, but less protection against alternative explanations for the improvement. Calling the extension “longer” doesn't make it automatically more decisive about causation than the blinded phase.

What did the long-term report find?

The 2024 extension report described an average 96.1-metre increase in six-minute walking distance at week 168 relative to the last predose measurement in the original double-blind treatment period, with p=0.003. Fatigue scores were below baseline at the reported extension visits. The study also evaluated strength, cardiac measurements and biomarkers, and injection-site reactions were the most common adverse events. 168-week TAZPOWER report

The walking improvement is an observed within-group change. It is not a 96.1-metre advantage over a placebo group followed for the same period. Those statements would require different data.

The p-value doesn't solve the missing-comparator problem. It concerns the statistical comparison performed under the study's assumptions. A small p-value for change from baseline cannot tell us how much of that change would have happened without treatment.

The findings can therefore be encouraging without being a definitive estimate of treatment effect. They support continued investigation and provide information about long-term follow-up in this small cohort. Their limitations should travel with the numbers whenever the results are discussed.

Why the denominator matters over time

Twelve people entered the randomized trial, ten entered the extension and eight remained through the later reported visit. That progression matters because later measurements describe a smaller selected group.

The reasons are reported here: two participants withdrew by week 36 because of injection-site urticaria, or urticaria with a drug eruption. Another stopped treatment at week 72 because he did not want to continue daily injections, but remained in follow-up and contributed assessments through week 168. Thus, eight people completing follow-up does not mean eight people continuously taking treatment. FDA integrated review

The extension paper reports that excluding the participant who stopped treatment did not materially change the mean efficacy findings in a post-hoc analysis. That sensitivity check is useful, but it does not restore a placebo comparison or make the remaining group representative of all twelve original participants. Extension report

This is an attrition question, not an accusation. In a very small study, even a few missing participants can materially affect the picture. A transparent summary should state the number assessed at each important time point rather than repeat the enrollment number beside every later result.

FDA's snapshot makes the small cohort and its limited demographic diversity explicit. All participants were male; the data were too limited for informative efficacy subgroup analyses. FDA trial snapshot

Could a treatment need longer than twelve weeks?

Yes, that is a possible explanation for different short- and long-term findings. Some biological or functional responses may develop gradually. But the extension cannot by itself prove that duration explains the difference, because duration and study design changed together.

Other possibilities include ordinary variation, repeated-test familiarity, changes in care or activity, and the characteristics of those who remained in follow-up. These are potential influences, not established explanations for this particular cohort.

A longer controlled study is useful precisely because it can test the delayed-benefit hypothesis while retaining a comparison group. If a benefit emerges after more time under those conditions, confidence in the duration explanation increases. Without that comparison, it remains one interpretation among several.

What did FDA approve, and on which endpoint?

The current Forzinity label specifies improvement in muscle strength in adults and children with Barth syndrome weighing at least 30 kg. It describes accelerated approval based on improvement in knee-extensor muscle strength, an intermediate clinical endpoint, with continued approval potentially contingent on confirming clinical benefit. Forzinity prescribing information

An intermediate clinical endpoint is a clinical measurement used to predict a further benefit that still needs confirmation. Here, the strength measurement is not merely a blood biomarker, but the regulatory pathway still requires verification of the anticipated benefit.

The FDA snapshot reports that strength increases were observed during the extension rather than the randomized phase. Its descriptive table gives a median change of 63 newtons at week 168 among eight participants. That is a median strength change, not the mean walking-distance change reported in the extension paper. FDA trial snapshot

Keeping those measurements separate prevents a common numerical error: combining different endpoints, baselines or summary statistics into one larger-looking result.

Why the baseline and summary statistic matter

For the reported 96.1-metre walking change, the reference was the last predose value in the original double-blind period, before the participants had been exposed to elamipretide. It was not their measurement when they entered the extension. Participants had already passed through the blinded treatment periods by then, so substituting the extension-entry visit would describe a different change. The paper’s baseline definition is essential to interpreting the number. Extension methods and Figure 1

The mean and median also answer slightly different descriptive questions. The mean averages all observed changes; the median identifies the middle of their ordered distribution. In a small group, a large response in one participant can influence the mean substantially. Neither statistic shows that every participant improved by the reported amount.

For TAZPOWER, this means preserving the paper's walking summary and FDA's strength summary separately. Converting one into a percentage does not make it comparable with the other. The endpoints still use different units and describe different abilities.

What does accelerated approval leave to establish?

Accelerated approval is intended for serious conditions where earlier access can be justified by a suitable predictive endpoint, with further study to verify clinical benefit. It is neither a declaration that all uncertainty has vanished nor a statement that the drug has no evidence behind it. FDA explanation of expedited programs

For Forzinity, FDA's approval letter requires a randomized, double-blind, placebo-controlled trial in patients aged five years and older with Barth syndrome to verify the benefit predicted by the strength finding. The letter gives a development timetable, but a timetable is not a completed result. FDA approval letter

This is the appropriate way to resolve the apparent contradiction. The initial primary endpoints were negative; the later observations supported a specific accelerated decision; confirmatory research is needed. None of those statements needs to be suppressed to make the story coherent.

How tolerability fits the longer picture

Long-term studies also help reveal the burden of repeated treatment. A medicine may be considered generally tolerable in a report while still causing frequent symptoms that matter to patients. The label identifies injection-site reactions and potentially serious hypersensitivity, so those should be part of any treatment conversation. Forzinity prescribing information

The extension's safety observations concern a small cohort, with treatment discontinuation and study retention needing separate interpretation. They cannot exclude uncommon harms or provide an exact prediction for every new patient. At the same time, their extended observation is useful information that a brief trial cannot supply.

A practical assessment considers both sides: what function is improving and what treatment is like to live with. A repeated symptom record and standardized clinical measurements can help describe that experience, while the broader evidence determines how confidently improvement can be attributed to treatment.

What this means in a clinical conversation

For someone with Barth syndrome, the most useful questions concern the approved goal, the expected time course and how the care team will judge response. Ask which outcome will be followed—strength, walking, fatigue or another disease-related measure—and what would count as a worthwhile change.

It is also reasonable to ask how the team interprets the blinded and extension findings together, and what confirmatory evidence is still awaited. A clear answer can acknowledge a promising long-term signal while preserving the limits of an uncontrolled comparison.

For someone without Barth syndrome, these results do not establish that longer use will eventually improve general fatigue or healthy aging. That would be a different population and a different clinical question. The SS-31 guide explains the other trials so those distinctions remain visible.

Sources

• Original TAZPOWER report, online 2020; journal issue 2021: randomized crossover and early extension.
• 168-week extension report, 2024: long-term within-group observations, original predose baseline and follow-up after treatment discontinuation.
• FDA integrated review, 2025: participant disposition and reasons for stopping treatment.
• FDA Forzinity trial snapshot: participant counts, strength findings and limitations.
• Forzinity prescribing information: indication, intermediate endpoint and safety.
• FDA approval letter, 2025: required confirmatory study.
• FDA expedited-program explanation: accelerated-approval framework.