In a controlled trial of competitive distance runners, four weeks of 1 g/day urolithin A didn’t change 3000 m performance (in a tested subset) or VO2max, but effort felt easier and muscle-damage markers were lower. Across other human studies, mitochondrial-related signatures shift repeatedly. How can upstream biology move while race outcomes stand still?
Mechanistic insight
Oral urolithin A consistently modulates molecular signatures linked to mitochondrial quality control and metabolism in humans, and in several trials lowers systemic proxies such as plasma acylcarnitines and inflammatory/damage markers. These signals are compatible with enhanced ‘maintenance and recovery’ capacity. However, direct, flux-sensitive increases in mitophagy in human muscle and consistent gains in muscle mitochondrial function have not been demonstrated. Because endurance and strength are emergent outcomes shaped by multiple factors (including cardiovascular delivery, neuromuscular function, substrate handling, and recovery) changing an upstream mitochondrial process does not necessarily change performance. One possible explanation is that mitochondrial modulation matters most when mitochondrial quality control or recovery is functionally limiting, and when exposure lasts long enough for remodeling. Current human trials, however, have not tested this bottleneck hypothesis directly. Other factors may therefore explain why perception and molecular signals can change without measurable gains in VO2max or race performance.
When effort eases but the clock won’t budge
Competitive male distance runners supplemented with 1 g/day urolithin A during a 4-week altitude camp showed no improvement versus placebo in VO2max, and 3000 m performance (assessed in a subset of participants (approximately 11 per group)) also showed no between-group improvement. Yet they reported lower perceived exertion during the time trial and had lower post-exercise creatine kinase area-under-the-curve. Skeletal-muscle proteomics indicated upregulation of mitochondria-related pathways and downregulation of inflammatory pathways, while direct measures of muscle mitochondrial function did not change. This dissociation (engaged biology and easier-feeling effort without faster performance) raises a precise question: if mitochondria matter, why didn’t the finish time move? [Whitfield 2025]
Why we expected urolithin A to help
Mitochondria sit at the heart of muscle energetics, and mitophagy is credible cellular housekeeping. Early human work established safety and showed that 4 weeks of 500–1000 mg/day urolithin A alters skeletal-muscle mitochondrial gene expression and lowers plasma acylcarnitines, signals compatible with changes in substrate handling and quality control. In middle-aged adults over 4 months, urolithin A increased proteins linked to mitophagy and mitochondrial metabolism and improved some strength and endurance indicators, though the prespecified primary peak-power endpoint did not change. In older adults over 4 months, primary endpoints (6-minute walk distance and muscle ATP production by MRS) were not different from placebo, despite interim gains in muscle endurance tests and favorable biomarker shifts. Put together, there is real human pathway engagement and enough positive signals to make a performance benefit feel imminent. [Andreux 2019; Singh 2022; Liu 2022]
Performance is an emergent outcome, not a single switch
Race time, VO2max, 6-minute walk distance, maximal voluntary torque, or repetitions-to-failure emerge from many coordinated systems. Cardiac output and oxygen delivery, neuromuscular recruitment, local mitochondrial capacity and quality control, fuel selection, and recovery all contribute. Changing one upstream process therefore doesn’t guarantee movement in the whole. This is why lower perceived exertion or lower CK can coexist with an unchanged time trial: perception and damage are part of the physiological picture, but the available trials do not establish which factors were limiting performance.
What the human biology actually shows
Quality-control signaling: Multiple trials show increased expression of proteins and genes linked to mitophagy and mitochondrial metabolism in human muscle after urolithin A. These signatures indicate engagement but are not flux-sensitive; a demonstrated increase in mitophagy flux in human skeletal muscle remains unproven. Substrate proxies: Plasma acylcarnitines tend to decrease with urolithin A, consistent with altered fatty-acid handling; this is a systemic proxy, not a direct measure of tissue performance. Recovery milieu and perception: C-reactive protein reductions have been reported, and in trained runners CK exposure and perceived exertion fell. These shifts can make work feel easier or recovery smoother without obligating a change in VO2max or race results. The mechanistic texture fits a picture of complementary modulation, biologically meaningful, yet not automatically performance-defining. [Andreux 2019; Singh 2022; Whitfield 2025; Liu 2022]
Context, ceilings, and time
Across RCTs, functional outcomes are mixed. In middle-aged adults, muscle strength improved by about 12% and some endurance indicators rose, while the primary peak-power endpoint did not. In older adults, prespecified primary endpoints were unchanged, despite interim endurance gains and favorable biomarkers. In trained runners, performance did not improve over 4 weeks, despite easier effort and lower CK. Small athlete pilots report gains in selected field tests, higher maximal voluntary isometric contraction and more repetitions-to-failure in resistance-trained men after 8 weeks, and greater Yo-Yo intermittent recovery distance and jump height in academy soccer players after 6 weeks in a single-blind randomized pilot, findings that warrant confirmation. A meta-analysis pooling 6-minute walk distance found a directionally positive but statistically inconclusive mean difference (~+17 m; low certainty). One hypothesis that could help explain this heterogeneous landscape is that mitochondrial quality control or recovery matters more for function in some physiological contexts than in others. Ceiling effects in highly trained athletes, baseline mitochondrial impairment, training stress, duration of exposure, and other performance-limiting systems are all plausible contributors, but current human trials have not established these factors as moderators of response. Exposure variability (formulation, bioavailability, microbiome phenotype) may also matter, but current trials did not stratify outcomes by exposure. Short-term safety appears favorable at 500–1000 mg/day for 4 weeks and 1 g/day up to 4 months; long-term safety and special-population data are limited. [Singh 2022; Liu 2022; Whitfield 2025; Zhao 2024; Monsalve Acevedo 2025; Dao 2026; Andreux 2019]
One hypothesis that fits the pieces
Urolithin A can move upstream signals (mitochondrial-quality-control signatures, substrate proxies, inflammatory and damage markers) and even how effort feels without necessarily moving performance. One possible explanation is that functional benefits emerge when mitochondrial quality control or recovery is limiting and sufficient time is available for remodeling. Current human trials have not established when (or whether) this bottleneck condition applies. The important distinction is therefore not simply whether the pathway responds, but whether that response becomes large and relevant enough to alter the whole-system outcome.
What this means for interpretation and decisions
What was studied: Most human trials used about 1 g/day for 6–16 weeks (functional endpoints) or 500–1000 mg/day for 4 weeks (biomarkers/safety). What was observed: Reproducible pathway/biomarker engagement; mixed functional outcomes with some secondary gains in non-elite cohorts; a null performance result in trained runners over 4 weeks alongside lower RPE and CK; small athlete pilots with positive signals that require replication. What may be hypothesized: Functional benefits could be more likely in contexts where mitochondrial quality control or recovery meaningfully constrains function, and exposure variability may influence response. Neither proposition has yet been established as a human moderator effect. What is not established: A consistent improvement in prespecified primary functional endpoints across populations; mitophagy flux in human muscle; food-to-supplement exposure equivalence; long-term safety; who benefits most. Translation today: Promising but not prescriptive.
Back to the runners
Seen through a bottleneck-and-timescale lens, the runners’ trial becomes less puzzling. Over four weeks, in already well-adapted athletes, mitochondrial-related signals shifted and effort felt easier without improved performance. One possibility is that the processes affected by urolithin A were not limiting performance in this setting; another is that four weeks was insufficient to reveal functional remodeling. Other explanations are also possible, and the trial cannot distinguish among them. What it does show is the boundary between engaging a biological pathway and changing an emergent human outcome.
Mechanistic pathways can respond without a corresponding change in performance.
Whether performance changes depends on more than pathway engagement; the relevant limiting factors in humans remain to be established.
Treat signatures, proxies, and perception as informative, but not as proof of performance change.
Curious to discover more?
Explore more stories about biology, nutrition and health.
Discover more articlesReferences
- Whitfield J et al. Evaluating the Impact of Urolithin A Supplementation on Running Performance, Recovery, and Mitochondrial Biomarkers in Highly Trained Male Distance Runners. Sports medicine (Auckland, N.Z.). 2025. PMID: 40839339
- Singh A et al. Urolithin A improves muscle strength, exercise performance, and biomarkers of mitochondrial health in a randomized trial in middle-aged adults. Cell reports. Medicine. 2022. PMID: 35584623
- Liu S et al. Effect of Urolithin A Supplementation on Muscle Endurance and Mitochondrial Health in Older Adults: A Randomized Clinical Trial. JAMA network open. 2022. PMID: 35050355
- Andreux PA et al. The mitophagy activator urolithin A is safe and induces a molecular signature of improved mitochondrial and cellular health in humans. Nature metabolism. 2019. PMID: 32694802
- Zhao H et al. Assessment of Urolithin A effects on muscle endurance, strength, inflammation, oxidative stress, and protein metabolism in male athletes with resistance training: an 8-week randomized, double-blind, placebo-controlled study. Journal of the International Society of Sports Nutrition. 2024. PMID: 39487653
- Monsalve Acevedo A et al. Effects of Urolithin A supplementation on performance and antioxidant status in academy soccer players during preseason: a pilot randomised controlled trial. Frontiers in nutrition. 2025. PMID: 41245402
- Dao T et al. Effects of Urolithin A supplementation on muscle health outcomes in humans from randomized controlled trials. Frontiers in nutrition. 2026. PMID: 42559151