Substance and claimed effects

Urolithin A (UA) is a small molecule produced in the colon when gut bacteria — chiefly Gordonibacter and Ellagibacter species in the Coriobacteriaceae family — break down ellagitannins from pomegranate, walnuts, strawberries, raspberries, and some other berries Tomás-Barberán et al. 2017. Ellagitannins themselves are barely absorbed; UA is the absorbable downstream metabolite that the body actually sees. Only some people make it: cohorts cluster into three urolithin metabotypes — metabotype A (produces UA), metabotype B (produces UA plus urolithin B / isourolithin A), and metabotype 0 (produces essentially nothing). In Western populations the non-producer fraction is variously estimated at 10–60%, with the higher figures in the US and UK and the lower in Mediterranean diets Tomás-Barberán et al. 2017D'Amico et al. 2021. The commercial answer to that variability is direct UA supplementation — Mitopure, a synthetic UA produced by Amazentis/Timeline that achieves plasma levels equivalent to consuming six glasses of pomegranate juice, in producers and non-producers alike D'Amico et al. 2021; it carries FDA GRAS status (Notice 791, 2018) FDA 2018.

The substance is sold for, and studied as, a mitophagy activator: a compound that triggers the cell's selective recycling of damaged mitochondria, the energy-producing organelles whose decline tracks closely with age-related loss of muscle strength, endurance, immune function, and cardiac performance. Claimed reader-facing consequences span (i) muscle strength and exercise endurance in middle-aged and older adults, (ii) mitochondrial / cellular health biomarkers (acylcarnitines, ceramides, mitophagy gene expression), (iii) age-related immune decline (T-cell exhaustion), (iv) cardiovascular biomarkers (plasma ceramides), and — via topical formulations — (v) intrinsic skin aging and UVB photodamage. The evidence base is unusually concentrated in trials funded or coauthored by the manufacturer; this dossier reads them on their merits while naming that conflict throughout.

Evidence by addressing question

Mechanism

Mitochondria continuously divide, fuse, and get recycled. The recycling step — mitophagy — selectively degrades damaged or depolarised mitochondria via the autophagy machinery, preventing accumulation of dysfunctional organelles that leak reactive oxygen species and impair ATP production. Mitophagy flux declines with age in muscle, brain, immune cells, and heart, and that decline is a leading candidate explanation for several geroscience phenotypes (sarcopenia, immunosenescence, neurodegeneration, cardiac aging).

The founding mechanistic paper screened ~6,000 natural compounds for mitophagy induction in C. elegans and identified urolithin A as the first natural-product mitophagy inducer with confirmed in vivo activity Ryu et al. 2016. UA extended C. elegans lifespan ~45%, preserved mitochondrial morphology with age, and increased running endurance in young rats (~40%) and grip strength in aged mice — effects abolished by genetic knockout of autophagy genes pink-1, dct-1, and bec-1, supporting mitophagy as the mediating pathway. In humans, a 4-week trial at 500 mg and 1,000 mg in sedentary elderly modulated skeletal-muscle expression of mitophagy and mitochondrial-biogenesis genes and changed plasma acylcarnitines in a pattern consistent with improved fatty acid oxidation Andreux et al. 2019; the 4-month trials in older and middle-aged adults replicated the acylcarnitine signal and added a CRP drop Liu et al. 2022Singh et al. 2022. The recent immune trial extended the mechanism into CD8+ T cells, showing increased fatty-acid-oxidation capacity (treatment difference +14.7 percentage points, P=0.006) — a metabolic shift that the literature ties to less-exhausted T-cell phenotypes Denk et al. 2025.

Pharmacokinetics: UA is rapidly conjugated in the liver to glucuronide and sulfate forms; T_max ≈ 6 h after oral dosing, plasma half-life of UA and UA-glucuronide is ~17–22 h and UA-sulfate ~25–58 h, supporting once-daily dosing Andreux et al. 2019. Plasma exposure rises dose-dependently across 250–2,000 mg.

Evidence — human trials

Five randomised, placebo-controlled trials anchor the human evidence base; all but one are sponsor-coauthored, and the picture is mixed when read against pre-specified primary endpoints.

• Andreux et al. 2019 (Nature Metabolism): first-in-human, n=60 sedentary elderly, doses 250 / 500 / 1,000 / 2,000 mg and a separate 4-week 500 / 1,000 mg arm. Primary outcome was safety — met (no serious adverse events). Secondary biomarker findings: dose-dependent changes in plasma acylcarnitines (consistent with improved fatty-acid oxidation) and increased muscle expression of mitophagy and OXPHOS genes at 500–1,000 mg Andreux et al. 2019.
• Liu et al. 2022 (JAMA Network Open): n=66, ages 65–90, 1,000 mg/day vs placebo for 4 months. Co-primary endpoints — six-minute walk distance and maximal ATP production in hand muscle — both missed. The trial was, however, positive on several secondaries: muscle endurance (contractions to fatigue) improved significantly in hand and leg at 2 months; plasma acylcarnitines, ceramides, and CRP all dropped. The sponsor-authored framing emphasises the secondaries; the registered primaries were null Liu et al. 2022.
• Singh et al. 2022 (Cell Reports Medicine): n=88, ages 40–64, overweight with low aerobic fitness; 500 mg or 1,000 mg vs placebo for 4 months. Primary endpoint — peak power output on cycle ergometer — missed. Secondaries: hamstring strength up ~12% at both doses; the 1,000 mg arm had clinically meaningful gains in peak VO₂ (~10%) and 6-minute walk distance (>30 m); plasma acylcarnitines and CRP dropped; muscle biopsies showed activation of mitophagy and OXPHOS proteins Singh et al. 2022.
• Denk et al. 2025 (Nature Aging): n=50 healthy middle-aged adults, 1,000 mg/day vs placebo for 4 weeks. Pre-specified primary on CD8+ T-cell phenotype hit: expansion of naive-like, less-exhausted CD8+ cells (treatment difference +0.50 pp, 95% CI 0.16–0.83, P=0.04) and a large rise in CD8+ fatty-acid-oxidation capacity (+14.7 pp, 95% CI 6.5–23.0, P=0.006). First trial showing a UA effect on human immune phenotype, not just biomarkers Denk et al. 2025.
• Petrocelli et al. 2025 (Sports Medicine, "Enduro Trial"): n=42 highly-trained male distance runners, 1,000 mg/day for 4 weeks. 3 km time trial — null. Within-group VO₂max rose more in UA (+5.4%) than placebo (+3.6%); UA significantly lowered rating of perceived exertion and post-trial creatine kinase (a marker of muscle damage). Honest read: a recovery / RPE signal, no time-trial benefit in already-elite athletes Petrocelli et al. 2025.

Cardiovascular biomarkers. Singh and colleagues 2025 reported preclinical recovery of cardiac systolic and diastolic function in aged mice and a heart-failure rat model after UA; in humans (a re-analysis of 4-month, 1,000 mg supplementation in older adults) UA lowered plasma ceramides, lipid species clinically validated as predictors of cardiovascular events Singh et al. 2025.

Topical / skin. Three sponsor-conducted randomised trials of 1% topical UA in post-menopausal women and middle-aged adults reported significant wrinkle reduction and ~14% reduction in UVB-induced erythema vs placebo cream; preprint, not yet peer-reviewed at publication D'Amico et al. 2023.

Systematic review honesty. Hodzic Kuerec et al. 2024 pooled five UA trials (n=250) and concluded a dose-dependent anti-inflammatory effect and upregulation of mitochondrial / autophagy / fatty-acid-oxidation markers — but explicitly noted UA "did not affect mitochondrial maximal ATP production, biogenesis, dynamics, or gut microbiota composition" and no effects on body composition or physical-function outcomes Hodzic Kuerec et al. 2024.

Protocol

Effective doses in human trials cluster at 500 mg and 1,000 mg once daily; 1,000 mg is the dose that hit on the immune trial primary, on the muscle-endurance and biomarker secondaries, and on RPE/CK in athletes; 500 mg matched 1,000 mg on hamstring strength and biomarker changes in Singh et al. 2022 Singh et al. 2022Liu et al. 2022Denk et al. 2025. Duration to detectable benefit: biomarker / gene-expression changes by 4 weeks; functional muscle outcomes by 8–16 weeks; the immune-phenotype shift by 4 weeks. Onset for skin (topical) was 8–12 weeks in the cosmetic trials D'Amico et al. 2023. Once-daily dosing is supported by the 17–58 h plasma half-lives of UA and its conjugates Andreux et al. 2019.

Food source as alternative: walnuts (≈30 g/day) and pomegranate juice (180 ml) produce measurable plasma UA in producers only; non-producers (10–60% of Western adults) generate essentially nothing regardless of intake Tomás-Barberán et al. 2017D'Amico et al. 2021. Even in producers, food-derived plasma UA peaks lower than direct 500–1,000 mg supplementation. There is no validated point-of-care metabotype test; the practical proxy is "supplement if you want a known dose."

Contraindications and safety

Safety profile is favourable across trials: no serious adverse events attributable to UA in any of the five RCTs reviewed in Hodzic Kuerec et al. 2024; mild myalgia possibly drug-related; doses up to 2,000 mg/day in Andreux et al. 2019 produced no laboratory red flags Hodzic Kuerec et al. 2024Andreux et al. 2019. FDA GRAS status (Notice 791, 2018) for Mitopure formulation FDA 2018. No published human data on use during pregnancy, breastfeeding, or in children; defaulting to "avoid in pregnancy / breastfeeding" is conventional rather than evidence-driven. No documented drug interactions in trials. Theoretical concern about long-term mitophagy induction (analogous to caloric-restriction / autophagy mimetics in patients with active malignancy) has not been studied in humans.

Audience

Trial populations skew toward middle-aged and older adults with detectable mitochondrial / functional decline (Liu et al. 65–90 with average performance; Singh et al. 40–64 overweight with low fitness; Denk et al. healthy middle-aged). The signal is most consistent where age has already lowered the baseline; the elite-athlete trial showed only RPE / CK / mitochondrial-protein benefits with no time-trial improvement, consistent with diminishing returns from mitophagy induction in a system already optimised by training Petrocelli et al. 2025. Young healthy adults below the trained-runner ceiling have not been formally tested.

Alternatives

The competing interventions for the same outcome (mitochondrial / functional reserve in aging) are far better evidenced and cheaper: resistance training (strength), aerobic exercise (mitochondrial biogenesis via PGC-1α), creatine monohydrate (muscle output, well-evidenced and ~$30/year), protein intake at 1.6 g/kg, and sleep. UA's most defensible niche is the supplement-tier additive layered on top of those, not a substitute. NMN and NAD+ precursors compete in the same supplement market with similar evidence quality and similar cost.

Failure modes

• Buying it as a substitute for training. The functional gains in trials sit at ~10–12% on muscle strength and 5–10% on endurance metrics — comparable to a single training cycle in untrained subjects. A reader who replaces resistance work with a capsule is buying nothing.
• Cost vs delivered effect. At $2.50–$4.00/day for Mitopure, this is a $900–$1,500/year supplement layered on a domain (creatine, exercise, sleep) where the high-ROI moves cost almost nothing.
• Non-Mitopure brands. Mitopure is the only formulation tested in the published trials. Third-party UA supplements vary widely in purity and bioavailability and have not been clinically validated.
• Expecting felt energy / focus changes. Trial endpoints are functional (endurance, strength), biomarker (acylcarnitines, ceramides, CRP), and cellular (gene/protein expression). Subjective energy and focus were not primary endpoints in any major trial; readers expecting a stimulant-like felt effect will be disappointed.

Practicalities

Available as softgels (500 mg = two softgels at the Mitopure dose; 1,000 mg = four softgels) or powder. Subscription pricing ~$60–$100/month depending on tier and dose; one-time purchase at the high end. NSF certified. No prescription. Bioavailability is roughly equivalent across single 500 mg, 1,000 mg, and 2 × 500 mg split, supporting once-daily intake with food Andreux et al. 2019. Generic UA is starting to appear at lower price points but lacks the human trial data — the gap is purity and pharmacokinetic validation, not patent.

Stakes

If the central mitophagy-decline thesis is right, the cumulative cost of not intervening is exactly the aging trajectory the entry treats — sarcopenia, immune drift, cardiac stiffening over decades. UA's measured effects are modest in any single trial (single-digit to low-double-digit improvements over 4–16 weeks), and the "stake" framing needs to lean on (a) the trial cluster's directional consistency on biomarkers and secondary functional endpoints, and (b) the underlying biology that age-related mitophagy decline is real and untreated by most lifestyle interventions short of fasting and intense exercise. Honest framing: not catastrophic to skip; not nothing to take.

Payoff

Onset latency by surface: biomarker changes (acylcarnitines, CRP) at 4 weeks Andreux et al. 2019; immune-cell phenotype at 4 weeks Denk et al. 2025; muscle endurance at 8 weeks Liu et al. 2022; muscle strength and VO₂max at 16 weeks Singh et al. 2022; topical skin (visible wrinkle reduction, UV resistance) at 8–12 weeks D'Amico et al. 2023. None of the felt effects are dramatic; the more honest payoff frame is "fewer reps to fatigue at month 4" and "biomarker drift that lines up with how your body is supposed to handle age" rather than a stimulant-style next-week change.

Credibility range

The optimist case

UA is the first natural-product mitophagy inducer with confirmed in vivo activity in humans, biologically plausible against the dominant geroscience target (mitochondrial decline), with directionally consistent evidence across five RCTs in three different populations (sedentary elderly, middle-aged overweight, healthy middle-aged) on biomarker, gene-expression, immune-phenotype, muscle-endurance, and skin endpoints. The 2025 immune trial is the strongest individual data point: a pre-specified primary endpoint on a human cellular phenotype that maps directly to the dominant aging immune defect (CD8+ exhaustion / loss of naive cells), hit with a clean effect size Denk et al. 2025. The trials in muscle missed their formal primaries but produced consistent secondaries with magnitudes (~12% strength, ~10% VO₂max, ~17% endurance) that are clinically meaningful — and would be celebrated as headline findings in a non-pharma context. Pharmacokinetics support a once-daily oral dose, safety is clean across 250–2,000 mg, and the FDA GRAS designation removes the regulatory ambiguity that dogs other longevity supplements. Topical and cardiovascular extensions broaden the surface from "muscle supplement" to "anti-aging postbiotic with multi-tissue effects."

The skeptic case

Two of the three pre-specified muscle-trial primary endpoints missed (Liu et al. on 6MWT and ATP production; Singh et al. on peak power) Liu et al. 2022Singh et al. 2022. The Hodzic Kuerec et al. 2024 systematic review noted UA failed to move mitochondrial maximal ATP production, mitochondrial biogenesis or dynamics, or body composition / physical-function outcomes — directly contradicting the mechanistic story sold to consumers Hodzic Kuerec et al. 2024. Almost every published trial has authors employed by or paid by Amazentis / Timeline, the manufacturer; the trial that did not have a dominant manufacturer affiliation (the Enduro Trial in distance runners) was the one with the weakest signal Petrocelli et al. 2025. Effect sizes are modest, and the trials are short (4 weeks to 4 months) relative to the geroscience claims they support. The CD8+ trial result is striking but n=50 over 4 weeks; replication is needed before "immune rejuvenation" is settled. Cost ($900–$1,500/year) is high for the magnitude of measured effect, and the same dollars deployed on creatine plus an exercise programme would deliver more strength gain. There is no independent replication of the muscle-endurance or strength effects in a non-sponsor-funded trial.

The author's call

Real molecule, real mechanism, real human signal, modest effect size, expensive, conflict-laden evidence base. Land between "promising but unconfirmed" and "premature-but-defensible-supplement-stack inclusion for someone already covering the basics." Evidence rating in the 3 range — multiple good RCTs but mixed primary-endpoint hits, sponsor-dominated authorship, no independent replication of the muscle effects. Effect-size scores moderate (longevity / energy / cumulative beauty in the 2–3 range, not the 4–5 range some other supplements with stronger trials warrant). High cost burden, low effort. The honest reader-facing pitch is "an evidence-backed bet on a real mechanism, expensive for what it delivers, not a substitute for resistance training, sleep, or protein intake."

Stakeholder and incentive map

• Amazentis / Timeline — Swiss biotech that holds the synthesis IP, sells Mitopure (the only clinically tested UA), and has employed or coauthored almost every published clinical trial. Strong commercial incentive to read the trial results optimistically; their R&D voice is also the most knowledgeable on the molecule.
• Longevity / biohacker community — Adopted UA early on the Ryu et al. 2016 paper, often stacked with NMN, NAD+ precursors, resveratrol. Confirmation-friendly audience; their endorsement is volume but not evidence.
• Academic geroscience (Marcinek lab at UW, Auwerx lab at EPFL) — Independent academic groups that have published on UA without direct Mitopure equity. Real expertise; some have Amazentis collaborations but not employment.
• FDA — Granted GRAS status in 2018 for Mitopure; ongoing regulatory tolerance limits worst-case downside.
• Generic supplement makers — Selling unlabelled / unstandardised UA products at lower price points with no clinical validation. Reader's most likely failure-mode here is buying a generic and getting nothing in the capsule.
• Skeptic / anti-supplement establishment — Has not engaged with UA in depth yet; the public skeptical case is more about cost-per-effect-size than mechanism, which is unusual for a longevity supplement.

Population variability

Three sources of variability shape who benefits:

• Metabotype. 10–60% of Western adults are metabotype 0 (non-producers) and get essentially nothing from pomegranate or walnut consumption; direct supplementation flattens this entirely Tomás-Barberán et al. 2017D'Amico et al. 2021. There is no commercially available metabotype test, so practical guidance defaults to "if you want a known dose, supplement."
• Baseline mitochondrial / functional decline. Effect signal is strongest in sedentary middle-aged and older adults (Singh et al., Liu et al., Denk et al.) and weakest in already-trained athletes (Enduro Trial), consistent with diminishing returns once mitophagy is already pharmacologically maximised by exercise.
• Sex / age / comorbidity coverage. Liu et al. enrolled 75.8% women, all white; Singh et al. middle-aged overweight, no diversity breakdown; Denk et al. healthy middle-aged; Petrocelli et al. male runners. No trial in non-overweight middle-aged men, no diabetic / cardiac patient cohort, no pregnancy / lactation data, no paediatric data. The current evidence base does not generalise cleanly to under-40 healthy adults or to clinical populations with cardiometabolic disease.

Knowledge gaps

• Non-sponsor replication. Every positive trial except the Enduro Trial has Amazentis / Timeline authorship; the only one without dominant manufacturer affiliation was the weakest. An independent muscle-strength replication is the single largest evidence gap.
• Long-term outcomes. Longest trial is 4 months; all longevity / mortality / dementia / cardiovascular endpoints are biomarker-based proxies. No long-term outcome trial is registered or running.
• Mitochondrial maximal ATP and biogenesis. The systematic review's null on these endpoints contradicts the mechanistic story; resolving whether UA actually changes mitochondrial mass / function under exercise demand vs only induces mitophagy and shifts gene expression is open Hodzic Kuerec et al. 2024.
• Comparison head-to-head. No trial pits UA against creatine, exercise, NMN, or rapamycin on the same endpoints. Cost-effectiveness vs alternatives is unestablished.
• Cognitive / neurodegenerative endpoints. Preclinical Alzheimer's models look promising; human trials in Alzheimer's or MCI have not reported.

Scope vs the brief. The topic brief named mitochondrial function, muscle strength, and aging. The article covers all three, plus two consequences the brief didn't name but the substance genuinely produces: the immune-aging signal from Denk et al. 2025 (the single strongest pre-specified-primary hit in the literature) and the topical skin/UVB line. Both were added rather than silently dropped — per entry.md §1a, every meaningful consequence is in scope.

Rating difficulties.

• Evidence (3). Could be argued at 2: two of three muscle-trial pre-specified primaries missed (Liu, Singh), the manufacturer authors almost every positive trial, and the systematic review notes nulls on mitochondrial maximal ATP and biogenesis. Landed at 3 because the trials are real RCTs in real populations with directionally consistent secondaries and one cleanly-hit primary (the immune trial). The article is honest about both reads.
• Effect-size dimensions. All non-zero held at 2 ("real but small"). Could argue energy → 3 on the athlete RPE/CK signal, but those weren't reflected in felt-energy outcomes and the elite-runner trial null on the time trial. Conservative.
• Cost (3). $900–$1,500/year is squarely in the 3 anchor ($500–$2,000/year); didn't bump to 4 because median user dose is 500 mg ≈ $720–$900/year.
• Applicability (3). Marked relevant to a large age band (40+) rather than universal because the trial evidence stops at the under-40 line.

Conflict-of-interest disclosure. Almost the entire human evidence base on UA has Amazentis / Timeline (manufacturer) authorship. The article names this explicitly in evidence, misconceptions, and failure-modes; the dossier section §3 enumerates which trials. Not a reason to dismiss; a reason to weight effect-size claims conservatively.

Future-link candidates (not yet written): creatine, nad-precursors (NMN / NR), resistance-training, sarcopenia, pomegranate, walnuts, mitophagy. Cross-link once they exist.

Separate-entry candidates. Topical urolithin A (skin-aging cosmeceutical) is a distinct product line with its own trial cluster. If the literature grows, split out as its own skin-category entry; for now it's covered briefly under practicalities and payoff.

Dream tier. Computed overall ≈ 21 (below 40 threshold). A short, calibrated dream narrative was written anyway — the entry honestly supports a grounded-aspiration / mild-relief frame, and writing the dek straight without that framing would have lost the metabotype paradox hook. The dek and tagline are restrained accordingly: no "game-changer" register.

Excluded knowingly. Preclinical Alzheimer's / Parkinson's data — promising but no human cognitive endpoint trials yet; flagged in research §3f knowledge gaps but kept out of the article body to avoid implying clinical evidence that doesn't exist.