CoQ10: Lyons et al., 2026
A Cochrane meta-analysis of eight randomized trials testing CoQ10 (alongside idebenone, omaveloxolone, and other compounds) in 574 people with Friedreich ataxia found that pharmacological treatments probably make little or no difference to ataxia severity after 12 months, though upper limb dexterity may improve slightly, a finding the authors call "unexpected" given the lack of effect on the primary outcome.
Key takeaways
- Meta-analysis of seven trials showed no meaningful change in ataxia rating scale scores after 12 months (SMD 0.02, 95% CI -0.23 to 0.26; 513 participants).
- Upper limb dexterity showed probable improvement (SMD -0.42, 95% CI -0.73 to -0.11; 166 participants), but the clinical significance is unclear given the lack of overall benefit.
- Evidence quality ranged from very low to moderate, the review was downgraded for small sample sizes, inconsistency across studies, and suspected publication bias.
- This is a Cochrane systematic review, the gold standard for synthesizing trial evidence, but the included trials vary widely in intervention, dose, and outcome measures.
- Four of eight studies were industry-controlled, and studies with null results are less likely to publish full data, both are red flags for reliability.
The study
Lyons and colleagues published a Cochrane Database Systematic Review in 2025 (PubMed), the third update of a review first published in 2009. They searched CENTRAL, MEDLINE, Embase, CINAHL, and trial registries through February 2025 for randomized controlled trials testing pharmacological treatments in people with genetically confirmed Friedreich ataxia (FRDA). Inclusion criteria: RCTs or quasi-RCTs lasting at least 12 months. Eight trials met criteria, seven contributed to meta-analysis (574 participants total; sample sizes 29-232 per study). Mean participant age ranged from 18-35 years (pooled mean 25.9 years); 21-58% female across studies. Sixty-eight percent had severe ataxia, 32% moderate. Interventions tested: epoetin alpha, CoQ10 plus vitamin E, idebenone, leriglitazone, omaveloxolone, RT001. One trial tested pioglitazone but published no results. Four studies were pharmaceutical-industry-controlled. Outcomes assessed at 12 months: ataxia rating scale score, interventricular septal thickness in diastole (IVSTd), activities of daily living (ADL), upper limb dexterity, cardiopulmonary exercise testing (CPET), and adverse events.
How to read this study
What this review does well
This is a Cochrane systematic review, the most rigorous synthesis methodology in evidence-based medicine. The authors searched multiple databases, included only randomized trials lasting at least 12 months (long enough to detect disease-modifying effects in a slowly progressive condition), and used GRADE to transparently assess certainty of evidence. They assessed risk of bias with RoB 2, reported confidence intervals alongside all point estimates, and acknowledged inconsistency and publication bias by downgrading evidence quality. The 12-month minimum is critical: shorter trials in FRDA often show transient biomarker changes that don't translate to clinical benefit.
What this review is missing or weak on
The pooled interventions are pharmacologically unrelated, CoQ10, idebenone (a synthetic CoQ10 analog), omaveloxolone (an Nrf2 activator), leriglitazone (a PPAR-gamma agonist), epoetin alpha. Lumping them together assumes a shared mechanism, which is speculative. The authors report moderate heterogeneity (I² = 42%) for the primary outcome and high heterogeneity (I² = 80%) for IVSTd, meaning the studies don't agree. Four of eight trials were industry-controlled, and one trial (pioglitazone) published no results, both are classic markers of publication bias. The authors suspect this bias and downgrade accordingly, but we don't know what unpublished negative data exist. Finally, only 104 of 545 participants contributed to the adverse event analysis, so the safety profile is underreported.
How I'd weight this review
I weight Cochrane reviews heavily when they synthesize homogeneous interventions in well-defined populations. This one doesn't meet that standard, it's a meta-analysis of dissimilar compounds tested in overlapping but not identical patient populations. The null result on ataxia severity is probably real, but the "unexpected" improvement in upper limb dexterity without overall benefit is a red flag for measurement noise or selective outcome reporting. I treat this as strong evidence that no single compound tested here, including CoQ10, produces clinically meaningful benefit in FRDA at 12 months, but not as evidence against mechanistic hypotheses in healthier populations.
What they found
Pooling seven trials (513 participants) showed no meaningful difference in ataxia rating scale scores after 12 months of treatment (SMD 0.02, 95% CI -0.23 to 0.26; I² = 42%; moderate-certainty evidence). The confidence interval crosses zero and is narrow enough to exclude a clinically important effect. Evidence for interventricular septal thickness (IVSTd) was very uncertain (MD -0.51 mm, 95% CI -1.10 to 0.09; I² = 80%; 2 studies, 72 participants), as was evidence for activities of daily living (MD -0.59, 95% CI -1.39 to 0.21; I² = 24%; 3 studies, 167 participants). Three trials testing upper limb dexterity (166 participants) found probable improvement (SMD -0.42, 95% CI -0.73 to -0.11; I² = 0%; moderate-certainty evidence). The authors note this result is "unexpected" given the lack of effect on overall ataxia severity. Cardiopulmonary exercise testing showed very uncertain effects (SMD -0.16, 95% CI -0.46 to 0.13; I² = 0%; 3 studies, 181 participants; very low-certainty evidence). Treatment-related adverse events (2 studies, 104 participants) showed no clear difference (RR 0.88, 95% CI 0.63 to 1.22; very low-certainty evidence), and treatment-emergent adverse events leading to cessation or death (6 studies, 313 participants) may be no more common than placebo (RR 1.24, 95% CI 0.44 to 3.48; low-certainty evidence).
What it means for the average man
This review synthesizes evidence for CoQ10 (and related compounds) in a rare mitochondrial disease characterized by progressive neurological decline and early death from cardiomyopathy. The findings do not support pharmacological intervention, including CoQ10 supplementation, as disease-modifying in FRDA after 12 months of treatment. The upper limb dexterity finding is intriguing but isolated and inconsistent with the lack of overall benefit, which limits confidence in its clinical relevance. For men without FRDA using CoQ10 for cardiovascular support or energy metabolism, this review does not apply directly, FRDA is a specific genetic disorder with severe mitochondrial dysfunction, and the disease context differs meaningfully from age-related mitochondrial decline or exercise performance. The review does reinforce that CoQ10 interventions require long durations (12+ months) and objective outcome measures to assess efficacy, not surrogate biomarkers or short-term subjective reports.
The caveats
The interventions pooled in this meta-analysis are pharmacologically dissimilar, which limits the interpretability of the pooled estimate. Moderate to high heterogeneity across trials suggests the studies do not agree, yet the authors synthesized them anyway. Four of eight trials were industry-controlled, and one trial published no results, both are known sources of positive-result bias. The authors downgraded evidence quality for imprecision, inconsistency, and suspected publication bias, which is appropriate. The upper limb dexterity finding contradicts the null result on overall ataxia severity, and the authors themselves call it "unexpected," which raises the possibility of measurement artifact or selective reporting. The safety analysis included only 104 of 545 participants, so the adverse event profile is incomplete. Finally, this is a population with severe mitochondrial dysfunction from a known genetic defect, not a model for general CoQ10 use in healthy or aging populations.
Frequently asked questions
Is Cochrane the same as a regular meta-analysis?
Cochrane reviews are the gold standard for systematic reviews and meta-analyses. They follow a pre-registered protocol, search exhaustively across multiple databases and trial registries, assess risk of bias systematically, and use GRADE to rate certainty of evidence. Most meta-analyses don't follow this rigorous methodology. A Cochrane review is not automatically correct, but it is methodologically transparent and more likely to capture unpublished or negative trials that other reviews miss.
Why did upper limb dexterity improve if overall ataxia didn't?
The authors themselves call this result "unexpected." One possibility is measurement noise, dexterity tests are sensitive to practice effects and test-retest variability. Another is selective outcome reporting: if studies tested many outcomes and only reported the ones that showed benefit, you can get isolated positive findings by chance. A third possibility is that the treatment genuinely improved one motor subdomain without affecting overall disease progression, but that would be unusual. The confidence interval is narrow (SMD -0.42, 95% CI -0.73 to -0.11), which suggests the finding is statistically robust, but the clinical significance is unclear without knowing if participants noticed any functional improvement.
Does this review mean CoQ10 doesn't work for cardiovascular health?
No. This review tested CoQ10 and related compounds in Friedreich ataxia, a rare genetic disease with severe mitochondrial dysfunction, progressive neurological decline, and early-onset cardiomyopathy. The population, disease context, and endpoints (ataxia severity, interventricular septal thickness) are not generalizable to healthy men or age-related cardiovascular decline. The review does not address CoQ10's effects on blood pressure, endothelial function, or exercise capacity in non-FRDA populations, where separate trial evidence exists.
What does "industry-controlled" mean for bias?
Industry-controlled means the pharmaceutical company funding the trial also designed the protocol, controlled the data, and decided what to publish. Studies with industry control show positive results 4-5 times more often than independent studies, not because the data are fabricated, but because of subtle design choices (favorable comparators, short durations, surrogate endpoints) and selective publication of positive findings. Four of eight trials in this review were industry-controlled, and one trial published no results at all, both are red flags for publication bias. The authors acknowledged this and downgraded evidence quality accordingly.
Sources
- Lyons S, Kearney M, Fahey M, Janjal P, Pandolfo M, Patton P. Pharmacological treatments for Friedreich ataxia. Cochrane Database Syst Rev. 2026. PubMed.
- Higgins JPT, Thomas J, Chandler J, et al. Cochrane Handbook for Systematic Reviews of Interventions. 2nd ed. Chichester (UK): John Wiley & Sons; 2019.
These statements have not been evaluated by the Food and Drug Administration. This article is for educational purposes only and is not intended as medical advice. Consult a physician before starting any supplement or exercise program.
