Hair Loss Science: DHT, Follicles and What Actually Works

Key Takeaways

Almost every conversation about hair loss goes wrong at the same point: it assumes the problem is too much DHT. It is not. Men with pattern baldness do not, as a rule, have higher circulating androgens than men with full heads of hair. What they have is scalp follicles that respond to a normal amount of androgen as if it were a signal to shut down — and that response is inherited.

Once that is clear, the evidence table reads differently. Treatments that reduce the signal at the follicle (finasteride) or bypass it entirely (minoxidil) have the strongest data. Treatments that reduce the signal weakly (saw palmetto) or through mechanisms nobody has measured in a human scalp (rosemary oil) have weak data. And treatments that address a nutrient the follicle was not short of (biotin) have none.

This article lays out the mechanism, then the hierarchy, then the two conditions that are routinely mistaken for pattern loss and treated with the wrong tools.

The Mechanism: Receptor, Enzyme, Cycle

Where the sensitivity lives

Every hair follicle is governed by a small cluster of cells at its base, the dermal papilla. Papilla cells from balding scalp are measurably different from papilla cells taken from the occipital scalp of the same man — the fringe that never goes.

Three differences matter. Balding papilla cells express more type II 5-alpha-reductase, the enzyme that converts testosterone into the more potent dihydrotestosterone, at levels comparable to beard follicles and higher than other body sites. They express more androgen receptor, demonstrated both by hormone-binding assays and by transcript measurement. And they express more of an androgen receptor co-activator (Hic-5/ARA55) that amplifies the receptor's signal once it is bound.

So the same testosterone arriving at two follicles gets converted to DHT more efficiently at the balding one, meets more receptor when it arrives, and triggers a stronger transcriptional response. The follicle is not seeing more hormone. It is listening harder.

What the signal does

Androgen-receptor activation in the papilla does not kill the follicle directly. It changes what the papilla secretes. Balding papilla cells, under androgen stimulation, release TGF-β1, TGF-β2 and DKK-1 — paracrine factors that suppress the proliferation of the keratinocytes that actually build the hair shaft. In co-culture, androgen-stimulated papilla cells from balding scalp inhibit keratinocyte growth through TGF-β1; normal papilla cells do not.

The consequence is a shortened anagen phase. A scalp follicle normally grows for two to six years before resting and shedding. Under androgen pressure the growth phase contracts round after round, so each successive hair is shorter, finer and less pigmented than the last.

That is miniaturisation: the conversion of terminal hair to vellus hair, in place, over many cycles. A follicle that has completed the process is still there, but it produces a hair too fine to see, and once the papilla has shrunk past a certain point the process no longer reverses. This is why every treatment works better the earlier it starts, and why "wait and see" is not a neutral choice.

Why it is genetic

Male pattern loss is highly heritable, and the first genetic locus found was in and around the androgen receptor gene itself. Genome-wide studies in large cohorts have since added risk variants across androgen-pathway components and hormonal pathways beyond it. The point for a reader is simple: the level of DHT is roughly the same across men; the sensitivity of the follicle is set by inheritance. Measuring serum DHT tells you almost nothing about whether you will bald, and "lowering DHT naturally" through diet addresses a variable that was never the problem.

The patterns

In men, the loss follows the Hamilton–Norwood sequence: bitemporal recession, vertex thinning, then bridging. In women, the Ludwig pattern predominates: diffuse thinning of the central scalp with the frontal hairline preserved, in a widening-parting or "Christmas tree" distribution. The patterns differ because the distribution of androgen-sensitive follicles differs, and because oestrogen appears to be protective in women. The pathology underneath — miniaturisation driven by androgen-receptor signalling — is the same.

The Evidence Hierarchy

The table ranks treatments by the quality of their randomised evidence in androgenetic alopecia, not by popularity, cost or how natural they sound.

The prescription tier, in more detail

Finasteride. The pivotal data are as clean as any in dermatology. Baseline vertex count was 876 hairs in a one-inch circle; finasteride added 107 hairs relative to placebo at one year and 138 at two (P < 0.001 on every measure), while placebo lost hair steadily. Sexual side-effects are the sticking point, and the meta-analyses genuinely disagree: one pooled 15 placebo-controlled trials and found a 1.57-fold risk of sexual dysfunction (95% CI 1.19–2.08); another pooled the AGA trials separately and found no significant increase (RR 1.21, 95% CI 0.85–1.72). Reasonable people read that differently, and it is a conversation for a prescriber, not a supplement label.

Dutasteride inhibits both 5-alpha-reductase isoforms and is licensed for hair loss in some countries. The network meta-analysis found it no different from finasteride in efficacy or sexual adverse effects — which is neither a reason to prefer it nor a reason to avoid it, but a reason to treat it as finasteride's equal rather than its upgrade.

Minoxidil works by a mechanism that has nothing to do with androgens — it prolongs anagen and enlarges follicles through pathways still not fully mapped — which is why it is the one agent that works in both sexes and stacks cleanly with a 5-ARI. The 5% concentration is the standard in men; in women, Cochrane found 2% and 5% performed the same.

Low-dose oral minoxidil (0.25–5 mg) has become common off-label practice on the strength of large case series rather than randomised trials; a meta-regression found each additional 1 mg/day associated with 47 more hairs/cm² of total density at six months, and also with a 17.9-point rise in hypertrichosis risk and 4.8 points in cardiovascular adverse events. Effective, dose-dependent, and not something to self-prescribe.

Microneedling is the cheapest adjunct with real data. In the 12-week randomised trial, 82% of men receiving weekly microneedling plus minoxidil reported more than 50% improvement against 4.5% on minoxidil alone. The 13-trial meta-analysis found no increase in adverse events with combination therapy (RR 0.83). The mechanism — wound-response growth factors and improved drug penetration — is plausible and the follow-up durations are short.

Laser devices and PRP both have positive meta-analyses and both carry heterogeneity that should temper enthusiasm. The FDA-cleared home-laser analysis pooled seven double-blind RCTs and found a large standardised effect on density. The PRP meta-analysis found a mean difference of 25.6 hairs/cm² but noted that the half-head designs used might have inflated the result. Neither is first-line; both are reasonable additions when the first line is already in place.

The supplement tier

Saw palmetto is the only supplement with a long comparative trial, and the comparison is not flattering: 38% improved versus 68% on finasteride over two years, with the effect largely confined to the vertex. It is weaker at the same enzyme. The full profile — including why its prostate evidence is essentially null — is at Saw Palmetto.

Rosemary oil matched minoxidil 2% over six months in one trial in which nothing happened in either arm at three months and no placebo arm existed to say whether anything happened at six. That trial is dissected in Rosemary Oil vs Minoxidil: What the 2015 Trial Actually Showed, and the profile is at Rosemary Oil.

GHK-Cu has a genuine skin literature and a thin hair one; it appears in the protocol as a topical layer and is covered fully in GHK-Cu for Skin and Hair.

Biotin deserves a specific sentence because it is the best-selling hair supplement in the world. Of three studies of oral biotin for hair growth, the one with a double-blind placebo-controlled design found no difference between biotin and placebo. Biotin corrects biotin deficiency — rare, and usually inherited or drug-induced — and does nothing measurable in people who are not deficient.

Deficiencies: Correct Them, Do Not Chase Them

Two micronutrients belong in a hair protocol conditionally.

The logic is the same in both rows. A deficiency is worth fixing because it is a deficiency. Neither nutrient is a hair-loss treatment in someone who is replete, and "hair vitamins" that bundle both at high doses with biotin are selling the association as if it were a mechanism.

What Is Not Pattern Loss

Two conditions get treated with DHT blockers every day and respond to none of them.

Telogen effluvium is diffuse shedding — hair coming out by the handful in the shower — that begins three to four months after a trigger: illness, surgery, childbirth, rapid weight loss, a new medication, severe stress. The follicles are not miniaturising; a large cohort of them has been pushed into the resting phase simultaneously and is now shedding together. It usually resolves on its own once the trigger is gone, and the work-up is history plus blood tests for thyroid, iron and nutritional causes. A saw palmetto capsule does nothing except delay the diagnosis.

Alopecia areata is autoimmune: T-cells infiltrate around the bulb of growing follicles after a breakdown of the follicle's immune privilege, producing well-defined smooth patches, sometimes total scalp or body loss. It affects about 2% of people at some point in life, is polygenic, and is now treatable with JAK inhibitors under a dermatologist. It has nothing to do with androgens.

Scarring alopecias — lichen planopilaris, frontal fibrosing alopecia, central centrifugal cicatricial alopecia — destroy the follicle permanently. Scalp pain, burning, scaling or redness alongside loss is the flag. These are dermatological emergencies in the sense that every month untreated is irreversible.

The rule: sudden, patchy, painful, or accompanied by scalp symptoms means a dermatologist, not a supplement.

Women: A Different Order of Operations

Female pattern hair loss is the same miniaturising process in a different distribution, and its treatment order differs in three ways.

First, assessment comes first. The multidisciplinary androgen-excess task force calls evaluation for androgen excess mandatory in every woman with pattern loss — irregular cycles, hirsutism and acne alongside thinning point to PCOS or another endocrine cause — and lists vitamin D, iron, zinc, thyroid hormones and prolactin as optional but recommended. Ferritin is worth a sentence of honesty: iron deficiency is common in women, and the largest controlled comparison (381 women with pattern loss or chronic effluvium against 76 controls) found it not more common in hair-loss patients than in controls. Check it because it is common and correctable, not because it is proven to be the cause.

Second, minoxidil is first-line and finasteride is not. Cochrane pooled 47 trials in 5,290 women: minoxidil nearly doubled the proportion reporting moderate-to-marked regrowth (RR 1.93, 95% CI 1.51–2.47) and added about 13 hairs/cm² over placebo, with no difference between 2% and 5%. Finasteride 1 mg was no more effective than placebo (RR 0.95). Once-daily 5% foam gave 9.1 hairs/cm² over vehicle at 24 weeks in a 404-woman trial.

Third, antiandrogens are physician territory. Spironolactone (typically 80–110 mg in the series that use it) and cyproterone acetate have open-label data — in 80 women on at least a year of either, 44% regrew, 44% held, 12% continued losing — but no placebo-controlled trial, and both require monitoring and contraception. They are added when loss is severe or androgen excess is present, not bought online.

Reading the Table Honestly

Three patterns fall out of the evidence.

The treatments that work best are the ones that act on the mechanism directly — reduce the enzyme's output, or bypass the androgen signal altogether. The adjuncts that work are the ones that improve delivery or stimulate the follicle physically. The supplements that trail the field are the ones that act on the mechanism weakly, or on a mechanism that was never demonstrated in a human scalp.

The second pattern is time. Nothing in the table shows a result before three months, most need six, and the pivotal finasteride data were still improving at two years. Any product judged in six weeks was judged on noise.

The third is irreversibility. The placebo arms of the finasteride trials lost hair progressively over two years. Those men were not harmed by a treatment; they were harmed by the absence of one. Where a supplement layer sits in relation to that — as a starting point for early, slow loss, and as an adjunct rather than a substitute for anyone losing hair actively — is set out in The Hair Retention Protocol.

Frequently Asked Questions

If my DHT is normal, why am I losing hair?

Because pattern loss is not a DHT-level disease. Follicles on balding scalp express more 5-alpha-reductase and more androgen receptor than follicles on the fringe of the same head, so they respond to a normal DHT level as if it were high. The sensitivity is inherited; the hormone level is a bystander.

Does finasteride cause permanent sexual side-effects?

The randomised evidence disagrees with itself: one meta-analysis found a 1.57-fold risk of sexual dysfunction across 15 placebo-controlled trials, another found no significant increase when the hair-loss trials were pooled alone. Trial safety reporting has been criticised as inadequate. The honest answer is that the risk is real for some men, uncommon, and best discussed with a prescriber who can stop the drug early if it appears.

Is oral minoxidil better than the foam?

It is more convenient and the case-series response rates are high, but the evidence is weaker — large series rather than large randomised trials — and the side-effects scale with dose: more hair growth per extra milligram, but also more hypertrichosis and more cardiovascular adverse events. It is a prescriber's decision.

Can saw palmetto replace finasteride?

Not on the evidence. In the only long head-to-head, 38% improved on saw palmetto against 68% on finasteride over two years, and the trial had no placebo arm to show how much of the 38% was noise. It is a weaker inhibitor of the same enzyme, with a cleaner side-effect record, and for someone with early slow loss who has declined the drug it is a defensible choice — as long as they know what they are choosing.

Does biotin do anything?

In someone deficient in biotin, yes. In anyone else, the only double-blind placebo-controlled study found no difference from placebo. High-dose biotin can also interfere with laboratory assays, including thyroid and troponin tests, which is a real harm for a supplement with no demonstrated benefit.

How do I know if my hair loss is pattern loss or something else?

Pattern loss is gradual, symmetrical, and follows a recognisable distribution — temples and crown in men, a widening central parting in women — with fine hairs replacing thick ones in the same place. Shedding by the handful, smooth bald patches, scalp pain or scaling, or loss that started suddenly all point elsewhere and need a dermatologist before any treatment.

Related Research

• The Hair Retention Protocol
• Rosemary Oil vs Minoxidil: What the 2015 Trial Actually Showed
• The Hair Retention Protocol: Topical and Systemic Support Built in Evidence Order
• GHK-Cu for Skin and Hair
• Saw Palmetto