Hair Loss — Evidence Map

• Classical twin studies: heritability around 0.81. Monozygotic twins show striking concordance in pattern and timing.
• Large pedigree studies: ~0.60. Suggests older twin numbers were slightly inflated and environment/lifestyle accounts for around 40% of phenotypic variance in timing and severity.
• SNP-based heritability (GWAS): common variants capture ~38–39% of variance in the 2017 mega-GWAS. The gap between pedigree (60%) and SNP (39%) is "missing heritability" — rare variants, epigenetics, gene-gene interactions.
• What's left for environment/lifestyle: smoking (OR ~1.8), chronic scalp inflammation, severe metabolic disease, stress, nutrition. See §Lifestyle.

Nyholt 2003 twin study; Heilmann-Heimbach 2017 Nat Commun GWAS; Ramos 2025 Genetic Landscape of AGA review PMC12837269.

The androgen receptor gene sits on Xq12, with a nearby EDA2R (ectodysplasin A2 receptor) region showing independent signal. Multiple GWAS have converged on this region as the largest single-gene risk factor — estimated to carry up to ~40% of total genetic risk, which is extraordinary for a polygenic trait. The AR gene has a CAG-repeat polymorphism in exon 1; fewer CAG repeats means a more transcriptionally-active receptor, which has been associated with earlier and more severe AGA, though replication is mixed.

Hillmer 2005 Am J Hum Genet; Ellis 2001 CAG repeat; Prodi 2008.

A region near the genes PAX1 and FOXA2 on 20p11 was the second locus to reach genome-wide significance (Hillmer 2008). Because it's autosomal, it passes from either parent. Carriers of the risk haplotype have roughly 2× the odds of AGA; homozygotes have even higher odds. Replicated in European and Han Chinese cohorts. Mechanism is not fully worked out — PAX1 is involved in morphogenesis, which fits the GWAS theme of morphogenesis/apoptosis pathways.

Hillmer 2008 PMC2672151; Li 2013 Chinese Han replication.

The 2017 meta-GWAS tied AGA to 71 independent genome-wide-significant loci. Pathway enrichment analysis pointed to:

• Androgen metabolism — AR, SRD5A1/2, HSD3B, HSD17B2.
• WNT/β-catenin — the core "grow a hair" pathway; DKK1 is a key inhibitor up-regulated in AGA scalp.
• TGF-β signalling — inhibitory growth factor induced by DHT in AGA dermal papilla.
• Apoptosis — premature follicle regression.
• Morphogenesis — follicle development and cycling genes.

Polygenic risk scores built from these loci correlate with severity stage, but are not clinically useful yet — the PRS explains a small fraction of individual-level variation, and family history is about as predictive in practice.

Heilmann-Heimbach 2017 Nat Commun; Yap 2018 UK Biobank replication.

James B. Hamilton's 1942 paper remains the cleanest in-vivo demonstration of the dual requirement. Observations:

• Men castrated before puberty never developed pattern baldness.
• Men castrated as adults with existing loss showed no further progression.
• When Hamilton gave testosterone to prepubertally-castrated men, only those with a family history of balding developed it. The others grew normal scalp hair.

The takeaway: androgens are necessary but not sufficient. You need genetic susceptibility plus adequate DHT exposure for AGA to unfold. Hamilton's original phrasing — "male hormone stimulation is prerequisite and an incitant in common baldness" — is still the correct framing.

Hamilton 1942 Am J Anat; Ayob 2015 Exp Dermatol historical review.

Given the polygenic + X-linked structure, useful family-history heuristics:

• Both maternal and paternal lineages matter. The 20p11 locus and most autosomal loci come from either side.
• Age of onset concordance is high. If your father/uncles started Norwood-III at 27, your timing correlates.
• Pattern (vertex vs hairline) shows some familial trend but isn't fully heritable — local scalp factors also shape presentation.
• Female pattern (FPHL) tracks the same loci partially. AR and 20p11 signals contribute, but the FPHL picture has a substantial non-androgen component — see §FPHL Biology.

A standard confusion: "do hair-loss genes make me produce more DHT, or make DHT worse for my hair?" The honest answer is overwhelmingly the second, with a small contribution at the local-enzyme level and almost none at the circulating-hormone level. Three layers, in order of importance:

• Follicle sensitivity to DHT (biggest factor). Balding-scalp dermal papilla cells have roughly 2× the androgen receptor density of occipital ("safe") scalp. Same DHT in blood, more binding at the follicle, more signal. The AR locus at Xq12 is doing most of the work here.
• Downstream response program (also huge). When DHT binds AR in balding-scalp papilla, the cell secretes growth-inhibiting factors (TGF-β1, DKK1, IL-6). Same binding in beard papilla: growth-promoting IGF-1. Wired in during embryogenesis. Most of the 71 GWAS loci are in these downstream pathways (Wnt, TGF-β, apoptosis).
• Local scalp DHT production (smaller but real). Balding scalp has more 5α-reductase type 2 activity in the dermal papilla (>14× other compartments), producing more DHT locally from circulating testosterone. The gradient that matters is at the follicle, not in the bloodstream.

What the genes are not doing: setting your circulating DHT level. Serum DHT is roughly similar between balders and non-balders — a bald man with "low DHT" on a blood test and a full-headed man with "high DHT" are both entirely expected. This is why:

• Measuring serum DHT is useless for predicting who'll bald or how fast.
• Hamilton's experiment had to work with androgen presence vs absence (castrates vs not) rather than "too much".
• Non-susceptible men can have elevated DHT and keep their hair; susceptible men can have modest DHT and lose it.
• "Natural DHT blockers" that mildly lower serum DHT (10–20%) don't work well — the follicle is still saturated.
• Drugs that work do so by lowering DHT enough to matter at the saturated receiver (finasteride drops scalp DHT ~64%) or by blocking the receiver itself (clascoterone, spironolactone).

Hibberts 1998 AR density; Sawaya 1997 scalp vs follicle DHT; Inui & Itami 2013 Exp Dermatol; Vierhapper 2001 serum DHT in AGA; Kaufman 1996 finasteride DHT kinetics.

Balding-scalp and beard both have androgen-responsive follicles. The difference is in the dermal papilla — the mesenchymal signalling hub at the base of each follicle:

• Higher AR expression. Balding scalp dermal papilla cells have roughly 2× the androgen receptor density of occipital ("safe") scalp cells, so the same DHT signal hits harder.
• Higher 5α-reductase type 2 activity. Local DHT production is elevated — type 2 activity in the dermal papilla of balding scalp exceeds occipital levels, concentrating DHT right where it matters.
• Flipped downstream signalling. DHT-stimulated beard papilla secretes IGF-1 (pro-growth). DHT-stimulated balding scalp papilla secretes TGF-β1, TGF-β2, DKK1, and IL-6 — all growth inhibitors, DKK1 being a potent Wnt/β-catenin antagonist. Over successive cycles, this shortens anagen (growth phase), shrinks the follicle, and drives miniaturisation.

These are embryologically programmed differences — the follicle's behaviour is set during development based on its anatomical location, and then expressed lifelong. It's why transplanted occipital follicles continue growing normally on a balding forehead.

Inui & Itami 2013 Exp Dermatol; Kwack 2008 DKK1 data; Hibberts 1998 AR density.

DHT activity in the dermal papilla is documented at >14× other follicle compartments in the scalp — type 2 concentrates right at the mesenchymal signalling hub. This is why finasteride, despite being "just" a type 2 inhibitor, nails the biology: it shuts down the enzyme where it does the damage.

Dutasteride's extra potency comes from catching type 1 (sebaceous glands and periphery) plus type 3. Whether the extra shutdown matters clinically depends on serum vs local scalp DHT — see §Dutasteride.

Bayne 1999; Thigpen 1993; Eicheler 1995; Uemura 2008 SRD5A3.

Norman Orentreich established in 1959 that grafted follicles carry their genetic programming with them. An occipital follicle transplanted onto a balding forehead keeps its occipital behaviour — continuing to cycle normally in a scalp that would otherwise miniaturise it. This is the entire basis of hair transplantation.

Orentreich 1959 Ann NY Acad Sci; Unger & Shapiro 5e textbook.

Wnt/β-catenin signalling is the canonical "anagen on" pathway — turning it up extends growth phase, turning it down shortens it. DHT in balding scalp papilla up-regulates DKK1, which is a direct Wnt antagonist. Knock DKK1 out in mice and you get exuberant hair growth. This is why:

• Minoxidil has a Wnt-agonist component (separate from its K-channel action).
• Microneedling triggers a wound-healing Wnt cascade.
• PP405 (pipeline) shifts follicle metabolism toward glycolysis, which correlates with Wnt-driven stem-cell activation.
• Several pipeline agents (SCUBE3, WAY-316606, CHIR99021) target Wnt directly.

Kwack 2008, 2012 DKK1 work; Soma 2005 Wnt miniaturisation; Choi 2017 Wnt review.

The dose-response is flatter than the label implies. Drake 1999 tested 0.01, 0.05, 0.2, 1, and 5 mg doses:

Translation: scalp DHT suppression is essentially maximal at 0.2 mg/day. The 1 mg label dose is a margin-of-safety choice to ensure consistent suppression across genetic outliers. Going from 1 mg to 5 mg buys ~5 percentage points of extra suppression — and 5 mg is the Proscar/BPH dose, which triples some side-effect rates.

Practical implication: patients who've struggled with side effects sometimes try 0.25 mg/day or alternate-day 1 mg. The 0.2 mg dose-finding data supports this as a rational dose-reduction strategy, with efficacy only marginally below the label dose.

Drake 1999 PMID 10495374; Roberts 1999 dose-finding PMID 10495375.

The three original registration trials (Kaufman 1998; extension 2002) are still the best data set:

• Hair count at target area (vertex): finasteride group +15 hairs/cm² at 2 y; placebo −72 hairs/cm² at 2 y (i.e., placebo group lost meaningful hair while finasteride group gained). The meaningful clinical signal is not "gains" so much as "absence of the 70-hair decline".
• Global photographic assessment at 2 y: 48% of finasteride-treated men showed moderate-to-great improvement, 42% showed slight improvement, 10% no change/worse. Placebo: 7% improvement, 70% worsened.
• At 5 years: 90% of finasteride men had no further visible loss; 65% had visible regrowth. Placebo group: 75% had progressed.
• Frontal hair loss (Leyden 1999): effect is real but smaller than vertex. ~50% of men saw slight-to-moderate improvement vs 25% placebo. Don't expect a hairline rewind.

Headline: ~90% of men halt or reverse visible loss at 5 y; "halt" is most of the benefit. Regrowth is a bonus.

Kaufman 1998 J Am Acad Dermatol; Finasteride MPHL Study Group 2002 5-yr extension; Leyden 1999 frontal study PMID 10365924.

In the 1 mg AGA RCTs, absolute sexual AE rates:

Meta-analysis (Pooled 5-ARI RCTs) shows an overall RR ~1.55 for sexual dysfunction, RR ~1.87 for any sexual AE. Absolute excess is small (~2 percentage points) but real.

A separate detail that matters: at the 5 mg Proscar dose for BPH, ED rises to 3.4–15.8% (vs 1.7–6.3% placebo) and libido loss to 2.4–10.0%. Most of these studies aren't in AGA populations (older men with baseline BPH), but they confirm the dose-dependence.

Nocebo is a real component. Sham-controlled trials with blinded arms consistently show placebo groups reporting sexual AEs in the 2–6% range. Recent evidence shows giving patients a written patient information leaflet listing sexual AEs roughly doubles reported sexual AEs vs no leaflet. This doesn't mean PFS isn't real — it means the baseline "finasteride makes me feel weird" reporting rate has a meaningful nocebo contribution.

Gupta 2018 meta-analysis PMID 30206635; Belknap 2015; Mondaini 2007 nocebo trial.

PFS refers to a cluster of persistent sexual, mood, cognitive, and somatic symptoms that some men report after stopping finasteride, sometimes lasting months to years.

• Prevalence (low-quality data): one large retrospective series estimated 0.8% of treated men developed sexual dysfunction, with ~33% persistence after drug discontinuation (roughly ~0.25% persistent). Duration >7 months was the main independent risk factor.
• Regulatory recognition: NIH added PFS to its Rare Diseases list in 2015. FDA updated the Propecia label in 2022 to include suicidal ideation, persistent sexual dysfunction, and psychoneurocognitive AEs.
• Leading mechanistic hypothesis — allopregnanolone. 5α-reductase doesn't only make DHT; it also converts progesterone and deoxycorticosterone into neuroactive steroids (allopregnanolone, THDOC) that modulate GABA-A. Finasteride lowers these. In PFS patients, CSF levels of pregnenolone, progesterone, allopregnanolone, DHT, and 17β-estradiol are all decreased vs controls. Chronic suppression may induce epigenetic changes in the AR promoter and brain, which would explain persistence after stopping.
• Genetic susceptibility signal: AR CAG-repeat polymorphisms and pre-existing vulnerability traits (anxiety, prior depression) show up in PFS cohort studies.

PFS Foundation; Irwig 2012 persistent sexual AE series; Melcangi 2017 neuroactive steroid CSF data; Diviccaro 2022 gut/allopregnanolone rat model PMC9687671.

The state of evidence is genuinely split:

• FDA FAERS (pharmacovigilance): Finasteride-related suicidality reports peaked at 363 in 2024, with a rising trend since 2019 among young male users. Disproportionality analyses find a reporting signal for finasteride + depression + suicidal ideation (ROR well above threshold).
• Large cohort studies: a 2,213,600-patient meta-analysis found no statistically significant association between 5-ARI exposure and depression/suicide. A 2025 cohort study plus Mendelian randomisation analysis (Gupta 2025 JOCD) also came back null.
• MHRA April 2024 safety review concluded the signal was plausible enough to warrant continued enhanced surveillance and improved patient information.

Two ways to reconcile: (1) rare idiosyncratic susceptibility in a small subgroup that doesn't move large-cohort means; (2) self-selection — men who are already struggling are more likely to seek treatment and more likely to seek attribution. Both can be true simultaneously.

Gupta 2025 JOCD 10.1111/jocd.70102; Nguyen 2023 Sci Rep cohort; MHRA April 2024 Public Assessment Report.

Gynecomastia. At the 5 mg BPH dose, gynecomastia rose from ~2.8% (placebo) to ~4.5%. At the 1 mg AGA dose, the absolute rate is ~0.4% vs 0.1% — small but non-zero. It's reversible in most cases on discontinuation.

Prostate cancer signal (PCPT). The Prostate Cancer Prevention Trial (n=18,000, finasteride 5 mg) showed:

• Overall prostate cancer incidence reduced by ~25%.
• But high-grade (Gleason 7–10) cancers increased: RR 1.27, 95% CI 1.07–1.50.
• Long-term follow-up: no increase in prostate cancer mortality (HR 0.75, 95% CI 0.50–1.12).
• Consensus interpretation: improved biopsy detection in smaller finasteride-shrunk prostates created a detection artefact, not a true high-grade induction. But the label warning persists.

Fertility / pregnancy.

• Finasteride 1 mg has minimal impact on sperm parameters in most men, but a subgroup (~5%) shows reduced count/motility that normalises on discontinuation.
• Pregnancy category X for the drug itself. Pregnant women should not handle crushed/broken tablets. Semen transfer of finasteride at 1 mg is considered negligible (concentrations ~400× below the no-effect dose), but the conservative move is to pause 1–3 months before attempting conception if the patient is anxious.

Thompson 2003 PCPT NEJM; Thompson 2013 long-term follow-up; Samplaski 2013 sperm data.

Mechanism. Dutasteride inhibits both 5-AR isoforms (and partially type 3). It's ~3× more potent at type 2 than finasteride and ~100× more potent at type 1. Half-life is ~4 weeks vs finasteride's ~6 hours — dutasteride accumulates, and after stopping, serum DHT suppression lingers for 4–6 months.

Olsen 2006 Phase 2 dose-finding PMID 17110217; Harcha 2014 PMID 24411083.

The Olsen 2006 Phase 2 (n=416, 24-week) head-to-head is the cleanest comparison:

2019 meta-analysis (Zhou, 5 trials, 649 men): dutasteride superior for hair count, global photo assessment, investigator assessment, subject satisfaction. 2024 RCT: thrice-weekly dutasteride 35% moderate-to-marked improvement vs daily finasteride 21%. 2024 network meta-analysis (Gupta): dutasteride 0.5 mg was the most effective monotherapy. 2025 Korean low-dose Phase 3: 0.2 mg dutasteride (half the standard dose) is already efficacious for male AGA.

Headline: expect roughly 1.3–1.5× the hair count gain of finasteride at 0.5 mg/day.

Olsen 2006; Zhou 2019 PMID 30863034; Gupta 2024 JOCD 10.1111/jocd.16362; JAAD Int 2025 S2666-3287(25)00101-4; Low-dose dutasteride Korean Ph3 anndermatol 10.5021/ad.25.048.

In head-to-head AGA trials, dutasteride's sexual AE profile is roughly equivalent to finasteride. Olsen 2006 found sexual AE rates in dutasteride 0.5 mg (~5%) comparable to finasteride 5 mg (~4%); long-term Korean post-marketing data confirms this. Gynecomastia is also similar.

Twice- or thrice-weekly dutasteride is a useful dose-reduction strategy. The 2025 RCT used thrice-weekly 0.5 mg and still beat daily finasteride on efficacy. Given the 4-week half-life, any steady-state plasma difference is small.

Clark 1999 Proscar safety; Olsen 2006; Jung 2014 Korean post-marketing PMC4969473.

Piraccini 2022 J Eur Acad Dermatol Venereol PMC9297965; Chinese Ph3 2025 PMID 40090937.

Systemic absorption is reduced, not eliminated. Serum DHT still drops ~35%. For a subset of patients susceptible to PFS-type symptoms, this may still be enough to produce AEs. Cases of sexual dysfunction, anxiety, and depression on topical finasteride have been reported, though at lower rates than oral.

FDA posture as of 2024. FDA flagged a growing number of adverse-event reports attached to compounded topical finasteride products (where concentration, vehicle, and even presence of active ingredient are not reliably controlled). The agency has not approved any topical finasteride product in the US as of April 2026. Compounding pharmacies fill most US prescriptions; Almirall's formulation is available in parts of Europe.

Compounding variability. A 2019 analysis of compounded hormone preparations found potency deviations of ±30% not uncommon. The "0.25%" you're applying may actually be 0.15% to 0.35%. This is the main argument for standardised pharmaceutical topical finasteride when available — or for staying with oral.

Who it makes sense for. Patients who want 5-ARI effect and have a strong aversion to systemic exposure; patients planning near-term conception; patients who had sexual AEs on oral fin and want to try a lower-exposure route. Expect ~80–90% of oral efficacy at ~60% of the DHT suppression.

Piraccini 2022; FDA 2024 compounding warnings; Labrecque 2019 compounding potency.

Minoxidil is a prodrug. The scalp follicle sulfotransferase enzyme SULT1A1 converts minoxidil to minoxidil sulfate, the active form. Mechanisms of minoxidil sulfate on the follicle:

• K_ATP channel opening in the dermal papilla → vasodilation, improved nutrient delivery.
• Prolongs anagen via direct signalling — extends growth phase from ~3 years to ~5.
• Wnt/β-catenin activation (Wnt10b pathway) — separate from vasodilation, directly mitogenic.
• VEGF up-regulation — perifollicular angiogenesis.

Buhl 1990; Goren 2014 SULT1A1 assay PMID 24283387; Kwack 2015 Wnt10b; Ramos 2021 booster data.

• 5% foam/solution (Olsen 2002, n=393): +18.6 terminal hairs/cm² at 16 wk. Placebo: −1.4. Effect roughly doubles at 48 weeks.
• 2% (Price 1999): +8–12 hairs/cm²; about half the 5% effect.
• Plateau: hair count tends to peak at ~12 months. Beyond that, minoxidil mostly preserves — loss resumes only slightly if at all while continuing.
• Non-response rate: 30–50% of users (consistent with SULT1A1 distribution).
• Vertex > hairline. Effect at frontal hairline is smaller — consistent with the 5-ARI pattern.

Olsen 2002 JAAD 5% vs 2% RCT; Price 1999; Tsuboi 2009 long-term Japanese data.

• Scalp irritation/contact dermatitis. ~5–15% of solution users, driven mostly by the propylene glycol vehicle. Switch to minoxidil foam (PG-free) for most cases.
• "Dread shed" at weeks 2–8. Minoxidil synchronises follicles from telogen into anagen, which briefly accelerates shedding of weakly-anchored telogen hairs. Common, normal, usually resolves by week 10. Don't stop in the shed phase — that's the rookie mistake.
• Unwanted facial hypertrichosis from application spill/runoff or pillow transfer. More common with solution than foam. Particularly bothersome for women.
• Withdrawal hair loss. Benefits start to reverse 2–8 weeks after stopping. By 3–6 months you're back to the trajectory you'd have had without treatment, which means visible loss of the gains. In ~40% of men in discontinuation studies, hair count dropped below baseline (the drug was masking ongoing underlying AGA). Practical implication: this is a forever drug, or you need a plan to transition to another agent.
• Hyperaldosteronism / edema / tachycardia from topical minoxidil are rare but reported, especially at higher concentrations (10%+, which aren't standard) or when used with compromised skin barrier.

Cats & dogs. Minoxidil is highly toxic to cats (a drop on a paw can be fatal) and dangerous to dogs. Applied hair, shared pillows, ungloved hands. This is a genuine emergency if your pet licks treated skin or a treated surface.

Olsen 2002 foam vs solution; Shin 2007 shed data; Friedman 2002 cats; Price 1999 hypertrichosis.

Minoxidil was originally a 1970s antihypertensive. Oral doses of 10–40 mg treat resistant hypertension; at those doses hypertrichosis is near-universal. Low doses (0.25–5 mg/day) give efficacy for hair without clinically meaningful BP reduction.

• 2024 Iranian RCT (Asilian): 1 mg oral = 5% topical on hair diameter over 24 weeks in men with AGA.
• Vañó-Galván 2021 multicentre safety series (n=1404): effective across AGA, FPHL, FFA, CCCA. Discontinuation rate only 1.7%.
• BP meta-analysis (2024): LDOM did not significantly alter systolic (−0.13 mmHg) or diastolic BP (−1.25 mmHg) across doses studied.
• Responds when topical failed: particularly useful for SULT1A1 non-responders (oral absorption bypasses the scalp sulfotransferase step, as other body tissues activate it).

Asilian 2024 JOCD 16086; Vañó-Galván 2021 PMID 33639244; JAAD 2024 BP meta-analysis S0190-9622(24)03100-1.

Based on the 2024 Sinclair-led international consensus:

Sinclair 2024 consensus statement; Dhurat 2020 initial Indian data.

Common adverse events (Vañó-Galván 2021, n=1404):

Contraindications / high-caution. Heart failure (even mild), severe CKD, concurrent other vasodilators (nitrates, PDE5 inhibitors chronic), pregnancy.

Vañó-Galván 2021; Gupta 2024 FAERS pericardial PMC11743066 & PMC11837233; JAAD Case Rep pericardial effusion report PMC9478873.

Mechanism. Dual action — suppresses Malassezia-driven seborrheic dermatitis inflammation (itself associated with AGA progression), plus direct androgen-receptor antagonism at the follicle (independent of 5-AR inhibition).

• 2% > 1% for hair density over 6 months. 1% is effective for dandruff, 2% is the AGA-relevant dose.
• Effect size: roughly 16–17% reduction in shedding at 6 months; stand-alone effect smaller than 5% minoxidil.
• Frequency: 2–3× weekly, lather and leave on 3–5 minutes before rinsing. Rinsing immediately loses most of the effect.
• Additive to the finasteride + minoxidil stack; low cost; virtually no systemic exposure at this use pattern.

Contact dermatitis ~1–3%. Some users report scalp dryness — alternate with a regular shampoo on non-treatment days.

Piérard-Franchimont 1998; Inui 2007; Gupta 2025 JEADV CP review.

A non-feminising stereoisomer of estradiol used topically. Up-regulates scalp aromatase, increasing local conversion of testosterone to estradiol rather than to DHT. Clinical effect is mainly stabilisation of existing hair, with modest density gains over 6–18 months. In the Lannas 2020 head-to-head:

• Topical finasteride 0.5% + 2% minoxidil outperformed 17α-estradiol 0.05% + 2% minoxidil for postmenopausal FPHL.
• 17α-estradiol combinations still had clinically useful effects — not a failed drug, just not the strongest lever.

Systemic estrogenic effect is debated. 17α-estradiol has ~1/200 the estrogenic potency of 17β-estradiol, but chronic use has non-zero systemic signal. Avoid in personal/family history of hormone-sensitive cancers.

Blume-Peytavi 2007; Lannas 2020 PMC7190559.

FDA-approved for eyelash hypotrichosis (Latisse = bimatoprost) and as glaucoma eye drops. Applied to scalp in small pilot RCTs:

• Blume-Peytavi 2012: latanoprost 0.1% once daily, 24 weeks, pilot in men with AGA: significant hair density vs baseline and vs placebo.
• 2023 meta-analysis (6 trials): prostaglandin analogs significantly improved hair density vs placebo; bimatoprost showed earlier, stronger hypertrichosis than latanoprost.
• Mechanism: PGF2α receptor agonism on the dermal papilla; induces telogen→anagen transition.

Concerns. Ocular use causes iris hyperpigmentation and periorbital fat atrophy; unclear whether scalp application carries similar local-pigmentation risks long-term. Cost is high relative to established options. Usually reserved for patients who've failed minoxidil or want an adjunct.

Blume-Peytavi 2012 PMID 21875758; Frontiers 2023 meta-analysis PMC10043424.

In vitro, caffeine at 0.001–0.005% reverses testosterone-induced hair follicle growth suppression, increases anagen duration, and boosts matrix keratinocyte proliferation. Mechanism: phosphodiesterase inhibition → ↑cAMP → counter-regulation of DHT-induced TGF-β2.

Clinical, the story weakens. A non-inferiority trial of caffeine-based 0.2% topical vs minoxidil 5% showed non-inferiority at 6 months, but sample quality is modest. Most commercial caffeine shampoos don't have RCT support and mix caffeine with surfactants that rinse most of the active away in normal use. Caffeine shampoos aren't a bad addition; they're a poor primary therapy.

Fischer 2007 in vitro; Brandner 2007 follicle penetration; Dhurat 2017 non-inferiority trial PMC5804833; 2025 systematic review Healthcare.

Topical dutasteride (0.01–0.05%) entered a 2025 Phase 2 (n≈ small) — showed dose-responsive hair count gains with minimal systemic DHT suppression in some studies, though absorption is higher than topical finasteride because dutasteride is more lipophilic.

Mesotherapy dutasteride (intradermal 0.01% injections, 1 mL per session, 3–12 sessions) is widely used off-label in Spain, Latin America, the Middle East. Open-label series report significant density and diameter gains; a recent case series (JAAD Case Reports 2024) and an RCT comparing 12 sessions to control found 62.8% improvement vs 17.5% control. Serum DHT minimally affected vs oral dosing.

Compounding variability + injection infection/nodule risk. Unstandardised. Reserve for clinics with transplant-grade protocols.

Sobhy 2013 mesotherapy trial; Aguilar-Calderón 2024 FFA mesotherapy; Mayo Clinic 2024 JAAD CR series; Ph2 topical dut PMC12405733.

• Same AR and 20p11 GWAS loci show up in FPHL genetic studies — the susceptibility architecture overlaps with MPHL.
• Hyperandrogenic conditions produce male-pattern loss in women. PCOS, congenital adrenal hyperplasia, androgen-secreting ovarian/adrenal tumours, and iatrogenic androgens (exogenous T, anabolic steroids) all trigger temporal recession + vertex thinning in a Norwood-like pattern.
• Anti-androgens work. Spironolactone ~57% improvement; bicalutamide long-term retrospective positive; finasteride 2.5–5 mg and dutasteride 0.5 mg show benefit in postmenopausal FPHL. If androgens had nothing to do with it, none of these would move the needle.
• DHT is elevated in balding scalp of women with FPHL, same as men — local 5α-reductase activity plus AR density produce the same miniaturisation signal.

• Most women with FPHL have completely normal serum androgens. You can biopsy a thinning scalp with no hyperandrogenism anywhere on labs. Classic FPHL is often "normoandrogenic AGA-like miniaturisation" — the signalling cascade is running in women with unremarkable total T, free T, DHEAS, and DHT.
• The pattern is different. Usual FPHL: diffuse central thinning with a preserved frontal hairline (Ludwig I–III or Sinclair 1–5). Male Norwood: vertex + bitemporal recession. Different vulnerability map suggests different underlying biology in at least part of the disease.
• Minoxidil works well as monotherapy. Topical 2–5% produces real density gains without any anti-androgen. If androgen signalling were the whole story, blocking it would be necessary for meaningful response — it isn't.
• Age-related follicle fragility piles on. Follicle stem-cell depletion, telomere attrition, declining perifollicular vasculature, and matrix stiffening all accumulate with age and interact with whatever androgen vulnerability is present.

Aromatase converts testosterone to estradiol (instead of to DHT). Key fact: female frontal scalp follicles have ~6× higher aromatase activity than male frontal scalp follicles. This is a meaningful anatomical difference, not a small one.

Functional consequences:

• A fraction of circulating T that would become DHT in a male scalp becomes estradiol in a female scalp — locally reducing DHT exposure and adding a pro-hair estradiol signal.
• This protection concentrates at the frontal hairline, which helps explain why women usually retain hairline even as the central part widens. The frontal aromatase shield holds.
• Aromatase is not infinite. With enough androgen (PCOS, exogenous T), the shield is overwhelmed and male-pattern presentation appears.

Sawaya & Price 1997 J Invest Dermatol 5-AR/aromatase/AR levels; Hoffmann 2002.

FPHL commonly appears or worsens around perimenopause and menopause. Mechanism is multi-hit:

• Ovarian estrogen collapses → less estradiol reaching follicles → less pro-anagen signal.
• Aromatase activity declines in peripheral tissues including scalp → the 6× shield weakens.
• SHBG drops in some women → more free (bioavailable) testosterone.
• Adrenal androgens (DHEAS, androstenedione) become the dominant androgen source and are still converted to testosterone and DHT peripherally.

The shift in estrogen-to-androgen ratio matters more than the absolute level of either. A woman who was androgen-vulnerable her whole life but estrogen-protected often decompensates within 2–5 years of menopause.

HRT and hair: systemic estrogen therapy has mixed effects on FPHL. Most formulations are slightly protective against progression; some (especially those with androgenic progestins like levonorgestrel) can worsen shedding. Transdermal estradiol + micronised progesterone has the cleanest hair profile of the common HRT regimens.

Clinical practice reality: most women with FPHL fall on the normoandrogenic end; workup routinely finds nothing. This doesn't mean treatment is hopeless — it means the evidence base for combining minoxidil with empirical anti-androgens is strong enough that many dermatologists use it regardless of lab results.

Olsen 2005 FPHL classification; Yip 2011 hyperandrogenic FPHL; Vujovic & Del Marmol 2014.

A reasonable synthesis of the above:

• FPHL has a real but partial androgen component, mediated by the same AR / 5α-R / DHT machinery as MPHL, with the same genetic susceptibility loci.
• It is superimposed on age-related follicle fragility (stem-cell depletion, telomere/matrix/vascular change) that doesn't need androgens at all.
• The aromatase shield dampens androgen impact, especially at the frontal scalp, and the estrogen withdrawal at menopause removes that buffer.
• Therefore: treatment is almost always a stack — minoxidil to stimulate the fragile follicle, optional anti-androgen to block whatever androgen contribution is active. Monotherapy with either alone underperforms combo.

Drug choices and dosing for women are in §Women-Specific Systemics.

• Mechanism. Androgen receptor antagonist + inhibits adrenal androgen biosynthesis. Also a potassium-sparing diuretic.
• Efficacy. Meta-analysis: 57% overall improvement, 66% in combination with minoxidil. Monotherapy 43%. Prevents further thinning in 85–100%; visible regrowth in 33–49%.
• Typical dose: start 50 mg/day, titrate to 100–150 mg/day.
• Time to judge: 6–12 months.

Concerns. Menstrual irregularity ~33%; breast tenderness; postural dizziness; hyperkalaemia (check K+ in patients with renal impairment, ACEi/ARB, K-sparing combos). Pregnancy category C (contraindicated near conception in men, but this is a women-specific therapy so moot). Dose-dependent breast tenderness/enlargement from AR blockade.

Famenini 2015 review; Sinclair JAAD 2020; 2023 meta-analysis PMC10502763; 2025 premenopausal RCT PMC12448166.

A pure, non-steroidal AR antagonist developed for prostate cancer. Used off-label at 25 mg/day (often lower than prostate-cancer doses) for FPHL, hirsutism, and androgen-related acne.

• 2025 RCT (n=74): bicalutamide 25 mg + minoxidil 1 mg vs minoxidil 1 mg alone: mean hair count gain +18.1 vs +21.5 (no significant difference at 24 weeks). Short duration likely limits finding — retrospective data show gains accrue over 12–24 months.
• Largest retrospective (n=138): progressive Sinclair-stage improvement from 0.18 at 3 months to 0.80 at 24 months.
• Systematic review: 17–29% reduction in hair-loss severity.
• Improves minoxidil-induced facial hypertrichosis in women (retrospective n=35) — useful combined.

Concerns. Hepatotoxicity — liver enzyme monitoring at baseline and 3-month intervals is recommended. Rare severe hepatic injury. Avoid with pre-existing liver disease or alcohol misuse. Menstrual changes. Doesn't have the diuretic / postural / K+ profile of spironolactone, which some patients prefer.

Fernandez-Nieto 2020 introductory; JAAD Int 2025 bicalutamide vs min RCT; Martínez-Velasco 2021; JAAD Int 2024 time-dependent improvement PMC12493240.

A steroidal AR antagonist + progestin, usually combined with ethinyl estradiol as Diane-35 in parts of Europe and elsewhere. Effective in FPHL, but carries a venous thromboembolism (VTE) signal that's driven more EU regulators toward alternatives for cosmetic indications. Not available in the US.

VTE risk, especially in smokers >35 or with thrombophilia history. Liver tumour risk at high cumulative doses (pharmacovigilance signal in MtF HRT literature).

Small RCTs in postmenopausal FPHL (Price 2000 negative with 1 mg; later studies with 2.5–5 mg positive). Used off-label in dermatology practice, usually as an adjunct. Strict contraception or postmenopausal-only because of pregnancy teratogenicity. Some clinicians prefer dutasteride 0.5 mg every other day for its deeper DHT suppression.

Not for premenopausal women without robust contraception. Mood/libido AEs seem qualitatively similar to men's, at broadly similar rates.

Price 2000; Iorizzo 2006; Boersma 2014; Yeon 2011.

The single well-cited RCT: pumpkin seed oil 400 mg/day vs placebo for 24 weeks in men with mild-moderate AGA. ~40% hair count increase in the PSO group vs ~10% placebo. Proposed mechanism: weak 5-AR inhibition (phytosterols). Single trial, moderate N, self-funded/confounded by broader supplement ingredients in some formulations. Replication hasn't happened.

One trial is not proof. Treat as a reasonable adjunct if you want something OTC and oral, not as a substitute for fin/min in moderate-severe loss.

Cho 2014 PMC4017725.

Rosemary oil and 2% minoxidil produced equivalent hair count at 6 months in men with AGA. No significant between-group difference at months 3 or 6. Rosemary had less scalp itching.

Panahi 2015 PMID 25842469.

Weak 5-AR inhibition in vitro. Multiple small RCTs (320 mg oral, or topical 2% shampoo) show modest positive signal on shedding and density. The 2025 network meta-analysis found saw palmetto less effective than 5% minoxidil but better than placebo. Useful if someone insists on an OTC AR-axis approach; not a substitute for finasteride.

Rare cholestatic hepatitis. GI upset common (~5%). Mild sexual AEs reported in some trials, unsurprising given the mechanism.

Prager 2002; Rossi 2012; Ablon 2025 RCT JOCD 70585.

• Peppermint oil: mouse data strong (comparable to minoxidil in shaved-mouse model); no human RCTs of quality.
• Topical melatonin 0.1%: small RCTs (Fischer 2004, 2012) show modest density gains in men and women, mechanism via anti-oxidant + hair-cycle signalling. Reasonable niche adjunct.
• Onion juice topical: one small RCT (Sharquie 2002) in alopecia areata showed high regrowth rates vs tap water — possibly because alopecia areata sometimes regrows on its own. Not evidence for AGA.

Oh 2014 peppermint mouse; Fischer 2012 melatonin; Sharquie 2002 onion (AA only).

Biotin deficiency causes hair loss. In biotin-replete people (>99% of the Western population), supplementing biotin does not increase hair growth. High-dose biotin (5,000–10,000 µg) interferes with immunoassays — skewing thyroid, troponin, and hormone labs. Dermatologists routinely ask patients to stop biotin 72 hours before labs. "Hair, skin, nails" supplements generally bundle biotin, zinc, silica — benign but usually not evidence-backed unless correcting a specific deficiency.

Patel 2017 JAAD biotin review; Soleymani 2017 biotin assay interference.

Evidence. Pei 2024 meta-analysis: 12 RCTs, n=631 AGA patients, microneedling + minoxidil vs minoxidil alone gave a standardised mean difference of 1.32 (large effect) for hair count. A 2025 RCT comparing minoxidil alone vs min+micro vs min+micro+fin showed 80% of the triple-stack group scored ≥3 on Norwood improvement.

Infection risk with shared/re-used devices. Pinpoint bleeding and post-procedure redness are expected; anything more should be evaluated. Avoid on active seborrheic dermatitis or scalp infections.

Dhurat 2013 pilot RCT; Pei 2024 meta-analysis JOCD 16186; 2025 "Big 3" stack RCT PMID 39954111.

Evidence. Multiple 2023–2024 meta-analyses consistently show PRP increases hair density vs placebo. Hair thickness signal is weaker. Activated PRP (calcium chloride or thrombin activation before injection) outperforms non-activated in head-to-head data. Combined with minoxidil: additive.

Gentile & Garcovich 2024; Yuan 2024 FPHL SRT; Aesthetic Plast Surg 2023 meta-analysis.

Photobiomodulation with 630–680 nm red light stimulates cytochrome c oxidase → ATP production → cell proliferation. FDA-cleared devices (HairMax LaserComb/LaserBand, iRestore, Capillus, Theradome) have positive sham-controlled RCTs.

Adherence is the killer. The 20-minute, 3× weekly time cost defeats most users within 6 months. Non-adherent use explains most "LLLT didn't work for me" anecdotes. Also: cheap no-name devices don't deliver the stated irradiance — stick with FDA-cleared models if you buy one.

Afifi 2017 meta-analysis; Lanzafame 2013; Jimenez 2014.

Injected into galea-adjacent muscles (frontalis, occipitalis) to reduce scalp tension → theoretically reduces mechanotransduction-driven miniaturisation and DHT accumulation via improved blood flow. Early pilot (Freund 2010) was positive. 2024 triple-blind RCT: no significant benefit in male AGA using objective hair-count technology. Stays in "adjunct" / experimental territory.

Eyelid ptosis if injection sites drift; forehead muscle weakness; headaches; cost ($400–$1,500 per round every 3–4 months). Not a defensible primary therapy given the negative recent data.

Freund 2010 pilot; 2024 triple-blind RCT JAAD S0190-9622(24)02542-8; systematic reviews 2022, 2023.

The galea-aponeurotica tension theory (English 2018, 2022) argues that chronic scalp muscle tension contributes to the pattern of AGA via mechanotransduction. Supporting data: Koyama 2016 showed standardised 4-minute scalp massages daily for 24 weeks increased hair thickness. A 2019 self-survey of people doing ~20 min/day "standardised scalp massage" reported improvement in ~75% over 8 months. The survey is confounded (self-selection, self-assessment), but the theoretical link to DKK1-suppressing mechanotransduction is interesting.

Practical take: free, near-zero risk, unclear magnitude of benefit. Not a replacement for drugs, but a plausible low-effort adjunct. 5–20 minutes daily of firm scalp finger-pad massage.

Koyama 2016 PMC4740347; English 2018 mechanical stress hypothesis PMC4639964; English 2019 self-assessment survey PMC6380978.

ISHRS 2020 guidelines; Rose 2018 FUE vs FUT head-to-head PMC28/5/179; Unger & Shapiro 5e.

Shock loss (recipient effluvium). Surrounding native hairs, stressed by surgery + anaesthetic + local trauma, shed 2–6 weeks post-op. Usually regrows by 3–6 months. Higher risk in women, patients with advanced miniaturisation, and large sessions into dense native hair. Mitigations: pre- and post-op minoxidil, possibly systemic low-dose steroids.

Donor shock loss. Rare; resolves within 3–4 months in almost all cases.

Donor depletion / overharvesting. The most important long-term complication. Over-aggressive FUE harvesting (>25% of follicular units in a zone) produces a permanently see-through donor area. This is a common failure mode at "high-yield" discount mills and is irreversible. The safe donor zone is ~30–40% of the full occipital/temporal field.

Cobblestoning / pluggy appearance. Older punch-graft era problem, rare with modern micrografts, but still seen with inexperienced FUE surgeons placing grafts at wrong angle/depth. Impossible to fully repair.

Other. Infection <1%; folliculitis 5–10% transient; hypoaesthesia of donor area common temporarily; keloid formation (rare).

Turkey / discount mills. Turkey is home to both world-class clinics and low-cost, high-volume mills. Mill-pattern failures: excessive graft counts beyond survival limits, non-physician technicians doing extraction, inappropriate candidate selection (NW-II patients with minimal loss), no pre-op medical stabilisation, over-packed recipient zones causing graft death. Verify: ISHRS membership, photo evidence of 3+ year-post results (not just freshly-implanted "wow" shots), physician-led (not tech-led) extraction and placement.

Konior 2013 Facial Plast Surg Clin shock loss; Shin 2020 donor depletion; PMC6371733 overview; Harris 2011 FUE complications.

• Stabilise AGA medically first, 6–12+ months on finasteride ± minoxidil. Operating on a rapidly-progressing scalp produces islands of hair surrounded by receded native hair by year 5.
• Age. Transplants on men <25 are generally discouraged — you don't yet know their final Norwood pattern, and the donor zone isn't yet tested.
• Donor density & quality. Low native density, low caliber hair, early diffuse thinning in the donor zone = poor transplant candidate.
• Realistic patterning. A transplant designed for a 25-year-old hairline may look weird on the same person at 50. Mature, conservative hairline placement is almost always the right call.
• Scarring alopecias. Generally contraindicated while disease is active. See §Scarring.

Mechanism. Topical cortexolone 17α-propionate — a pure, local androgen-receptor antagonist. Competes with DHT at the follicle AR without systemic androgen blockade.

If approved, it's the first new AGA mechanism in 25+ years and the first topical that meaningfully blocks androgen action at the follicle without systemic effects. Particularly attractive for patients who can't or won't take oral 5-ARIs, for women of reproductive age, and for PFS-cautious users.

Cosmo Pharmaceuticals Dec 2025 topline release; SCALP study (Hordinsky presentation); Dermatology Times coverage.

Mechanism. First-in-class dual mitochondrial pyruvate carrier (MPC1 + MPC2) inhibitor. Forces a metabolic shift in hair follicle stem cells from oxidative phosphorylation to aerobic glycolysis — the metabolic state of activated, proliferating HFSCs. Builds on Lowry lab work (UCLA) demonstrating lactate-mediated HFSC activation.

Why this matters: all existing AGA drugs slow loss. PP405 could reactivate bald follicles — a categorically different outcome. If Phase 3 holds, it changes the endgame for heavily-miniaturised scalps.

Phase 2a isn't Phase 3. Other follicle-activating candidates (Histogen HSC, FOL-005, SCUBE3 predecessors) have crashed at this transition. The mechanism is novel enough that unknown long-term concerns (sustained HFSC activation without malignancy? Effect on other mitochondrial-dependent tissues?) are worth tracking.

Pelage 2024 press release; Flores 2017 Lowry lab lactate; NCT06393452; Dermatology Times coverage 2024.

Mechanism. A proteolysis-targeting chimera that binds the androgen receptor and an E3 ubiquitin ligase, tagging AR for proteasomal degradation. Unlike antagonists that occupy the AR binding pocket, PROTACs remove AR protein entirely, then dissociate and catalytically degrade more AR. Doses can therefore be lower.

Phase 2 (n=180, Chinese men, Hamilton-Norwood III-V, 12 weeks): all 4 active arms had significant within-group TAHC increases; GT20029 0.5% QD and 1.0% BIW significantly beat respective placebos. No sexual AEs, low systemic exposure, mild local irritation.

Kintor 2025 Phase 2 tandfonline 10.1080/09546634.2025.2574304; PMID 41328006.

Monoclonal antibody that blocks prolactin receptor signalling at the hair follicle. Prolactin inhibits hair growth (explains some postpartum hair loss). Licensed by Hope Medicine from Bayer in 2019.

• Phase 1b (Sinclair, Australia, n=16): +14 non-vellus hairs/cm² at trial end.
• Phase 2 (NCT06118866): 180 participants, 24 weeks, ongoing.
• Dosing: subcutaneous, likely monthly — different patient experience vs a daily pill or topical.

The appeal: an androgen-axis-free mechanism, useful for patients who can't tolerate 5-ARIs or AR antagonists, and possibly useful in combination with them. Concerns: antibody therapies are expensive; prolactin has broader physiological roles (lactation, immune modulation, bone), so safety profile needs long follow-up.

HopeMed pipeline; NCT06118866; Frontiers Pharmacol 2024 antibody review PMC11362041.

Another Kintor topical AR antagonist, not cortexolone-based. Status:

• US Phase 2 (2023) missed primary endpoint (TAHC at 24 weeks vs placebo), though within-group improvements were seen.
• Chinese long-term Phase 3 (52-week): 53% of men / 48% of women improved on 0.5% BID.
• New expanded Phase 3 (2025): 1% concentration, 750+ participants.

The inconsistent results (plus a narrower safety margin vs clascoterone) have made Western analysts skeptical. Worth watching but probably second-tier behind clascoterone in the topical AR category.

Kintor 2024 filings; Follicle Thought 2024; NCT06126965.

Built around the osteopontin observation — hair growing on top of moles (osteopontin-rich) is longer and thicker. AMP-303 is an osteopontin-derived peptide delivered by intradermal injection. Amplifica first-in-human completed in early 2024; the company reported statistically significant frontal-temporal hair growth after one treatment cycle. SCUBE3 (Amplifica's second, preclinical programme) is a separate signalling-molecule approach.

Intradermal delivery is a real patient-experience hurdle. Injection compliance issues killed a number of peptide hair-loss programmes historically.

Amplifica press releases 2023–2024; Follicle Thought coverage.

Small, mostly open-label, wildly unstandardised. Density/diameter gains reported in:

• Placental-MSC exosomes (P-MSC): Phase 1/2 single-arm, n=12, +67 hairs/cm² at 6 weeks.
• Foreskin-MSC exosomes: prospective, density gain in months 1–3.
• Adipose-derived (ASC): n=30, 24 weeks, density increase.
• Rose-stem-cell exosomes + electroporation: small Thai study.

Kwon 2024 Aesth Plast Surg; Lueangarun 2024 JOCD 16463; FDA Tissue Products warnings 2019–2024.

The "cure for baldness" has been projected for the 2020s repeatedly. Status as of 2026:

• Stemson Therapeutics (iPSC-derived follicles): created de novo human hair follicles in humanised mice. Company closed operations December 2024. Technology may be acquired/licensed.
• Replicel / Shiseido (RCH-01, autologous dermal sheath cup cells): Phase 2 data showed +11.8% density in 10/16 patients at 6 months — modest. Shiseido launched a Japanese market version. US development has been slow.
• Tsuji lab (RIKEN, then OrganTech): long-running bioengineered follicle programme. Feb 2026 paper identified the "third cell type" (dermal condensate cells) that reportedly enables 100% efficient lab-grown follicles. Commercialisation timeline unclear.
• HairClone (UK): storing follicles for future cell-expansion therapy. Banking service live; actual implant therapy not yet clinical.

Cell therapies face a universal challenge: the autologous model is expensive per-patient; the allogeneic model has immune-rejection and regulatory barriers; the engineered follicle model has to produce natural-looking hair (direction, density, cycling) at scale. Every one of these has tripped multiple companies. Clinical availability realistic for this decade seems low.

Stemson closure reports Dec 2024; Replicel RCH-01 data; Tsuji 2026 Yokohama; HairClone company materials.

Topical osteopontin-derived peptide. Phase 2a (n=200+, 2021): hair count +6.6 hairs/cm² in highest-dose group vs +5.6 placebo. Not significant. Company ceased development. Restructured into Coegin Pharma; "follicopeptide gel" plans remain. Illustrative: good Phase 1 does not predict Phase 2 success.

Follicum 2021 Phase 2a readout; company restructuring news.

Wnt-containing "hair stimulating complex" from hypoxia-cultured cells. Early Phase 1/2 (2012) reported 86% responder rate and significant density gains. Multiple subsequent trials mixed to negative; 2020 Phase 1b/2a missed primary endpoint at week 18, met it at week 26, but the company suspended HSC development in 2022 after 12 years. Cautionary tale of publication-to-replication drop-off.

Allergan's Phase 2 bimatoprost scalp solution missed its primary endpoint; programme shelved. Smaller independent studies continue to show positive signal (see §Other Topicals), but the large pharmaceutical pivot is off.

Based on Cotsarelis 2012 finding that PGD2 is elevated in balding scalp. Setipiprant (Kythera/Allergan) entered Phase 2 for AGA but the programme stalled without convincing readout. The PGD2 hypothesis remains plausible but hasn't produced a drug. Zenagen and similar small-molecule plays also quiet.

AA is an autoimmune attack on anagen follicles. Completely distinct biology from AGA — 5-ARIs and minoxidil don't treat the underlying cause. Until 2022, treatment was corticosteroids (topical/intralesional/systemic), contact sensitisers (DPCP, squaric acid), and off-label immunosuppressants. Outcomes were mediocre, especially for totalis/universalis.

SALT (Severity of Alopecia Tool) ≤20 = ≤20% scalp involvement, a meaningful "mostly regrown" threshold for severe AA. Placebo response is generally <10%.

King 2022 BRAVE-AA1/2 NEJM baricitinib; King 2023 ALLEGRO Lancet ritlecitinib; FDA 2024 deuruxolitinib NDA materials.

JAK inhibitors carry FDA boxed warnings from the tofacitinib ORAL-Surveillance study (RA population, older/comorbid patients):

• Serious infections (TB, opportunistic)
• Venous thromboembolism (VTE)
• Cardiovascular events (MACE)
• Malignancy (lymphoma, skin cancers)
• Gastrointestinal perforation (rare)

How much these translate to younger AA populations is an active debate. Dermatology cohorts have so far shown substantially lower event rates than the ORAL-Surveillance RA population. Still, pre-treatment workup is standard: TB screen (IGRA), hepatitis panel, lipid panel, CBC, cancer screening appropriate to age.

Ytterberg 2022 ORAL-Surveillance NEJM; King 2024 AA JAK safety pooled analysis.

Pre-JAK options that still have a role for moderate AA or when JAKi is contraindicated:

• Intralesional triamcinolone (3–10 mg/mL, 4–6 week intervals) — first-line for limited patchy AA.
• Topical/oral corticosteroids — widely used; skin atrophy limits long topical use.
• Diphencyprone (DPCP) / squaric acid dibutylester (SADBE) — deliberately induces contact dermatitis; about 30–50% regrowth in moderate-severe AA. Practitioners are few.
• Minoxidil 5% — adjunct; modest effect alone.

Stressor triggers synchronous follicle entry into telogen ~2–4 months later. Diffuse shed, no patterning, pull test positive across scalp. Self-limited if trigger removed; chronic TE (>6 months) is a distinct, often idiopathic entity.

Common triggers:

• Major illness (including COVID-19 — peak "long COVID TE" literature 2021–2023)
• Surgery / general anaesthesia
• Childbirth (postpartum TE, typical onset 3 months postpartum)
• Crash dieting / >10% weight loss in <2 months
• Severe iron deficiency (ferritin <30, possibly <70 for hair)
• Thyroid dysfunction (both hypo and hyper)
• Medications: isotretinoin, SSRIs, anticoagulants, beta-blockers, lithium, ACE-I, retinoids
• Severe psychological stress

Standard workup: ferritin, TSH, free T4, CBC, B12, vitamin D, zinc; in women consider hormones (prolactin, DHEAS, total/free T, ovarian ultrasound if PCOS suspected). 2024 case-control (Durusu Turkoglu): zinc, selenium, and Cu/Zn ratio were more predictive for chronic TE than B12 or vitamin D.

TE is diagnosed backwards. If you're shedding and a clear trigger happened 2–4 months ago, it's TE. If there's no trigger and the shed has been going for >6 months, work up for FPHL/AGA, chronic TE, or autoimmune issues.

Malkud 2015; Turkoglu 2024 JOCD 16512; StatPearls TE chapter.

Inflammatory destruction of the follicle stem-cell compartment (bulge) — regrowth isn't possible once the follicle is lost. The goal is to halt progression. Common entities:

• Lichen planopilaris (LPP) — multifocal, patchy, with perifollicular erythema and scale; burning/itch common.
• Frontal fibrosing alopecia (FFA) — band-like hairline recession in postmenopausal women (also increasingly in younger women and men); eyebrow loss common.
• Central centrifugal cicatricial alopecia (CCCA) — vertex-centered, most common in women of African ancestry.
• Discoid lupus erythematosus (DLE) scalp — autoimmune variant, often red-purple plaques.

Red flags for scarring (not AGA): loss of follicular ostia on dermatoscopy, persistent redness or scale, burning/itch/pain, rapid patchy loss centred on the vertex or frontal band, eyebrow loss concurrent with scalp loss. See a dermatologist early — tissue lost is lost forever.

Treatment stack (systemic):

• Hydroxychloroquine (LPP, DLE) — most frequently used first-line.
• Methotrexate (LPP) — highest response rates in 2024 systemic-retrospective data (~33% good response).
• Cyclosporine A (LPP, FFA).
• 5-ARIs (FFA specifically) — finasteride/dutasteride have indication-specific evidence.
• Emerging: topical tofacitinib 2%, oral brepocitinib (TYK2/JAK1 Phase 2a for cicatricials).

Villablanca 2024 MDPI JCM 13/16/4947; Ochoa 2024 systemic tx retrospective; Piraccini 2024 topical tofacitinib JAAD; Vañó-Galván PLACE registry.

Gupta 2024 meta-analysis: ever-smokers OR 1.82 (95% CI 1.55–2.14) for AGA vs never-smokers. For progression from Norwood I-III to IV-VII, smokers had 27% higher odds. ≥10 cig/day doubles the effect size vs light smokers.

Plausible mechanisms: microvascular dysfunction reducing perifollicular blood flow; oxidative stress on dermal papilla DNA; Langerhans-cell activation and low-grade scalp inflammation; endocrine effects increasing SHBG-free androgens.

Gupta 2024 JOCD 16132; Fortes 2017 systematic review PMC9069908.

• Chronic stress / elevated cortisol can exacerbate any hair loss (particularly TE); no clean AGA-specific effect size.
• Sleep deprivation correlates with AGA in observational studies but confounded heavily by shift work and stress.
• Metabolic syndrome (particularly insulin resistance): observational associations with earlier AGA onset. Biologically plausible via SHBG reduction, increased free androgens, and scalp microvascular dysfunction. Observational only — no RCT of "treat insulin resistance, reverse AGA".
• Chronic seborrheic dermatitis: associated with AGA progression. Treat it (ketoconazole 2% is the cheap win).

No specific diet causes or prevents AGA. What matters is avoiding deficiency:

• Protein — severe restriction triggers TE, not AGA. Minimum ~0.8 g/kg/day, higher during caloric deficits.
• Iron / ferritin — ferritin <30 clearly associated with increased shedding; target >70 in women with FPHL per some guidelines.
• Vitamin D — receptor expressed on hair follicle; deficiency association with TE and AA is real, AGA less clear. Correct if deficient.
• Zinc — deficiency causes hair loss; routine supplementation in replete people doesn't help.
• Crash dieting (>10% body weight / <2 months) is a reliable TE trigger.

Guo 2017 nutritional review Dermatol Pract Concept.

• Finasteride 1 mg/day oral (or dutasteride 0.5 mg/day or 3×/week if after 6 months the fin response is weak)
• Minoxidil 5% topical foam 1 mL once daily, OR LDOM 1.25–2.5 mg oral (preferred if SULT1A1 non-responder, topical irritation, or bad lifestyle fit)
• Ketoconazole 2% shampoo 2–3×/week, lather 5 min
• Microneedling 0.5–1.5 mm weekly at home (separate day from minoxidil)
• Optional: LLLT device 3×/week

Evaluate progress at 12 months. Most people see "the shed stopped" by month 3–4, density gains by month 6–9, full effect by month 12–18.

• Dutasteride 0.5 mg/day (deeper DHT suppression)
• LDOM 2.5–5 mg/day (men-dose)
• Ketoconazole 2% 2–3×/week
• Microneedling weekly
• Plan for transplant consultation once disease is stable for 12+ months, if donor density supports it

Advanced pattern with aggressive medical therapy still won't fully restore NW-VI scalps — manage expectations. Transplant is usually needed for acceptable cosmesis in this group.

• Minoxidil 5% topical (foam) OR LDOM 0.625–1.25 mg/day
• Spironolactone 100 mg/day (premenopausal) OR bicalutamide 25 mg/day (AE profile preference) if hyperandrogenic features, PCOS, or minoxidil alone insufficient
• Ketoconazole 2% 2–3×/week
• Microneedling weekly
• Topical 17α-estradiol 0.025% if androgen-blocking contraindicated (e.g., hormone-sensitive cancer history)
• Consider finasteride 2.5 mg/day or dutasteride 0.5 mg alt-day in postmenopausal women if above inadequate

Spironolactone/bicalutamide/finasteride are pregnancy-incompatible. Reliable contraception required in premenopausal use. Hepatic monitoring for bicalutamide; potassium monitoring for spironolactone.

For patients unwilling to take oral 5-ARIs (family history of depression/anxiety, concern over PFS):

• Topical finasteride 0.25% once daily (or wait for topical approval) — lower but non-zero systemic exposure
• Minoxidil 5% foam or LDOM 1.25 mg (no hormonal axis effect)
• Ketoconazole 2% + microneedling + LLLT
• Watch clascoterone approval timeline — likely 2027 — at that point add topical clascoterone 5% for proper AR blockade without systemic exposure

Expect ~60–70% of the efficacy of the standard oral-5-ARI stack, with materially lower AE risk.