MMPOWER-2 reported different patterns across walking performance and patient-reported symptoms. A summary should not let a favorable secondary outcome replace the primary result. Primary report
Trial findings
Thirty adults with genetically confirmed primary mitochondrial myopathy entered a randomized, double-blind crossover trial. Four-week treatment periods were separated by a washout. The primary six-minute-walk difference was 19.8 meters, with a 95% confidence interval from −2.8 to 42.5 meters, and did not reach statistical significance. Several fatigue-related patient-reported outcomes favored elamipretide. Injection-site reactions were common. The report discloses investigator relationships with Stealth BioTherapeutics. Primary report
Our appraisal: show the interval, not just the point estimate
The point estimate describes the center of the reported estimate, not the only effect compatible with the analysis. An interval spanning zero should not be presented as a precise established benefit. It also does not demonstrate exact equivalence.
Some patient-reported outcomes favored elamipretide. However, the prespecified testing hierarchy required a significant primary result before confirmatory testing of selected secondary outcomes. That requirement was not met, and the paper calls the secondary analyses exploratory. Describing the fatigue observations as proven benefit would remove the statistical condition that governed them. MMPOWER-2 statistical analysis and discussion
What the crossover adds to the review
Our checklist includes sequence, period effects, possible carryover, completion of both periods, and the method used for within-participant comparisons. A washout period is a design feature to assess, not proof that carryover is impossible.
MMPOWER-2 should also retain its place in a development program. Its findings can motivate a later trial without predicting that later trial's result. The separate MMPOWER-3 page should be consulted when examining subsequent evidence, rather than treating this earlier signal as the final answer.
Frequently asked questions
Did the primary walking endpoint reach statistical significance?
No. Its confidence interval included zero.
Does that mean the estimated effect was exactly zero?
No. The interval reflects uncertainty around the estimate.
Were fatigue findings the same endpoint as walking distance?
No. They came from different measurements.
Were adverse events absent?
No. Injection-site reactions were commonly reported.
Does crossover mean two independent participant groups?
No. Participants receive conditions in a sequence; the dependence must be handled analytically.
Does an earlier signal guarantee phase 3 success?
No. Later studies provide their own evidence.
What source was inspected?
The primary indexed abstract and disclosures. Complete methods and publication-history review remain release requirements.