The niche science, made simple

In papilla cells, minoxidil lowered androgen receptor, CYP17A1 and DHT. A 2023 study used network pharmacology to predict that minoxidil acts on the androgen receptor, CYP17A1 and aromatase, then tested it in human dermal papilla cells given testosterone. Minoxidil lowered AR and CYP17A1 expression and raised aromatase activity; culture-fluid DHT fell from 346.86 to 267.43 pg/mL and estradiol rose from 12.01 to 13.22 pg/mL against testosterone-only controls. It is cell culture with no hair outcome, so it does not show that minoxidil works as an antiandrogen in people.

Minoxidil cut 5-alpha-reductase type 2 gene expression to 0.22-fold in cells. In a 2017 cell-culture study, human keratinocytes (the HaCaT line) treated with minoxidil expressed the type 2 5-alpha-reductase gene at 0.22 times the untreated level (p < 0.0001). That is an antiandrogen-like effect on the same enzyme finasteride blocks. It is mRNA in a skin cell line, though, with no enzyme activity, DHT level or hair outcome measured.

Low follicular sulfotransferase predicted a better response to oral minoxidil. Among 41 oral minoxidil users, 63.4% improved, and people with low follicular sulfotransferase activity responded more often (85% versus 43%). That is the opposite of what the topical prediction implies, where more enzyme should mean more drug activation. It suggests topical response tests do not simply transfer to tablets.

Dutasteride 0.5 mg beat finasteride 1 mg on hair count and width in 917 men. In a 917-man trial, 0.5 mg dutasteride beat 1 mg finasteride on hair count, hair width and frontal photographs, with similar adverse events. Check who funded the trial before leaning on it.

Dutasteride lowered serum DHT 94.7 to 98.4%, finasteride 70.8%. In 399 men with prostate enlargement over 24 weeks, dutasteride 0.5 mg lowered serum DHT by 94.7% and 5 mg by 98.4%, against 70.8% for finasteride 5 mg (P < 0.001), with less variability. Testosterone rose but stayed in the normal range. This was not a hair trial, but it explains why dutasteride grows more hair per milligram.

Men told about side effects reported sexual problems 43.6% versus 15.3%. In 120 men on finasteride, those who were told about sexual side effects reported them far more often (43.6%) than those who were not (15.3%). That is a nocebo effect, but it does not mean the drug has no effect. Reported rates depend heavily on how the question is asked.

Men born without type 2 5-alpha-reductase rarely go bald and have sparse beards. Men born without the type 2 enzyme have low DHT, sparse beards, and rarely develop baldness or acne, according to a review of the 5-alpha-reductase family. The original 1974 descriptions of a Dominican kindred showed partial virilisation at puberty. This human genetic experiment is the proof of concept for finasteride.

In balding scalp, the goosebump muscle was replaced by fat in all 8 samples. In scalp biopsies (18 in total: 8 pattern baldness, 5 telogen effluvium, 5 normal), the arrector pili muscle was replaced by fat in all 8 balding samples. That makes it one more tissue change in pattern hair loss, beyond the follicle itself. Whether the muscle loss causes thinning or follows it is unknown.

Bald scalp keeps hair stem cells but lacks the progenitors that follow them. Flow cytometry of bald and non-bald scalp from men with pattern loss found similar numbers of KRT15-high stem cells, but the larger, faster-dividing CD200-high and CD34-high progenitor populations were markedly reduced in bald scalp. The defect looks like a failure to convert stem cells into progenitors, not a loss of stem cells.

Plucking enough hairs made up to five times more neighbours regrow in mice. In mice, plucking hairs at different densities regenerated up to five times more neighbouring unplucked resting hairs, via CCL2 from injured follicles and TNF-alpha-secreting macrophages over about 1 mm. Skin ignores mild injury but responds to stronger injury with cooperative stem cell activation. Whether microneedling triggers the same response in a human scalp is untested.

Cancer immunotherapy darkened white hair in 13 of 14 patients. A Spanish case series of 14 lung-cancer patients on anti-PD-1 or PD-L1 drugs reported diffuse hair darkening in 13, and 13 had at least stable disease. It was uncontrolled and reflects a drug with serious immune effects. It is of scientific interest for how immune signalling controls follicle pigment.

A scalp alpha-agonist raised plucking force 172% and cut shedding 42%. A related stimulant applied to the scalp of women aged 18 to 40 raised the force needed to pluck a hair by about 172% and cut shedding by about 42% on treatment days. That shows hairs held tighter by the arrector pili muscle, not more hair growing. Midodrine, a tablet, is studied in the same context.

Clascoterone matched finasteride in a receptor assay in papilla cells. In cultured human scalp papilla cells, clascoterone inhibited androgen-driven receptor activity about as well as finasteride in a reporter assay and blocked DHT-induced IL-6 release better than enzalutamide. The work was company-affiliated.

Cyclosporine's hairy side effect was traced to a Wnt brake called SFRP1. In human scalp follicles in organ culture, cyclosporine lowered SFRP1, a Wnt brake in the dermal papilla, and the blocker WAY-316606 reproduced the effect, boosting hair shaft production and delaying regression. It is a lab route to hair-growth drugs that skips cyclosporine's toxicity. It is lab work.

Finasteride 0.2 mg worked, and 0.01 mg did no better than placebo. A dose-ranging trial found 0.2 mg and above worked, with 1 mg and 5 mg similar to each other and better than 0.2 mg, while 0.01 mg was no better than placebo. The threshold for effect is therefore well below the 1 mg dose sold for hair loss. It is a reason some clinicians discuss lower or intermittent doses.

A bigger early shed went with better results on topical minoxidil. In 49 patients using 2% or 5% minoxidil for 24 weeks, shedding rose temporarily within the first 12 weeks, and for longer on 2% than 5%. The size of the shed correlated with improvement in hair-loss grade in both groups, and with trichoscopy improvement on 5%. It is a small retrospective study, but it supports the idea that the early shed is old hairs making way for new ones.

Most of the scalp's minoxidil-activating enzyme sits inside the follicle. Minoxidil has to be sulfated before it works. In scalp skin from nine stumptailed macaques the follicle held 47 to 84% (mean 61%) of the sulfotransferase activity, against a mean of 18% in the epidermis and 21% in the dermis. The activating machinery is concentrated where the drug has to act, which is why follicle enzyme levels have been studied as a predictor of response.

Minoxidil sulfate was 14 times more potent than minoxidil in cultured follicles. In cultured mouse whisker follicles, the sulfated metabolite was 14 times more potent than minoxidil itself. Drugs that block sulfation stopped minoxidil from working but did not stop minoxidil sulfate. That is the evidence behind calling minoxidil a prodrug and behind interest in SULT1A1, the follicle enzyme that makes the sulfate.

Minoxidil activated the Wnt effector beta-catenin in papilla cells. In depilated mice, 3% minoxidil applied daily for 10 days modestly extended the growth phase and delayed catagen. In cultured human dermal papilla cells it raised beta-catenin reporter activity, nuclear beta-catenin and the targets Axin2, Lef-1 and EP2. Beta-catenin is the signal that keeps follicles in growth, which fits minoxidil's anagen-prolonging effect, but this is mouse and dish data.

Oral minoxidil 1 mg raised serum VEGF in a small 2025 study. In a prospective study of 50 people, serum VEGF after 12 weeks was 217.88 pg/mL on oral minoxidil 1 mg/day versus 142.81 in controls, alongside better hair count and diameter. The reported correlation between VEGF and hair count (r = 0.9965) is implausibly tight for a sample that size, and the study had a single author, so treat it as a hypothesis. VEGF has long been proposed as a minoxidil pathway on the basis of cell work.

Minoxidil's potassium-channel mechanism had never been shown in the follicle. A 2004 review noted that opening of KATP potassium channels by minoxidil sulfate is the favoured explanation, yet there was no clear demonstration that these channels are expressed in the hair follicle. Cell culture shows more cell growth, less collagen synthesis and more VEGF and prostaglandin output, and the authors warned that applying such results to a whole follicle is uncertain. In animals it shortens telogen and probably prolongs anagen.

A potassium-channel gene mutation acts like a lifelong dose of minoxidil. Cantu syndrome, which includes excess body hair from birth, is caused by gain-of-function mutations in ABCC9, a gene for a KATP potassium-channel subunit. That is the channel family opened by the minoxidil metabolite and by diazoxide, two drugs that also cause hypertrichosis. It is the closest genetic support for the channel theory of minoxidil.

Tretinoin cream raised minoxidil absorption almost threefold in men. In 19 healthy men, 0.05% tretinoin cream applied before 2% minoxidil raised minoxidil absorption almost threefold, against 1.3-fold with a plain vehicle cream. More absorption means more drug in the skin, but also more irritation and more systemic exposure. Whether it means more hair is a separate question that this absorption study did not test.

Oral minoxidil showed a dose-response in a six-study meta-regression. A 2022 meta-regression of six studies found each extra 1 mg/day of low-dose oral minoxidil went with 47.1 more total hairs/cm2 and 9.1 more terminal hairs/cm2 at six months, but also a 17.9-point higher hypertrichosis risk and a 4.8-point higher cardiovascular adverse-event risk. These are cross-study associations, not a randomised dose comparison. The one direct 2.5 mg versus 5 mg trial found almost no difference.

Oral minoxidil 2.5 mg matched 5 mg in men, with fewer side effects. In 100 men randomised to 2.5 mg or 5 mg a day for 24 weeks, non-vellus hair density differed by only 0.9 hairs/cm2. Ankle swelling and dizziness were more common on 5 mg. In this range, a higher dose bought side effects rather than extra hair.

Hypertrichosis rises with oral minoxidil dose, and women get it at lower doses. Pooled data in a 2025 review showed hypertrichosis in 6.7% of people on 0.25 mg versus 56.1% on 5 mg a day. Women developed it at lower doses than men (mean 1.4 mg versus 4.1 mg), and fluid retention appeared in 1.3 to 10%, usually as ankle swelling. Dose, not just sensitivity, sets how much unwanted hair appears.

Low-dose oral minoxidil did not change blood pressure in a meta-analysis. Pooling studies of low-dose oral minoxidil showed no significant change in systolic or diastolic blood pressure. The only measurable effect was a small heart-rate rise of 2.67 beats per minute. This is a reassuring null, though individual sensitivity and interacting drugs still matter.

Four head-to-head trials found oral and topical minoxidil equal on hair density. A meta-analysis of four trials (279 people) found no difference in hair density (standardised difference 0.02), while oral minoxidil roughly doubled the risk of hypertrichosis (relative risk 2.01). An earlier network meta-analysis had ranked oral 5 mg highest, but only through indirect comparisons. For many people oral minoxidil is a convenience choice, not a stronger one.

In 393 men, 5% minoxidil grew about 45% more hair than 2%. Over 48 weeks, 5% solution beat both 2% and placebo, with about 45% more regrowth than 2% at week 48 and a faster onset. Itch and irritation were more common on 5%. The dose-response here is real in men, which contrasts with the result in women.

In women, 2% and 5% minoxidil performed the same in a Cochrane review. A Cochrane review covering 5,290 women found minoxidil added about 13.2 hairs/cm2 over placebo, with no difference between 2% and 5%. The 381-woman Lucky 2004 trial agreed on counts and found 5% better only on self-rated benefit, with more itch and facial hair. Women gain little from the stronger strength.

Minoxidil's gains were gone within 24 weeks of stopping. After 96 weeks of treatment, hair weight and counts fell back to placebo levels within 24 weeks of stopping. That trial tested topical minoxidil; the gains need continued use to hold.

Only about 1.4% of topical minoxidil is absorbed through healthy scalp skin. A 2025 review reports about 1.4% absorption through healthy scalp. By simple arithmetic, 1 mL of 5% solution carries 50 mg, so about 0.7 mg reaches the bloodstream. Swallowing 60 mL of 5% solution, by contrast, caused circulatory shock in several reports.

A 3% minoxidil lotion beat placebo for beard density in a 16-week trial. In 48 men, 3% lotion twice daily beat placebo on hair count, photographs and self-rating over 16 weeks. Use on the beard is off-label, and the trial ran only 16 weeks.

Minoxidil plus finasteride improved 94.1% of men, versus 80.5% and 59% alone. In 450 Chinese men followed for 12 months, 94.1% improved on the combination, 80.5% on finasteride alone and 59% on 5% minoxidil alone. The two drugs attack different arms of the problem, DHT blockade and growth stimulation.

Compounded minoxidil capsules held 10 to 100 times the prescribed dose. Twelve severe reactions (0.7% of about 1,700 prescriptions) came from capsules holding 10 to 100 times the intended dose. Where low-dose tablets are unavailable, pharmacies compound capsules. This is a pharmacy-quality risk rather than a drug risk.

Cats got sick from licking an owner's skin or pillowcase after minoxidil use. In a poison-centre series of 211 dog and cat cases, signs followed even a few drops or licks, often from an owner's skin or bedding. Of 62 cats with symptoms, 8 died. Wash hands after application and keep treated skin and pillows away from pets.

Finasteride targets type 2, found in the follicle's root sheaths. Type 1 5-alpha-reductase is the predominant scalp enzyme overall, which made finasteride's effect puzzling since it blocks mainly type 2. Immunostaining of scalp from men with pattern loss placed type 2 in the outer and inner root sheath of the follicle, while type 1 was in the sebaceous glands and not in the follicle. That supports a local follicle effect. The authors were at the drug's manufacturer.

Papilla cells in culture expressed mostly type 1, and beard cells the most. Cultured dermal papilla cells from beard and scalp expressed mostly type 1 5-alpha-reductase mRNA, highest in beard cells, with minimal type 2. Testosterone was converted mainly to androstenedione and only a little DHT, and a type 1 inhibitor beat finasteride in beard papilla cells. Tissue staining and cell culture disagree about which enzyme matters, which is part of the case for dual inhibitors.

Dutasteride lingers for weeks; finasteride was cleared far faster in animals. In a 1997 pharmacokinetic study, dutasteride's terminal half-life was 14 hours in rats and 65 hours in dogs, against 1 and 4 hours for finasteride, and about 240 hours after a single human dose. The Avodart label gives about 5 weeks. The long tail explains both the fuller DHT suppression and the slow recovery after stopping.

Dutasteride 0.2 mg matched 0.5 mg in a Korean trial. In 139 men, 0.2 mg dutasteride gained 21.5 hairs versus 6.0 on placebo and was as effective as 0.5 mg on all endpoints. Lower doses mean less drug exposure. In the US, dutasteride is labelled only for prostate enlargement, so use for hair there is off-label.

Untreated men lost about 26% of their target-area hair over five years. In the placebo arm of the pivotal finasteride trials, men lost about 239 hairs (26.3% of density) in the target area over five years, and 75% looked worse on photographs. Finasteride men held what they gained. The comparison puts the size of the treatment effect in context: it is largely stopping loss.

After 10 years of finasteride, 42.8% of men aged 20 to 30 had not improved. In an uncontrolled series of 118 men on 1 mg daily, benefit was maintained, and men over 30 and with higher grades of loss responded better. Of men aged 20 to 30, 42.8% had not improved by 10 years, and 6% reported side effects. Young men with early loss may be holding steady rather than regrowing.

Topical finasteride matched oral on hair, with 100-fold lower blood levels. In 458 men, a topical finasteride spray gained about 20.2 hairs versus 6.7 on placebo, similar to oral finasteride. Plasma finasteride was more than 100 times lower, and serum DHT fell 34.5% on topical versus 55.6% on oral. Approval status varies by country, so check locally.

Topical finasteride cut scalp DHT about 50% but serum DHT only 25%. In a 50-man trial, 100 to 200 microlitres lowered scalp DHT by about 47 to 52% and serum DHT by 24 to 26%. Larger volumes lowered serum DHT 44 to 48%, close to the oral dose. More liquid on the scalp means more of the drug enters the blood, so volume matters.

Dutasteride lowered serum DHT 94% and sperm count 29% at 26 weeks in a 99-man trial. In 99 healthy men given dutasteride 0.5 mg, finasteride 5 mg or placebo for a year, dutasteride lowered serum DHT 94% versus 73% on finasteride. On dutasteride, total sperm count fell 28.6% at 26 weeks, 24.9% at 52 weeks and was still 23.3% lower 24 weeks after stopping; semen volume fell about 30%. (On finasteride 5 mg, sperm count fell 34.3% at 26 weeks and 6.2% at follow-up.) The authors called the changes mild and reversible after stopping. Men trying for a baby should know this before choosing a 5-alpha-reductase inhibitor.

Sexual dysfunction risk: finasteride 1.66-fold, dutasteride 1.37 (n.s.). A meta-analysis of 4,495 participants found finasteride 1 mg raised sexual dysfunction risk (relative risk 1.66, 95% CI 1.20 to 2.30), while dutasteride's relative risk of 1.37 was not statistically significant. Dutasteride suppresses DHT more, so the difference runs against intuition. Few dutasteride trials exist, so wide uncertainty remains.

Men on finasteride make normal amounts of sebum. People with complete androgen insensitivity made no measurable sebum, yet men with very low DHT or taking finasteride made normal amounts. Sebaceous glands mainly use the type 1 enzyme, which finasteride inhibits poorly. So finasteride should not be expected to fix an oily scalp.

5-ARI depression risk: 1.61 versus non-users, 0.89 versus alpha-blocker users. A 2026 meta-analysis of over 2.5 million men found a depression hazard ratio of 1.61 against non-users but 0.89 against men taking alpha-blockers for the same prostate condition. Who the comparison group is changes the answer. Regulators nevertheless added suicidal ideation to finasteride tablet labels.

Finasteride 1 mg did no better than placebo in 137 postmenopausal women. In a trial of 137 women with female pattern hair loss, finasteride 1 mg showed no benefit over placebo on any endpoint. Higher doses are used in some women in small uncontrolled series, but the controlled evidence is thin. Finasteride is also unsuitable in anyone who could become pregnant.

DHT grips the androgen receptor twice as hard and lets go five times slower. In human cells, DHT bound the androgen receptor about 2-fold more tightly than testosterone and dissociated about 5-fold more slowly. That is why a small amount of DHT can do the work of much more testosterone, and why lowering DHT matters more than lowering testosterone for pattern loss. It also explains why a receptor blocker has to out-compete DHT continuously.

The same androgen enlarges beard follicles and shrinks scalp follicles. Reviews of papilla-cell work show androgens trigger IGF-1 release from beard papilla cells but TGF-beta1, TGF-beta2, DKK-1 and IL-6 from balding scalp papilla cells. The difference is thought to be set in the embryo and runs through the dermal papilla. These paracrine signals are the leading explanation for the paradox.

A variant near the AR gene: 98.1% of young bald men, 76.6% of others. In 553 men, a variant near the androgen receptor gene was found in 98.1% of young bald men and 76.6% of non-bald men. Most non-bald men carry it too, so it is a risk marker rather than a cause by itself.

Baldness heritability is about 0.62 and spread across 624 genome regions. A UK Biobank study of 205,327 men estimated pedigree heritability at 0.62, with 624 near-independent genetic regions, 26 of them on the X chromosome. Hundreds of small effects explain why no single gene test is decisive.

Papilla cells from balding scalp aged prematurely in the dish. Papilla cells from balding scalp underwent premature senescence in culture, expressed the ageing marker p16INK4a and secreted more of the hair-growth inhibitors TGF-beta1 and TGF-beta2 under oxidative stress. They also handled oxidative stress less well than cells from the same men's occipital scalp. This ties pattern loss to cell stress, not only to hormones.

Testosterone raised collagen output in scalp fibroblasts; finasteride blunted it. In cultured scalp fibroblasts, testosterone raised TGF-beta1 (by 81.9%) and collagen production, and finasteride blunted that effect. Fibrosis around follicles may be part of why late-stage loss responds poorly to drugs, but that link is a hypothesis.

Oral antiandrogens regrew hair in 44% of women; no hormone level predicted who. In 80 women with female pattern hair loss treated with spironolactone 200 mg or cyproterone acetate, 44% regrew, 44% stayed stable and 12% kept losing hair. Results were similar for both drugs and no hormone level predicted response. The series was open with no placebo.

Spironolactone improved 65.8% in combinations but only 43.2% as monotherapy. A meta-analysis of spironolactone in female pattern hair loss found 56.6% improved overall: 65.8% when combined with other treatments and 43.2% as monotherapy. Heterogeneity was high, and observational studies were mixed with RCTs. It suggests spironolactone is better as an add-on than as a stand-alone.

Adding spironolactone to minoxidil raised density, but microneedling added more. In 120 premenopausal women with normal hormones, 24 weeks of 5% minoxidil plus oral spironolactone 80 to 100 mg raised density over minoxidil alone (p = 0.009), but less than minoxidil plus microneedling. Most side effects were in the spironolactone arm. A cheap procedure matched a prescription drug here.

Hyperkalaemia on spironolactone was 0.72%, the same as the background 0.76%. In 974 healthy young women taking spironolactone for acne, the hyperkalaemia rate was 0.72%, matching the background rate of 0.76%. The authors judged routine potassium checks unnecessary in this group. Older patients and anyone with kidney disease or interacting drugs are different.

Cyproterone lost to minoxidil: hairs fell by 2.4 against a gain of 6.5. In 66 women over 12 months, hairs thicker than 40 micrometres fell by 2.4 per 0.36 cm2 on cyproterone acetate plus ethinyl estradiol and rose by 6.5 on minoxidil 2%. Cyproterone did better for oily skin, acne and excess facial hair. It matters which problem is being treated.

High-dose cyproterone carried a 6.6-fold meningioma risk, 21.7-fold above 60 g. In a French cohort of 253,777 women starting high-dose cyproterone acetate, the adjusted hazard ratio for meningioma surgery or radiotherapy was 6.6 (95% CI 4.0 to 11.1) and 21.7 above 60 g cumulative dose. European regulators restricted products at 10 mg or more to androgen-dependent conditions. The risk is dose-related.

RU58841 reached the blood in six men after a single scalp dose. In six men given a single 30 mg topical dose, a doping-control lab found the parent drug and a metabolite in dried blood spots. That means considerable systemic exposure, though steroid profiles stayed normal. Macaque studies reported no systemic effects, and long-term human safety data do not exist.

Pyrilutamide reported +22.39 hairs/cm2 against +8.73 on placebo in 90 men. A Chinese company reported a 24-week phase II stage in 90 men in which 0.5% and 1.0% pyrilutamide raised non-vellus hair count by 22.39 and 21.87 hairs/cm2, against 8.73 on placebo, and later said its pivotal trial met its primary endpoint. No peer-reviewed paper has appeared yet. It is an androgen-receptor blocker applied to the scalp.

A scalp PROTAC that destroys the androgen receptor beat placebo at 2 of 4 doses. GT20029 tags the androgen receptor for destruction instead of blocking it. In a 12-week phase 2 trial of 180 men, 0.5% daily and 1.0% twice weekly beat their placebo on hair count (p = 0.032 and p = 0.023), and hair width improved for 1.0% twice weekly. Two of four dose groups reached significance, so the signal is encouraging but modest.

Creatine did not change DHT or hair in a 12-week trial of 45 men. A 2025 randomised trial of 45 resistance-trained men over 12 weeks found no differences in testosterone, DHT, DHT:T ratio, hair density, follicular units or thickness (38 completed). The baldness worry rests on an earlier hormone marker, not a hair outcome.

Every woman with pattern loss should be screened for androgen excess. A guideline states every woman with female pattern loss should be assessed for androgen excess, but isolated hair loss with normal androgens is not a sign of hyperandrogenism. Acne, excess facial hair and irregular periods change which antiandrogen is chosen. Hormone levels alone did not predict response to oral antiandrogens in the series above.

Saw palmetto's enzyme blocking came from its free fatty acids. In cloned human enzymes, the dual 5-alpha-reductase effect of a saw palmetto extract came only from its free fatty acids (about 80% of the extract). Oleic acid inhibited type 1 (IC50 4 micrograms/mL) far more than type 2 (over 100), lauric acid inhibited both, and sterols and esterified fatty acids did nothing. Extract quality and fatty-acid content therefore decide potency.

Saw palmetto grew hair in 38% of men against 68% on finasteride. In an open-label 100-man comparison over 24 months, hair growth increased in 38% on saw palmetto versus 68% on finasteride, judged on photographs. Randomisation is not stated in the abstract. Saw palmetto did less than the drug head to head.

Ketoconazole competes for the androgen receptor, but only at high concentrations. In human skin fibroblasts, ketoconazole displaced a synthetic androgen from the receptor with 50% displacement at 6.4 x 10^-5 M, and the interaction was competitive. Of four imidazole antifungals tested, only ketoconazole bound. The author noted that this concentration is unlikely to be reached in plasma, which makes a direct receptor effect from a shampoo or tablet speculative.

2% ketoconazole shampoo rivalled 2% minoxidil in one small trial. In a 1998 controlled study of men with pattern hair loss, 2% ketoconazole shampoo improved hair density, hair size and the share of growing follicles about as much as 2% minoxidil, and sebum fell. Participant numbers and funding are unclear from the abstract. No large trial has followed, so it is an adjunct at best.

Weekly ketoconazole shampoo cut 6-month dandruff relapse from 47% to 19%. In a 575-person trial of 2% ketoconazole shampoo twice weekly for dandruff and scalp seborrhoeic dermatitis, 88% had an excellent response. In the prophylaxis phase, relapse within 6 months was 19% with weekly ketoconazole versus 47% with placebo. Dandruff recurs unless a maintenance wash continues.

Dandruff tracks fatty acids Malassezia releases, not yeast numbers. Procter & Gamble researchers found that Malassezia numbers did not correlate with dandruff severity. Instead the yeast splits sebum triglycerides into irritating free fatty acids (oleic acid is the model), and individual sensitivity to them decides who flakes.

Zinc pyrithione kills dandruff yeast by smuggling copper into it. Zinc pyrithione carries copper into yeast cells and inactivates iron-sulfur proteins, stopping growth. This was shown in baker's yeast and confirmed in the scalp fungus Malassezia globosa. It explains a mechanism that was unknown for decades of use, and why copper levels in a product matter.

Stop an anti-dandruff shampoo and scalp biomarkers worsen in about 3 days. In a double-blind study, people switched from an anti-dandruff to a plain shampoo showed rising Malassezia load, barrier disruption and inflammatory and oxidative markers after about 3 days. Visible flaking did not increase until about 3 weeks. The study was industry-funded.

A scalp treatment changed the water-repellency of the hair that grew out. In a 24-week trial in people with dandruff, hair grown while the scalp was treated with zinc pyrithione had better surface water-repellency than hair grown on a placebo-treated scalp. The scalp environment therefore shapes the fibre before it emerges. It is an industry-funded study.

Free fatty acids on the scalp doubled from day 1 to day 10 after washing. In 67 people, the share of free fatty acids in scalp and hair lipids rose from about 19% on day 1 after washing to about 39% on day 10. The lipid mix on the scalp shifts toward the irritant type as days pass.

Washing three times a week instead of once did not speed up re-greasing. Changing shampoo frequency from once to three times a week had no significant effect on how fast hair re-fatted. In this older study, washing habits did not change how fast the scalp re-greased, so the idea that washing more makes the scalp oilier was not supported.

Men with pattern loss had more sebum triglyceride and Malassezia restricta. In 118 Japanese men, those with pattern hair loss had more sebum triglyceride and palmitic acid, more Malassezia restricta and a shifted bacterial pattern. It is an association only, and the authors were employees of a pharmaceutical company. Whether the scalp chemistry causes loss or follows it is unknown.

Wounding grew brand-new follicles in adult mice, and only if Wnt was active. After large wounds, genetically normal adult mice formed new hair follicles that produced shafts and cycled; blocking Wnt abolished it and overexpressing Wnt increased follicle numbers. This is the best-known lab argument for why controlled skin injury, such as microneedling, might stimulate hair. It is mouse work with large wounds, not microneedles on a human scalp.

Microneedling plus minoxidil: +38.3 hairs/cm2, versus +18.8 on minoxidil alone. In 60 Chinese men over 24 weeks, density rose 18.8 hairs/cm2 on 5% minoxidil, 23.4 on electrodynamic microneedling alone and 38.3 on both. Small trial, one site.

0.6 mm microneedling beat minoxidil alone, but 1.2 mm did not on hair count. In 60 people over 12 weeks, 0.6 mm microneedling every two weeks plus 5% minoxidil beat minoxidil alone on hair count and thickness, while 1.2 mm did not significantly beat it on count. Deeper is not clearly better.

Microneedling alone gave no density gain in 30 men, despite new collagen. In a 30-man randomised trial of monthly microneedling alone, with a roller or tattoo-cartridge device across four sessions, hair density did not rise, though biopsies showed new collagen. The benefit seen in other trials may depend on the drug it is paired with. Collagen remodeling is not the same as hair regrowth.

In 245 women, microneedling added nothing to 2% minoxidil. In an open-label trial of 245 women with female pattern hair loss, minoxidil 2% alone or with 0.5 to 0.55 mm microneedling every 2 or 4 weeks gave the same improvement in all three groups over 24 weeks. Microneedling's added value is therefore not proven in women. The open-label design leaves room for bias in either direction.

Hair laser combs are cleared on 'substantial equivalence', not drug-style proof. The FDA classifies hair laser combs as class II devices needing a 510(k), meaning substantial equivalence to an earlier device rather than drug-style efficacy approval. A clearance is not a statement that the device beat a drug. Check the trials, not the clearance.

Light-therapy hair gain tracked energy per session, not diode count. A systematic review found hair gain tracked energy per session and total exposure time rather than diode count, and pulsed light did better than continuous. More diodes is a marketing number, not a dose.

A meta-analysis of seven trials found laser added about 6.6 hairs/cm2 to minoxidil. A meta-analysis of seven trials found about 6.6 more hairs per cm2 and better satisfaction when low-level laser was combined with minoxidil. Small, varied trials make the size of any extra gain uncertain.

PRP activated ERK, Akt, FGF-7 and beta-catenin in papilla cells. In lab work, activated PRP increased dermal papilla cell proliferation, stimulated ERK and Akt signalling and raised FGF-7 and beta-catenin; injected into mice it sped the telogen-to-anagen transition. These are the pathways behind the PRP idea. They do not show that human trial gains are caused by platelets.

In split-scalp PRP trials, the placebo side kept about 37 extra hairs/cm2. A systematic review found that placebo halves in split-scalp PRP trials kept about 37 hairs/cm2 more than placebo in whole-head trials, suggesting PRP effects spread to the placebo side. A split design can hide a real effect.

A 30-man placebo-controlled PRP trial found no improvement in hair counts. A 30-man placebo-controlled trial of five PRP sessions found no improvement over placebo in hair number or thickness. One trial is not the whole picture, and PRP trials are mixed.

Caffeine reversed testosterone's growth suppression in human follicles. In balding men's follicles in organ culture, testosterone suppressed hair-shaft growth and caffeine at 0.001 to 0.005% reversed it. It is follicle-in-a-dish work.

Caffeine reached the blood in 5 minutes via open follicles, 20 if plugged. In healthy volunteers, topical caffeine reached the bloodstream after 5 minutes when follicles were open versus 20 minutes when they were plugged; with open follicles the blood level peaked at 11.75 ng/mL after 1 hour. The follicle opening is therefore the fast route into the skin for caffeine.

0.2% caffeine liquid matched 5% minoxidil on anagen ratio in an open-label trial. In 210 men over six months, the anagen ratio rose 10.59% with 0.2% caffeine liquid and 11.68% with 5% minoxidil, within the 5% non-inferiority margin. There was no placebo arm, the trial was open-label, and the manufacturer sponsored it. Non-inferiority against an unblinded comparator is weak evidence.

All nine caffeine hair trials were positive, but certainty was mostly low. A systematic review of nine trials found every one favoured caffeine, yet certainty was medium in three, low in one and very low in five, and the caffeine concentration was often missing. Positive results from product-tied trials are common. Independent replication is missing.

Adenosine 0.75% and minoxidil 5% both did little at three months. In 110 men, relative recovery was only 1.9% for adenosine and 2.4% for minoxidil at three months, with higher satisfaction on adenosine. Adenosine's effects are small and short-term.

A rosemary compound inhibited 5-alpha-reductase by 82.4% in the lab. Rosemary leaf extract inhibited 5-alpha-reductase by 82.4% at 200 micrograms/mL in enzyme assays, and restored testosterone-blocked regrowth in mice; the active compound was 12-methoxycarnosic acid. This explains why rosemary is called an antiandrogen. It does not show the effect on a human scalp.

Rosemary oil and 2% minoxidil: no change at 3 months, equal gains at 6. In 100 young men, neither rosemary oil nor 2% minoxidil changed hair count at three months, and both rose at six months with no difference. Itching was more frequent with minoxidil. The comparator was the weaker 2% strength and there was no placebo arm, and the trial has not been replicated.

The hair-follicle lab data credited to GHK-Cu belongs to a different peptide. The only hair-follicle culture study often cited for copper peptides used AHK-Cu, not GHK-Cu: follicles grew longer and papilla cells multiplied at 10^-12 to 10^-9 mol/L. Reviews describe GHK-Cu as a skin-remodeling signal with hair a minor theme, and no controlled human hair trial exists. Serum concentrations are rarely disclosed.

Niacinamide's famous VEGF hair claim failed in a cell test. Niacinamide is often said to boost VEGF around the follicle. A 2020 letter from Procter & Gamble researchers found it could not stimulate VEGF across a wide dose range in vitro. The authors work for a company that sells competing products.

In mice, telogen splits into a refractory phase and a ready-to-grow phase. In mice, dermal BMP2 and BMP4 cycle out of phase with Wnt/beta-catenin, dividing resting telogen into a refractory phase (high BMP) and a competent phase (low BMP). Overexpressing the BMP blocker noggin shortened the refractory phase, and BMP4 made competent skin refractory.

Goosebump muscle and sympathetic nerves tune how awake follicle stem cells are. In mice, sympathetic nerves and the arrector pili muscle form a niche for hair follicle stem cells; without norepinephrine signalling the stem cells entered deep quiescence. Pattern loss may involve the same niche, since in scalp biopsies from 8 balding people the muscle was replaced by fat in all 8. This is part of why midodrine and related alpha-agonists attract hair-loss interest.

In 14 men, shedding peaked in late summer at over double the winter rate. In 14 healthy UK men counted over 18 months, shedding averaged about 60 hairs a day at its August-September peak, more than double the winter rate. A late-summer rise in shedding is therefore part of a normal seasonal rhythm. This is a small, older study.

After shedding, follicles can sit empty for 4 to 7 months before regrowth. In 10 men followed for 14 years, follicles stayed empty (kenogen) for about 4 months, rising to about 7, in about 80% of hair cycles. Shedding and the next growth phase are controlled separately. This long gap helps explain why a shedding episode takes many months to refill.

A 60-second comb count averaged about 10 hairs in healthy men. In 60 men, a standardised 60-second comb count averaged about 10 hairs and was stable on consecutive days and 6 months later. That gives a concrete baseline for judging a bad shedding spell.

Postpartum shedding began at 2.9 months, peaked at 5.1 and ended by 8.1. In a survey of 331 mothers, 91.8% reported postpartum hair loss, starting on average at 2.9 months, peaking at 5.1 and ending at 8.1 months. Only 21% responded, so it may overstate the problem. A predictable arc helps distinguish normal postpartum shedding from something else.

'Chronic telogen effluvium' may mostly not exist as a separate disease. A systematic review of 18 studies with 1,628 cases found none had excluded every other cause of shedding, and every biopsy study showed a normal ratio of thick to fine hairs. Of those diagnosed, 97.5% were women. The authors suspect early female pattern hair loss, an undetected trigger, or normal shedding noticed by anxious long-haired people.

In 18 case reports, biotin helped hair only when something underlying was wrong. A review of 18 reported cases found every one in which biotin helped hair or nails had an underlying cause, such as biotin deficiency, uncombable hair syndrome or brittle nails. There is no good evidence that extra biotin grows hair in people who are not deficient.

In a 10-man crossover, biotin alone did not change hair growth speed. In 10 healthy men, biotin 5 mg alone did not change hair growth speed (2.50 to 2.47 mm per week, p = 0.53); only the combination with minoxidil changed it significantly. The trial was tiny and short (14 days per arm). It is a negative signal for the idea that biotin speeds growth.

Biotin at 10 mg/day gave wrong results on 9 of 23 biotin-based lab assays. In six healthy adults taking 10 mg biotin a day for a week, 9 of 23 biotin-based immunoassays gave wrong results, while none of 14 non-biotin assays did. Anyone taking biotin should tell the clinician before blood work.

Only 6.2% of 194 women with chronic shedding had ferritin of 20 or less. In 194 women with scalp biopsies, only 6.2% had ferritin of 20 ng/mL or less. The five with normal biopsies who took iron until ferritin passed 20 did not stop shedding. Low ferritin is common in the population, so finding it does not prove it is the cause.

Ferritin was lower in pattern loss and alopecia areata, not telogen effluvium. In 117 women, mean ferritin was 37.3 ng/mL in androgenetic alopecia and 24.9 in alopecia areata against 59.5 in controls, but not lower in telogen effluvium. The control group was very small (11). It runs against the usual assumption that iron matters most for diffuse shedding.

Alternate-day iron gave the same ferritin with fewer gut side effects. In 150 iron-depleted women given 100 mg iron on alternate versus consecutive days, ferritin was the same at equal total dose, gut side effects were fewer, and iron deficiency at 6 months was 3.0% versus 11.4%. Check with a doctor before supplementing iron.

Black tea cut iron absorption by 79 to 94% in a radio-iron test. With a bread meal in adults, black tea cut iron absorption by 79 to 94%, peppermint tea by 84%, cocoa by 71% and camomile by 47%; milk added little.

Vitamin D ran low only in alopecia areata and female pattern loss. A meta-analysis found vitamin D deficiency in about 47 to 54% of people with alopecia, but levels lower than controls only in alopecia areata and female pattern loss, not in telogen effluvium. These are associations. No trial shows that supplementing vitamin D regrows hair.

Zinc was 96 vs 99 micrograms/dL in hair loss, a clinically minor gap. In a cross-sectional study of 23,975 Israeli health-fund members, including 2,279 with hair loss, median zinc was 96 versus 99 micrograms/dL: statistically significant but clinically minor. The authors advised against routine zinc testing.

Selenium excess affects thyroid hormones from about 300 micrograms a day, and high exposure thins hair. A review found chronic selenium excess affects thyroid hormone synthesis from about 300 micrograms/day, and nail and hair loss occur at high exposure. Selenium is essential, yet shedding can be a sign of excess, not only of deficiency.

GLP-1 drugs carried three times the hair loss risk of placebo. Across seven RCTs and two non-randomised studies (4,114 people), the pooled hair-loss risk ratio versus placebo was 3.25 (95% CI 1.44 to 7.36). Labels tie it to the weight lost, as seen in the stress-shedding that follows crash diets.

57% had hair loss after bariatric surgery, linked to low folate, ferritin, zinc. A pooled analysis of 18 studies with 2,538 people found 57% had hair loss after bariatric surgery. Lower folic acid, ferritin and borderline lower zinc went with loss, while B12 and iron did not. Loss is more common in younger women and tends to fade with longer follow-up.

Each cuticle cell is coated in 18-MEA, and alkaline treatments strip it. 18-MEA is a fatty acid bonded to the outside of cuticle cells that makes virgin hair water-repellent. Alkaline treatments strip it, leaving a water-attracting surface. Loss of this layer is one chemical explanation for why treated hair feels rough and tangles.

Losing 18-MEA makes wet hairs stick together, tangle and dry frizzy. A Japanese group found that removing 18-MEA from hair makes wet hairs stick together, tangle, and dry in a disordered, frizzy arrangement. The study removed it chemically in the lab.

Curly follicles are bent, with uneven cell division, and keep curving in culture. Curly follicles are curved, with a bent bulb and uneven cell division on the two sides, and keep producing curved hair in culture. Curl is therefore programmed in the follicle rather than made by the hair's surface treatment. Curl is set in the follicle bulb, not by the hair's surface.

Six people developed coarse, kinky hairs after puberty at the front and crown. Six people developed coarse, kinky hairs after puberty at the front, temples and crown, thought to be androgen-dependent. It is the same distribution as pattern hair loss. A case series, not a controlled study.

31 papers tie chemotherapy, anti-seizure drugs and retinoids to texture change. A systematic review of 31 papers found texture changes with chemotherapy drugs, antiepileptics, retinoids and antiretrovirals. New or stronger curling was most common, starting 4.5 to 17 months in and usually reversible. Curl is not fixed for life.

No structural difference fully explains tight-curl breakage; grooming is suspect. A review concluded no structural or chemical difference fully explains the higher breakage of tightly curled hair, and grooming practices are the main suspect. Technique, not texture, is the lever.

In one study, nobody's ten sampled hairs fell into a single curl class. In a study of 48 volunteers whose hairs were classified, nobody's ten sampled hairs fell into a single class, and raters agreed only modestly on eight curl groups (inter-rater kappa 0.418). A hair type is a distribution, not a label.

Bleach nearly tripled hair's internal surface area within the first minute. Gas-sorption measurements show bleach creates many small pores quickly, which later merge into larger ones. The internal surface area of hair nearly tripled within the first minute. This is a lab measurement on hair fibres, but it shows why porosity jumps quickly with the first bleaching.

In a lab test on bleached hair, Olaplex No. 0 raised strength 47.6%; K18 did not. On bleached tresses (three per group), Olaplex No. 0 raised single-fibre tensile strength by 47.6% (significant), while K18 raised it by 19.5% (not significant) and formed an external layer on the fibre. Single application, three tresses per group, so a small laboratory result.

No bond-builder tested increased disulfide content in the cortex. Spectroscopy of bleached hair treated with maleate-type and shikimic-acid repair products found no treatment increased disulfide content in the cortex and no direct evidence of the claimed thiol-maleate cross-linking, although the cuticle looked more regular. The marketing story that these products rebuild broken bonds is not shown here. They may still act as coatings or ionic binders.

About 17.5% of a bond-builder's active degraded during a 40-minute model bleach. Analysis of Olaplex No. 1 found 9.15% of the diamine ingredient, and about 17.5% of it degraded into a by-product during a 40-minute model bleach. The authors were affiliated with a cosmetics company. The active partly reacts in the bleach it is meant to protect against.

Conditioner cut snagging of wet bleached hair from every stroke to under 10%. In 231 women, hair was combed about 16 strokes a day, wet bleached hair snagged on every stroke, and conditioner cut that to under 10%. Breakage in the combing studies comes from tangling and friction. A slippery film is the practical benefit of conditioner.

Conditioner cut wet combing force but reversed none of the oxidative damage. In a 2026 lab study, behentrimonium chloride cut wet combing force significantly and out-conditioned behentrimonium methosulfate, yet infrared spectroscopy showed no reversal of oxidative damage. The smoothness is a coating, not a structural fix. A damaged fibre is the same fibre underneath.

Straightening damp hair caused far more structural damage than dry hair. Heat styling wet hair caused considerably more structural damage than heat on dry hair, probably because water evaporates rapidly inside the fibre. A water-based heat-protection spray did worse than an ethanol-based one. Drying hair fully before a hot tool is the cheapest protection.

Coconut oil gets inside hair, mineral oil stays outside. Mass spectrometry of hair cross-sections found coconut oil inside the shaft and mineral oil only on the outside. The work came from a coconut-oil marketer.

Topical topiramate went viral with no hair data we could find. We found no published study of topical topiramate on hair growth. The closest data are a 2% cream that sped wound closure in diabetic mice (98% versus 81% at day 10), which says nothing about hair.

H syndrome, a nucleoside-transporter defect, underlies the dipyridamole idea. H syndrome, with hypertrichosis and hyperpigmentation, is caused by mutations in the nucleoside transporter ENT3, which moves adenosine across cell membranes (11 families studied). Hair-loss hobbyists extend this to dipyridamole, but the cited study does not involve that drug and gives no hair data for it.

Ageing pigment cells in moles released a protein that woke hair stem cells. In mice, clusters of ageing pigment cells released a protein called osteopontin that activated neighbouring hair stem cells via CD44. Injecting it induced growth in mice, it stimulated human follicles in culture and it is over-expressed in human hairy moles. This is a mechanism study, not a treatment.

In mice, raising stem cell lactate triggered hair growth. In mice, hair follicle stem cells make more lactate than other skin cells. Raising it, by blocking the pyruvate carrier with UK-5099, triggered growth while untreated skin did nothing. It is a mouse mechanism, not a human result.

The same drug can thin hair when swallowed and thicken it on the scalp. Valproate commonly causes shedding or curly-hair change when taken for epilepsy or mood disorders. Yet an 8.3% sodium valproate scalp spray raised total hair count more than placebo in a 40-man Korean trial over 24 weeks (P = 0.047); only 27 completed (15 valproate, 12 placebo). It is a route-specific effect.

A convincing theory failed: setipiprant +7.1 hairs versus +6.7 on placebo. Prostaglandin D2 is higher in bald scalp and blocks hair growth in the dish and in mice. But an oral blocker of its receptor, setipiprant, tested in 169 men for 24 weeks, gave +7.1 hairs versus +6.7 on placebo (p = 0.92). Biology that looks right in the lab fails in people more often than not.

A Wnt-activating scalp drug failed in 675 men. In the 675-man phase 2/3 trial of the topical Wnt-pathway drug SM04554, the week-48 hair count ratio versus vehicle was 1.024 (P = 0.54) and 1.014 (P = 0.745), no significant benefit. Activating a growth pathway is not the same as growing hair.

Bimatoprost on the scalp gained +13.1 hairs/cm2, minoxidil +21.9. In a 307-man company trial, bimatoprost scalp formulations changed hair count by +13.1, +6.1 and +6.3 against +4.1 for vehicle and +21.9 hairs/cm2 for minoxidil 5%. Isolated follicles responded in the dish, and scalp levels reached at least 10 times the concentration needed. It shows a lash drug does not simply transfer to the scalp.

Bimatoprost lash gains held about 2 months after stopping and faded by 4 to 6. In a trial of adults with idiopathic or chemotherapy-induced lash loss, composite response at month 4 on bimatoprost 0.03% versus vehicle was 40.2% versus 6.8% in the idiopathic group and 37.5% versus 18.2% after chemotherapy. Gains held about 2 months after stopping and largely faded by 4 to 6 months. Lash serums have to be kept up.

An anti-prolactin-receptor antibody was tested open-label in 16 people, no control. HMI-115, an antibody against the prolactin receptor, was tested open-label in a phase 1b study of 12 men and 4 women. There was no control group, so the result is only a hint. It is a non-androgen route.

In alopecia areata, oral JAK inhibitors worked (OR 8.15); topical ones did not. A meta-analysis of seven randomised trials (1,710 people) found oral JAK inhibitors beat placebo for 90% improvement (odds ratio 8.15), while topical ones did not.

Stopping baricitinib: 80% of responders lost benefit, versus 7% who continued. In a randomised withdrawal study of baricitinib in alopecia areata, by week 152, 80% of responders switched to placebo had lost the benefit, against 7% who stayed on the drug; retreatment recaptured response in most. Remission needs maintenance. The same pattern appears with minoxidil.

EU regulators confirmed suicidal ideation as a finasteride tablet effect in 2025. In 2025, EU regulators confirmed suicidal ideation as a side effect of finasteride tablets, with frequency unknown. Causality remains debated.

The EU capped silicones D5 and D6 in cosmetics at 0.1%. Under Regulation (EU) 2024/1328, cosmetics may not contain D5 or D6 at 0.1% or more, with different start dates for rinse-off and leave-on products. Regional rules mean the same product can differ by market.

Ingredient lists hide the 1% line. EU law lets any ingredient under 1% appear in any order. So position 20 versus position 25 on a label tells you nothing about how much is inside.

Scalp cooling cut lasting post-chemo hair loss from 52% to 13.5%. In a Korean randomised trial of 139 breast cancer patients, only 13.5% of cooled patients still had alopecia six months after chemotherapy, versus 52.0% of controls. The trial was open-label. Cooling protects the follicles, not just the hair on the day.

After a 100-unit Botox session, men's frontal hairs were thinner at six months. In a 2025 randomised trial of 15 men with Norwood III to V hair loss, comparing two Botox injection methods (under the skin versus under the skin plus muscle), hair thickness fell significantly in the frontal area six months after a single 100-unit session in both groups, and density did not improve. There was no placebo or untreated group.

No exosome product is FDA-approved, and a 2019 warning followed patient harm. The FDA states there are no approved exosome products, and exosomes used to treat disease are regulated as drugs and biologics requiring premarket approval. Its December 2019 notice followed serious adverse events in patients given unapproved products marketed as exosomes.