Abstract

Background: Coenzyme Q10 has been studied in chronic heart failure for more than three decades. Successive meta-analyses report reductions in all-cause mortality, yet the compound appears in no major heart failure guideline. The size of the reported effect relative to established therapy, and the reasons for its exclusion, have not been examined together. Objectives: To pool published review-level estimates of the effect of coenzyme Q10 on all-cause mortality and heart failure hospitalisation; to set the pooled effect against effects commonly cited for guideline-mandated therapy; and to examine whether the internal structure of the evidence supports the magnitude reported. Methods: Systematic reviews with meta-analysis of randomised trials of coenzyme Q10 in chronic heart failure reporting all-cause mortality were eligible. Review-level estimates were pooled on the log scale in a DerSimonian-Laird random-effects model with Cochran’s Q, I², ² and 95% prediction intervals. Absolute effects were derived from reported event counts. Certainty was recorded as graded by the source reviews. Analyses were performed in Python 3. Results: Three syntheses contributing mortality estimates and representing 63 trials in more than 5,900 participants were pooled, giving a risk ratio of 0.65 (95% CI 0.54-0.79; z = –4.36, p < 0.001) with no detectable heterogeneity (Q = 0.19, df = 2, p = 0.910; I² = 0.0%). The 95% prediction interval was 0.43-0.99 and only just excluded the null. Heart failure hospitalisation was reduced to a similar degree (0.44, 95% CI 0.35-0.56). Derived from reported event counts the absolute risk difference was 2.9% points, a number needed to treat of 34. The single adequately powered mortality trial randomised 420 participants and reported a hazard ratio of 0.513 (0.30-0.89). Against this, the effect on left ventricular ejection fraction—the presumed mechanistic intermediate—was small (mean difference 0.51, 95% CI 0.31-0.71) and graded low certainty. Conclusions: The pooled mortality estimate for coenzyme Q10 in heart failure exceeds the effects commonly cited for betablockers, renin-angiotensin inhibitors, mineralocorticoid antagonists and SGLT2 inhibitors. That is not a reason to adopt it. An effect this large from trials this small, with a prediction interval touching the null, a mechanistic intermediate that barely moves, and no adequately powered confirmatory trial in thirty years, is exactly the profile that should prompt scepticism rather than adoption. The finding of this review is that the question remains unresolved, is resolvable by a single adequately powered trial, and that nobody has an obvious commercial incentive to run one.

Keywords: Coenzyme Q10, Chronic heart failure, All-cause mortality, Publication bias, Evidence thresholds, Nutraceuticals