Two decades of mitochondrial disease trials: what failed, what worked, and what's next

Twenty-five years of clinical trials for primary mitochondrial disease — a family of ultra-rare, genetically diverse disorders of cellular energy — have produced exactly two approved therapies. A new scoping review and horizon scan, led by Amel Karaa of Massachusetts General Hospital and Harvard Medical School and published in Therapeutic Advances in Rare Disease, catalogues all 71 human interventional trials from 2000 to November 2025 and asks why most failed, and what changed for the ones that worked.

Why so many trials failed

Most of the 71 trials were small, and open-label designs (37) outnumbered randomized, double-blind, placebo-controlled trials (28). The authors argue many early broad antioxidant and metabolic therapies failed because the drugs didn't directly correct the underlying defect — or because trial designs didn't fit the biology. The 218-patient MMPOWER-3 trial of elamipretide in primary mitochondrial myopathy missed both co-primary endpoints (a walking test and a fatigue score); the authors say enrolling mixed genotypes likely diluted any signal. At least 10 trials were terminated and 3 withdrawn.

What worked

Success came from precision: genetically defined diseases with homogeneous features. Doxecitine/doxribtimine (Kygevvi) won full FDA approval for thymidine kinase 2 deficiency with symptom onset at age 12 or younger, on a matched external-control comparison of 78 treated and 78 untreated patients showing 4% versus 36% mortality. Elamipretide (Forzinity) received accelerated FDA approval for Barth syndrome in September 2025 based on improved knee-extensor strength — with a randomized confirmatory trial required. Idebenone has been approved in Europe for LHON since 2015, though the FDA asked for more controlled data.

What's next — and the watchdog notes

The authors propose matching trial populations to molecular biology, building natural-history registries, validating biomarkers (GDF-15 and FGF-21 don't yet qualify as regulatory surrogates), using adaptive and Bayesian designs, and publishing negative results. They also flag concentration of trials in North America and Western Europe — with little activity in Latin America, Africa, or much of the Middle East — and acknowledge publication bias and English-only sources. Their verdict: the field is at a pivotal moment that depends on sustained investment in natural-history data, better endpoints, and equitable global access.

Where the evidence stands

• Established: Kygevvi's full FDA approval for TK2 deficiency and Forzinity's accelerated approval for Barth syndrome are regulatory facts; both used tailored evidence packages rather than large randomized trials.
• Preliminary: The review is a narrative synthesis of heterogeneous trials without formal risk-of-bias grading or meta-analysis; its lessons are interpretive, not statistical.
• Contested: Forzinity's accelerated approval rests on a surrogate endpoint (knee-extensor strength); the required confirmatory trial will determine whether the benefit holds.