Working Paper · Nutritional Biochemistry
Grey hair is a stem-cell problem, and baldness is mostly a hormone-sensitivity problem. Nutrition, thyroid and stress sit on top of both. When grey hair has come back in people, it's usually because a specific deficiency or disease was fixed. The antioxidant products have much weaker evidence than their marketing suggests.
Hair gets its colour from melanocytes in the follicle bulb. Each hair cycle, those melanocytes are restocked from a pool of melanocyte stem cells. Greying happens when that pool runs out, stops working, or is killed off. Oxidative stress (hydrogen peroxide building up as catalase fails), sympathetic-nerve bursts during stress, and mitochondrial failure are all proven ways to deplete it in mice or in human follicles. Genes set the schedule. Deficiencies and disease can push it earlier, including low B12, low iron stores and hypothyroidism, and copper matters in theory. Those are also the cases where grey hair has actually come back.
Baldness is a separate problem. Pattern baldness (androgenetic alopecia) is follicles shrinking under DHT in people genetically sensitive to it. Diffuse shedding (telogen effluvium) is the form driven by nutrition, thyroid, stress and crash diets. The drugs work for pattern baldness. Nutrients work for shedding when something is actually missing. Most of the confusion comes from mixing up those two problems, and from treating product trials run by the product's maker as independent evidence.
Seven parts
Hair is a readout of the terrain. The best-supported way to keep your colour or your hair is to find what's actually wrong. Stacking antioxidants hoping one hits is much weaker.
Parts I–III build the model. Part IV goes through the nutrients and compounds one at a time. Parts V–VI cover baldness, and Part VII is the practical summary. Supplements and nutrients that already have their own entries on this site are linked where they come up.
Strong Moderate Animal–in vitro Inferred Contested Speculative
Melanin is made from the amino acid tyrosine by tyrosinase, a copper enzyme, inside melanocytes that sit at the base of each growing hair. The melanocytes hand the pigment to the keratin cells building the shaft. Tyrosinase gets its copper from a copper pump (ATP7A, the Menkes protein) inside the cell. Without the pump, tyrosinase is made but stays inactive, which is why Menkes disease, a genetic copper-transport failure, produces pale, steely hair.3
Pigment is only made while the hair is growing (anagen). When the follicle rests and restarts, it needs a fresh set of melanocytes, and those come from melanocyte stem cells parked higher up the follicle.
Strongtextbook pigment biology; the copper-pump requirement is shown directly in cells.
The landmark 2005 Science paper by Nishimura and colleagues showed, in mice and in ageing human follicles, that greying is caused by melanocyte stem cells failing to maintain themselves. They either die when they should go dormant, or they mature in the wrong place and are used up.1 A 2023 Nature paper from NYU added a twist. These stem cells normally shuttle between compartments of the follicle, and with age more of them get stuck in the wrong compartment where they can't be triggered to make pigment cells.2 Stuck isn't dead. That's one reason some grey hair can come back (Part III).
Strongfor stem-cell exhaustion as the core mechanism, shown in mice and confirmed in human follicles; Animal–in vitrofor the "stuck cells" model, which so far is mostly mouse work.
Oxidative stress. Making melanin generates hydrogen peroxide, so pigment cells live under constant oxidative load. A University of Bradford group found hydrogen peroxide at millimolar levels in grey and white hair shafts. They also found catalase (the enzyme that breaks it down) and the methionine-sulfoxide repair enzymes almost absent in grey follicles. The resulting oxidation of a methionine in tyrosinase's active site slows the enzyme down.4 A Charité group found melanocyte death, a marker of mitochondrial DNA damage, and more oxidative stress in greying follicles. Adding an oxidant to cultured pigmented follicles killed melanocytes.5 A Wuhan study, part-funded by L'Oréal, found catalase switched off in both the pigment cells and their stem cells in grey follicles.6
Stress nerves. Harvard's Hsu lab showed in 2020 that in mice, acute stress greys hair through the sympathetic nerves wired to the stem-cell niche, not through cortisol and not through the immune system. A burst of noradrenaline makes the stem cells all activate at once, so the reserve is spent and gone for good.7 The stressor in that study was an injected pain-causing chemical, which is intense and acute.
Mitochondria. Deleting a mitochondrial DNA-maintenance enzyme in mice raised oxidative stress in the stem cells and greyed the coat. The surviving melanocytes still made plenty of pigment. The problem was the number of cells, not their output.32 NAC's effect in these mice is in Part IV.
Genes. The first genome-wide hit for greying, in 6,000 Latin Americans, was near IRF4, a gene that regulates melanin production.9 Family history is the most consistent risk factor in every case-control study below.
Strongthat oxidative damage, stress-nerve activity and mitochondrial failure can each deplete the pool experimentally; Inferredthat these are the main drivers of ordinary age-related greying in a given person, rather than the genetic clock.
A heat-sensitive tyrosinase. A common tyrosinase variant, R402Q (rs1126809), is carried by nearly half of people of European descent.11 It makes an enzyme that works normally at 31 °C but misfolds at body temperature and is trapped inside the cell instead of reaching the pigment granules.10 The variant is tied to lighter hair, skin and eyes. Whether it speeds greying hasn't been studied. One suggestion is that people with two copies may do better avoiding chronic heat, but that's untested.11
Cortisol suppressing the pigment hormones. ACTH and MSH, cut from the same precursor (POMC), drive pigmentation. Cortisol feeds back to suppress POMC, so chronically high cortisol could cut the pigment signal. MSH activation also needs vitamin C, copper, zinc and glycine.11 This is a different stress route from the nerve mechanism above, which worked without the adrenal glands.7 Both may be real.
The practical framing that follows is to find the weakest link and fix it. The candidates are energy signalling (thyroid, insulin sensitivity, protein, carbohydrate, stress), the energy machinery (B vitamins, iron, copper, magnesium) and antioxidant defence.11 Speculative
Almost everything here comes from small case-control studies of premature greying (before about 20–30, depending on ethnicity), mostly from India, Iran and Egypt. They compare blood levels in people who grey early with people who don't.
| Factor | What the studies found | Grade |
|---|---|---|
| Copper | Iran, 66 teenagers vs 66 controls: serum copper lower in the grey group (90.7 vs 105.3 µg/dL, p=0.048), and serum iron higher.12 Egypt, 60 vs 30: copper lower in greying patients, but it did not track with severity (p=0.74), while iron and calcium did.13 A 2020 meta-analysis of risk factors didn't list copper among the consistent ones.16 | Contested |
| Iron / ferritin | Ferritin lower in Indian schoolchildren with early greying (36.9 vs 60.9)14, and iron fell with severity in the Egyptian study13, but serum iron was higher in the Iranian group.12 The meta-analysis lists low ferritin as a consistent factor.16 | Moderate |
| Vitamin B12, folate | Listed as consistent risk factors in the meta-analysis16. Pernicious anaemia has been linked to premature greying for decades. B12 was not different in the Indian study.14 | Moderate |
| Vitamin D, calcium | All 35 Indian cases were deficient or insufficient, versus about two-thirds of controls. Serum calcium was lower.14 Calcium fell with severity in Egypt.13 | Moderate |
| Oxidative stress markers | Delhi, 50 cases vs 30 controls: MDA (a lipid-peroxidation marker) higher, glutathione and SOD lower. MDA and glutathione tracked with how grey people were.15 | Moderate |
| Thyroid | Thyroid disorders are on the meta-analysis list.16 Thyroid hormone directly raises pigment production in human follicles in culture.22 | Moderate |
| Smoking | First flagged in a 1996 BMJ survey18; among the most consistent factors in the meta-analysis.16 | Moderate |
| Heart disease | In Turkish men, the degree of premature greying independently predicted coronary artery disease on angiography, which the authors read as a marker of biological age.19 Obesity, high blood pressure, dyslipidaemia and inactivity also cluster with it.16 | Moderate |
These are cross-sectional snapshots of a few dozen people each, in populations where deficiency is common. They can't say whether low copper or low ferritin caused the greying, or whether both reflect something else like gut absorption, inflammation or diet quality. The copper case rests on one study where it was barely significant and one where it didn't track with severity. The pattern that does hold up across studies is broader than any one nutrient: people who grey early are, on average, less well nourished and less metabolically healthy. That fits the terrain framing of the Metabolic Health paper better than it fits a single-nutrient fix.
Reviews of the field open the same way: grey hair is generally thought permanent, and there's no proven treatment other than dye.24,17 Reversal does happen, though, and where it's been documented tells you what's worth trying. Ranked by strength of evidence:
Treating the underlying deficiency or disease. The 2020 meta-analysis concluded that "treating the accompanying pathologies led to the reversal of the disease in many cases."16 Examples: premature greying reversed on oral iron in a Korean patient20. White hair darkened in two patients given T3 for hypothyroidism, one of them in myxoedema coma.21 B12 deficiency is on the meta-analysis's list of treatable causes.16
A systematic review found 27 reports of drug-induced repigmentation.23 Most were side effects of anti-inflammatories (thalidomide, adalimumab, prednisone, cyclosporine) or drugs that stimulate melanin (imatinib, erlotinib, tamoxifen, latanoprost, and levodopa, which darkened a Parkinson's patient's white hair25). None of these would be taken for grey hair. They point to inflammation and pigment signalling as levers.
Columbia's Picard lab mapped pigment along single hairs and found white hairs that regained colour naturally, across sexes, ethnicities and body sites. In some people it tracked with stressful periods and their end.8 Their model suggests hairs near the greying threshold can flip back when the load comes off. Hairs far past it can't. This is the most important finding for expectations: early grey is the grey most likely to come back.
Moderatethat correcting a real deficiency or thyroid disease can restore colour, based on case reports and the meta-analysis, not trials; Speculativethat any supplement reverses ordinary age-related greying in a well-nourished person.
Tyrosinase can't make melanin without copper, delivered inside the cell by the ATP7A copper pump. When that delivery fails, as in Menkes disease, hair comes out pale.3 In one study, teenagers who greyed early had lower serum copper than controls.12 Getting enough matters. Beef liver and oysters are by far the densest food sources (see Copper and Freeze-Dried Oyster).
More than enough doesn't help. Free copper is itself pro-oxidant, which is the opposite of what a greying follicle needs.11 Long-term high-dose zinc drains copper. In mice, high-dose oral zinc lightened new hair and cut its melanin content.33 Injected copper peptides (GHK-Cu) are being tried for grey hair. They grow hair in mice, but there are no human studies on colour.37
Strongfor copper's role in pigment; Contestedthat low copper explains most premature greying (Part II).
These are the deficiencies most consistently linked to early greying (Part II), and the ones where fixing a real shortfall has brought colour back. Premature greying reversed on oral iron in a reported case.20 B12 and folate deficiency are on the meta-analysis list of treatable causes.16 They are worth testing because they are cheap to fix and matter well beyond hair. See B12, Folate, Iron and Vitamin D.
Moderatecase reports and case-control data, in people who were actually deficient.
An underactive thyroid is a recognised cause of early greying. White hair darkened in two patients treated with T3, one recovering from myxoedema coma and one given too much thyroid hormone.21 In cultured human follicles, T3 and T4 directly raised pigment production.22 In people with an underactive thyroid, two herbs raised T3 in small placebo-controlled trials:
Neither herb has been studied for hair.
Moderatefor treating hypothyroidism restoring colour; Inferredthat thyroid support helps hair colour when the thyroid is already normal.
NAC. In mice whose mitochondria were broken by deleting Dguok, NAC in drinking water (1 mg/mL from day 20 to day 70) restored the melanocyte stem cells and prevented most of the greying. Half that dose didn't work.32 By standard scaling, the effective dose works out to roughly 1 g a day for an adult. It was prevention, started before greying, in a genetic model. No human studies. See NAC.
Luteolin. A flavonoid found in celery, peppers and herbs. Fed at 0.5 mg/kg a day, or applied to the skin, it slowed greying in a mouse strain bred to grey early. It did this by keeping the follicle's stem-cell signalling (endothelin to its receptor) from fading with age. The work was part-funded by Hoyu, a Japanese hair-dye company. No human studies.31
Topical antioxidant serums. One placebo-controlled trial exists. In 44 men over four months, an antioxidant ingredient from Givaudan, which ran the trial, reduced the proportion of grey hair by 17% versus 8% on placebo.28 Two other serums reported less grey in open-label studies with no placebo arm. One was a multi-ingredient serum containing rosemary oil and the anti-grey ingredient MelanoGray, funded by its manufacturer.29,30
Diet. People who grey early have higher MDA, a marker of oxidised fat, and lower glutathione.15 Fewer seed oils (Linoleic Acid), enough protein, glycine and cysteine for glutathione (Glycine, Cysteine), and fruit all lower the oxidative load the follicle sits under.
Animal–in vitrofor NAC and luteolin; Moderatefor the one placebo-controlled topical trial, with the maker running it.
Both come from an older literature. In 39 young women, calcium pantothenate (B5) 200 mg a day combined with plucking the grey hairs reduced grey counts in most of those followed long-term. There was no control group, and the authors credit the plucking as much as the vitamin.27 PABA darkened grey hair in 5 of 20 patients, but at 12–24 g a day taken as a drug for other diseases.26 A 2020 meta-analysis gives both a low-grade recommendation.16 See B5.
Speculativeuncontrolled reports.
Stress greys hair through the sympathetic nerves wired to the stem-cell niche.7 In people, grey hairs near the threshold have regained colour as stress eased.8 Of everything on this list, managing stress is the lever with direct evidence for reversal in ordinary people, without a deficiency to fix.
Moderatemouse mechanism plus human hair-mapping data.
Androgenetic alopecia (AGA), meaning male-pattern or female-pattern hair loss, is follicles miniaturising over many cycles, driven by DHT in genetically sensitive scalp. Temple and crown in men, widening part in women. Each cycle the hair comes back thinner and shorter until it's vellus fuzz. The follicle is still there for years, which is why treatment works best early.38
Telogen effluvium (TE) is diffuse shedding, usually two to three months after a trigger: illness, surgery, childbirth, crash dieting, low protein, iron deficiency, thyroid disease, major stress. Many follicles switch to resting at once. It is usually reversible once the trigger is fixed.49 A small study of medical students before a major exam found stress shifted immune markers and hair-cycle parameters, though within the normal range.52
The distinction matters because the evidence splits cleanly. Nutrients fix shedding when something is missing. They don't reverse pattern loss. A 2024 systematic review found deficiencies in vitamin D, iron, zinc and B vitamins associated with AGA, but the evidence was "not entirely consistent."50 (Alopecia areata, patchy autoimmune loss, is a third condition and out of scope here.)
Strongfor the AGA/TE distinction and DHT's role in AGA.
This is where hair connects to the rest of the series. A Finnish case-control study found men who started balding before 35 had a "strikingly increased risk" of hyperinsulinaemia and insulin-resistance disorders: obesity, hypertension and dyslipidaemia.53 A meta-analysis of 37,000 men found severe baldness carried a 32% higher risk of coronary heart disease, rising to 48% for severe vertex (crown) baldness.54 Frontal recession wasn't linked. Premature greying shows the same pattern with heart disease (Part II).
Moderateobservational, but consistent across cohorts; Inferredthat improving insulin sensitivity slows hair loss. That hasn't been tested directly.
Early crown balding or early greying is worth reading as a prompt to check fasting insulin, lipids and blood pressure, the same terrain the Metabolic Health paper works on. The hair isn't the disease. It may be an early, visible readout of it.
DHT doesn't act on the cells that build the hair. It acts on androgen receptors in the dermal papilla, the small cluster of signalling cells at the base of the follicle that tells the hair-making matrix cells when to grow and when to stop. In balding scalp, DHT makes those papilla cells send out three "stop" signals:
Balding scalp is also mildly inflamed. Gene-expression data from bald versus haired scalp of the same men show more T cells, mast cells and neutrophils in the bald patches.58 And in dermal papilla cells, DHT raises reactive oxygen species and damages mitochondria. Three different antioxidant approaches blunted this in cells or mice and lowered the DKK-1, TGF-β1 and IL-6 that DHT induces:
So DHT is the trigger and oxidative stress and inflammation carry its signal. DHT is still upstream of all of it in genetically sensitive follicles, which is why lowering it with finasteride regrows hair.38 The downstream steps explain why red light, which has placebo-controlled trials, works, and why antioxidant and anti-inflammatory approaches are reasonable additions to the drugs rather than replacements.
Strongfor DHT acting through dermal papilla signals, replicated in human balding papilla cells; Animal–in vitrofor antioxidants blocking those signals. No human studies yet, apart from red light.
Linoleic acid, the main fat in seed oils, is easily oxidised into lipid hydroperoxides (see Linoleic Acid). Applied to mouse skin, linoleic-acid hydroperoxides pushed follicles into the regression phase early and caused cell death in the hair matrix. They also killed human skin keratinocytes in culture.62 Unoxidised linoleic acid did the opposite on dermal papilla cells: it switched on Wnt growth signalling and lowered DKK-1.63 The damage comes from the oxidation product, which is the same substrate-times-oxidant logic as the cardiovascular case. Linoleic acid is the raw material for that oxidation, and the more of it there is in tissue, the more peroxide an oxidative hit can produce. People who grey early have higher MDA, a breakdown product of oxidised linoleic acid, in their blood (Part II).15
Animal–in vitrofor lipid peroxides killing follicle cells; Inferredthat a diet high in seed oils speeds hair loss. There are no human studies of seed-oil intake and hair.
| Treatment | Evidence | Grade |
|---|---|---|
| Finasteride 1 mg (oral, men) | Blocks conversion of testosterone to DHT. In 1,553 men over two years, hair counts ran 107–138 hairs above placebo in a 1-inch circle, and the placebo group kept losing hair.38 The trials were Merck's. Sexual side effects occur in a minority, and some reports describe them persisting. The label also warns about depression. | Strong |
| Minoxidil (topical 2–5%, or low-dose oral) | Beats placebo in men (2% and 5%) and women (2%) on meta-analysis.39 Low-dose oral minoxidil (≤5 mg) didn't meaningfully change blood pressure on meta-analysis. Heart rate rose slightly, and unwanted body hair was the main reason people stopped.40 | Strong |
| Low-level red light (e.g. 630–660 nm devices) | Beats placebo in men on meta-analysis.39 Device-maker trials dominate. | Moderate |
| Microneedling + minoxidil | 100 men: weekly microneedling plus minoxidil grew 91 new hairs/cm² vs 22 on minoxidil alone at 12 weeks, with evaluator blinding.41 A pilot, but a large effect. | Moderate |
| Pumpkin seed oil 400 mg/day (oral) | 76 Korean men, double-blind, 24 weeks: hair count up 40% vs 10% on placebo.42 The product was a branded capsule (Octa Sabal Plus). One trial, not replicated. | Moderate |
| Rosemary oil (topical) | 100 men, 6 months: hair count rose equally with rosemary oil and 2% minoxidil, and neither changed at 3 months.43 No placebo arm, and 2% is the weaker minoxidil. A 2025 review of blinded non-prescription trials counts it alongside pumpkin seed oil, caffeine and saw palmetto as promising but thin.46 | Moderate |
| Caffeine (topical) | In 210 men, a caffeine liquid left on the scalp twice daily matched 5% minoxidil on the share of growing hairs over 6 months. The trial was open-label with no hair counts, and was sponsored by Alpecin's maker.44 | Moderate |
| Niacin derivatives (topical) | 60 women with pattern loss, double-blind: octyl and tetradecyl nicotinate improved hair fullness on photos over 6 months (p=0.04).45 A pilot. | Moderate |
| Iron, vitamin D, zinc, protein | Help when low, mostly for shedding.49,50 Check ferritin before assuming. In women with pattern loss, oral vitamin D added to minoxidil beat minoxidil alone.47 See Iron, Vitamin D, Zinc. | Moderate |
| Collagen + amino acids (oral) | 83 people: a tablet of fish collagen, taurine, cysteine, methionine, iron and selenium added to hair-loss drugs improved blinded photo scores more than the drugs alone over 3 months. The authors work for the company selling it.48 See Glycine & Collagen. | Moderate |
| Biotin | All 18 reported successes were in people with an underlying deficiency or disorder. No evidence in healthy people.51 It also skews thyroid blood tests (see Biotin). | Speculative |
| GHK-Cu | Cultured follicles and mice only.36,37 | Animal–in vitro |
The site's bias runs toward terrain over drugs, and the metabolic link in V.2 is real. But for pattern baldness, finasteride and minoxidil have multi-thousand-person placebo-controlled trials, and nothing in the natural column comes close. Pumpkin seed oil and rosemary each rest on one trial. The honest position is that fixing the terrain and the approved drugs aren't competing options. Someone losing hair to AGA who wants to keep it should know the drugs exist and what they cost in side effects, and decide.
Nutrients: Copper · Zinc · Iron · B12 · Folate · B5 · Biotin · Vitamin D · Vitamin A · Selenium · Cysteine · Glycine & Collagen
Supplements: NAC · Black Seed · Pomegranate · Freeze-Dried Oyster. No entry yet for luteolin, pumpkin seed oil, rosemary oil, topical caffeine or GHK-Cu.
What would move this paper most is a randomized trial in people with early, recent greying, testing a deficiency-correction protocol against placebo. That's the only intervention with a consistent human signal, and nobody has run it.