Substance and claimed effects

Finasteride (1 mg/day) and dutasteride (0.5 mg/day) are oral inhibitors of the 5-alpha reductase enzyme — finasteride blocks the type-II isozyme dominant in scalp and prostate; dutasteride blocks both type-I and type-II Mella et al. 2010. By lowering circulating and tissue dihydrotestosterone (DHT), they arrest the follicular miniaturisation that drives androgenetic alopecia (AGA, male-pattern hair loss). Finasteride 1 mg was FDA-approved in 1997 as Propecia for male AGA in men 18-41 FDA 1997; dutasteride 0.5 mg is approved for AGA in South Korea and Japan and used off-label elsewhere Eun et al. 2010. The brief lists six consequences: scalp density, hairline progression, sexual side effects, mood, prostate volume, and post-finasteride syndrome (PFS). The entry covers all six holistically. The substance also has a small benign-prostatic-hyperplasia footprint (the 5 mg finasteride / 0.5 mg dutasteride doses are the BPH indications; the BPH entry would be separate).

Evidence by addressing question

mechanism

Scalp DHT, not testosterone, is the proximate androgen that miniaturises genetically susceptible follicles on the vertex, mid-scalp, and temples. The conversion is catalysed by 5-alpha reductase; type-II dominates in scalp and prostate. Finasteride 1 mg/day suppresses serum DHT by roughly 70% and scalp DHT by about 60-70%; dutasteride 0.5 mg/day suppresses serum DHT by approximately 90% because it blocks both isozymes Eun et al. 2010 Gubelin Harcha et al. 2014. Lowering DHT lengthens the anagen (growth) phase and reverses miniaturisation in follicles that have not been fully fibrotic-replaced. The mechanism predicts the clinical findings: follicles still producing miniaturised hairs respond; long-dead follicles do not. The mechanism also predicts the side-effect profile — DHT and 5-alpha-reductase neurosteroid metabolites (allopregnanolone, 3α-androstanediol) act centrally and in genital tissue, so suppression has biological pathways to libido, erectile function, ejaculate volume, and mood Diviccaro et al. 2019 Trost et al. 2013.

evidence

Finasteride 1 mg has one of the largest evidence bases in dermatology. The pivotal phase-III programme randomised 1,879 men aged 18-41 with vertex or anterior-mid balding to finasteride 1 mg or placebo for 1-2 years; at 24 months, 48% of treated men showed improvement in vertex hair count vs 7% of placebo, with 83% of treated men maintaining or improving vs 28% of placebo on global photographic assessment Kaufman et al. 1998. Mean hair count rose by roughly 88 hairs/cm² above placebo at 1 year. A companion trial in frontal/anterior-mid scalp showed parallel benefit Leyden et al. 1999. A 5-year extension in men 41-60 confirmed durable benefit Whiting et al. 2003. A 3,177-man Japanese real-world cohort over 1-3.5 years reported improvement in 87.1% of patients Sato et al. 2012. A 2010 systematic review by Mella et al. pooled 12 RCTs (n=3,927) and confirmed efficacy across hair counts and patient/investigator assessment Mella et al. 2010.

Dutasteride 0.5 mg outperforms finasteride 1 mg on head-to-head metrics. The Gubelin Harcha phase-III trial randomised 917 men to dutasteride 0.02 / 0.1 / 0.5 mg, finasteride 1 mg, or placebo for 24 weeks; dutasteride 0.5 mg increased target-area hair count by 90 hairs/cm² vs 75 for finasteride 1 mg (mean difference ~14 hairs/cm² favouring dutasteride; p < 0.05) Gubelin Harcha et al. 2014. The Eun et al. Korean phase-III RCT in 153 men showed dutasteride 0.5 mg increased hair count by 23.43 hairs/cm² over placebo at 24 weeks with no serious adverse events Eun et al. 2010. A 2019 network meta-analysis of 1,025 patients across the head-to-head trials found dutasteride 0.5 mg gave roughly 18-20 more hairs/cm² and a meaningfully higher patient/investigator self-assessment score than finasteride 1 mg, with no significant difference in adverse-event rates Zhou et al. 2019. The 2018 European S3 guideline (EDF) recommends finasteride 1 mg as first-line for male AGA; dutasteride is described as effective but off-label in Europe Kanti et al. 2018. A 2022 network meta-analysis ranked oral 5-ARIs (especially dutasteride) above topical minoxidil for hair-count outcomes in men Gupta et al. 2022.

Effect magnitude in plain terms: roughly two of three users hold or regain density at 1-2 years; about half visibly improve; about one in six continues to lose hair despite the drug. Loss continues at the rate of untreated AGA the moment the drug is stopped — the effect is suppressive, not curative.

protocol

Dosing for hair loss is finasteride 1 mg PO daily, or dutasteride 0.5 mg PO daily (off-label for AGA in most jurisdictions). Both are taken with or without food; absorption is unaffected. Onset is slow: a paradoxical shed of telogen hairs in the first 2-6 weeks is expected and indicates conversion of dormant follicles to growth phase; visible density change emerges at 3-6 months; peak effect at 12-24 months Kaufman et al. 1998 Whiting et al. 2003. Discontinuation is followed by reversion to the untreated trajectory within ~12 months. Half-life: finasteride ~6-8 hours, dutasteride ~5 weeks (one missed finasteride dose matters more pharmacologically than one missed dutasteride dose). Some clinicians use 1 mg three-times-weekly finasteride to halve cost with modest efficacy loss, but the head-to-head data is thin. Topical finasteride spray (0.25%) has a phase-III RCT showing non-inferiority to oral 1 mg on hair-count change, with markedly lower serum-DHT suppression — proposed as a way to keep efficacy while reducing systemic exposure, though long-term real-world data is limited Piraccini et al. 2022.

contraindications

Sexual adverse events are the most common documented harm. In the original Kaufman trials, reduced libido (1.8% vs 1.3% placebo), erectile dysfunction (1.3% vs 0.7%), and ejaculate volume reduction (1.2% vs 0.7%) were the canonical findings Kaufman et al. 1998. The Mella 2010 systematic review found pooled sexual adverse events around 2.1% — small absolute, statistically significant excess over placebo Mella et al. 2010. The BPH-dose (5 mg/0.5 mg) population shows substantially higher sexual side-effect rates: a 2017 Corona meta-analysis of BPH RCTs found 5-ARIs roughly doubled risk of erectile dysfunction (OR ~1.5-2.5) and reduced libido (OR ~1.5) Corona et al. 2017. Belknap et al. 2015 documented poor adverse-event reporting in the original AGA trials — short follow-up windows, lack of standardised instruments — concluding the true population side-effect rate at 1 mg is likely higher than 2% but uncertain Belknap et al. 2015.

Pregnancy partner: finasteride is FDA pregnancy category X for women — handling crushed/broken tablets by a pregnant woman can cause hypospadias in a male fetus. Whole tablets are coated and safe to handle; men taking the drug do not need to avoid conception (ejaculate concentrations are negligible) but the FDA label is conservative FDA 1997. Liver impairment is a relative contraindication (hepatic metabolism). Concurrent depression / suicidality history warrants caution given the population mood signal (see mood section).

Mood: A 2017 Welk et al. population-based propensity-matched cohort (n=93,197 men over 66) found 5-ARI users had higher rates of self-harm (HR 1.88 in the first 18 months) and depression (HR 1.94) than non-users; the absolute risk increase was small (≈1 self-harm event per 1,800 men treated) Welk et al. 2017. A 2021 FDA Adverse Event Reporting System analysis (Nguyen et al., JAMA Dermatology) found a disproportionate suicidality signal in younger men on finasteride for AGA (reporting odds ratio elevated), though FAERS data cannot establish causation Nguyen et al. 2021. Diviccaro et al. 2019 demonstrated long-lasting depressive-like behaviour, reduced hippocampal neurogenesis, and gut-microbiota changes in male rats after finasteride exposure — plausible biological pathway via suppression of neuroactive 5α-reduced steroids (allopregnanolone) Diviccaro et al. 2019.

Prostate volume and PSA: At the 5 mg dose (BPH), finasteride reduces prostate volume by ~20-30% over 12 months Thompson et al. 2003. At the 1 mg AGA dose, prostate-volume effect is smaller but real; serum PSA is reduced by roughly 40-50% even at 1 mg, requiring doubling of the measured value when screening for prostate cancer in men over 40 on long-term finasteride D'Amico & Roehrborn 2007. The Prostate Cancer Prevention Trial (PCPT, n=18,882) found finasteride 5 mg reduced overall prostate cancer incidence by ~25% but initially showed a small absolute increase in high-grade tumours; 18-year follow-up showed no overall survival difference and the high-grade signal was largely a detection artefact Thompson et al. 2003 Thompson et al. 2013. The REDUCE trial (dutasteride, n=8,231) showed similar findings Andriole et al. 2010. Net interpretation: at the 1 mg AGA dose, prostate-cancer risk is approximately neutral; at any 5-ARI dose, clinicians must adjust PSA interpretation.

misconceptions

(1) "The early shed means it isn't working." The early telogen shed at 2-6 weeks is the drug shifting dormant follicles into the active phase synchronously — paradoxically, it is the strongest early predictor of subsequent response, not failure Kaufman et al. 1998. (2) "Side effects are permanent and inevitable." In trials, sexual side effects occurred in roughly 2% of users above placebo, and reversed in most when the drug was stopped — but a minority report persistent symptoms (post-finasteride syndrome, PFS) Irwig & Kolukula 2011 Healy et al. 2018. The signal is real and biologically plausible (5-alpha-reductase produces neurosteroids active in CNS and genital tissue) but the population incidence is unknown; trial follow-up was too short and used non-validated instruments Belknap et al. 2015. (3) "Finasteride causes prostate cancer." The 18-year PCPT follow-up shows no excess mortality; the original high-grade signal was largely detection bias Thompson et al. 2013. (4) "Dutasteride is more dangerous than finasteride." Head-to-head AGA trials show similar adverse-event rates at AGA doses (0.5 mg dutasteride, 1 mg finasteride); dutasteride has greater DHT suppression but not proportionally greater sexual side-effect rates Zhou et al. 2019 Gubelin Harcha et al. 2014. (5) "It regrows lost hairline." The temple recession trial data is weaker than the vertex/mid-scalp data; some frontal regrowth occurs but the deeply receded hairline is largely beyond the drug's reach.

failure-modes

(1) Starting too late. By Norwood VI-VII the follicles in the bald zones are largely fibrotic; the drug preserves the zones still producing miniaturised hairs. The pivotal trial limited enrolment to men with vertex/mid-scalp loss not below Norwood V — the data does not generalise to severe baldness Kaufman et al. 1998. (2) Inconsistent dosing. The drug suppresses DHT only while taken; intermittent use produces intermittent suppression and gives the underlying AGA process windows. (3) Stopping after partial improvement. The hair gained reverses within a year of stopping Whiting et al. 2003. (4) Mismatched expectations. Roughly half of users see clear regrowth; another third hold steady; about one in six continues to lose. The honest framing is preservation, not restoration. (5) Ignoring the early shed and stopping at 6 weeks. (6) Combining with high-protein/anabolic supplementation that may load testosterone substrate has no clinical efficacy data either way.

alternatives

Topical minoxidil 5% (separate mechanism — potassium channel opener, increases follicle blood flow) has independent efficacy and is the standard add-on; the combination outperforms either monotherapy Gupta et al. 2022. Oral minoxidil (low-dose, off-label) is emerging as effective. Topical finasteride 0.25% has phase-III non-inferiority data with reduced systemic exposure Piraccini et al. 2022. Hair transplant (FUE / FUT) is the only restorative intervention but does not stop ongoing miniaturisation — most transplant surgeons require concurrent finasteride to protect the surrounding native hair. Platelet-rich plasma, low-level laser, microneedling have signal but weaker evidence. The 2022 Gupta network meta-analysis ranked combination therapy (5-ARI + minoxidil) highest Gupta et al. 2022.

stakes

Untreated AGA progresses on a known trajectory. Norwood-stage progression averages ~0.5-1 stages per decade in genetically susceptible men; first-degree relatives' pattern is the strongest predictor. The window where 5-ARIs preserve the most hair is the window before substantial loss — the conversation usually happens after the recession is visible, by which point a year-or-two delay translates directly into Norwood-stage delta that cannot be recovered. There is no longevity cost to losing hair, but the cosmetic, social, and identity costs are well-documented (the published quality-of-life literature on AGA shows effects on self-esteem and social anxiety comparable to several visible chronic skin conditions). The substantive stakes are aesthetic and identity — the felt-experience prose in the article projects from this.

payoff

For responders, the felt experience is asymmetric. Density does not visibly increase week-to-week — it stops decreasing, then slowly fills in. The reader notices the shed slowing first (around month 3-4), then the part width holding (around month 6-9), then occasionally that someone hasn't asked when they grew hair back (month 12-24). Patients in 5-year follow-up report photographic improvement at 5 years over baseline Whiting et al. 2003. The honest forecast is a different aging trajectory — a man who would have been Norwood IV at 40 holds at Norwood II-III instead — rather than a transformation. Onset is slow; the discipline is staying on through the early-shed phase.

The credibility range

Optimist case

30 years of evidence, multiple large RCTs, two meta-analyses, real-world cohort data on tens of thousands of men, generic-pricing under $10/month, and a mechanism so clean it predicts both the efficacy and the side-effect profile. The drug does what it says: preserves and modestly regrows the hair in the zones it can reach. Side effects in the AGA trial population were modest (~2% above placebo, mostly reversible). Persistent harms are documented (PFS exists) but rare; the trade is favourable for most men who care about their hair. The drug is one of the cleanest substance-to-effect-to-mechanism stories in dermatology.

Skeptic case

The pivotal AGA trials were short (1-2 years), in narrow populations (men 18-41 without severe loss), with poorly standardised adverse-event capture Belknap et al. 2015. Persistent post-finasteride syndrome is documented in case series and FAERS signals but the population incidence is unknown Irwig & Kolukula 2011 Healy et al. 2018 Nguyen et al. 2021. A 2017 population cohort found nearly doubled risk of self-harm and depression in older 5-ARI users Welk et al. 2017. Mechanism for persistence is biologically plausible — finasteride suppresses neuroactive steroids that act on GABA-A and androgen receptors; animal models show long-lasting CNS, behavioural and microbiome effects Diviccaro et al. 2019. The PFS Foundation, a patient advocacy group, has documented thousands of self-reported cases of post-discontinuation sexual, mood, and cognitive symptoms. The drug treats a cosmetic problem; "rare but real" persistent harm is a hard trade for a non-vital indication. The bar for risk-tolerance is higher than for a chronic-disease drug.

Author's call

For the typical young-to-middle-aged man with active AGA, oral 5-ARIs are the highest-evidence intervention in the catalogue for hair preservation — the efficacy data is unambiguous and the mechanism is clean. The post-finasteride syndrome signal is real, biologically plausible, and not characterisable in incidence from current data — uncommon (low single digits at most by trial data; possibly higher) but not zero. The decision is genuinely a trade-off: a near-certain modest cosmetic effect against a small risk of persistent harm that the field has not yet measured. The right framing is decide, not do — informed consent, clinician involvement, a trial period during which sexual function and mood are monitored, and a clear plan to discontinue if symptoms emerge. Evidence rating is high (5); controversy is moderate-to-high (3-4) because efficacy is settled but safety is actively contested.

Stakeholder and incentive map

Commercial: Merck (Propecia, original patent now expired; generic finasteride is the dominant product). GlaxoSmithKline (Avodart / dutasteride). Telehealth hair-loss platforms (Hims, Keeps, Roman) — large user-acquisition spend, prescription-by-questionnaire models, financial incentive to maximise enrolment and minimise stopping. Hair-transplant surgeons routinely co-prescribe to protect transplanted-graft outcomes. Professional: Dermatology guideline bodies (EDF S3, AAD, BAD) endorse finasteride as first-line for male AGA Kanti et al. 2018. Urology guideline bodies endorse 5-ARIs for BPH and have a less aesthetic stance. Counter-incentive: PFS Foundation (patient advocacy organisation, raises awareness of persistent harm), several plaintiff-side law firms with active mass-tort proceedings, a vocal online community organised around persistent post-discontinuation symptoms. Regulatory: FDA has updated the Propecia label multiple times since 2011 to add information on persistent sexual dysfunction; European medicines bodies have done similarly. The pull is uniformly toward more disclosure of persistent risk, not less.

Population variability

Response is bimodal but skewed favourable. About half of users see clear regrowth, a third hold steady, and a sixth continue to lose despite the drug Kaufman et al. 1998. Predictors of response: younger age at treatment start, less advanced Norwood stage, vertex/mid-scalp pattern (frontal hairline responds less). Asian populations (Japanese, Korean cohorts) show comparable or slightly higher response rates and similar adverse-event profile Sato et al. 2012 Eun et al. 2010. Female AGA: finasteride 1 mg in premenopausal women showed no benefit in one early trial; higher-dose finasteride or dutasteride in postmenopausal women has been studied off-label with mixed results — outside the scope of this male-AGA entry. Older men over 60 are under-represented in the AGA-dose trials; the safety signal in older men is largely extrapolated from BPH-dose data, which involves higher exposure. Men with pre-existing erectile dysfunction or depression have higher baseline rates of relevant symptoms and may be more sensitive to attribution effects.

Knowledge gaps

(1) True population incidence of persistent post-finasteride syndrome — the trial population was selected, the follow-up was short, the instruments were not standardised. No properly powered post-marketing cohort has yet measured persistence of sexual or mood symptoms with adequate follow-up. (2) Long-term (10+ year) outcome data on AGA-dose users — most data is 1-5 years. (3) Mechanism of persistence: whether epigenetic or receptor changes account for the post-discontinuation symptoms in susceptible individuals. (4) Genetic predictors of response and of side effects — both could enable risk-stratified prescribing. (5) Whether topical finasteride genuinely uncouples efficacy from systemic risk over years of use (the phase-III data is 6 months) Piraccini et al. 2022. (6) Comparative-effectiveness data for pulsed (3×/week) vs daily dosing. Evidence that would change the author's call: a rigorous prospective post-discontinuation cohort showing persistent symptoms in >5% of users would shift the entry's framing toward avoid for cosmetic-only indication.

Brief coverage. The brief named six consequences — scalp density, hairline progression, sexual side effects, mood, prostate volume, and post-finasteride syndrome. All six are covered: density and hairline in evidence, stakes, and payoff; sexual side effects, mood, and PFS in contraindications; prostate volume / PSA in contraindications plus the prostate-cancer misconception. No silent narrowing.

Excluded by design.

• Female-pattern hair loss. Overlapping biology, different evidence base, different dosing (some clinicians use higher-dose finasteride or topical spironolactone in postmenopausal women, with thin RCT support). Warrants its own entry.
• BPH / prostate-volume reduction at the 5 mg finasteride / 0.5 mg dutasteride BPH dose. Different indication, different exposure, different risk-benefit. The 1 mg AGA dose still produces measurable PSA artefact, which is covered.
• Topical minoxidil and oral minoxidil — covered briefly under alternatives and out-of-scope; their own evidence base earns a separate entry.
• Hair transplantation — pointed at, not detailed. Separate decision tree, surgical, different cost tier.

Hard scoring calls.

• beauty_direct = 0. The ladder anchors short-term visible change within days to weeks. The substance is not topical; the only visible-within-weeks signal is the paradoxical shed in months one and two, which reads as worsening. Scoring 1 would have over-claimed direct visible effect. The substance's visible effect lives entirely in beauty_cumulative.
• mood = 0. The meta ladder scores positive mood lift; there is no signal that finasteride lifts mood, and there is a small population signal of worse mood in a minority Welk et al. 2017 Nguyen et al. 2021. The negative is captured in the contraindications section rather than as a negative score (no such concept in the rubric). Zero is honest; the catch is in the warning callout.
• controversy = 3. Efficacy is settled; PFS / persistent sexual and mood symptoms after discontinuation is an active expert debate but not yet a paradigm-fight tier. 4 felt over-stated given the dermatology / urology / patient-advocacy split is still tractable.
• evidence = 5. The trial base, three meta-analyses, the 2018 European S3 guideline endorsement, and 30 years of FDA experience clear the "name 2+ rigorous trials" bar comfortably.

Contraindication vocabulary gap. None of the closed-vocabulary tokens (pregnancy, breastfeeding, cardiac-condition, etc.) cleanly apply to a male AGA patient. The actual clinical contras — personal/family history of depression or self-harm, planning conception with a pregnant partner who might handle broken tablets, liver impairment — are surfaced in the contraindications addressing section and warning callout but cannot be tagged. A future depression-history or 5ari-pregnancy-partner token would be useful; flagging for the bundle authors.

Future-link candidates.

• Minoxidil for Hair Loss — the standard add-on, additive effect, different mechanism.
• Hair Transplantation (FUE / FUT) — the restorative option; almost always co-prescribed with finasteride.
• Female-Pattern Hair Loss — same biology, different population, different protocol.
• PSA Screening and Prostate Cancer — the screening interaction (halved PSA on 5-ARI) is a concrete cross-link.
• Post-Finasteride Syndrome — if the evidence base matures enough to warrant a stand-alone entry on persistent symptoms.

Knowledge gap that could shift the call. A properly powered prospective post-discontinuation cohort with standardised sexual-function and mood instruments would either confirm PFS as rare (current dermatology view) or measurably more common than rare (current PFS-community view). The current entry lands on "uncommon but real" with controversy 3; a measured incidence above ~5% would shift the framing toward avoid for cosmetic indication and the controversy score upward.