Mitochondria, Part 3: Supplements and Tools for Mitochondria
Written By Kevin Wallace.
Trial by trial. Primary endpoint first.
1. The short answer
Part 2 left a hard question. Can anything in a bottle, a cold plunge, or a sauna match what exercise does for mitochondria? After the human trials, the honest answer is no. Not yet.
Creatine has strong evidence as a training aid. It is not a mitochondrial repair agent. CoQ10 reduced cardiac events in one heart-failure trial that has not been replicated. Magnesium matters when intake is low, which is common. Urolithin A has real but early signals, all from studies tied to its maker, and the two muscle trials missed their primary goals. Most of the rest move a marker more often than they change how people function. Methylene blue carries a documented FDA interaction warning.
None of this is a reason to dismiss the tools. It is a reason to put them in their place: possible additions to the pillars, never substitutes. Doses below are what the studies used, not recommendations. Talk with your physician before starting any of them, especially with a prescription drug.
2. How to read a supplement study
Was the main goal met? A trial names its primary endpoint before it starts. If that endpoint failed and the headline is a secondary result, treat it as a lead, not a finding.
Did it change how people function, or only a marker? A blood level is not strength, endurance, or fewer heart attacks.
Who paid? Company-funded trials can be sound. They need independent replication before we lean on them.
3. Urolithin A
Urolithin A is made by gut bacteria from ellagitannins in pomegranate, and also in walnuts and some berries. It is studied because it can switch on mitophagy in cells and animals. People do not make it equally. In 100 adults, only 12 percent had detectable urolithin A at baseline. After pomegranate juice, about 40 percent converted the precursors into meaningful amounts. A 500 mg supplement produced more than six times the exposure of the juice. (1) That paper, like the trials below, involved the company that sells it.
Three randomized, placebo-controlled trials are the record. All three have an Amazentis author or sponsor. Amazentis sells it as Mitopure.
Singh et al., 2022, 88 overweight sedentary adults aged 40 to 64, 500 or 1,000 mg for 4 months. Primary endpoint, peak power output, was not met. Hamstring strength rose about 12 percent at 500 mg and about 10 percent at 1,000 mg versus placebo. Gains described as clinically meaningful in peak oxygen use and the 6-minute walk at the high dose were not statistically significant against placebo. CRP and acylcarnitines fell. (2)
Liu et al., 2022, 66 adults, mean age 72, 1,000 mg for 4 months. Both primary endpoints failed: the 6-minute walk, and maximal ATP production in a hand muscle. A secondary endurance test at 2 months improved: hand contractions until fatigue rose about 95 versus 12 on placebo, leg contractions about 41 versus 6. The scatter was wide. Acylcarnitines, ceramides, and CRP were lower. (3)
Denk et al., 2025, 50 adults aged 45 to 70, 1,000 mg for 4 weeks. Naive-like CD8 T cells rose by half a percentage point versus placebo. That is a small absolute change. An Amazentis scientist is an author. (4)
Verdict. Consistent secondary signals. Neither muscle trial met its primary goal. No independent replication. Promising, not proven.
4. NAD boosters: NR and NMN
Nicotinamide riboside and nicotinamide mononucleotide are precursors meant to raise NAD. The pitch rests on NAD falling steadily with age. As Part 1 noted, the human evidence for that decline is thin and inconsistent. (5)
NR, 1,000 mg a day for 6 weeks, 13 overweight or obese adults, crossover. No improvement in muscle mitochondrial function or insulin sensitivity. Fat-free mass rose about 1.3 percent. (6)
NMN, 250 mg a day for 10 weeks, 25 postmenopausal women with prediabetes, overweight or obese. Insulin-stimulated glucose disposal in muscle rose about 25 percent per kilogram of fat-free mass. A rise in muscle NAD was not detected. Blood sugar, blood pressure, lipids, liver fat, and inflammation markers did not change. Government-funded. The authors did not make a clinical recommendation. (7)
In September 2025 the FDA decided NMN is not excluded from the dietary-supplement definition. (8) That is a legal status, not evidence that it works.
Verdict. They raise NAD metabolites more reliably than they raise mitochondrial function. Small trials, narrow effects.
5. CoQ10, ubiquinol, and MitoQ
Coenzyme Q10 ferries electrons in the energy chain. It is sold as ubiquinone or as ubiquinol, the reduced form. MitoQ is a modified CoQ10 marketed as mitochondrially targeted.
Heart failure, Q-SYMBIO, 2014. 420 patients with moderate to severe heart failure took 100 mg three times a day or placebo for 2 years, on top of standard care. The 16-week primary endpoints did not change. At 2 years, major cardiovascular events occurred in 15 percent on CoQ10 versus 26 percent on placebo (hazard ratio 0.50, 95 percent CI 0.32 to 0.80). Cardiovascular death was 9 versus 16 percent. All-cause death was 10 versus 18 percent. (9) This is the strongest outcome result in this article. It is one unreplicated trial, in a specific disease, not evidence for healthy people.
Statin muscle aches. As Part 2 covered, a 2021 pooled analysis found no significant benefit. Do not stop a statin over it.
MitoQ and arteries, 2018. Twenty healthy adults aged 60 to 79, already selected for impaired artery dilation, took 20 mg a day for 6 weeks in a crossover. Flow-mediated dilation was 42 percent higher after MitoQ than placebo. Stiffness fell only in the subset with stiffer arteries. It measured blood vessels, not muscle mitochondria. (10)
Verdict. A real case in diagnosed heart failure, to discuss with a cardiologist. For healthy people, thin.
6. Creatine
Creatine is not a mitochondrial repair agent. Stored as phosphocreatine, it regenerates ATP quickly in hard, short efforts. It works alongside the mitochondria. It earns a place here because the evidence is the strongest of anything in this article. The International Society of Sports Nutrition’s 2017 position stand calls creatine monohydrate the most effective nutritional supplement for high-intensity capacity and for lean mass gained with training, with no compelling signal of harm in healthy people at the doses studied. Typical study doses are 3 to 5 grams a day, sometimes after a short loading phase. (11)
Verdict. A partner to strength and interval work, which are the real builders. It helps you train. It does not replace the training. People with kidney disease should not start it without their physician.
7. Alpha-lipoic acid, PQQ, and L-carnitine
Alpha-lipoic acid is a cofactor for mitochondrial enzymes. Its best evidence is diabetic nerve pain, and it is the intravenous form. A 2016 American Family Physician review found 600 mg a day intravenously for 3 weeks reduced neuropathy symptoms. Oral ALA, the supplement form, at 600 mg or more for 3 to 5 weeks gave a minimal improvement that did not reach clinical significance. No long-term evidence. (12)
PQQ is promoted for building mitochondria. In 23 untrained men, 20 mg a day or placebo for 6 weeks during endurance training raised PGC-1 alpha protein. Aerobic performance improved no more than placebo. Both groups gained from the training. (13)
L-carnitine carries fatty acids into mitochondria. The body makes it. Gut bacteria can convert it to TMAO, linked to atherosclerosis. In 2,595 heart patients, higher carnitine predicted risk only when TMAO was also high. Vegetarians and vegans produced little TMAO after a carnitine meal. Meat-eaters produced much more. (14) No clear benefit for healthy people, and an open safety question.
8. Magnesium and omega-3
Magnesium is required for oxidative phosphorylation and is a cofactor in hundreds of enzyme systems. About 48 percent of Americans get less than the estimated average requirement from food. (15) That is a gap to fill, ideally with food. The upper limit for magnesium from supplements, not from food, is 350 mg a day for adults. The usual sign of too much is diarrhea.
Omega-3 fats enter mitochondrial membranes. Eighteen young men took 2 grams of EPA plus 1 gram of DHA daily for 12 weeks. Membrane omega-3 rose several-fold. Maximal respiration did not change. The mitochondria became more sensitive to the signal to make ATP. Oxidant-production capacity rose, without signs of oxidative damage. (16) A real change in construction. Not a demonstrated gain in capacity or health.
9. Methylene blue
Start with safety. In 2011 the FDA warned that methylene blue can cause serotonin syndrome in people taking serotonergic psychiatric drugs, including SSRIs, SNRIs, and some tricyclics. Most reported cases were intravenous doses around surgery. The FDA said the risk at lower or oral doses was uncertain. (17) Uncertainty is a reason for caution, not a green light.
The human function evidence is one randomized trial: 26 healthy adults, a single low oral dose, more activity on brain imaging in attention and memory regions, and a small gain on a memory task, on the order of 7 percent. The authors called it groundwork. (18)
Verdict. Thin evidence and a documented interaction. Not for self-prescribing. Never with an antidepressant unless the prescribing physician is directing it.
10. Cold, heat, and red light
Cold. Eight adults with type 2 diabetes spent 10 days in a room at 14 to 15 °C, about 57 to 59 °F. Peripheral insulin sensitivity rose about 43 percent. The gain was mostly skeletal muscle moving more glucose transporters to the cell surface, with only a minor increase in brown-fat glucose uptake. (19) Eight people, mild room cold, not an ice bath. I did not find comparable plunge trials that measured mitochondria in healthy people. Heart disease or uncontrolled blood pressure is a reason to ask a physician first.
Heat is two different bodies of evidence. Twenty adults had one thigh heated with pulsed shortwave diathermy, 2 hours a day for 6 days, raising muscle temperature about 3.9 °C. Respiratory capacity rose, along with PGC-1 alpha, two energy-chain complexes, and heat-shock proteins (about 45 and 38 percent). (20) That is direct heating of one muscle, not a sauna.
Sauna and heart death are observational. In 2,315 Finnish men aged 42 to 60, followed a median of 20.7 years, men who used a sauna 4 to 7 times a week had a hazard ratio of 0.37 for sudden cardiac death versus once a week, after adjustment for risk factors. That is about a 63 percent lower rate. The frequent group was 201 men, and 10 sudden deaths. (21) Frequent sauna users may differ in other ways. Nothing in the study measured mitochondria. Heat is not for everyone. Low blood pressure and heart disease belong in a physician’s office first.
Red and near-infrared light, roughly 650 to 1,000 nanometers, is proposed to be absorbed by mitochondria and to nudge ATP production. The best-known human test is the eye. Twenty-four people aged 28 to 72 looked into a 670 nm light for 3 minutes a day for 2 weeks. In those about 40 and older, color-contrast vision improved by up to 20 percent. Younger participants did not change. There was no placebo group. The custom light matched no commercial product. (22) It does not tell us whether a consumer panel works. Do not look into bright lights or lasers without guidance.
11. The drug that is not a supplement
Part 2 mentioned elamipretide, also called SS-31, a peptide designed to bind cardiolipin. It is an investigational drug, given by injection in the trials, not a bottle on a shelf. The primary mitochondrial-myopathy trial and a heart-failure trial did not meet their primary endpoints. An accelerated approval in 2025 for Barth syndrome is a rare-disease drug decision. It is not evidence for healthy aging, and it is not a supplement.
12. The evidence at a glance
| Tool | Best human result | Grade |
|---|---|---|
| Creatine | High-intensity capacity and lean mass with training | Strong, as a training aid |
| Magnesium | Required for energy metabolism. About 48 percent below the EAR. | Strong, if intake is low |
| CoQ10 | Fewer cardiac events in heart failure over 2 years. One unreplicated trial. | Moderate, heart failure only |
| Urolithin A | Secondary endurance and marker signals. Primaries missed. Company-linked. | Low to moderate |
| Cold room, 10 days | Insulin sensitivity up about 43 percent in 8 people with diabetes. | Low to moderate |
| Local muscle heat | Respiration up after 6 days of diathermy. Not a sauna. | Low to moderate |
| Sauna | Observational lower cardiac death. 10 events in the frequent group. | Low, observational |
| MitoQ | Artery dilation 42 percent higher in 20 selected older adults. | Low |
| NR and NMN | No muscle mitochondrial gain. One narrow insulin effect. NAD not shown up in that muscle. | Low |
| Omega-3 | Membranes changed. Maximal respiration did not. | Low, for mitochondria |
| PQQ | Biogenesis signal up. Performance no better than placebo. | Low |
| Red light | Color contrast up in 24 people. No control group. | Low |
| Oral ALA | Minimal, not clinically significant, for diabetic nerve pain. | Low |
| L-carnitine | No clear benefit. TMAO question. | Low, with a caution |
| Methylene blue | One single-dose memory study. FDA interaction warning. | Low, safety warning |
| Elamipretide | Myopathy and heart-failure primaries missed. Rare-disease approval only. | Not a supplement |
13. Where we land
Nothing in this article rivals the training result in Part 2. Creatine can help you do that training. Magnesium is worth getting from food if the diet is short. CoQ10 has one unreplicated heart-failure trial, which is a cardiology conversation, not a longevity stack. Urolithin A is the most interesting of the mitochondrial-targeted supplements, and it has not cleared the bar it set for itself.
Cold rooms, local heat, and saunas are early or observational. Red light has a mechanism and a study without a control group. Methylene blue has a warning that outruns its evidence.
October in zone 5 is a good month to keep the order straight. Walk, lift, sleep, eat food you can name. A bottle comes after that, and only with a physician if you already take a prescription. The research frontier is the next article: what is still a mouse, and what might earn a human trial.
14. Sources
Singh et al., European Journal of Clinical Nutrition, 2021. https://doi.org/10.1038/s41430-021-00950-1
Singh et al., Cell Reports Medicine, 2022. https://www.cell.com/cell-reports-medicine/fulltext/S2666-3791(22)00158-6
Liu et al., JAMA Network Open, 2022. https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2788244
Denk et al., Nature Aging, 2025. https://doi.org/10.1038/s43587-025-00996-x
Peluso et al., Nutrients, 2021. https://www.mdpi.com/2072-6643/14/1/101
Remie et al., American Journal of Clinical Nutrition, 2020. Nicotinamide riboside in overweight and obese adults.
Yoshino et al., Science, 2021. https://www.science.org/doi/10.1126/science.abe9985
FDA response on NMN, September 29, 2025. https://www.regulations.gov/document/FDA-2023-P-0872-2754
Mortensen et al., JACC: Heart Failure, 2014. https://pubmed.ncbi.nlm.nih.gov/25282031/
Rossman et al., Hypertension, 2018. https://pubmed.ncbi.nlm.nih.gov/29661838/
Kreider et al., Journal of the International Society of Sports Nutrition, 2017. https://jissn.biomedcentral.com/articles/10.1186/s12970-017-0173-z
American Family Physician, Diabetic Peripheral Neuropathy, 2016, summarizing ALA trials. https://www.aafp.org/pubs/afp/issues/2016/0501/p786.html
Hwang et al., Journal of the American College of Nutrition, 2020. PQQ with endurance training in untrained men.
Koeth et al., Nature Medicine, 2013. https://pubmed.ncbi.nlm.nih.gov/23563705/
NIH Office of Dietary Supplements, Magnesium fact sheet. https://ods.od.nih.gov/factsheets/Magnesium-HealthProfessional/
Herbst et al., Journal of Physiology, 2014. Omega-3 and muscle mitochondrial membranes.
FDA Drug Safety Communication, methylene blue and serotonergic drugs, 2011, updated. https://www.fda.gov/drugs/drug-safety-and-availability/fda-drug-safety-communication-updated-information-about-drug-interaction-between-methylene-blue
Rodriguez et al., Radiology, 2016. Single-dose methylene blue and a memory task.
Hanssen et al., Nature Medicine, 2015. https://www.nature.com/articles/nm.3891
Hafen et al., Journal of Applied Physiology, 2018. Local heat and muscle mitochondria.
Laukkanen et al., JAMA Internal Medicine, 2015. https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2130724
Shinhmar et al., Journals of Gerontology Series A, 2020. Deep red light and color contrast.
