Mitochondria and longevity

Mitochondria are the parts of your cells that turn food and oxygen into ATP, the molecule that powers almost everything your body does, from a heartbeat to a thought. They also carry their own small set of DNA, recycle their broken members and build new ones when your muscles ask for more, and each of those jobs matters for how they age.

How they make energy

The energy work happens on an assembly line called the electron transport chain, which sits in the inner membrane of each mitochondrion. Food is broken down into electrons, and these flow along four stations in that membrane. CoQ10 is the courier that carries them from the early stations to the next one, and oxygen catches them at the very end, which is the whole reason you breathe. As the electrons move, the chain pumps charged particles to one side of the membrane and builds up something like a battery. A final machine, ATP synthase, lets that charge flow back and spins like a turbine to make ATP. Fat yields several times more ATP per molecule than sugar, which is why the body stores its reserves as fat.

That assembly line leaks a little: some electrons slip off early and form reactive oxygen species, or ROS, best pictured as sparks.^[1] Researchers have mapped at least 11 sites inside mammalian mitochondria that can throw them.^[2] CoQ10 helps with the cleanup too, because in its reduced form it is the membrane's main fat-soluble antioxidant, so the courier also mops up part of the mess it works next to.

A small genome of their own

Mitochondria descend from bacteria that moved into the cells of our distant ancestors, and they still keep a tiny ring of their own DNA. That ring holds 37 genes, 13 of which build parts of the energy chain, and you inherit it only from your mother.^[48] Each cell carries hundreds to thousands of copies, so symptoms usually appear only once a large share of them carry a fault.^[48]

Those copies do pick up mutations with age. In human brain tissue, single-letter mutations rose about fivefold over 80 years, and their pattern points to copying errors rather than oxidative damage.^[78] That pattern matters, because the old idea that sparks from the chain wreck the DNA and drive aging rests on the opposite assumption.

Cleanup and renewal

Mitochondria constantly merge and split, which lets them share parts and wall off damage. When one loses its charge, a tag-and-remove system marks it and the cell recycles it, a process called mitophagy. The key cleanup protein in that system, Parkin, was first caught homing in on damaged mitochondria in 2008.^[10] Faults in the Parkin gene cause an inherited early-onset form of Parkinson's disease, which shows how much nerve cells depend on this cleanup.

Recycling is one half of the system, and the other half builds fresh mitochondria. Its master switch is a protein called PGC-1α, first described in the brown fat of cold-exposed mice.^[9] Two sensors flip it: AMPK, the cell's low-fuel alarm, and SIRT1, an enzyme that needs a helper molecule called NAD⁺ to work. In daily life, those sensors respond to muscle contraction, cold, fasting and eating less.

Mitochondria and the machinery that renews them slow down with age in nearly every tissue that has been studied. The pattern is solid, while the reason behind it is still argued over. Two molecules show the decline most clearly. CoQ10 in the human heart roughly halves between age 20 and 80,^[22] and NAD⁺ in human skin falls steadily from birth to old age.^[53] Taurine was once added to that list of falling molecules, but a 2025 study found that its blood levels held steady or rose with age in people.^{[36, 37]}

NAD⁺ gets the most attention, because SIRT1, DNA-repair enzymes and many other enzymes need it, and its drop seems to have two sources. In mice, an NAD⁺-burning enzyme called CD38 rises with age and drives much of the loss,^[79] partly because worn-out (senescent) cells switch it on in nearby immune cells.^[80] In human muscle, at least in people with age-related muscle loss, the making and recycling of NAD⁺ are reduced as well.^[81] So both faster use and slower production seem to contribute, depending on the tissue.

The theory that did not survive

For decades the favourite explanation was the free radical theory of aging: sparks from the mitochondria were thought to chew up DNA, fats and proteins until the body wore out. Its strong form did not survive the experiments. Mice built with mitochondria that cannot proofread their own DNA pile up mutations and age fast, with grey fur, a hunched spine and early death.^{[3, 4]} Yet the later of those studies found no overall rise in oxidative-stress markers,^[4] so the DNA was wrecked and aging sped up without the spark damage the theory predicted.

Other evidence points the same way. A review of 18 genetic changes to antioxidant defences in mice found that only one, deleting the antioxidant enzyme Sod1, shortened lifespan.^[54] In people, Cochrane reviews of antioxidant pills, which pool all the available trials, found no survival benefit and possible harm at high doses.^[56]

What replaced it

The idea that took over is called mitohormesis. Small, brief bursts of ROS, the kind that exercise, fasting and eating less produce, work as a signal: they tell cells to strengthen their own defences and build more mitochondria. In worms, this kind of signal extended lifespan.^[5] The same idea explains an awkward human result, in which high-dose vitamins C and E blunted the gain in insulin sensitivity that young men got from exercise.^[55] A steady flood of ROS still marks broken mitochondria and feeds disease, but it looks more like a symptom than the spark that starts the fire.^[59]

That shift has a practical side. If small bursts of ROS are useful signals, blanket antioxidant pills work against them, while training, which produces exactly those bursts, has far better evidence behind it.

Failing mitochondria are one of the twelve hallmarks of aging, the core breakdowns that show up as we get older. A hallmark, though, is a steady companion of aging rather than a proven single cause.^[7] The list started with nine hallmarks in 2013 and grew to twelve in a 2023 update.^{[6, 7]}

Mitochondria touch most of the other eleven. When they fail, proteins misfold, cells send out inflammatory signals (partly from leaked mitochondrial DNA), fuel sensing goes off balance and stem cells wear out faster. The framework judges each hallmark by three questions: does it appear with age, does making it worse speed aging up, and does fixing it slow aging down?^[7] Mitochondrial decline fits that pattern, but it is also one of twelve breakdowns that lean on each other, which is why "mitochondria cause aging" says more than the evidence does.

A repair system that extends life in worms

Mitochondria run their own emergency response when too many proteins inside them go wrong, and in animals, nudging that response can extend life. In the worm C. elegans, a workhorse of aging research, gently turning down the mitochondria's own protein-making machinery set off this repair response. It stretched lifespan by roughly 40 to 60 percent, depending on the gene and the experiment,^[8] and raising NAD⁺ reproduced part of the effect in worms and mice.^[8]

Where it shows up in people

The tissues that lean hardest on mitochondria, the heart, brain, slow-twitch muscle and liver, are the ones that fail first in inherited mitochondrial diseases.^[48] They are also the organs that fade with age. Nerve cells and heart-muscle cells barely renew themselves, so over decades they collect damaged mitochondria and a growing mix of healthy and faulty mitochondrial DNA.^[59]

That overlap between mitochondrial disease and ordinary aging fits the hallmark idea, and the worm experiments add the strongest piece, because there, boosting mitochondrial repair extended life.^[8] Whether fixing mitochondria slows aging in people has not been shown, so for now they count as a close companion of aging rather than its proven cause.

Training is the best-proven way to improve your mitochondria, because the system that builds them (PGC-1α and its sensors AMPK and SIRT1) answers mostly to muscle contraction and energy stress. Sleep, when you eat and how much you eat add to that. Together they are the foundation: none of the supplements later in this guide has matched their effect, and the useful ones work best on top of it.

Endurance training

The founding experiment dates from 1967, when rats that ran for 12 weeks roughly doubled the activity of their muscles' energy enzymes.^[11] Decades of follow-up work in people sort into a simple rule. How much you train drives mitochondrial quantity (more mitochondria, more enzymes), while how hard you train drives mitochondrial quality (each one works better).^[12] You want both, which is why practical advice splits into plenty of easy work plus a little hard work.

For the easy part, a common coaching target is zone 2: 3 to 4 sessions a week of 30 to 60 minutes at an effort of about 5 to 6 out of 10. The talk test works well, since you can hold a conversation but not sing. In trained adults this usually sits around 60 to 70 percent of maximum heart rate, just below the first lactate threshold, the point where lactate starts to build up in the blood. These figures are practical guidance rather than trial results, and the zone 2 and VO₂ max training guide explains them in detail. The case for zone 2 is that you can do a lot of it without wearing yourself out.

Intervals and strength

Interval training raises mitochondrial enzymes and content about as much as steady endurance work, in less time.^[13] A classic protocol is 10 rounds of one hard minute at around 90 percent of maximum heart rate, each followed by one easy minute.^[13] Intervals add to easy endurance work rather than replace it, and one to two sessions a week is a sensible ceiling for most adults.

For older readers, the trial that matters most compared intervals, weights and a mix of both in young and older adults.^[14] Twelve weeks of high-intensity intervals switched on the broadest set of mitochondrial genes. In the 65- to 80-year-olds, they also moved the protein makeup of their muscle mitochondria back toward a younger profile. Weights built muscle and strength but did not match intervals for the mitochondrial signal. So a plan for aging well uses both: lifting for muscle and strength, endurance and intervals for mitochondria.

When you eat

Eating windows have a mixed record, and the timing of the window seems to matter. In a small trial, 8 men with prediabetes ate within a 6-hour window ending before 3 p.m. for 5 weeks. Their insulin sensitivity, blood pressure and oxidative stress improved without any weight loss.^[15] The popular late version did less. In the TREAT trial, a noon-to-8 p.m. eating window brought no extra weight loss over 12 weeks compared with three regular meals.^[16] Among the participants measured in the clinic, the eating-window group also lost more lean mass, meaning muscle and other tissue that is not fat.^[16]

A head-to-head trial then found that an early window beat a midday one for insulin sensitivity.^[58] So if you try time-restricted eating (the intermittent fasting guide compares the options), the earlier window is the one with the signal.

How much you eat

Eating less overall has the longest human trial behind it. In CALERIE 2, healthy adults without obesity cut their calories by about 12 percent for two years, and their heart and metabolic risk factors improved more than weight loss alone would explain.^[17] A later analysis of the same people looked at DunedinPACE, an epigenetic blood test that estimates how fast the body is aging. That pace ran 2 to 3 percent slower in the calorie-restriction group.^[18] The authors call that effect small, two other aging clocks did not change, and whether it changes disease or lifespan is unknown.

Inside the muscle, an earlier 6-month trial in overweight adults pointed the same way. Eating less brought more mitochondrial DNA and higher activity of the genes that build mitochondria, though key mitochondrial enzyme activities did not change.^[46]

Cold and sleep

Cold has a real but narrow mitochondrial angle. Adults still have active brown fat, which burns energy as heat in the cold,^[88] and 10 days of mild cold made that brown fat more active.^[89] Experienced winter swimmers regulated body temperature differently and produced more heat in the cold, but the study did not measure whether they had built more mitochondria.^[87] Cold therefore helps at the edges, and an ice bath, however bracing, is no stand-in for endurance training.

Sleep loss shows up in muscle quickly. After one night without sleep, young men's muscle turned down genes for the energy chain,^[47] and five nights of short sleep shifted muscle gene activity toward inflammation and immune pathways.^[60] Both are gene-activity signals rather than direct measures of mitochondrial function, but they make 7 to 9 hours of sleep the cheapest item on this page.

Removing the obvious damage

Smoking delivers carbon monoxide and cyanide, both of which block the last station of the energy chain. Heavy drinking disturbs mitochondrial protein building and the NAD⁺ balance in the liver. A diet heavy in ultra-processed food goes with more inflammation, although its link to mitochondria is inferred rather than shown in cells.

Together, endurance work, some intervals, strength training, enough sleep and a sensible eating pattern move the measures that matter. Those are mitochondrial enzymes, the lactate threshold and VO₂ max, the most oxygen your body can use during hard exercise.^{[12, 14]} Some of these moves carry real risks for certain people, though.

Who should be careful

Fasting and calorie restriction are not for pregnancy or breastfeeding, a history of eating disorders, underweight, children and teenagers, or anyone on insulin or sulfonylureas without medical supervision, because of the risk of low blood sugar. People with a diagnosed mitochondrial or fatty-acid oxidation disorder should plan fasting and exercise with their specialist team. If you are over about 40 and untrained, or have heart disease, high blood pressure or diabetes, get a check-up before hard intervals (in Germany, from your Hausarzt or a sports-medicine exam). Cold water is risky with heart disease or uncontrolled blood pressure, and you should never swim in cold water alone.

As an addition to training, four supplements have the strongest mitochondrial rationale: CoQ10, riboflavin (vitamin B2), an NAD⁺ precursor (NR or NMN) and taurine. Each sits at a load-bearing point of the energy machinery, so the biology is solid. Each also has a group that profits most, from people with heart failure (CoQ10) to adults with migraine (riboflavin), and the clinical evidence ranges from one strong heart-failure trial for CoQ10 to a long trial for NR that found no benefit.

CoQ10 (ubiquinone and ubiquinol)

CoQ10 is the courier that carries electrons along the energy chain, and nothing else can do its job. Your body makes it, but heart levels roughly halve between age 20 and 80.^[22] Statins, the common cholesterol drugs, also lower CoQ10 in the blood, though whether that matters for health is unclear.^[85]

The best evidence comes from Q-SYMBIO, which gave 420 people with moderate to severe heart failure 300 mg a day or placebo for two years, on top of their usual medicines.^[19] It met its long-term main goal: a major heart event hit 15 percent of the CoQ10 group, versus 26 percent on placebo. It did not meet its 16-week goals for symptoms, walking distance and a heart-strain blood marker, and the drop in deaths was a secondary result of one mid-sized trial.^[19] A Swedish trial of selenium plus CoQ10 in older adults also found fewer heart deaths, still visible years after the capsules stopped.^{[20, 61]} The selenium muddies that picture, though, and the authors themselves call it a small study that raises a hypothesis rather than settling one. In heart failure with preserved pumping function, the evidence is too thin to say.^[69]

Outside heart failure, migraine is the most promising use. In a small migraine trial, 300 mg a day for three months halved how often attacks came in 48 percent of people, versus 14 percent on placebo.^[70] Trials on statin muscle pain have been inconclusive, and there is no outcome trial for general prevention in healthy people over 60. In early Parkinson's disease, the large QE3 trial at 1200 or 2400 mg a day found no evidence of benefit and was stopped early,^[21] so CoQ10 is not a Parkinson's treatment.

Ubiquinone and ubiquinol are the two forms you will see on labels. In a comparison of seven 100 mg products in healthy adults, the carrier oil and how the CoQ10 was dissolved mattered more than which form it was.^[86] The CoQ10: ubiquinol vs ubiquinone guide compares products in detail.

Studied doses run from 100 to 300 mg a day with a meal that contains some fat, and trials have used up to 2400 mg a day.^[21] CoQ10 can interact with the blood thinner warfarin.^[97] If you take warfarin or another vitamin K antagonist such as phenprocoumon (Marcumar in Germany), talk to your doctor before starting, stopping or changing CoQ10 at any dose, and have your INR (the clotting test) checked. In heart failure, CoQ10 is an add-on to heart medicines and never a replacement, so start it only with your cardiologist. On the label side, EFSA, the EU's food safety agency, reviewed CoQ10 in 2010 and rejected all six proposed health claims, from energy to blood pressure. As a result, no health claim for CoQ10 is authorised in the EU.^[23]

Riboflavin (vitamin B2)

Riboflavin is vitamin B2, and your cells turn it into two helper molecules (FMN and FAD) that the energy chain cannot run without. They work at its first and second stations and feed the fat-burning pathway into the chain. Riboflavin's clearest medical use is in two rare inherited disorders, both treated under specialist care. In riboflavin-responsive MADD, a disorder of the fat-burning enzymes that most often traces to one gene, riboflavin is the treatment.^[25] In a disorder of the riboflavin transporter (Brown-Vialetto-Van Laere syndrome), high-dose riboflavin can be life-saving.^[26]

For everyone else, migraine has the best evidence. In a trial of 55 adults, 400 mg a day for three months reduced migraine attacks.^[24] Headache days fell by at least half in 59 percent of people on riboflavin, versus 15 percent on placebo.^[24] Two small trials in children, at 200 mg and 50 mg a day, found no benefit,^{[94, 95]} so the adult result does not simply carry over to children. A narrower finding concerns blood pressure. In people already on blood-pressure drugs who carry a common gene variant called MTHFR 677TT, 1.6 mg a day lowered the top number by a further 5.6 mmHg.^[27]

Riboflavin is cheap, and EFSA found too little data on harm to set an upper limit.^[102] The usual sign of a high dose is bright yellow urine, which is harmless and mostly proof that you absorbed it. Riboflavin also carries nine authorised EU health claims, including "contributes to normal energy-yielding metabolism" and "contributes to the reduction of tiredness and fatigue".^[68]

NAD⁺ precursors: NR and NMN

NAD⁺ falls with age, so two precursors, building blocks the body turns into NAD⁺, offer a direct way to top it up: nicotinamide riboside (NR) and nicotinamide mononucleotide (NMN). The idea is that more NAD⁺ feeds the sirtuins, the enzyme family that includes SIRT1, which switch on PGC-1α and tune up the energy chain. That is solid molecular biology, and both precursors do raise NAD⁺ in people. The open question is what that extra NAD⁺ does for your health.

NR has the most data on that question. A single 1000 mg dose raised blood NAD⁺ in a small pilot,^[28] and 1000 mg a day for 6 weeks raised it in immune cells by about 60 percent.^[29] Higher NAD⁺ did not reliably bring better metabolism, though: 2000 mg a day for 12 weeks did not improve insulin sensitivity or any other metabolic measure in obese, insulin-resistant men.^[30] In a twin study where most participants knew they were taking NR (an open-label design), muscle mitochondria increased in number, but body fat did not fall and insulin sensitivity got worse. With only four placebo-controlled twin pairs, cause and effect stay unclear.^[82]

The biggest test so far found no benefit. In NOPARK, a phase 3 trial (the large final stage of testing), 410 people with early Parkinson's disease took 1000 mg of NR a day for a year. They gained no benefit, and their motor and non-motor scores ended slightly worse than on placebo.^[62] NR is not a Parkinson's treatment.

NMN has smaller and shorter trials, some with real early signals. In postmenopausal women with prediabetes, 250 mg a day for 10 weeks improved muscle insulin sensitivity.^[31] In another trial, 300 to 900 mg a day for 60 days improved walking distance, with the gain levelling off at 600 mg.^[32] So NAD⁺ boosters are a promising field with early functional gains such as a longer walk; a lasting health benefit has not shown up so far, and both share an open safety question.

That question concerns what is left over, because part of every NAD⁺ precursor ends up as two breakdown products of niacin, called 2PY and 4PY. In a study of more than 4,000 heart patients, those with the highest levels had roughly 1.6 to 2 times the three-year risk of a heart attack, stroke or death.^[33] That link held even after accounting for the usual risk factors.^[33] In mice, 4PY at physiological levels switched on markers of vascular inflammation, which suggests a mechanism.^[33] In people, though, the link is observational, a pattern in patient data rather than a tested effect. NR and NMN do raise these breakdown products in small human studies,^{[28, 63]} and no study has yet shown whether such supplement-driven rises lead to more heart attacks or strokes.

The two precursors also differ in law (as of October 2026). NR chloride is authorised in the EU as a novel food, an ingredient that needs EU approval before sale. It may be used in food supplements up to 300 mg a day in adults and 230 mg a day in pregnancy and breastfeeding.^{[64, 65]} The original applicant's data protection (ChromaDex, brand Niagen) ended on 20 February 2025, so the authorisation is no longer limited to one company.^[64]

NMN is not on the EU's list of authorised novel foods. EFSA judged it safe in May 2026 for supplements up to 300 mg a day in adults, though not in pregnancy or breastfeeding.^[57] That opinion is not a sales permission, because only a Commission regulation adding NMN to the Union list makes it legal to sell. Until that regulation exists, NMN supplements may not be placed on the market as food in Germany or Austria.^[66] Switzerland has its own rules, and neither NR nor NMN has an authorised health claim.^[68]

Taurine

Taurine has a precise job inside mitochondria: it is attached to the small adapter molecules (mitochondrial tRNAs) that read the mitochondrial genetic code. Without that taurine tag, the adapters misread certain code words, and the mitochondria cannot build their own machinery properly.^[34] MELAS, a severe inherited disease with stroke-like episodes, comes from mutations that leave exactly this tag missing.^[35]

That makes MELAS the strongest test case. In a small open-label trial without a control group, 10 patients took 9 to 12 g a day for a year. Their stroke-like episodes fell from 2.22 to 0.72 a year, and 6 of 10 had none at all.^[35] Japan approved taurine for MELAS in 2019. Outside Japan it is used off-label, and a 2026 consensus of European and US mitochondrial medicine experts rated its benefit as unproven.^[73] High-dose taurine for MELAS belongs in a specialist clinic.

For healthy aging, the picture is less clear. A 2023 study reported that blood taurine drops with age in several species. In the same study, a very high dose, 1 g per kg of body weight a day, extended the median lifespan of middle-aged mice by about 10 to 12 percent.^[36] A 2025 study then undercut the first finding: taurine rose or held steady with age in three human groups, and in monkeys and mice too.^[37] The MELAS biology stands, but the idea that you are running low and only need to refill has not held up.

Trials in healthy adults use 1 to 6 g a day. The observed safe level, the highest intake with strong evidence of no harm, is 3 g a day, and above that, long-term safety data are thin.^[52] Taurine is an ordinary food ingredient in the EU, familiar from energy drinks, but it has no authorised health claim.^[68]

The four, compared

Side by side, the four split by what they are good for. CoQ10 has the strongest single trial, as an add-on in one kind of heart failure, and riboflavin is the cheapest, with clear roles in rare mitochondrial diseases and adult migraine. Taurine's best case is MELAS under specialist care. NR and NMN reliably raise NAD⁺, and if you want to try that, NR is the one that may be sold as food in the EU, best tested over a defined period with a checked product, since a proven health benefit is still missing and NMN may not be sold as food in the EU. Beyond these four, the evidence thins out, though several of the next ones still have a clear use.

Magnesium, creatine, vitamin B12, curcumin and PQQ each earn their place in a specific situation, and the parts below say which one. All five have real mitochondrial biology behind them, with weaker human evidence than CoQ10, riboflavin, the NAD⁺ precursors and taurine.

Magnesium

Magnesium is the partner mineral that makes ATP usable, and several steps of the energy chain depend on it. EFSA's tolerable upper intake level for magnesium from supplements is 250 mg a day, which is scientific advice rather than a sales limit.^[102] More can cause diarrhoea, and because labels often give the weight of the whole salt rather than the elemental magnesium, the small print matters. With kidney disease, ask your doctor before taking magnesium.

The form matters mostly for how your gut copes. Oxide is the most laxative, while bisglycinate tends to be gentler, and its glycine carrier is a small bonus for sleep (see the deep sleep guide). Magnesium carries ten authorised EU claims, including "contributes to normal energy-yielding metabolism" and "contributes to the reduction of tiredness and fatigue".^[68]

Creatine monohydrate

Creatine works like a rechargeable battery for ATP rather than building new mitochondria, storing energy and releasing it where demand spikes, in contracting muscle and in busy nerve cells. Its strongest evidence is in muscle performance, where the EU authorises two claims at 3 g a day or more.^{[50, 68]} One covers better performance in repeated short bursts of high-intensity exercise, and the other a stronger effect of resistance training on muscle strength in adults over 55.

For the brain, a review of six small trials found that creatine may improve short-term memory and reasoning, with hints of more benefit in older or stressed people and in vegetarians.^[90] Those trials used quite different amounts and schedules, so there is no established dose for brain effects.^[90] For muscle, 3 to 5 g of creatine monohydrate a day is the usual dose. With kidney disease, ask your doctor first, and because creatine can raise creatinine on a blood test without any kidney damage, tell whoever reads your results.

Vitamin B12

B12 powers an enzyme inside the mitochondria that feeds certain carbon fragments into the central energy cycle. When B12 runs low, a substance called methylmalonic acid (MMA) builds up, so a raised MMA level is a useful sign of a possible deficiency. It is read together with B12 levels and kidney function, because MMA also rises with kidney disease. Low B12 is common in older people, vegans and people on the diabetes drug metformin or on acid-blocking PPIs, so those groups gain the most from checking and topping up. Methylcobalamin, adenosylcobalamin and the cheap cyanocobalamin are all on the shelf, and for most people the dose matters more than the form. B12 carries eight authorised EU claims, including energy metabolism and the reduction of tiredness and fatigue.^[68]

Curcumin

Curcumin's main strength is inflammation; it nudges PGC-1α only indirectly, and direct mitochondrial measurements in people are thin. To get the most out of it, absorption is the key, since plain curcumin barely reaches the blood. Boosted forms get many times more into the blood, though the gain differs by product and by how it was measured.^[101] Rare liver injuries have been reported with these forms,^[110] so ask your doctor first if you have liver disease or take blood thinners or other regular medication. Curcumin has no authorised EU health claim,^[68] and its better fit is joints and inflammation rather than mitochondria.

PQQ (pyrroloquinoline quinone)

In cell cultures and rodents, PQQ switches on PGC-1α and grows new mitochondria. The human evidence is still small but shows early signals: a 10-person biomarker study plus small placebo-controlled trials at 20 mg a day that reported memory gains.^[84] None of them shows mitochondrial building in humans, and the results need independent confirmation. PQQ (as its disodium salt) is an authorised novel food in the EU up to 20 mg a day for adults, excluding pregnancy and breastfeeding.^[49] It has no health claim,^[68] and a "mitochondrial biogenesis nutrient" label on the box promises more than the human data show so far.

Several ingredients in mitochondrial formulas do real work elsewhere in the body, and knowing that job tells you when one is worth taking. None of those jobs is to improve mitochondrial function in healthy adults, and some of these ingredients carry safety limits worth knowing before you buy.

Citicoline, Rhodiola and myo-inositol

These three are simple to place, because their main uses lie outside mitochondria. Citicoline (CDP-choline) helps build nerve-cell membranes and touches mitochondria mainly after injury. It is authorised as a novel food in supplements up to 500 mg a day, and labels must say it is not intended for children.^[108] Rhodiola rosea, sold as an adaptogen (a herb marketed to help the body cope with stress), shows mitochondrial signals only in lab and animal studies. Myo-inositol is an insulin-signalling molecule studied mainly in polycystic ovary syndrome. None of the three has an authorised EU health claim,^[68] so if one of them fits you, take it for its own use rather than for your mitochondria.

Vitamin B6 (P5P)

P5P, the active form of vitamin B6, helps many enzymes, some of them inside mitochondria, so it matters most for people who are actually low; in people who are not deficient, topping it up does not raise mitochondrial output. EFSA lowered the upper limit for B6 to 12 mg a day in 2023, from 25 mg, because long-term high intake damages nerves.^[96] Do not assume P5P is safer than pyridoxine, the usual form. That 12 mg limit applies to total vitamin B6 from all sources combined, including ordinary food, fortified foods and supplements, whatever the form.^[96] B6 carries ten authorised EU claims.^[68]

5-MTHF (active folate)

5-MTHF is the ready-to-use form of folate. It sidesteps a common gene variant (MTHFR 677TT) that slows folate processing. Its main work is methylation, the body's way of attaching small chemical tags called methyl groups to DNA and other molecules, and that happens mostly outside the mitochondria. Its best-known result comes from depression: 15 mg a day added to an antidepressant helped some people whose depression had not responded.^[92] That trial used a dose far above supplement levels, under psychiatric care, so do not treat depression yourself. Keep in mind, too, that folate can mask a B12 deficiency.

Betaine (TMG)

Betaine (TMG, trimethylglycine) is a methyl donor, meaning it hands on those methyl groups, and it works mostly in the liver, outside the mitochondria. It is the usual partner suggested for NR and NMN, because breaking down the extra nicotinamide they deliver uses up methyl groups. Yet the NR trial that checked found methyl-group metabolism undisturbed,^[83] and no trial shows that adding TMG helps, so the pairing is a theoretical precaution rather than a rule. Its authorised EU claim is "contributes to normal homocysteine metabolism", for a daily intake of 1.5 g.^[68]

Piperine

Piperine (black pepper extract, sold as BioPerine) has one job in supplements. It blocks the enzymes and pumps that clear substances from the body, which raises blood levels of whatever is taken with it. In an early study, 20 mg of piperine raised curcumin absorption by about 2000 percent,^[93] at a dose ten times the German guidance below. The same effect raises blood levels of several drugs, including tacrolimus, cyclosporine, simvastatin, midazolam, calcium channel blockers, theophylline and phenytoin.

Germany's Federal Institute for Risk Assessment (BfR) advises no more than 2 mg of isolated piperine a day from supplements, and advises pregnant women to avoid such products.^[75] Many curcumin-plus-piperine products contain more. Avoid piperine entirely if you take any of the drugs above. Each ingredient in this group has its own use, so when a label sells one as a mitochondrial booster, judge it by that use, since the mitochondrial claim runs ahead of the science.

Whole dietary patterns have better long-term outcome data than any single mitochondrial compound, and the strongest of those data come from the Mediterranean diet.

The Mediterranean pattern

In the corrected analysis of the PREDIMED trial, about 7,400 adults at high heart risk were randomised to a Mediterranean diet with extra-virgin olive oil or nuts, or to a low-fat control diet. Major cardiovascular events fell by about 30 percent over roughly five years.^[44] No single-compound supplement on this page has shown an outcome like that. The diet changes many things at once (fibre, fats, polyphenols, fish), so no single ingredient can take the credit. That is a fair argument for eating the pattern rather than buying its parts.

Omega-3 from fish

Fish, one part of that pattern, supplies the omega-3 fats EPA and DHA. They are one of the few entries here with authorised EU claims at a set intake. EPA and DHA contribute to normal heart function at 250 mg a day combined, and DHA to normal brain function and vision at 250 mg a day.^[68] Claims on blood triglycerides and blood pressure need 2 to 3 g a day.^[68] Their mitochondrial link runs through the membranes around the energy chain and comes mainly from lab work. The human evidence is about the heart, and it is more mixed than the claim suggests.

In a large observational study, a red-blood-cell fatty-acid pattern that includes the omega-3 index predicted death from any cause about as well as cholesterol and blood pressure did.^[43] That is a link, not proof. The trials split. REDUCE-IT gave statin-treated patients with high triglycerides 4 g a day of purified EPA (icosapent ethyl) and cut major heart events by about a quarter.^[41] That dose is a prescription drug, though, not a fish-oil supplement. STRENGTH gave 4 g a day of a different EPA-plus-DHA product and found no benefit over corn oil.^[104] Some researchers read part of REDUCE-IT's win as an effect of its mineral-oil placebo, and the two results are still debated. At an ordinary supplement dose, the VITAL trial's daily 1 g fish-oil capsule brought no overall benefit for general prevention, with hints of benefit in people who ate little fish.^[42]

EFSA found that supplements of up to 5 g a day of EPA and DHA combined raise no safety concerns for adults.^[103] High doses can raise the risk of atrial fibrillation, an irregular heartbeat: in REDUCE-IT, more people on the high dose were hospitalised for it.^[41] So if you take blood thinners or have atrial fibrillation, discuss gram-level fish oil with your doctor.

Sprouts and green tea

Sulforaphane from broccoli sprouts switches on the body's own antioxidant defences in lab studies. A daily handful of sprouts with a pinch of mustard powder, whose enzyme activates it, is the food route. Raw or lightly cooked sprouts can carry Salmonella, Listeria or E. coli, so children, older adults, pregnant women and people with weakened immunity should eat sprouts only thoroughly cooked, never raw or lightly cooked.^[99]

Green tea as a brewed drink is generally considered safe, and its rare liver reactions are thought to be idiosyncratic, meaning they strike a few people unpredictably.^[98] Its main catechin, EGCG, is a different matter in concentrated extracts. EFSA found that 800 mg EGCG a day or more from supplements may cause early signs of liver damage and could not name a safe supplement dose.^[98] Since June 2023, EU law caps a daily portion of green tea extract supplements below 800 mg EGCG. It also requires warnings: not on an empty stomach, not together with other green-tea products that day, and not for pregnant or breastfeeding women or anyone under 18.^[67]

Urolithin A

Isolated food compounds are a separate question from the foods themselves, and urolithin A is the one with the most specific mechanism. It is made by gut bacteria from plant compounds (ellagitannins) in pomegranate, walnuts and berries, and many people make little of it because they lack the right bacteria. A supplement, sold as Mitopure, skips that step. In lab studies, urolithin A switched on mitophagy, the recycling of worn-out mitochondria, extended lifespan in worms and improved muscle function in rodents.^[51]

In people, the trials show real signals in secondary measures, though they are fewer and more modest than the marketing suggests. After a first study that met its safety goal,^[38] both efficacy trials missed their main goals.^{[39, 40]} In adults aged 65 to 90 taking 1000 mg a day for 4 months, the walking test and muscle ATP production did not improve, though muscle endurance and some blood markers did.^[39] In middle-aged adults taking 500 or 1000 mg a day for 4 months, the main peak-power goal was missed, while muscle strength rose about 12 percent more than on placebo.^[40] Those trials were run with the maker, Amazentis. The first independent trial, in adults with obesity taking 1000 mg a day for 4 weeks, also missed its main goal, on blood-vessel function.^[71]

Studied doses are 500 to 1000 mg a day.^{[39, 40]} In the EU, Amazentis applied for novel food authorisation in 2018,^[109] and the application is still listed as pending, with no authorising regulation (as of October 2026). Because a novel food needs an authorisation before it can be sold,^[66] urolithin A may not be sold as food in the EU, and it has no health claim.^[68]

Resveratrol and fisetin

Resveratrol started with the biggest promise in this group, and the follow-up has been modest. The original idea that it directly activates the anti-aging enzyme SIRT1 did not survive later testing, which traced much of the effect to the lab assay itself.^[105] In one small trial in obese men, 150 mg a day for 30 days improved how well muscle mitochondria burned fat-derived fuel,^[106] but later trials have been mixed. Resveratrol has no authorised EU health claim.^[68] As a novel food, trans-resveratrol is authorised for adults only, up to 150 mg a day, and the label must say that people taking medicines should use it only under medical supervision.^[107]

Fisetin cleared worn-out (senescent) cells and extended both median and maximum lifespan in mice.^[45] As of October 2026, ClinicalTrials.gov still lists the main human trials, the Mayo Clinic's AFFIRM frailty trials, without results. Until they report, "senolytic" (cell-clearing) branding is ahead of the human evidence, so the best time to decide on fisetin is once those results are out.

EU law decides two separate things about these supplements: whether a product may be sold as food at all, and what its label or advertising may claim. Everything below is as of October 2026.

Who the claim rules cover

The EU health-claim rules (Regulation (EC) No 1924/2006) apply to commercial communications such as packaging, adverts and sponsored content. Independent editorial content without a commercial purpose falls outside them (Article 1(2) and Recital 4). Scientific wording alone does not create that exemption, because the context matters, such as who pays and whether a product is sold next to the text. So a report that 300 mg of CoQ10 a day was followed by fewer major heart events in heart failure.^[19] is a different thing from a label line such as "CoQ10 supports cardiovascular function".

Authorised claims

Authorised claims are listed in Regulation (EU) No 432/2012, and a label must keep their meaning and conditions of use. The wording may be adapted, but only if it means the same to consumers and does not make the claim stronger^[68] For the nutrients in this guide:

• Riboflavin: nine claims, including normal energy-yielding metabolism and the reduction of tiredness and fatigue.

• Magnesium: ten claims, including normal energy-yielding metabolism, normal muscle function and the reduction of tiredness and fatigue.

• Niacin, pantothenic acid, vitamin B6 (ten claims) and vitamin B12 (eight claims): each carries "normal energy-yielding metabolism" and "reduction of tiredness and fatigue".

• Folate: the fatigue claim and others (homocysteine metabolism, blood formation, cell division), but no energy-metabolism claim.

• Vitamin C and vitamin D: claims including the immune system, and for vitamin D also bones and muscles.

• EPA and DHA: normal heart function at 250 mg a day combined; DHA alone for normal brain function and vision at 250 mg a day.

• Creatine: better performance in repeated short, high-intensity bursts at 3 g a day or more, and, together with resistance training, muscle strength in adults over 55.^{[50, 68]}

• Betaine: normal homocysteine metabolism at 1.5 g a day.

No health claim is authorised for CoQ10, whose six proposed claims EFSA rejected,^[23] or for NR, NMN, taurine, resveratrol, curcumin, urolithin A, PQQ, citicoline, fisetin, L-theanine and glycine.^[68] A missing claim means no regulator accepted the evidence for a label statement, which does not automatically mean the compound is useless.

Which ones may be sold as food

Ingredients that were not eaten to a significant degree in the EU before May 1997 count as novel foods, and they need an authorisation before they can be sold.^[66] For this guide:

• NR chloride: authorised for food supplements up to 300 mg a day for adults and 230 mg a day in pregnancy and breastfeeding.^{[64, 65]}

• PQQ (disodium salt): authorised up to 20 mg a day for adults, excluding pregnancy and breastfeeding.^[49]

• Resveratrol (trans-resveratrol): authorised for adults up to 150 mg a day, with a label warning that people taking medicines should use it only under medical supervision.^[107]

• Citicoline: authorised in food supplements up to 500 mg a day, with a label stating it is not intended for children.^[108]

• NMN: not authorised. A positive EFSA safety opinion from May 2026.^[57] is not a sales permission, so NMN supplements may not be placed on the market as food in Germany or Austria^[66]

• Urolithin A: an application has been pending since 2018, and no authorising regulation exists.^{[66, 109]}

• Green tea extracts: allowed, but capped below 800 mg EGCG per daily portion, with mandatory warnings.^[67]

Upper intake levels

EFSA's tolerable upper intake levels are scientific advice on the highest daily intake not expected to cause harm, not binding sales limits. These are the adult values relevant to this guide.^[102]:

• Magnesium: 250 mg a day from supplements and added magnesium salts.

• Vitamin B6: 12 mg a day from all sources combined, including ordinary food^[96]

• Vitamin D: 100 µg (4,000 IU) a day.

• Selenium: 255 µg a day.

• EPA plus DHA: no upper level set; supplements up to 5 g a day combined raise no safety concerns.^[103]

• Riboflavin and vitamin B12: no upper level set.

Germany: when a supplement becomes a medicine

In Germany, a product advertised as treating or preventing disease can count as a medicine by presentation (§2(1) no. 1 AMG). A product with a relevant pharmacological effect at its dose, meaning it acts on the body like a drug, can count as a medicine by function (§2(1) no. 2 AMG). In either case, the drug regulator BfArM becomes responsible instead of food law.

So a label raises three separate questions: whether the ingredient may be sold as food at all, whether its claim is on the authorised list, and whether its daily dose stays within the upper intake level. Checking each one on its own is a quick way to spot a product that promises more than the law or the evidence allows.

The order that follows from the evidence is lifestyle first, then food, then basic cofactors, with everything beyond that as optional fine-tuning. The tiers below rank the options by how much evidence you get for your effort and money.

One ground rule first. The doses in Tiers 2 to 5 are those commonly studied or sold over the counter, not personal recommendations. Talk to your doctor (in Germany, your Hausarzt) before starting a stack, especially if you take medication, are pregnant or have a chronic condition.

Tier 0: foundation, not supplements. As practical targets: zone 2 endurance 3 to 4 times a week, intervals 1 to 2 times, resistance training 2 to 3 times, 7 to 9 hours of sleep, and an earlier eating window if it fits your life. If you lift weights, aim for about 1.6 g of protein per kg of body weight a day, since more than that added no further muscle in trials.^[91] Needs vary, and with kidney disease, ask your doctor. Not smoking, little or no alcohol and less ultra-processed food complete the tier. This is the tier that moves VO₂ max (the most oxygen your body can use during hard exercise), the lactate threshold and mitochondrial content,^{[12, 14]} and every tier after it works as an addition to it, not a replacement.

Tier 1: foods with the best evidence.

• A Mediterranean-style diet with vegetables, legumes, whole grains, fish, extra-virgin olive oil and nuts.^[44]

• Fatty fish 2 to 3 times a week, or EPA plus DHA of at least 250 mg a day, the intake behind the EU heart claim.^[68]

• Berries, walnuts and pomegranate, which supply the plant compounds gut bacteria turn into urolithin A; how much you make depends on your gut bacteria.

• Broccoli sprouts with a pinch of mustard powder; children, older adults, pregnant women and people with weakened immunity should eat them only thoroughly cooked.^[99]

• Green tea as a brewed drink rather than a concentrated extract.^[98]

Tier 2: basic cofactor coverage. This tier matters most for older adults, vegans, people on metformin, PPIs or statins, and anyone whose diet does not cover the basics. For them, closing a gap here is a cheap, well-understood win.

• A multivitamin with riboflavin, B12, folate and the rest of the B group at roughly the recommended daily intake.

• Magnesium: up to 250 mg elemental magnesium a day from supplements, which is EFSA's tolerable upper intake level.^[102] More can cause diarrhoea, and labels often state the weight of the whole salt rather than elemental magnesium. With kidney disease, ask your doctor first.

• Vitamin D: on this list for general health, not as a mitochondrial cofactor. The Endocrine Society's 2024 guideline found no trial evidence for a target blood level and advises against routine testing in healthy people.^[74] It suggests supplements without testing for children and teens, adults over 75, pregnancy and high-risk prediabetes, and not above the recommended intake for healthy adults under 75.^[74] EFSA's upper level is 100 µg (4,000 IU) a day.^[102] In Germany, a vitamin D test without a medical reason is usually a self-pay service (IGeL).

• B12: commonly used doses are 500 to 1000 µg a day, swallowed or under the tongue, in at-risk groups (vegans, metformin or PPI users, people over 60). Blood tests can help assess a possible deficiency: besides the B12 level itself, doctors look at methylmalonic acid (MMA) and homocysteine, two substances that build up when B12 runs short.

Tier 3: optional additions with a mitochondrial angle.

• Creatine monohydrate: 3 to 5 g a day for muscle, with two authorised EU claims.^{[50, 68]}; there is no established dose for brain effects^[90] With kidney disease, ask your doctor first, and remember that creatine can raise creatinine on a blood test without kidney damage.

• CoQ10: studied doses of 100 to 300 mg a day with some fat; Q-SYMBIO used 300 mg.^[19] The one strong result is in heart failure with reduced pumping function, as an add-on to heart medicines and only with your cardiologist.^[19] In heart failure with preserved pumping function, the evidence is too thin to say.^[69] Statins lower blood CoQ10,^[85] but trials have not shown that topping it up relieves statin muscle pain, and there is no outcome trial for general prevention in healthy over-60s. If you take warfarin or another vitamin K antagonist, talk to your doctor before any change and have your INR checked.^[97]

• Omega-3 EPA plus DHA: a typical consumer dose is 1 to 2 g a day if you rarely eat fatty fish. With blood thinners or atrial fibrillation, discuss gram-level doses with your doctor.^[41]

Tier 4: higher cost, higher uncertainty.

• NR: a typical consumer dose is 300 mg a day, the EU maximum for adults (230 mg in pregnancy and breastfeeding).^[64] It raises NAD⁺, but the long phase 3 trial in Parkinson's was negative,^[62] so if you want to try it, use a tested product for a defined period and treat it as an experiment with an open outcome. Adding TMG is optional at most, since no trial shows it helps.^[83]

• Urolithin A: the studied dose is 500 mg a day, mostly of interest to adults over 60 worried about muscle loss. The two efficacy trials missed their main goals, and the positive results are in secondary measures.^{[39, 40]} It is not authorised as a food in the EU (as of October 2026).^[66]

• Curcumin in a boosted form: a typical consumer dose is 500 to 1000 mg a day for joint or inflammation uses, not as a mitochondrial agent. Rare liver injury has been reported,^[110] so ask your doctor first if you have liver disease or take blood thinners or other regular medication. If the product contains piperine, the BfR's advice of no more than 2 mg of piperine a day applies.^[75]

• Taurine: the studied dose is 1 to 3 g a day, within the observed safe level of 3 g.^[52] The MELAS evidence is real but small, and the longevity case weakened in 2025.^[37]

Tier 5: preclinical or contested.

• NMN: its mechanism is comparable to NR's, with early signals in small trials,^{[31, 32]} but it is not authorised in the EU (as of October 2026), so it may not be sold as food in Germany or Austria.^[66]; NR is the legal way to try the same idea.

• PQQ: a typical consumer dose is 10 to 20 mg a day (adults only, not in pregnancy or breastfeeding^[49]), with small memory trials and no human evidence of mitochondrial building^[84]

• Fisetin: the main human trials, at the Mayo Clinic, have not published results yet, so it makes sense to wait for them.

• Resveratrol: later testing did not support the idea that it directly activates SIRT1,^[105] so if you want to try it, stay within the authorised adult maximum of 150 mg a day.^[107] and treat it as an experiment.

If you only do one thing from this guide, make it Tier 0. If you only buy one thing, a decent multivitamin with riboflavin, B12 and folate, plus magnesium, covers the basics, and anything beyond that is easier to judge side by side. A short stack you have checked against the table below is a better stack than a long one you have not.

The main supplements at a glance

The table puts the main supplements side by side. Compare the evidence column with the dose column, because several trials used far more than a typical capsule.

Sources: the trials and regulations cited in each row.

The mechanism stories behind these supplements are real, and you get the most out of them by matching each one to the outcome it has shown in people, because a mechanism is not an outcome. A label that turns "needed for the energy chain" into "extends your healthspan" skips the randomised trial in people, the step at which NR failed in the long Parkinson's trial^[62] So before you buy anything from Tiers 3 to 5, a reasonable check is to find its row in the table above. Then compare the dose and the claim on the box with the dose studied and the evidence column.