Libido Deep Dive

Status: draft complete (compiled 2026-04-28)

Libido is not one thing, and most of the trouble in the popular literature comes from pretending that it is. What gets called "sex drive" in everyday speech is actually a stack of motivational, hormonal, vascular, neurochemical, relational, and contextual systems that interact in ways that no single variable cleanly summarises. A person can have an intact dopaminergic reward circuit and still report low desire because their hormonal substrate has been pharmacologically suppressed; another person can have textbook hormones, normal vascular function, and a perfectly preserved orgasmic reflex, and still meet criteria for hypoactive sexual desire disorder because their relational context has flattened motivation independently of biology. The folk-pharmacology framing in which "more testosterone equals more libido" survives mostly because it is short, intuitive, and wrong in roughly the same direction as "more sleep equals less depression" — pointing in a defensible direction, useful as a slogan, useless as a clinical model. What the actual literature contains is a series of partial drivers whose effect sizes are real but modest, whose mechanisms overlap, and whose net contribution to any individual's desire is mediated by variables that pharmacology cannot reach. Sources: Basson 2000 J Sex Marital Ther, PMID 10693116, Wåhlin-Jacobsen et al. 2017 J Sex Med, PMID 28117267, Roney & Simmons 2013 Horm Behav, PMID 23601091, Cappelletti & Wallen 2016 Horm Behav, PMC4720522. Confidence: C2.

The "two axes" framing that organises a lot of online and trans-community discussion — a masculine T-driven libido that is always somewhat on, genitally localised, and crashes sharply after orgasm, versus a feminine E/P-driven libido that is often off, slow to ignite, body-wide, and decays gradually — is folk theory rather than sexology. Rosemary Basson's 2000 paper in the Journal of Sex and Marital Therapy proposed a circular, intimacy-driven model in which a "responsive" component of desire often follows arousal rather than preceding it, in explicit contrast with the linear Masters-and-Johnson sequence. Basson nowhere claimed that responsive desire was driven by estrogen-and-progesterone or that spontaneous desire was driven by testosterone, and Chivers and Brotto's subsequent reviews stress that responsive-versus-spontaneous is best understood as a continuum that crosses sexes rather than mapping onto a male/female or T-versus-E binary. The phenomenological details inside the two-axes claim — that men on average report more frequent spontaneous sexual cognitions, that the postejaculatory refractory time is on average longer in men than in women, that genital-and-subjective arousal concordance is famously lower in women than in men — are real population-level effects, but none of them require the hormonal grouping the folk model insists on. The right read is that the phenomenology is partly correct and the hormonal taxonomy is invented. Sources: Basson 2000, PMID 10693116, Baumeister, Catanese & Vohs 2001 Pers Soc Psychol Rev, Chivers et al. 2010 meta-analysis, PMC2811244. Confidence: C2.

Once the framing is set aside, the within-cycle physiology in healthy cycling women is the place where the testosterone-supremacy story collides hardest with the evidence. Wåhlin-Jacobsen and colleagues' 2017 study of 428 premenopausal women, awarded the 2016 ISSM Female Sexual Dysfunction Prize, is one of the largest and best-controlled endogenous-androgen datasets in this population, and it found that none of the androgens — total testosterone, free testosterone, DHEA-S, androstenedione — were associated with hypoactive sexual desire disorder in the structured sense; only a soft correlation between free T or androstenedione and a low-desire subgroup survived adjustment, with psychosocial variables explaining as much variance as androgens. Roney and Simmons' 2013 daily-saliva diary study went further and disentangled the three steroids on a within-person, within-cycle basis: estradiol positively predicted daily desire, progesterone negatively predicted it, and testosterone dropped out as a significant predictor when E and P were modelled together. The Melbourne Women's Midlife Health Project found that the menopause-related decline in desire correlated with falling estradiol rather than falling testosterone, and Cappelletti and Wallen's 2016 review concluded that the likelihood of an androgen-only treatment meaningfully increasing women's sexual desire is "minimal," with clinical efficacy in the trial record largely tied to supraphysiological androgen exposure rather than restored endogenous physiology. The honest summary is that endogenous T levels in healthy cycling women are weak predictors of clinically meaningful low desire, and any sweeping claim that "T is the master libido hormone in women" cannot rest on within-person physiology. Sources: Wåhlin-Jacobsen et al. 2017, PMID 28117267, Roney & Simmons 2013, PMID 23601091, Cappelletti & Wallen 2016, PMC4720522, Dennerstein, Lehert & Burger 2005, PMID 16414328. Confidence: C1 for the underlying datasets, C2 for the synthesis.

What testosterone does have, and what justifies its place in clinical practice, is RCT evidence in postmenopausal women. The 2019 Lancet Diabetes & Endocrinology meta-analysis by Islam, Davis, Bell and colleagues pooled 46 reports of 36 RCTs in 8,480 participants and found that testosterone increased satisfying sexual events by a mean of 0.85 per month (95% CI 0.52–1.18), with a standardised mean difference of 0.36 for desire, 0.28 for arousal, and 0.25 for orgasm. Those effect sizes are statistically robust and clinically modest — roughly one extra satisfying sexual event a month above placebo — and they are derived almost entirely from postmenopausal women, often on background estrogen. Shifren's 2000 NEJM trial of transdermal testosterone in surgically menopausal women, Braunstein's 2005 follow-on, and Davis's 2008 APHRODITE trial in postmenopausal women not on estrogen all reported the same direction with similarly modest absolute increments. The 2019 Global Consensus Position Statement and the 2021 ISSWSH Clinical Practice Guideline endorse off-label physiological-range T (typically one-tenth of male dose, 5–10 mg/day of AndroFeme cream, ~4 drops/day of male transdermal gel) with the explicit instruction that total testosterone not exceed the upper premenopausal reference range. The "upper female range without masculinization" target is real clinical aim, just not the one that compounding clinics tend to hit; pellet and IM regimens routinely overshoot, and that is precisely where masculinization risk concentrates. The FDA has rejected every product application — Procter & Gamble's Intrinsa patch in 2004 on cardiovascular and breast-cancer safety grounds, BioSante's LibiGel in 2011–2012 because the placebo arm responded too strongly — and only Australia's AndroFeme (TGA-licensed 2020) is licensed anywhere as a women's testosterone product. Sources: Islam, Davis et al. 2019, PMID 31353194, Davis et al. 2019 Global Consensus, PMC6821450, Parish et al. 2021 ISSWSH CPG, PMC8064950, Shifren et al. 2000 NEJM, Davis et al. 2008 APHRODITE NEJM. Confidence: C1 for trial outcomes, C2 for the framing.

Estrogen is the hormone whose role most often gets caricatured in both directions. Systemic estrogen-only or estrogen-plus-progestin therapy at typical menopausal-replacement doses produces, in the 2023 systematic review of hormone therapy for sexual function, "no effect to small benefit" on libido as a clinical endpoint. But endogenous estradiol fluctuations strongly track within-cycle desire in cycling women (Roney–Simmons positive E coefficient), supraphysiological-T-on-top-of-E is more effective than either alone in postmenopausal trials, and Cappelletti–Wallen explicitly write that "estrogen-only therapies that produce periovulatory levels of circulating estradiol increase sexual desire in postmenopausal women." Compounding all of this, oral ethinyl estradiol — and to a lesser extent oral 17β-estradiol — is a potent inducer of hepatic SHBG, and Zimmerman et al.'s 2014 meta-analysis showed combined oral contraceptive users sustain a 40-60% reduction in free testosterone and substantial declines in the free androgen index, with elevation persisting at least six months after discontinuation in Panzer's 2006 sexual-dysfunction cohort. The clinical consequence is that prescribing oral estrogen as a generic "libido boost" can drop bioavailable T enough to worsen sexual function in some women, and that part of the perceived benefit of switching from oral COC to transdermal estradiol is probably an SHBG correction rather than estrogen action per se. "Estrogen is weaker than testosterone for libido" is therefore an oversimplification: true for some clinical endpoints in postmenopausal women on HT, false at the within-cycle physiology level, and complicated by the SHBG mechanism that turns oral estrogen into a free-T suppressor. Sources: Hormone therapy systematic review 2023, PMC10227948, Cappelletti & Wallen 2016, PMC4720522, Zimmerman et al. 2014, PMC3845679, Panzer et al. 2006, PMID 16409223. Confidence: C1 for the underlying datasets, C2 for the synthesis.

Progesterone is the single most over-claimed and under-evidenced node in popular hormone-and-libido writing, and it deserves to be treated harshly. Folk-trans and bioidentical-HRT cultures both promote oral micronized progesterone as a libido enhancer; the published trial record does not support that. Saadat's 2002 Fertility and Sterility comparison of 200 mg oral micronized progesterone versus 10 mg medroxyprogesterone acetate in estrogen-treated postmenopausal women found neither agent influenced mood or libido, and Cumming's 2002 randomised comparison of micronized progesterone versus MPA in 27 estrogen-treated women was small, short, and only argues "progesterone is no worse than MPA," not "progesterone improves libido." Wren's 2003 transdermal progesterone cream trial reported no effect on sexual desire. Roney and Simmons 2013 found endogenous luteal progesterone was negatively associated with daily desire on a within-person basis, and Grebe et al. 2016 confirmed that within-cycle progesterone fluctuations negatively predict in-pair and extra-pair desire. Saraf and colleagues' 2023 trans-feminine cohort found no difference in patient-reported libido satisfaction between progesterone-using and progesterone-naïve groups. Pharmacologically, progesterone's 5α-reduced metabolite allopregnanolone is sedating and a positive allosteric modulator of GABA-A; high-dose progestins (MPA, CPA) are explicitly used clinically to suppress sexual desire in men with paraphilic disorders. The honest summary is that progesterone, as a class effect, is at best neutral and at worst suppressive for libido, and the "natural progesterone raised my desire" anecdotal pattern — real but heterogeneous — most plausibly reflects mood improvement, sleep improvement via allopregnanolone, or partner-context confounding rather than primary desire effect. There is no high-quality RCT showing exogenous micronized progesterone reliably increases libido in any defined population. Sources: Saadat et al. 2002 Fertil Steril, Cumming et al. 2002, Roney & Simmons 2013, PMID 23601091, Grebe et al. 2016, PMID 27049465, Saraf et al. 2023, PMID 37549733, Aufrere & Sirota TransfeminineScience review. Confidence: C1 for the trial record, C2 for the synthesis; the "progesterone raises libido in some" anecdotal pattern is C4.

DHEA, often promoted as a dark-horse libido hormone, also fails the harsh-fact-check bar in cis women with normal adrenal function. Elraiyah and colleagues' 2014 meta-analysis of 23 RCTs in JCEM found systemic DHEA in postmenopausal women with intact adrenal function did not significantly improve libido or sexual function and was associated with mild androgenic side effects. The Cochrane review concluded similarly, with at most a small effect on sexual function and a clear signal of androgenic side effects. Labrie's intracrinology framework and the 2016 FDA approval of intravaginal DHEA (prasterone, Intrarosa) reframed DHEA as a local genital-tract therapy for vulvovaginal atrophy and dyspareunia rather than a systemic libido drug, with intravaginal 6.5 mg/day improving dyspareunia symptoms while the systemic-libido story was essentially abandoned. The fact that vaginal but not oral DHEA works mainly tells us that the local androgen-and-estrogen tissue effect on genital atrophy is real, not that DHEA is a hidden libido lever. Sources: Elraiyah et al. 2014, PMC5393492, Labrie 2018, PMID 30130283. Confidence: C1.

The progestin question — does the user's instinct that "progestins lower libido by lowering testosterone" actually hold up? — has a more precise answer than the popular literature usually concedes, and it is worth working through because the answer reveals four or five distinct mechanisms rather than one. "Progestin" is a regulatory and pharmacological category, not a single molecule; the accepted classification splits the class into progesterone-derived pregnanes (MPA, megestrol, chlormadinone, cyproterone, dydrogesterone) and norpregnanes (nomegestrol acetate, trimegestone), testosterone-derived estranes (norethisterone) and gonanes (levonorgestrel, desogestrel, gestodene, etonogestrel, norgestimate), the hybrid dienogest, and the spironolactone-derived drospirenone. Lumping every member under one libido prediction is like predicting how much a "diuretic" lowers blood pressure without specifying loop versus thiazide versus MR antagonist. In the highest-dose forensic-psychiatric and oncologic settings, the bulk of the libido suppression really does come from HPG-axis collapse and hypogonadal testosterone — oral MPA at 60 mg/day suppresses serum T by 45–75%, depot at 150 mg pushes T into the castrate range, and ENIGI's Kuijpers 2021 dose-finding work showed cyproterone acetate is maximally suppressive in trans-feminine HRT at just 10 mg/day. But the Michael and Zumpe 1993 primate experiment is the cleanest experimental dissociation in the entire literature: castrated cynomolgus monkeys with subcutaneous testosterone implants holding plasma T in the upper-normal intact-male range still showed reduced ejaculatory and mounting behaviour after 40 mg/week intramuscular MPA, demonstrating that progestins can suppress sexual behaviour through direct hypothalamic action even when systemic androgen is preserved. Sources: García-Sáenz 2023, PMC10218893, Michael & Zumpe 1993, PMID 8511186, ENIGI Kuijpers 2021, PMC8571811. Confidence: C2.

The full mechanism map for progestin-driven libido suppression is therefore at least five-fold. First, HPG-axis suppression with downstream reduction in endogenous testosterone — the dominant pathway for high-dose MPA, CPA, GnRH analogues, and any progestin regimen that reliably suppresses ovulation including DMPA, the etonogestrel implant, the etonogestrel ring, and desogestrel-based progestin-only pills. Second, direct AR antagonism — cyproterone acetate is a high-potency AR antagonist with IC50 around 7 nmol/L, which is why it works at lower dose than MPA for trans-feminine androgen suppression; chlormadinone and dienogest are weaker AR antagonists; drospirenone is a mild AR antagonist with antimineralocorticoid activity inherited from its spironolactone parent. Third, SHBG elevation when ethinyl estradiol is on board — combined oral contraceptives roughly double SHBG, which means free testosterone falls disproportionately to total testosterone, and the Pastor 2013 review of 36 studies covering 13,673 women reported 85% of COC users with increased or unchanged libido and 15% with decrease, with the very-low-dose 15 µg EE pill the most reliably libido-suppressive formulation. Fourth, ovarian estradiol suppression — Boozalis's 2016 cross-sectional analysis of the Contraceptive CHOICE Project found 37.3% of DMPA users reported lack of interest in sex versus 18.3% of copper-IUD users, with the etonogestrel implant and ring elevated as well, and the gradient survived the ECHO randomised trial in 7,829 sub-Saharan African women, suggesting that flattening endogenous estradiol does at least as much work as direct androgen suppression for the systemic ovulation-suppressors. Fifth, central nervous system effects via PR-expressing hypothalamic and limbic neurons and via allopregnanolone-driven mood/affect changes — the Bitran 1995 anxiolytic neurosteroid story and the PMDD literature's biphasic luteal-phase effects intersect with sexual desire even when no peripheral hormone has changed. The compact answer: lowering testosterone is the dominant mechanism in high-dose anti-androgen and trans-feminine contexts, but it is incomplete for COC use, arguably wrong as the primary explanation for DMPA/implant decrement, and supplemented by AR antagonism, SHBG-mediated free-T suppression, and central neurosteroid effects depending on which progestin and which dose. Sources: Pastor 2013, PMID 23320933, Boozalis et al. 2016, PMID 26855094, Hofmeyr et al. ECHO 2024, PMID 38722832, Zethraeus et al. 2017 RCT, PMID 27525531, Hantsoo & Epperson 2020, PMC7231988. Confidence: C2.

The PDE5 inhibitors occupy a different conceptual slot, and the popular framing is mostly correct: they are blood-flow drugs, not central libido modulators. Sildenafil, tadalafil, vardenafil, and avanafil all competitively inhibit phosphodiesterase type 5, the enzyme that hydrolyses cyclic GMP in vascular and corpus-cavernosum smooth muscle, allowing the cGMP that nitric-oxide-driven sexual stimulation generates to persist long enough to maintain helicine-artery dilation and venous compression. None of that pathway is dopaminergic, melanocortinergic, or noradrenergic in the brain, and the FDA labels are explicit that the drugs require upstream sexual stimulation to act. The pharmacokinetic numbers are tighter than community writing usually treats them: sildenafil's terminal half-life is "about 4 hours" with Tmax 30–120 minutes, vardenafil sits around 4–5 hours, avanafil around 5 hours with the fastest 30–45 minute Tmax, and tadalafil's terminal half-life is 17.5 hours (5th–95th percentile 11.5–29.6, rising to 21.6 hours in elderly men) with food-independent absorption — the basis of the "weekend pill" framing and the daily-dosing route. Tadalafil is structurally distinct (a methylenedioxyphenyl-pyrazinopyrido-indoledione rather than a sulfonyl-substituted pyrazolopyrimidinone) with the cleanest PDE6 selectivity in the class (~550-fold versus PDE5, compared with sildenafil's ~16-fold), which is why blue-tinted vision is rare on tadalafil and common on sildenafil. Calling tadalafil "basically a longer-lasting sildenafil" is a useful first approximation that understates two real differences — action curve and selectivity — but does no serious harm to clinical understanding. Sources: Viagra label 2014, Cialis label 2018, Forgue tadalafil PK, PMC1885023, PDE5 inhibitors StatPearls, Wang et al. avanafil selectivity, PMID 22759639. Confidence: C1.

The contraindication map is short but emphatic, and "few side effects" undersells the seriousness of the entries that exist. Co-administration with any nitrate or nitric-oxide donor — isosorbide mononitrate or dinitrate, sublingual or transdermal nitroglycerin, recreational alkyl nitrites ("poppers"), nicorandil — is an absolute contraindication because the multiplicative cGMP excursion produces severe and sometimes fatal hypotension; the nitrate-free interval should be at least 24 hours after sildenafil or vardenafil and 48 hours after tadalafil. Riociguat, the soluble guanylate cyclase stimulator, hits the same final pathway from another direction and is also absolutely contraindicated. Doxazosin, the older non-selective alpha-blocker, dropped 28% of healthy normotensive men below 85 mmHg standing systolic when co-administered with tadalafil 20 mg in Kloner's 2004 study; tamsulosin is much better tolerated. Common adverse events run headache 10–16%, flushing ~10%, dyspepsia 4–7%, nasal congestion ~4%, back pain ~6% (tadalafil), myalgia ~5% (tadalafil), and dose-dependent visual disturbances most prominent on sildenafil; rare serious events include non-arteritic anterior ischemic optic neuropathy (NAION) with a contested but real signal in vasculopathic men with small cup-to-disc ratios, sudden sensorineural hearing loss prompting the 2007 FDA label update, and priapism at well below 1 in 10,000 in standard prescribing. None of this contradicts "few side effects" as a summary for a healthy outpatient, but it should be paired with the labelled warnings rather than treated as "no risk." The reassuring news is that tachyphylaxis is not real in the strict pharmacological sense; six-month tadalafil RCTs specifically tested for tolerance and did not find it, and what patients perceive as "Cialis stopped working" is almost always organic-disease progression, deteriorating sleep or testosterone status, or anxiety rebound rather than receptor adaptation. Sources: Viagra label, Adempas (riociguat) label, Kloner et al. 2004, PMID 15540759, Lui WHO pharmacovigilance 2023, PMID 36905319, Penedones NAION meta-analysis 2020. Confidence: C1-C2.

The all-cause-mortality literature on tadalafil is the place where a dismissive "weak evidence" framing has aged worst. The pivotal Andersson 2021 JACC study used the Swedish Patient Register and Prescribed Drug Register to compare 16,548 men on PDE5 inhibitors with 1,994 on alprostadil after prior MI or revascularization, with mean follow-up of 5.8 years; PDE5i exposure was associated with a hazard ratio of 0.88 (95% CI 0.79–0.98) for all-cause mortality, with dose-response across quintiles. Anderson's 2017 Heart cohort of 5,956 men with type 2 diabetes reported a 31% lower adjusted all-cause mortality hazard. Kloner and colleagues' 2024 US commercial-claims cohort of 8,156 tadalafil-exposed and 21,012 unexposed men, propensity-matched, reported HR 0.56 (95% CI 0.43–0.74) for all-cause mortality, falling to HR 0.40 in the top exposure quartile. Jehle's 2025 TriNetX analysis of over 500,000 men with ED and over a million with lower-urinary-tract symptoms reported relative risks of 0.66/0.76 for all-cause mortality, 0.73/0.83 for myocardial infarction, 0.66/0.78 for stroke, 0.79/0.80 for venous thromboembolism, and 0.68/0.75 for dementia at three years (tadalafil/sildenafil), and Mostafa's 2024 EHJ-CVP meta-analysis pooled 16 studies in the same direction. The Xiao 2025 Mendelian-randomization analysis showed genetically predicted increased PDE5A expression correlated with elevated coronary heart disease and MI risk, providing the cleanest causal-inference foothold the field has. Healthy-user bias remains real — men who fill these prescriptions tend to be sexually active, mobile, well-engaged with primary care, and on secondary-prevention medication that independently lowers mortality — but the active-comparator design (Andersson) addresses the worst of it, and the consistency across designs and databases makes "weak evidence" a genuine undersell. The better framing is "consistent observational evidence with residual confounding, dose-responsive, mechanistically plausible through endothelial and platelet effects, no RCT confirmation forthcoming because no funder will randomize older men to placebo for years to test an off-label survival benefit." Sources: Andersson et al. JACC 2021, PMID 33766260, Anderson Heart 2017, PMID 27465053, Kloner Clin Cardiol 2024, PMID 38377018, Jehle AJM 2024, PMID 39532245, Mostafa systematic review, PMC11323371, Xiao MR analysis EHJ-CVP 2025. Confidence: C2-C3.

PT-141 (bremelanotide, sold as Vyleesi) is the central drug of the melanocortin pathway, and the popular framing of it as an "FDA-approved libido drug" is wrong by an important margin. The actual FDA-approved indication, granted June 21, 2019 under NDA 210557, is the treatment of premenopausal women with acquired, generalized hypoactive sexual desire disorder; the label explicitly excludes postmenopausal women, men, and "enhancement of sexual performance" use. The drug is a structural cyclic-lactam α-MSH analog descended from Mac Hadley and Victor Hruby's 1980s University of Arizona work on Melanotan II, with the C-terminal amide replaced by a free carboxylic acid to shift the receptor potency away from cutaneous MC1R and toward central MC4R/MC3R. The receptor potency order at the four bound receptors in vitro, per the label, is MC1R > MC4R > MC3R > MC5R > MC2R — meaning bremelanotide is still ranked highest at MC1R, but the absolute MC1R activity is much lower than MT-II's, which is exactly why focal hyperpigmentation appears on the FDA label rather than disappearing with the redesign. The dose is 1.75 mg subcutaneously into abdomen or thigh via single-use autoinjector, no sooner than 45 minutes before anticipated activity, with no more than one dose in any 24-hour period and no more than eight doses in any month; pharmacokinetics give Tmax around 1 hour and a mean terminal half-life of 2.7 hours (range 1.9–4.0). The user's "2–4 hour half-life" claim is correct as a stated range; the central tendency sits at the lower end. Sources: Vyleesi prescribing information at DailyMed, FDA Vyleesi label PDF, Hadley & Dorr 2006, PMID 16412534, Wessells 1998 J Urol, PMID 9679884. Confidence: C1.

The efficacy data from the RECONNECT phase-3 program need to be read with effect-size honesty. Studies 301 and 302 enrolled 1,267 premenopausal women with acquired generalised HSDD across 24 weeks of on-demand 1.75 mg bremelanotide or matching placebo; both co-primary endpoints — change from baseline on the Female Sexual Function Index–desire domain and on item 13 of the Female Sexual Distress Scale–DAO — were statistically significant, with integrated FSFI-D improvement around 0.30–0.35 points on a 1.2–6.0 scale and FSDS-DAO item 13 reduction around 0.29–0.37 points on a 0–4 scale. The responder analysis using a dynamic anchor reported approximately 58% bremelanotide responders versus approximately 36% placebo responders on FSFI-D — a drug-placebo delta of about 22 percentage points. That is a meaningful effect at the population level and a clearly small effect at the individual level. The placebo response of 36% is genuinely substantial: women in HSDD trials respond to attention, structured assessment, and protocolized expectation-setting at high rates, and the absolute drug-versus-placebo effect compresses into a modest band as a consequence. The total adverse-event load is qualitatively heavier than the standard PDE5i profile: nausea 40.0% (versus 1.3% placebo) with median onset around 30 minutes and median duration 2.4 hours, flushing 20.3%, headache 11.3%, injection-site reactions ~13%, and discontinuation due to adverse events of 18.7–23.4% in the open-label extension's original-bremelanotide arms and 24–26% in the placebo-to-bremelanotide crossover groups. Sources: Kingsberg et al. 2019 Obstet Gynecol RECONNECT, PMID 31599840, RECONNECT integrated analysis, PMC6819021, Clayton et al. 2022 safety profile, PMID 35147466. Confidence: C2.

Focal hyperpigmentation deserves its own treatment because it is bremelanotide's most distinctive adverse event and the one whose mechanism is genuinely "interesting." The label-quoted rate is approximately 1% in patients receiving up to eight doses per month — the labelled monthly maximum — but the rate climbs sharply with daily or near-daily dosing: the Vyleesi label specifically reports focal hyperpigmentation in approximately 38% of subjects who received up to 16 consecutive daily doses in dedicated dose-frequency studies, with affected sites including face, gingiva, areolae, and breasts. The mechanism is direct: bremelanotide is a structural constrained-cyclic α-MSH analog acting on the same MC1R that endogenous α-MSH uses to drive cAMP-mediated transcription of tyrosinase and the eumelanin synthesis pathway. The label's potency-order ranking lists MC1R first among bound receptors, so residual MC1R agonism is part of bremelanotide's pharmacology even at the lowered MC1R activity that distinguishes it from MT-II. Label-compliant on-demand dosing keeps cumulative MC1R exposure low enough that pigmentation usually does not become clinically apparent; off-label or daily dosing turns the same residual activity into visible pigmentation, especially in patients with darker skin who already carry higher constitutive eumelanin. The user's "1-in-100, interesting reasons" framing is accurate at labelled dosing and substantially understated for any off-label-frequency use. Sources: Vyleesi prescribing information, FDA Vyleesi label PDF. Confidence: C1.

Melanotan II is not the same regulatory or safety object as bremelanotide, and any reading that treats them as interchangeable is wrong. MT-II has never been approved by the FDA or EMA for any indication; it is sold online as research-grade lyophilized powder, often imported from gray-market suppliers, and used by consumers primarily for tanning and secondarily for libido. Independent analyses of online-sourced peptide vials have repeatedly shown wide variation in actual peptide content, contamination with bacterial endotoxin, presence of unrelated peptides, and underdosing or overdosing relative to label, none of which a buyer can verify at home. The case-report literature has accumulated several distinct flavours of harm: melanocytic concerns, with the Cardones and Grichnik 2014 case of a 20-year-old woman with a histologically confirmed melanoma in a previously inconspicuous gluteal lesion diagnosed during MT-II use; systemic sympathomimetic toxicity with rhabdomyolysis and acute kidney injury after a 6 mg dose roughly six times typical starting amounts (Nelson 2012); priapism case reports; and a Swedish case of renal infarction after cumulative 27 mg over six months. Case reports cannot establish causality, and the underlying epidemiology of melanoma in young tanning-bed users is already non-trivial, but constitutive MC1R activation has shown tumorigenic potential in animal models and the case-report cluster is biologically plausible. The honest framing for the user's "if you want to use either, use PT-141" recommendation is a tightening: MT-II is not an acceptable substitute for bremelanotide for any libido use case; if you are considering melanocortin agonism for libido, only bremelanotide has the regulatory and manufacturing record to be defensible, and it should be evaluated against PDE5 inhibitors and against flibanserin within its FDA-approved population, with the contraindication and side-effect profile fully internalised. Sources: Cardones & Grichnik 2014, PMID 24355990, Nelson et al. 2012, PMID 23121206, DermNet NZ Melanotan II overview. Confidence: C2.

Flibanserin (Addyi) deserves to be treated alongside bremelanotide because it is the original FDA-approved drug for HSDD — granted approval in August 2015 under NDA 022526, four years before bremelanotide — and any libido-pharmacology summary that misses this gets the chronology of the field backwards. Flibanserin's pharmacology is fundamentally different from bremelanotide's: it is a daily, chronic, oral 100 mg bedtime tablet whose effect emerges over weeks, classified as a "multifunctional serotonin agonist and antagonist" with full agonism at 5-HT1A in the prefrontal cortex and dorsal raphe, antagonism at 5-HT2A, and very weak partial agonism at dopamine D4. The downstream consequence in animal microdialysis is increased prefrontal dopamine and norepinephrine and decreased 5-HT — the corrective re-balancing that Stahl framed as the mechanistic claim the regulator credited. The approval was contentious: the FDA rejected the drug twice (2010, 2013) before Sprout Pharmaceuticals reorganised the dossier and engaged the "Even The Score" advocacy campaign, which claimed the FDA had approved 26 drugs for male sexual dysfunction and zero for women — a claim that was not actually true and which critics including the Hastings Center described as pharmaceutical marketing weaponising feminist framing to lower the regulator's risk-benefit threshold. Jaspers and colleagues' 2016 JAMA Internal Medicine meta-analysis of eight trials showed a mean increase of about 0.49 satisfying sexual events per month over placebo — half of one extra event monthly — with dizziness roughly four times placebo, somnolence and fatigue and nausea each roughly doubled, and the quality of evidence rated "very low." The original 2015 label included a near-absolute alcohol contraindication based on a Phase 1 study with a near-all-male population; in April 2019 the FDA softened it to a 2-hour-wait window for one or two drinks, and in October 2019 removed the alcohol contraindication and the REMS certification entirely while keeping the boxed warning about hypotension and syncope. Commercially the drug had already foundered — low uptake, daily dosing, the alcohol restriction during the most uptake-sensitive years, and a small effect size that was hard to brand convincingly all combined to make Addyi a textbook example of an FDA-approved drug that lost its market. Sources: Joffe 2016 NEJM Perspectives, Stahl 2015 mechanism review, Jaspers et al. 2016 JAMA Intern Med, PMID 26927498, FDA Addyi 2015 label. Confidence: C1.

The post-orgasmic refractory period is one of the more under-studied aspects of human sexuality, and the popular pharmacological story attached to it has been substantially undermined since 2020. Roy Levin's 2009 review in the Journal of Sexual Medicine opens with the blunt observation that "the paucity of data for most aspects of human PERT is remarkable; even the generally accepted statement that the duration of PERT increases with age has no published support data," and Turley and Rowland's 2013 BJU International review reaches the same conclusion. The canonical cabergoline study quoted in popular writeups is Krüger et al. 2003 in the Journal of Endocrinology (PMID 14656205, DOI 10.1677/joe.0.1790357), not a Lancet paper as the popular telling sometimes claims; the design was a single-blind, placebo-controlled, balanced cross-over with n=10, in which prolactin levels were pharmacologically increased by intravenous protirelin or reduced by oral cabergoline 0.5 mg, after which sexual arousal and orgasm were induced by erotic film and masturbation. The verbatim abstract reports "administration of cabergoline decreased prolactin levels and significantly enhanced all parameters of sexual drive (P<0.05), function (P<0.01) and positive perception of the refractory period (P<0.01). Administration of protirelin increased prolactin concentrations and produced small, but not significant reductions of sexual parameters. The sexual effects observed from cabergoline were completely abrogated by coadministration of protirelin." That is real and quote-accurate, but the design's statistical power for the protirelin arm was inadequate, the outcome was self-reported "positive perception" of the refractory period rather than measured shortening of time to second erection, and cabergoline acts at multiple dopamine targets (tuberoinfundibular D2 that suppress prolactin, but also striatal and limbic D2 sites that directly modulate sexual desire), so the design cannot distinguish "prolactin caused refractory" from "central D2 agonism enhanced subjective sexual reports." Sources: Levin 2009 J Sex Med, PMID 19515210, Turley & Rowland 2013 BJU Int, PMID 23470051, Krüger et al. 2003 J Endocrinol, PMID 14656205. Confidence: C4-C5.

The Valente, Marques and Lima 2021 paper in Communications Biology is the experiment that broke the prolactin hypothesis cleanly. They used domperidone — a peripheral D2 antagonist that does not cross the blood-brain barrier and therefore raises prolactin without confounding central D2 effects — and bromocriptine in two strains of mice, and showed that domperidone-induced prolactin elevation did not lengthen the refractory period and bromocriptine-induced prolactin suppression did not shorten it. The verbatim conclusion is that they "show compelling evidence refuting the idea that prolactin released during copulation is involved in the establishment of the refractory period, a long-standing hypothesis in the field of behavioral endocrinology." The cleanest objection to Krüger 2003 is therefore now experimentally addressed: the cabergoline result almost certainly reflected central D2 agonism affecting sexual drive directly, not an unloading of prolactin's "feedback brake." The mouse-to-human extrapolation is not airtight, but combined with Levin's 2009 explicit rejection and the Turley–Rowland 2013 critique, the field has moved decisively away from the prolactin-causes-refractory-period model. The post-orgasmic prolactin spike is real; it is probably a downstream marker rather than a causal upstream regulator of post-orgasmic sexual quiescence. Truitt and Coolen's 2002 Science paper identified the lumbar spinothalamic neurons in rat segments L3–L4 expressing neurokinin-1 receptors as a "spinal ejaculation generator" whose ablation by SSP-saporin completely disrupts ejaculatory behaviour while leaving other sexual behaviour intact — anatomical confirmation that orgasm (supraspinal) and ejaculation (spinal) are mechanistically separable, and the framework that explains why PDE5 inhibitors restore the erection during what would otherwise be a flaccid post-orgasmic window without resolving central refractoriness. Cabergoline's cardiac valvulopathy risk, often raised as a generic warning, is dose-dependent in a way that matters for off-label dosing: the original signal came from Parkinson's-disease cohorts on cumulative doses of 3 mg/day or more, where the Stiles 2018 meta-analysis confirmed elevated mild tricuspid regurgitation but no clinically significant valvular dysfunction at prolactinoma-scale doses (0.25–2 mg/week), and the 2021 CATCH primary-care cohort using hard endpoints found no increase in valve replacement or heart failure. Sources: Valente et al. 2021 Commun Biol, PMC7782750, Truitt & Coolen 2002 Science, PMID 12202834, Stiles et al. 2018 JCEM, PMID 30215804, Drake et al. CATCH 2021 JCEM, PMC7823250, Dostinex FDA label 2025. Confidence: C4.

The wider dampener landscape that the popular hormone-and-libido draft barely covers is enormous, and the SSRIs sit at its centre. The Spanish multicentre prospective study of 1,022 outpatients by Montejo and colleagues found sexual dysfunction emerged in 57.7% on fluoxetine, 62.9% on sertraline, 62.3% on fluvoxamine, and 70.7% on paroxetine, with bupropion and mirtazapine far lower; Serretti and Chiesa's 2009 meta-analysis pooled treatment-emergent sexual dysfunction rates ranging from roughly 25% with bupropion to over 70% with the most serotonergic agents. The mechanism is not a single lesion but a pile of 5-HT-mediated effects on top of dopaminergic and nitrergic systems: increased synaptic serotonin acts on 5-HT2A and 5-HT2C receptors to inhibit mesolimbic dopamine release, on 5-HT3 receptors to dampen orgasmic reflex pathways, and reduces central nitric-oxide signalling. Bupropion, a norepinephrine-dopamine reuptake inhibitor, sits cleanly outside this pathway, which is why its baseline sexual-dysfunction rate is near placebo and why it is the most-used switch and adjunct strategy; Clayton's 2004 trial of sustained-release bupropion 150 mg twice daily and Safarinejad's 2010 male SRI-treated cohort both confirmed the rescue effect, and a 2022–2024 wave of systematic reviews finds bupropion the most effective pharmacological treatment for antidepressant-induced sexual dysfunction. Mirtazapine's 5-HT2A and 5-HT2C antagonism gives it a similarly low rate (around 25% versus 60% for escitalopram in head-to-head), and vortioxetine in pooled phase-3 data showed treatment-emergent sexual adverse events of 1.6–1.8% versus 1.0% on placebo, although clinicians should treat the "placebo-like" framing cautiously because depression itself depresses libido and pre-treatment 35–50% of people with untreated major depression report reduced sexual desire prior to any drug exposure. Drug holidays work for short-half-life agents like sertraline and paroxetine but not for fluoxetine, whose long-lived norfluoxetine metabolite makes pharmacokinetic dips impractical. Sources: Montejo et al. 2001, PMID 11229449, Serretti & Chiesa 2009, PMID 19440080, Clayton et al. 2004, PMID 11894796, Clayton et al. 2019 vortioxetine vs paroxetine RCT, PMID 31405765, Safarinejad 2010 BJU Int. Confidence: C2.

Post-SSRI sexual dysfunction (PSSD) is the harder problem, and it is the place where the antidepressant story stops looking like a routine pharmacology chapter and starts looking like a contested clinical entity. The phenomenon — genital anaesthesia, reduced or absent libido, anorgasmia, pleasureless orgasm, and emotional blunting that persist for months to years after SSRI/SNRI discontinuation — was first described in case reports in the early 2000s, and after a citizen petition the European Medicines Agency's Pharmacovigilance Risk Assessment Committee concluded in May 2019 that sexual dysfunction can be long-lasting after SSRI/SNRI discontinuation in some patients and required label changes across the class; Health Canada issued a similar safety communication in 2021. The mechanism remains genuinely unknown — hypotheses include persistent epigenetic 5-HT1A downregulation, durable serotonin-dopamine circuit alteration, neurosteroid disruption, and possibly peripheral small-fibre neuropathy in genital tissue. Estimating prevalence is hard because spontaneous-reporting databases are biased; Healy and colleagues' 2023 estimate of around one in 216 patients prescribed serotonergic antidepressants is heavily caveated. Post-finasteride syndrome is the analogue, a contested entity in which patients report persistent sexual, neuropsychiatric, and somatic symptoms after stopping finasteride; Melcangi's group at Milano-Bicocca has documented altered cerebrospinal fluid neurosteroid profiles and abnormal pudendal somatosensory evoked potentials in PFS patients with severe ED, and the FDA's August 2022 PROPECIA label revision added depression and suicidal ideation under nervous-system/psychiatric adverse reactions, while explicitly stating that available data did not establish causation for persistent sexual dysfunction beyond what was already labelled. The honest read on both PSSD and PFS is that they have cleared the bar of "real enough to label-warn" but not the bar of "well-defined incidence and pathobiology"; the discrepancy between original-trial sexual side-effect rates (low single digits for finasteride) and the loud anecdotal cohort is unresolved, and pre-existing depression, body-image distress in alopecia patients, and a real but rare neuroendocrine phenotype probably each contribute. Sources: EMA PRAC 2019 SSRI label update, Healy et al. 2023, PMC10122283, Melcangi et al. 2017, PMID 28408350, PROPECIA 2022 FDA label. Confidence: C2-C4.

Hyperprolactinemia-inducing drugs, opioids, alcohol, cannabis, and antihypertensives round out the dampener landscape. Antipsychotics that potently and persistently block D2 — risperidone, paliperidone, haloperidol, depot phenothiazines — disinhibit pituitary lactotrophs from tuberoinfundibular dopaminergic tone and produce decreased libido in 37.8%, erectile dysfunction in 32%, and ejaculatory disorders in 33% of risperidone users; aripiprazole's D2 partial-agonist profile rescues prolactin in around half of cases in adjunctive trials and reduces sexual side effects in parallel; clozapine often does not raise prolactin at all. Opioid-induced androgen deficiency is the cleanest example of HPG-axis suppression as a libido mechanism, with chronic mu-agonism suppressing kisspeptin and GnRH signalling and producing secondary hypogonadism in 19–86% of chronic-pain and maintenance cohorts; methadone consistently produces more profound hypogonadism than buprenorphine in head-to-head studies, with meta-analytic odds ratios for sexual dysfunction running around 4. Alcohol's acute disinhibitory phase raises self-reported desire while reliably impairing erectile capacity, and chronic heavy intake suppresses Leydig-cell testosterone synthesis and accelerates hepatic catabolism — Santi 2024 confirmed reductions in total and free testosterone with chronic heavy use, and 90% of men with alcoholic cirrhosis have low testosterone. Cannabis is genuinely mixed: occasional use frequently increases self-reported sexual interest in survey data (Sun & Eisenberg 2017 NHANES) but heavy chronic use shows negative associations with testosterone, erectile function, and sperm parameters. Beta-blockers split cleanly by lipophilicity: propranolol especially shows consistent decrement; nebivolol, the third-generation β1-selective agent that releases NO via β3 activation, breaks the pattern with roughly 69% improvement at 3 months in Brixius's switch study. Sources: Park et al. 2012 antipsychotics review, PMC3623530, Brennan 2013 OPIAD review, Yee et al. 2014 meta-analysis, Santi et al. 2024 Andrology, Sun & Eisenberg 2017, PMID 28395129, Brixius et al. 2007, PMID 17593175. Confidence: C2.

The trans HRT literature is one of the cleanest natural experiments human endocrinology has on which sex steroid actually drives libido in adults, but the popular "huge drop / huge surge" framing oversimplifies the result. In Wierckx's 2014 one-year ENIGI prospective study of 53 trans women on cyproterone acetate plus estradiol, "low sex drive" emerged as a statistically significant new symptom within the first year (P ≤ 0.02), and the cross-sectional cohort of 214 trans women reported 62.4% with decreased desire, 73% rarely or never experiencing spontaneous and responsive sexual desire, and a DSM-HSDD prevalence of 22% — more than four times the 5% prevalence in trans men, but importantly not universal. Defreyne's 2020 ENIGI longitudinal follow-up of 401 trans women using the Sexual Desire Inventory showed total, dyadic, and solitary SDI scores all dropping during the first three months and then partially rebounding such that by 36 months total and dyadic SDI scores actually exceeded baseline, while solitary scores returned to baseline. The transmasculine surge is also more textured than the popular telling. Wierckx's trans-men cohort showed a statistically significant rise in desire (P ≤ 0.01) within weeks, with 71% of 138 men reporting increased desire and 5% HSDD prevalence, and Defreyne's longitudinal data showed sharp rises over the first three months and then plateaus, with total/dyadic SDI returning to baseline by 36 months and only solitary desire remaining above baseline. Crucially, both papers report no within-person correlation between absolute serum testosterone and SDI scores; Wierckx 2011 in post-SRS trans men found the same null T-desire correlation but did find an LH-desire correlation, suggesting that adequacy of replacement matters more than absolute level. The cleanest summary: testosterone is necessary for typical adult desire in both sexes within reasonable limits, but it is not sufficient, and it does not titrate desire in a clean dose-response way once a person is in their replacement window. Sources: Wierckx 2014 PMID 24828032, Wierckx 2014 PMID 24165564, Defreyne 2020 PMID 32008926, Wierckx 2011 PMID 21602316. Confidence: C1 for population direction, C2 for the absence of within-person T-libido titration.

Within trans care, the choice of androgen blocker is a libido variable that is often underplayed. Burinkul's 2021 RCT comparing CPA 25 mg/day to spironolactone 100 mg/day in trans women found CPA achieved female-range total testosterone (<50 ng/dL) in 90% versus 19% on spironolactone, which means CPA-based regimens drive deeper androgen suppression and the deepest end of the libido drop. Spironolactone, even at 100–200 mg/day, often leaves total T at 100–200 ng/dL while still producing some receptor blockade, which correlates with milder reported drops in subjective desire. Bicalutamide is a peripheral AR antagonist that does not lower androgen levels (and may raise total T modestly via feedback) and is therefore the antiandrogen with the lowest expected effect on libido and erectile function — although the trans-specific evidence base is small. GnRH agonists produce near-complete pituitary-gonadal axis shutdown after the initial flare and are functionally analogous to medical orchiectomy. Estradiol level is its own modulator: at supraphysiological levels (>300–400 pg/mL chronically) several mechanisms converge to deepen libido suppression — strong gonadotropin suppression drives endogenous T even lower, hepatic SHBG production rises (especially with oral routes), and in some patients high estradiol itself appears to produce a sedating or mood-flattening effect. Some patients reporting persistent low desire on transfem HRT are over-suppressed, not under-suppressed, and dose reduction toward the physiological female range can sometimes rescue desire without reversing feminization. The progesterone-for-trans-women hypothesis advanced by Jerilynn Prior's 2019 JCEM editorial has not been supported by the empirical evidence on libido: the 2023 randomised trial of progesterone added to feminizing HRT and the 2024 retrospective cohort both found no statistically significant difference in libido satisfaction between standard GAHT and GAHT-plus-progesterone arms. Sources: Burinkul 2021, PMID 34274044, Angus et al. 2021 ENIGI low-dose CPA, PMC8571811, Patel 2023 JAPhA progesterone GAHT trial, TransfeminineScience progestogens-and-sexual-desire review. Confidence: C2.

The dopamine system is the part of the human-libido apparatus that the popular hormone-centric framing systematically underweights. Apomorphine, a non-selective D1/D2 agonist, was marketed in the EU as Uprima from May 2001 for ED with onset within 15–25 minutes after sublingual dosing, but its Marketing Authorisation lapsed on 28 May 2006 because dose-limiting nausea, dizziness, and occasional syncope made it commercially unviable against the cleaner-profile PDE5 inhibitors. The Parkinson's-disease literature provides the clean natural experiment: pramipexole and ropinirole induce impulse-control disorders, with hypersexuality affecting roughly 2–8% of pramipexole-treated patients and aggregate ICD prevalence around 32%; pramipexole is relatively D3-selective compared with other agonists, and D3 receptor density is high in mesolimbic regions associated with reward, making this the strongest naturalistic evidence in humans that artificially boosting D3-skewed dopaminergic transmission can drive a clinically recognisable hypersexual phenotype. The conceptual frame the popular literature most needs is Berridge and Robinson's wanting-versus-liking dissociation: mesolimbic dopamine is necessary for incentive salience or "wanting" but not for the hedonic impact or "liking" of consumption, and pathological dopamine elevation can crank wanting without proportionally cranking pleasure — applied to libido, this predicts that drugs that boost dopamine produce hypersexual urges that are not necessarily accompanied by greater pleasure, and it explains why "more dopamine" is not a clean win. Stimulant-induced hypersexuality is the population-scale shadow: methamphetamine acutely raises desire, lengthens activity, and lowers risk-perception, with chronic use producing tolerance and frequent sexual dysfunction; the chemsex literature triangulates heavily on this point. Sources: EMA Uprima EPAR, Voon et al. ICD review, PMC2924724, Berridge 2007 Psychopharmacology, PMID 17072591, Robinson & Berridge 2024 Annu Rev Psychol. Confidence: C2.

Kisspeptin is the peptide that any forward-looking libido map should flag as the most credible candidate for the next pharmacological class. The KISS1 gene encodes kisspeptin, which signals through KISS1R (formerly GPR54) to drive GnRH neurons; loss-of-function mutations cause hypogonadotropic hypogonadism. Comninos and colleagues at Imperial College London have built an unusually clean translational programme: Comninos 2017 in J Clin Invest gave intravenous kisspeptin or vehicle to 29 healthy heterosexual men in double-blind crossover with fMRI and found enhanced limbic and paralimbic activity to sexual cues correlated with reduced sexual aversion; the 2018 PNAS paper showed altered limbic connectivity at rest; the 2022 JAMA Network Open trial in men with HSDD showed enhanced sexual brain processing and increased penile tumescence, and the parallel 2022 trial in women with HSDD showed modulation of sexual and facial-attraction processing in the predicted direction. Kisspeptin is administered intravenously, orally bioavailable analogues are early-stage, and the responder-level effects in HSDD are not yet well-characterised, but mechanism (a central GnRH-axis modulator with limbic effects) is more upstream than anything currently FDA-approved. Sources: Comninos et al. 2017, PMID 28112678, Comninos et al. 2022 men HSDD, PMID 36735255, Mills et al. 2022 women HSDD, PMID 36287566. Confidence: C2.

The botanical and supplement landscape rewards a willingness to call out hype. Yohimbine, an alpha-2 adrenoceptor antagonist from Pausinystalia johimbe bark, has the cleanest non-prescription evidence: Ernst & Pittler's 1998 J Urol meta-analysis pooled seven trials and reported an odds ratio of about 3.85 over placebo, with effect concentrated in psychogenic and mixed-aetiology ED. Maca (Lepidium meyenii) is the most defensible non-prescription option, with Shin's 2010 systematic review finding "limited evidence" of benefit and Dording's 2008/2015 dose-finding trials showing significant improvements at 3.0 g/day for SSRI-induced sexual dysfunction in women — modest, plausibly real, mechanism unclear, side-effect profile benign. Saffron (Crocus sativus) is the surprise: Modabbernia's 2012 RCT of saffron 30 mg/day in women with fluoxetine-induced sexual dysfunction found significant improvements in arousal, lubrication, and pain, and saffron has independent evidence for mild-to-moderate depression, with the open question being whether the sexual benefit is a saffron-specific effect or downstream rescue from depression. Tribulus terrestris is the supplement-aisle counter-example: cumulative evidence shows it does not meaningfully raise testosterone in eugonadal men despite decades of marketing claims, and Lopes's 2025 Nutrients systematic review continues to find a low-quality evidence base. Korean red ginseng sits between maca and Tribulus on the evidence hierarchy — Jang's 2008 meta-analysis found a pooled risk ratio of about 2.40 for ED but the 2021 Cochrane review concluded only a "trivial effect" relative to placebo with low-certainty evidence. L-arginine and L-citrulline are nitric-oxide precursors with peripheral vascular effect rather than a libido mechanism; fenugreek is a small-effect testosterone modulator with industry-funded libido data; icariin (horny goat weed) has real but ~80-fold weaker PDE5 inhibitory activity than sildenafil and poor oral bioavailability; oysters and chocolate are aphrodisiacs in the cultural-symbolic sense, which is the placebo literature speaking. Sources: Ernst & Pittler 1998 J Urol, PMID 9649257, Shin et al. 2010, PMID 20691074, Dording et al. 2015, PMC6494062, Modabbernia et al. 2012, PMID 23280545, Lopes et al. 2025 Nutrients, PMC11990417, Lee et al. 2021 Cochrane, PMID 33871063. Confidence: C2-C3.

The single most important methodological point that the popular literature on this topic systematically skips is that placebo response in sexual-dysfunction trials is enormous. In HSDD and FSAD trials, 35–45% of women on placebo meet responder criteria, which is approximately the active-arm responder rate of flibanserin and not far below bremelanotide's 58%. The mechanisms are not mysterious: enrolling in a trial requires self-identification of distress, attention from clinicians, behavioural changes (regular tracking, partner conversations, mandatory sexual encounters), and the simple expectancy effect of being treated. Bradford and Meston have published methodological work on correlates of placebo response; the FSD trial literature has actively grappled with this for years. The implication is severe: any drug whose active-minus-placebo delta is one extra satisfying sexual event per month is operating in the noise floor of expectancy and behaviour change, and the case for prescribing it has to be built on individual responders rather than on aggregate effect sizes. The DSM-5 in 2013 collapsed the older HSDD and female sexual arousal disorder diagnoses into a single Female Sexual Interest/Arousal Disorder (FSIAD), arguing that desire and arousal in women are highly comorbid in long-term partnered contexts; ISSWSH and the International Consultation in Sexual Medicine continue to maintain a separate nosology that retains HSDD and FSAD, on the argument that the new construct is too heterogeneous for trials and treatment-matching. From a libido-pharmacology standpoint this matters because it determines what is being measured: a drug effective for women whose problem is genuinely low spontaneous desire may look identical in trial data to a drug effective for women whose problem is responsive arousal failure. Sources: Bradford 2017, Bradford & Meston correlates of placebo response, Brotto FSIAD review. Confidence: C2.

The Coolidge effect, finally, is the relational-novelty variable that out-explains most pharmacology in real-world libido trajectories. First robustly demonstrated in rodents and then across many mammalian species, it is the phenomenon in which sexual interest and behavioural arousal are renewed when a sated male is exposed to a novel female partner; the mechanism involves dopamine in mesolimbic reward circuits, with sexual habituation reducing nucleus accumbens dopamine release to the same partner over repeated exposures and novel-partner exposure restoring it. Translation to humans is methodologically harder because partner novelty is confounded with relationship investment, but Klusmann's 2002 Archives of Sexual Behavior work on relationship-duration declines in female sexual desire, and the broader literature on partnered libido trajectories, are consistent with the rodent finding: sexual interest declines with relationship duration in both sexes, more steeply in many women, controlling for age. The libido-pharmacology punchline is that for many people the dominant variable in their sexual-interest trajectory is relationship duration and relational context, not neurochemistry; expecting a daily 5-HT1A agonist or an on-demand melanocortin agonist to fix a Coolidge-effect-driven plateau is asking the drug to do something it cannot do. Basson's responsive-desire model is the conceptual move that the DSM-5 was reflecting: for many women in long-term partnered contexts, desire is not a spontaneous drive that precedes engagement but a responsive state that emerges after the decision to engage and the experience of arousal, which means treating low desire by trying to pharmacologically generate spontaneous urge is conceptually incoherent for many women whose ordinary baseline never included spontaneous urge in the first place. Trial enrolment in this domain is itself a relational intervention, which is why placebo responses are so large. Sources: Klusmann 2002 Arch Sex Behav, Basson 2000, PMID 10693116, Basson 2001 CMAJ, PMID 11224952. Confidence: C2-C3.

Pulling it together: libido is androgen-anchored at the population level, weakly hormone-titrated at the individual level, and substantially modulated by neurotransmitter systems (dopamine, serotonin, melanocortin, kisspeptin), vascular substrate (nitric oxide and the cGMP/PDE5 axis), circadian and metabolic substrate (sleep, cortisol, insulin), and by relational context that is irreducible to chemistry. Testosterone has the most consistent RCT signal as a clinical intervention in postmenopausal women on background estrogen, with effect sizes that are real but modest. Estrogen does substrate work for vaginal/vulvar tissue and within-cycle desire in cycling women; progesterone is at best neutral and at worst suppressive for libido despite the persistent folk claim otherwise. PDE5 inhibitors are blood-flow drugs with a respectable but partly confounded all-cause mortality association for tadalafil specifically. Bremelanotide is FDA-approved for premenopausal acquired generalized HSDD with a modest drug-versus-placebo delta and a non-trivial side-effect cost; Melanotan II is its unregulated counterpart and is not an acceptable substitute. Flibanserin came first in 2015, was ferociously contested, and is now a commercial failure despite being the original FDA-approved HSDD drug. The refractory period is real but poorly characterised, the prolactin hypothesis has been substantially refuted since 2020, and the only clinical lever is PDE5 inhibitors restoring peripheral erection during what remains a central refractory state. The dampener landscape is broad — SSRIs, opioids, antipsychotic-driven hyperprolactinaemia, alcohol, beta-blockers, glucocorticoids, depression itself — and the trans-HRT data confirm population-level androgen direction without supporting clean within-person hormone titration. Kisspeptin is the most exciting forward-looking pathway. Placebo response and relational context dominate trial interpretation. Anyone reading the popular hormone-and-libido literature should take its single-variable framings as starting points to interrogate, not as conclusions to act on.

The companion chapters below break the subject into the pieces that the single-file overview could not carry without collapsing into shorthand:

The hormonal-foundations chapter, hormones-foundations.md, is the harsh fact-check of the popular "T is the master libido hormone" framing. It walks through Wåhlin-Jacobsen 2017, Roney–Simmons 2013, Cappelletti–Wallen 2016, the Islam 2019 meta-analysis, the Shifren/Braunstein/Davis postmenopausal trial record, the SHBG mechanism that turns oral estrogen into a free-T suppressor, and the regulatory history of Intrinsa and LibiGel. It is the right place to start when someone asks "but isn't testosterone the libido hormone?" and you need to answer that the truthful response is more textured than the question.

The progestins-and-progesterone chapter, progestins-and-progesterone.md, addresses the "do progestins lower libido by lowering testosterone?" question directly with a five-mechanism map. It covers the pharmacological taxonomy of pregnanes, norpregnanes, estranes, gonanes, dienogest, and drospirenone; HPG suppression as the dominant pathway in high-dose forensic and trans-feminine contexts; the Michael & Zumpe 1993 primate experiment that dissociates central from peripheral mechanism; the Pastor 2013 review of 36 COC studies; the Boozalis 2016 and ECHO 2024 ovulation-suppression data; the Zethraeus 2017 RCT showing levonorgestrel COC depressed three sexual-function domains; and the harshly negative empirical case against the trans-community claim that micronized progesterone reliably raises libido.

The PDE5-inhibitors chapter, pde5-inhibitors.md, is the labels-and-pharmacokinetics reference. It establishes the 4-hour sildenafil and 17.5-hour tadalafil half-lives, the 5-hour avanafil and 4–5-hour vardenafil numbers, the food-effect and selectivity differences, the absolute nitrate/riociguat contraindication, the alpha-blocker/CYP3A4 interactions, and the headache/flushing/dyspepsia/visual-disturbance/back-pain side-effect map. It then walks through the all-cause-mortality literature in detail — Andersson 2021, Anderson 2017, Kloner 2024, Jehle 2024, Mostafa 2024, and the Xiao 2025 Mendelian-randomization analysis — and explains why "weak evidence" undersells the consistency while "RCT-grade" oversells the inferential confidence.

The melanocortin chapter, melanocortin-pt141.md, is the bremelanotide/Vyleesi reference and the melanotan-II safety chapter. It traces the University of Arizona Hadley/Hruby α-MSH-analog work into the Wessells 1998/2000 libido trials, the Palatin Technologies C-terminal-modification programme into PT-141, the intranasal-to-subcutaneous pivot, the FDA approval on June 21, 2019 for premenopausal acquired generalized HSDD only, the 1.75 mg subcutaneous dose with 2.7-hour mean half-life, the RECONNECT efficacy data with its 22-percentage-point drug-placebo delta, the 40% nausea rate and 18–23% extension-phase discontinuation, the 1%-at-labelled-dosing versus 38%-at-daily-dosing focal hyperpigmentation, and the case-report harms that make MT-II a non-substitute rather than a peer.

The refractory-period chapter, refractory-period-prolactin.md, corrects the popular Krüger-2003-Lancet attribution to the actual Journal of Endocrinology source (PMID 14656205, single-blind n=10), summarises the Truitt & Coolen 2002 Science work establishing the spinal ejaculation generator's separability from supraspinal orgasm, and walks through the Valente/Marques/Lima 2021 Communications Biology paper that broke the prolactin-causes-refractory-period hypothesis using peripherally-restricted domperidone. It also covers cabergoline's pharmacokinetics (~65-hour half-life) and the dose-stratified valvulopathy story (real at Parkinson's-scale doses, modest mild-tricuspid-regurgitation signal at prolactinoma-scale doses), plus PDE5 inhibitors' role as the only clean clinical lever and post-orgasmic illness syndrome as an adjacent but distinct entity.

The dampeners chapter, libido-dampeners.md, covers the broader iatrogenic landscape that the popular hormone-centric framing skips: SSRI rates from 25% (bupropion) to 70% (paroxetine) and the rescue-by-bupropion literature, PSSD's regulatory recognition in 2019 and unresolved mechanism, finasteride and post-finasteride syndrome's August 2022 FDA label revision, antipsychotic-driven hyperprolactinaemia and the aripiprazole rescue, opioid-induced androgen deficiency with methadone-versus-buprenorphine differential, alcohol's acute-versus-chronic split, cannabis's mixed signal, beta-blocker lipophilicity differences with nebivolol as the standout, glucocorticoid HPG suppression, chemotherapy and pelvic radiation, and major depression as a major libido suppressor independent of any treatment.

The trans-specific chapter, trans-specific-libido.md, reads the ENIGI cohort and the Wierckx/Defreyne longitudinal data as the cleanest natural experiment on which sex steroid drives adult libido. It covers the 22% HSDD prevalence in trans women versus 5% in trans men, the 36-month SDI rebound that exceeds baseline for total/dyadic desire, the explicit no-correlation-with-absolute-T finding in both directions, the antiandrogen-choice dose-response (CPA versus spironolactone versus bicalutamide), the supraphysiological-estradiol over-suppression failure mode, the empirical case against the Prior 2019 progesterone-restores-libido hypothesis, and the cis-women testosterone supplementation regulatory history (Intrinsa rejection, LibiGel placebo-arm failure, AndroFeme as the only licensed product anywhere).

The missed-topics chapter, missed-topics.md, is the catch-all for the territory the original draft skipped entirely. It corrects the flibanserin chronology (Addyi 2015, four years before Vyleesi 2019) and details the "Even The Score" approval controversy and the Jaspers 2016 meta-analysis showing 0.49 extra satisfying events per month. It pulls in the dopamine system the draft underweighted: apomorphine's regulatory history, the pramipexole/ropinirole impulse-control-disorder literature as the natural experiment for dopamine-driven hypersexuality, bupropion as the libido-friendly antidepressant, and Berridge's wanting-versus-liking framework. It covers oxytocin's popularity-versus-evidence gap (Carmichael 1987's robust orgasm-induced peripheral spike versus MacDonald 2013 and Muin 2015's underwhelming intranasal trials), kisspeptin as the most credible forward-looking class with the Comninos Imperial-College programme, yohimbine for psychogenic ED, the modest signals for maca and saffron, the Tribulus debunking, the placebo-response problem in HSDD trials, the MMAS and TTrials/TRAVERSE epidemiology of aging-and-libido, the DSM-5 HSDD-to-FSIAD merger and Basson's responsive-desire model, and the Coolidge effect as the relational-novelty variable that out-explains most pharmacology.