Four independent research agents checked every substantive claim in the 17:53 audio overview against primary sources, plus a separate evidence review of the 15-marker lab panel. Every citation below carries a DOI or PMID retrieved from PubMed or an FDA database, not written from memory. Where a number could not be verified, that is stated rather than smoothed over.

The short version

The podcast is roughly two-thirds sound. Every statistic it lifted directly from the JACC Advances prescribing paper is accurate. What fails is everything in the connective tissue: the numbers it invented, the study designs it upgraded, the mechanism claims it overstated, and, most seriously, one source that says the opposite of what the podcast uses it for.

Ten claims are wrong. Four are outright false, three are misleading in ways that reverse the point being made, and three are unsourced.

Scorecard

---

The four outright false claims

1. "Millions of women are suddenly taking this medication"

Off by roughly an order of magnitude. The study counted 90,482 prescriptions in the Epic Cosmos network in 2025, a population rate of 130.8 per 100,000 women. Applied nationally that is about 174,000 women. Applying the single highest observed subgroup rate (357 per 100,000, White women aged 45 to 64) to every adult woman in the country still lands under 500,000.

The paper's own summary line is "up to 1 in 280," which is a ceiling in one subgroup, not an average. "Suddenly" is also loose: rates were flat through 2021 and the acceleration ran 2022 to 2025.

Accurate version: somewhere in the low hundreds of thousands of US women are now prescribed testosterone, up sharply since 2022, reaching as many as 1 in 280 White women in midlife.

2. "Zero testosterone products carry an FDA indication for use in women"

False, and it overstates the podcast's own source. Two currently marketed testosterone enanthate labels do carry an approved female indication: Eugia US (ANDA212659, label effective 2026-06-12) and Hikma Pharmaceuticals USA (ANDA091120, effective 2026-05-08). Both include a "Females" subsection covering palliative use in advancing inoperable metastatic mammary cancer in women one to five years postmenopausal.

Accurate version: no testosterone product is FDA-approved for menopausal symptoms, libido, energy, or cognition in women. The only female indication on any current US testosterone label is oncologic.

3. "46.8 percent improvement in sexual desire"

A transposition, but false as spoken. In the Glynne 2024 paper, 46.8 percent is the moderate-to-severe improvement figure for loss of interest in most things, the same symptom the podcast already quoted as 56 percent. The equivalent figure for sexual desire was 41.6 percent.

The related "far outpacing" framing also fails. The 56 percent figure did top the ten symptoms, but crying spells came in at 55 percent and loss of interest in sex at 52 percent. No statistical test compared symptoms against each other, and the denominators differ (325 versus 510, since low libido was an entry criterion).

4. "Testosterone suppresses hepcidin, so your body absorbs and hoards excessive iron"

The direction is backwards. Hepcidin suppression by testosterone is well established in men. But the measured downstream result is that ferritin falls, soluble transferrin receptor rises, functional iron stores are reduced, and iron incorporation into red cells increases. Iron is mobilized and consumed, not hoarded. In iron-deficient mice, testosterone actually worsened anemia through ineffective erythropoiesis.

Two related overstatements in the same passage:

• "Testosterone directly stimulates the bone marrow." Human data point to indirect routes (increased erythropoietin, a recalibrated EPO/hemoglobin set point, increased iron utilization). Mouse bone-marrow reconstitution work found the erythroid effect requires androgen receptor DNA binding in non-hematopoietic cells.
• Hepcidin as the causal driver. Hepcidin-knockout mice and liver-specific androgen-receptor-knockout mice still raise hematocrit on testosterone. Hepcidin suppression is a contributor, not the master switch.

---

The three misleading claims that reverse their own point

5. The Australian study says the opposite of what the podcast uses it for

This is the most serious problem in the audio.

The podcast correctly reports the AMY study (n=1,104, LC-MS/MS): testosterone declines to a nadir around age 58 to 59. It then says that age bracket "perfectly correlates with a period where a lot of women report profound debilitating brain fog, severe fatigue, and general apathy."

The paper correlates the nadir with sexual symptoms only. It then states, in its own words, that the nadir should not be interpreted as justification for treating women of that age for other purposes, having just named cognition and mood among those purposes. Its headline conclusion is that these data do not support menopause itself as an indication for testosterone supplementation.

Three further problems with how the study is used:

• The nadir is a model-predicted artifact of adding an age-squared term, and that model explained 2.6 percent of the variance in testosterone.
• Testosterone at ages 60 to 64 and 65 to 69 was not statistically different from ages 40 to 44.
• The paper concludes the change is a consequence of age, not menopause. Testosterone did not vary by menopausal stage at all.

The podcast's mechanism ("the ovaries and adrenal glands fundamentally shift their output") is invented. The paper calls the post-nadir rise unexplained, and notes it happens while the precursors keep falling (DHEA down 33 percent, androstenedione down 51 percent across the age range).

A 2026 AMY substudy went further: no association between any hormone, including testosterone, and sexual desire in midlife women, concluding these steroids should not be part of routine clinical assessment of women with sexual concerns.

6. "Over 90 percent of this use is off-label"

The 8.2 percent HSDD figure is correct. The inference inverts the regulatory reality.

Since no testosterone product is FDA-approved for any female sexual or menopausal indication, 100 percent of this prescribing is off-label, including the 8.2 percent coded as HSDD. HSDD is the indication endorsed by the 2019 Global Consensus, not by the FDA.

The paper also explicitly cautions against the podcast's reading: encounter diagnoses "may not accurately reflect clinical indications, and a low HSDD coding rate should not be interpreted as evidence of inappropriate prescribing."

7. "The risk was highest in the very first year"

The largest rise in hematocrit does occur in year one (0.39 to 0.45 L/L). But cumulative first-time erythrocytosis keeps climbing: 8 to 10 percent at one year, 38 percent at ten years, 50 percent at fourteen. The authors' own conclusion is that monitoring stays warranted for as long as testosterone is used.

A listener hears "the danger passes after year one." The paper says the opposite.

---

The three unsourced claims

8. "68 marketed prescription testosterone products"

No FDA database, publication, or statement matches this number. Queries run 2026-08-10:

65 is closest, and it counts product entries rather than products a patient would recognize. The figure also shifts as FDA data refreshes, which is an argument against stating any such number as fact.

9. The anhedonia versus depression distinction

The podcast's central framing (testosterone "cures anhedonia" but "totally fails to treat clinical depression," acting as a "pro-vitality agent") has no primary human literature behind it.

• A PubMed title search for testosterone plus anhedonia returns exactly one paper: a 2012 rat study whose finding is closer to the reverse. Testosterone prevented anhedonia in middle-aged male rats when given before chronic stress but "failed to reverse it once installed."
• Searches for testosterone with anticipatory or consummatory anhedonia in women return nothing.
• In the Glynne series, "feeling unhappy or depressed" also improved 43 percent. So nothing dissociated. Everything moved together, which is exactly what an unblinded pre-post series produces.
• In the null RCT, MADRS includes an anhedonia item, and there was no separation. That trial also found no separation on fatigue, which directly undercuts "pro-vitality agent."

10. Hepcidin suppression in women

Every human hepcidin study located was in men, at doses from 25 mg to 600 mg of testosterone enanthate weekly. The animal work used masculinizing or supraphysiologic doses. No human study at female-physiologic doses exists.

---

The neurobiology, checked against this project's own research

The podcast's mechanistic core is a two-step claim at 02:53 and 03:19.

"Testosterone increases the expression of tyrosine hydroxylase, widening the bottleneck"

Partially supported, overstated. Testosterone did raise nigral TH protein 70 to 106 percent in gonadectomized male rats. But:

• No androgen response element in the TH promoter has been confirmed in vivo or by ChIP. One putative ARE was mapped in transfected cell lines in 2006 and never replicated. The effect appears to be indirect, routed through trophic signaling (BDNF/TrkB, GDNF). [See the correction note below.]
• In the same experiment, DHT (a pure, non-aromatizable, more potent androgen) did nothing to TH, while reproducing testosterone's effect on COMT and MAO perfectly. If TH were a simple androgen target, DHT should have worked.
• Castration studies flatly contradict each other on whether TH-positive cell number goes up or down.
• Humans have four TH isoforms; rats have one, and they differ precisely in the N-terminal regulatory domain that governs phosphorylation and feedback inhibition.

The "bottleneck" analogy also skips the enzyme's actual control logic. TH is held in a low-activity state by end-product feedback inhibition from dopamine itself and unlocked by activity-dependent phosphorylation at Ser40. More TH protein does not straightforwardly mean more dopamine.

Correction, 2026-08-11. This section originally stated flatly that there is no androgen response element in the TH promoter. That was too absolute and is now fixed above. Jeong et al. 2006 (Neuroscience Letters 396(1):57-61, 10.1016/j.neulet.2005.11.011) reports that the androgen receptor transactivates the TH promoter in a ligand-dependent way, and maps a putative ARE at -1562 to -1328 base pairs. So an ARE has been reported, and saying otherwise was wrong.

What did not change: that evidence is a transient co-transfection reporter assay in two neuroblastoma-derived cell lines, the authors themselves call the element putative, and no ChIP or in vivo confirmation has followed in the twenty years since. We hold the abstract only, not the full text, so this page makes no claim about how strong the deletion mapping is. Every other objection in this section stands untouched, including the DHT result, which remains the hardest thing for a simple-androgen-target model to explain.

"It simultaneously boosts the sensitivity of the dopamine receptor"

Not supported. No study has measured what receptor sensitivity means pharmacologically under androgen manipulation: no shift in receptor affinity (Kd), no change in the D2 high-affinity state fraction, no agonist-stimulated GTP-gamma-S binding.

What has been measured is receptor counts, and they move in several directions at once:

The anatomy explains why: the androgen receptor is abundant on dopamine-producing neurons and in cortex, and only patchy and weak on the striatal neurons that receive dopamine.

The receptor-sensitivity evidence belongs to estradiol. A single physiologic dose converts about 40 percent of striatal D2 receptors from high to low agonist affinity within 15 to 30 minutes, with no change in total density. There is no equivalent measurement for testosterone.

What the podcast got right, and it is the best part of the audio

The effort discounting and uncertainty section is accurate and well told. Chronic testosterone in rats made them more willing to work harder for a larger reward, less sensitive to physical effort and to punishment, but more sensitive to uncertainty, shifting back to the smaller guaranteed reward faster than controls when the big reward became unpredictable. That is a real and correctly reported finding.

Worth adding what the podcast does not say: that uncertainty aversion is dopamine-mediated in the hypofunction direction. A D2 agonist restored large-uncertain preference in testosterone-treated rats; a D1 agonist eliminated the group difference. The interpretation is that chronic androgen exposure reduces D1 and D2 receptor function, which is the opposite of the podcast's "amplifier" framing in the same episode.

One caveat on the Iowa Gambling Task claim: that study is 12 women, single dose, and the actual finding was a shift toward disadvantageous decks, not a neutral "recalibration."

---

What holds up

Everything lifted straight from the JACC Advances paper is accurate:

Also correct: the AMY sample size of 1,104 and the use of mass spectrometry; the Glynne sample of 510 women already on HRT with persistent symptoms; the 56 percent figure for loss of interest; the trans men cohort at n=1,073 over 20 years; the 11 percent exceeding the hematocrit threshold; smoking, high BMI and route as risk factors; and the null depression trial.

Two things the podcast understated. Midlife women aged 45 to 64 rose 79.1 percent year over year, not 58.7, so the podcast undersold the number it was trying to make dramatic. And the null depression trial is bigger than implied: Dichtel 2020, n=101, randomized, double-blind, placebo-controlled, MADRS p=0.91, with no moderation by baseline free testosterone or menopausal status.

---

Three corrections the podcast needs even where its numbers are right

The 51.2 percent needs its comparator. The paper immediately notes those rates were "similar to the corresponding non-testosterone-prescribed population," and that established ischemic heart disease was actually lower in the testosterone group (3.1 versus 4.7 percent). Quoted alone, 51.2 percent implies an unusually high-risk group. It is not.

"Zero cardiovascular outcome trials" should be "zero published ones." BLISS existed, enrolled postmenopausal women with HSDD and elevated cardiovascular risk, used a composite cardiovascular endpoint, accrued more than 7,300 woman-years, and completed. It was never published because the sponsor's companion efficacy trials failed and the company went into liquidation.

The formulation mechanism is wrong. The podcast blames "massive spikes" from long-acting intramuscular injections. The risk comparison is right (OR 2.9 crude, 3.1 adjusted versus gel). But short-acting esters are the ones that spike, and short-acting esters showed no elevated risk (OR 1.1, 95% CI 0.7 to 1.6). Long-acting undecanoate is the formulation designed for stable levels. At the higher 0.52 threshold, the long-acting signal disappeared entirely (OR 1.0, 0.3 to 3.4).

---

Study design: what the podcast upgraded

This is where a listener is most likely to be misled, because nothing in the audio is factually false. The designs are simply presented as stronger than they are.

A percentage of women reporting improvement on a self-report item after four months of open-label treatment is not an effect size. There is no comparator, so "56 percent improved" has no counterfactual attached to it.

---

Monitoring: what the guideline actually says

The podcast says "regular blood monitoring is absolutely essential, you have to watch those hematocrit levels."

The 2019 Global Consensus Position Statement never mentions hematocrit. Not hemoglobin, not complete blood count, not erythrocytosis. The full recommended schedule is:

That is the entire list. Everything else on a monitoring panel is extrapolation or clinical judgment, and should be labeled as such rather than presented as guideline-directed.

Hematocrit monitoring lives in male guidance: the Endocrine Society recommends it at baseline, 3 to 6 months, then annually in men, with action above 54 percent. Quoting 54 percent at a woman is quoting a male number.

The defensible version of the podcast's advice, which needs one sentence of attribution rather than deletion: hematocrit monitoring is genuinely warranted on masculinizing doses, and is a reasonable precaution for any route risking supraphysiologic levels, which is exactly the pellet and injection scenario the consensus already advises against.

---

Part 2: The lab panel

Fifteen markers, ranked by whether they would actually change management for a woman in her 50s with high ApoB, APOE4 homozygosity, on tirzepatide and menopausal hormone therapy, considering or taking testosterone.

This is information, not medical advice. Several items turn on facts only a clinician has.

Ranking

The testosterone assay problem

The consensus is unusually blunt, and this is verbatim:

"Direct assays for the measurement of total and free testosterone are highly unreliable in the female range (Grade A)."

Grade A is the strongest evidence grade in that document. Three failures compound at female concentrations, which run ten to thirty times lower than the male range these platforms were built for: antibody cross-reactivity with steroids circulating at far higher molar concentrations, a sensitivity floor above the lower half of the female normal range, and calibration bias between platforms.

This is not a fringe position. JCEM made mass spectrometry a publication requirement for sex steroid assays in low-concentration populations. A journal will not print an immunoassay testosterone value in women.

What to insist on when ordering:

• Total testosterone by LC-MS/MS, named explicitly on the requisition. Ordering plain "total testosterone" usually returns the immunoassay, because it is the default and cheaper.
• Same lab, same method, every time. Inter-laboratory bias at female concentrations is large enough that changing labs can look like changing dose.
• Morning draw, consistent interval from the last dose, opposite arm from any application site. Transdermal transfer contamination is a real cause of implausible results.
• Biotin interferes with streptavidin-biotin immunoassays. Another reason to prefer mass spec.

On free testosterone. The consensus says research should focus on total testosterone "because evidence that 'free' testosterone is the biologically active testosterone fraction is lacking." The free-hormone hypothesis is an assumption, never demonstrated for testosterone in women. So: direct free-T immunoassay, ignore. Calculated free T (Vermeulen), secondary interpretive aid only, and only if the total was LC-MS/MS. Free Androgen Index, do not use in women at all. Equilibrium dialysis with LC-MS/MS is the only accurate method and is expensive enough that whoever orders it should say what decision it changes.

Where SHBG earns its place: on oral estrogen, SHBG rises through first-pass hepatic effect, so the same total testosterone means something different than it would on transdermal. That is an interpretive use, not an endorsement of chasing a free-T target.

Iron, ferritin, hepcidin and inflammation

Four numbers plus hs-CRP and the CBC form one system. Read individually they mislead.

Three forces act on hepcidin here in different directions. Inflammation raises it (via IL-6), so iron and transferrin saturation fall while ferritin is pushed up by the acute-phase response. Menopause lowers iron losses, so ferritin drifts upward over the following decade. Testosterone suppresses it, opening absorption and mobilizing stores while expanding erythropoiesis that consumes iron.

Three rules fall out:

1. Transferrin saturation distinguishes overload from inflammation. Ferritin alone cannot. A ferritin of 400 with a TSAT of 25 percent in metabolic syndrome is dysmetabolic hyperferritinemia, not hemochromatosis. Treating a raised ferritin as an overload diagnosis is the most common misread on this panel.
2. hs-CRP tells you whether to believe the ferritin. That is why both belong on the same draw.
3. TSAT at or above 45 percent with elevated ferritin, on two fasting morning draws, is the guideline trigger for HFE genotyping.

The testosterone overlay, which is where it gets subtle. Because testosterone suppresses hepcidin, it can produce a high TSAT alongside a falling ferritin. That is not hemochromatosis. It is increased iron flux feeding expanded erythropoiesis, and reading TSAT alone would send someone into a hemochromatosis workup for a drug effect. The distinguishing features are ferritin falling rather than rising, soluble transferrin receptor rising, and hemoglobin drifting up rather than down.

And a trap most explanations miss. On oral estrogen, the hs-CRP you were going to use to arbitrate the ferritin is itself elevated by first-pass hepatic synthesis, and the oral estrogen raises transferrin and TIBC, which mechanically depresses calculated TSAT. Both tie-breakers are distorted, in directions that make inflammation look more likely and overload less likely than they are. The workarounds are soluble transferrin receptor (not an acute-phase reactant), the sTfR/log-ferritin index, and reticulocyte hemoglobin content. None is on Quest 5616, and they are the right add-on if the panel comes back ambiguous rather than repeating the same four numbers.

Preanalytics: Quest 5616 specifies morning collection and 8 to 12 hours fasting. Serum iron swings up to 30 percent diurnally and spikes after any iron-containing supplement, so hold those 24 to 48 hours. Because TSAT is calculated from iron, a badly timed draw corrupts the most important number on the panel.

Hemoglobin and hematocrit, reframed

Keep them, because a CBC costs almost nothing. But the usual reason given is imported from male guidance, and the useful direction is the opposite of the one everyone watches.

Testosterone suppresses hepcidin and expands erythropoiesis, which consumes iron. In a woman on a physiologic transdermal dose, the realistic finding is falling ferritin and iron-restricted erythropoiesis, not polycythemia. Watch MCV and ferritin trends rather than a hematocrit ceiling.

Two practical notes. There is no established female threshold for testosterone therapy; the nearest defensible triggers to investigate are the WHO female criteria (Hb above 16.0 g/dL, Hct above 48 percent). A rise above her own baseline is more informative than any population cutoff, which is what makes the baseline draw the valuable one. And Boulder is not at sea level: Front Range reference intervals run roughly 0.5 to 1.0 g/dL higher for hemoglobin, so a lab reporting sea-level ranges will over-flag. Before attributing any rise to testosterone, sleep apnea, smoking, dehydration and diuretics are all more common explanations.

Three markers to drop or demote

TMAO: drop. The observational signal is real. The causal claim is not. TMAO is renally cleared, so adjustment for kidney function attenuates or reverses the cardiovascular association. Bidirectional Mendelian randomization found that diabetes and kidney disease raise TMAO, pointing to confounding or reverse causation. And decisively: no trial shows that measuring or lowering TMAO improves any outcome. The only actions it prompts are dietary changes a person makes or does not make regardless of the result. A fish meal in the preceding 24 hours spikes it, because fish contains preformed TMAO.

Free testosterone: drop, per the assay section above.

NT-proBNP: low yield, and it has a specific interpretation problem here. Testosterone lowers NT-proBNP, dose-dependently, roughly 4.3 percent lower per 1 g increase in dose. Estrogen raises it. Obesity suppresses it, which means weight loss on tirzepatide will raise it for reasons unrelated to cardiac status. Three opposing vectors act on one number, and a serial NT-proBNP in this situation is close to uninterpretable without accounting for all of them. The guideline support (Class 2a) applies to patients at risk for heart failure and is paired with team-based follow-up including echocardiography. It is a care-pathway recommendation, not a standalone number on a wellness panel.

Two things worth knowing about markers you are keeping

ApoB was not measured in the testosterone meta-analysis. The full text contains zero occurrences of "apolipoprotein" or "apoB." What it reported was total cholesterol, LDL-C, HDL-C and triglycerides: oral testosterone worsened lipids, non-oral was neutral. So the neutrality claim for transdermal rests on cholesterol fractions, not particle count.

The GLP-1 trap matters more here. As weight falls on tirzepatide, triglycerides drop and LDL particles become larger and more cholesterol-rich, which can make LDL-C look worse while ApoB improves. Watching LDL-C through that transition gives the wrong signal. ApoB stays honest.

ALT: most US female upper limits are too high. The recalculated healthy upper limit for women is 19 U/L, and the ACG operating threshold is 25 U/L. An ALT of 32 reported as normal against a lab range of 7 to 40 is, on the evidence, abnormal for a woman. Two caveats specific to this situation: tirzepatide reduced ALT by about 43 percent in SYNERGY-NASH, so normalization on a GLP-1 is expected and does not settle the fibrosis question. And ALT is a PLP-dependent enzyme, so a B6-depleted person can show a falsely low ALT, which is a direct interaction with another marker on this panel.

What is missing

Lp(a), and it is the most obvious gap. Measure once, ever. Roughly 80 to 90 percent genetically determined, stable for life, causal for atherosclerotic disease and aortic stenosis, elevated in 20 to 25 percent of people. Both the EAS 2022 consensus and the NLA 2024 focused update now recommend measuring it at least once in every adult.

Why it matters here specifically: with already-high ApoB, an elevated Lp(a) changes how aggressively that ApoB is lowered, it explains part of the ApoB value (each Lp(a) particle carries one apoB-100, and Lp(a) does not respond to statins the way LDL does), and it is heritable information relevant to children. Report in nmol/L, not mg/dL, because isoform size distorts mass-based assays. Note that oral estrogen lowers Lp(a) by 15 to 25 percent while transdermal does so much less, so record the route alongside the result.

HbA1c, with a hidden interaction nobody mentions: HbA1c is biased by red cell lifespan. Testosterone-driven erythropoiesis produces a younger red cell population, reading HbA1c falsely low. Iron deficiency reads it falsely high. With testosterone, postmenopausal iron accumulation and hemoglobin changes all in play at once, the number may not mean what it appears to. Fructosamine, glycated albumin, or CGM sidestep the problem.

uACR. Cheap, detects kidney disease before eGFR moves, independent cardiovascular risk marker, and needed to interpret NT-proBNP. A far better use of a test slot than TMAO.

eGFR with cystatin C. Creatinine-based eGFR is distorted in both directions here: testosterone increases creatinine generation (falsely lowering eGFR) while GLP-1 lean mass loss does the reverse. Cystatin C is muscle-mass independent and de-confounds exactly this.

Coronary artery calcium score, which is not a lab but is the highest-yield addition on the page. For a woman in her 50s with high ApoB and APOE4 homozygosity, a CAC score resolves a statin decision better than TMAO, PLP, NT-proBNP and the Omega-3 Index combined. It measures the disease rather than a risk factor for it, and zero versus 150 leads to genuinely different conversations.

Not recommended as reflex additions: fasting insulin and HOMA-IR (insulin immunoassays are not standardized, differing more than twofold between platforms, with no validated cutpoint; TG/HDL-C is already free in a lipid panel). Vitamin D screening in asymptomatic adults is a USPSTF I statement, and if bone is the real question, a DXA changes management far more.

On APOE4 specifically

Homozygous APOE4 status does not change any lipid target in any current guideline, and no marker on this panel needs a different threshold because of it. What it reasonably changes is emphasis: earlier and more sustained ApoB lowering, blood pressure control, and knowing the Lp(a).

It does not justify TMAO. And the APOE4-specific omega-3 dosing argument rests entirely on mouse data; there is no human outcome trial establishing that APOE4 carriers should be dosed differently or targeted to a different index. Anyone recommending an APOE4-specific lab threshold is going beyond the evidence.

On the Omega-3 Index

The 8 percent target comes from a 2004 paper whose title ends in a question mark, and that question mark is doing real work. It is an observational risk-marker cutoff, not a trial-validated treatment target.

As a marker it holds up: pooled analysis of 17 prospective cohorts found blood omega-3 levels inversely associated with mortality. As a target to treat toward, the outcome trials are the problem.

In the REDUCE-IT placebo arm, LDL-C rose about 10 percent and hs-CRP rose about 32 percent, alongside increases in IL-1beta, IL-6 and oxidized LDL. The cleanest read: the trial with an inert comparator was null; the trial with a possibly non-inert comparator was positive. That asymmetry is not a small detail.

---

What a careful writer should not repeat from this podcast

• "Millions of women are taking testosterone." It is low hundreds of thousands.
• "No testosterone product has any FDA female indication." Two do, for metastatic breast cancer.
• "68 marketed products." Unsourced.
• "Testosterone boosts dopamine receptor sensitivity." Never measured.
• "Testosterone makes your body hoard iron." The opposite happens.
• "The nadir at 58 to 59 explains brain fog." The source paper explicitly refuses that inference.
• "Testosterone cures anhedonia but not depression." No human study tests this.
• "Hematocrit monitoring is essential" without saying it comes from male guidance.
• "Risk is highest in the first year." Cumulative risk reaches 50 percent by year fourteen.
• Any figure from a 12-person or uncontrolled study stated without its n and design.

What can be said confidently

• Testosterone prescribing to US women rose 2.6-fold from 2016 to 2025, and 79.1 percent year over year in midlife women in 2025.
• All of it is off-label. There is no FDA-approved female product for any menopausal or sexual indication.
• The only evidence-based indication is postmenopausal HSDD after formal biopsychosocial assessment, at Level I, Grade A.
• Androgens decline with age from the twenties. Natural menopause is not the inflection point; surgical menopause is.
• Testosterone raises hematocrit dose-dependently, and the effect is real but characterized at masculinizing doses. Route matters, and pellets and injections are recommended against.
• Long-term cardiovascular and breast safety in women is unknown. Trials were mostly under two years and excluded women at high cardiometabolic risk.
• In rodents, testosterone increases willingness to work for reward and decreases willingness to accept unpredictable reward, and that uncertainty aversion runs through reduced D1 and D2 receptor function.

---

Sources

Every DOI and PMID below was retrieved from PubMed or an FDA database during this audit.

Prescribing and regulation

1. Avivi I, Stuenkel CA, Sampath-Kumar R, Ben-Yehuda O (2026). Accelerating Testosterone Prescribing for U.S. Women: Implications for Cardiovascular Safety. JACC Adv 5(6 Pt 1):102724. 10.1016/j.jacadv.2026.102724. PMID 42095794.
2. US FDA, Drugs@FDA, openFDA NDC directory, and openFDA SPL label endpoints. Queried 2026-08-10.
3. National Library of Medicine, DailyMed SPL web services. Queried 2026-08-10.
4. White WB, Grady D, Giudice LC, et al. (2012). A cardiovascular safety study of LibiGel (testosterone gel) in postmenopausal women with elevated cardiovascular risk and hypoactive sexual desire disorder. Am Heart J 163(1):27-32. 10.1016/j.ahj.2011.09.021. PMID 22172433.

Clinical studies

5. Wang Y, Islam RM, Bond M, Davis SR (2025). Testosterone and pre-androgens by age and menopausal stage at midlife. EBioMedicine 121:105972. 10.1016/j.ebiom.2025.105972. PMID 41106025.
6. Wang Y, Islam RM, Hodge A, et al. (2026). Associations between testosterone and pre-androgens and sexual function; findings from the Australian Women's Midlife Years Study. Fertil Steril. 10.1016/j.fertnstert.2026.05.156. PMID 42167520.
7. Glynne S, Kamal A, Kamel AM, Reisel D, Newson L (2024). Effect of transdermal testosterone therapy on mood and cognitive symptoms in peri- and postmenopausal women: a pilot study. Arch Womens Ment Health 28(3):541-550. 10.1007/s00737-024-01513-6. PMID 39283522.
8. Dichtel LE, Carpenter LL, Nyer M, et al. (2020). Low-Dose Testosterone Augmentation for Antidepressant-Resistant Major Depressive Disorder in Women. Am J Psychiatry 177(10):965-973. 10.1176/appi.ajp.2020.19080844. PMID 32660299.
9. Herrera-Perez JJ, Martinez-Mota L, Chavira R, Fernandez-Guasti A (2012). Testosterone prevents but not reverses anhedonia in middle-aged males. Horm Behav 61(4):623-30. 10.1016/j.yhbeh.2012.02.015. PMID 22373497.
10. Islam RM, Bell RJ, Green S, Page MJ, Davis SR (2019). Safety and efficacy of testosterone for women: a systematic review and meta-analysis of randomised controlled trial data. Lancet Diabetes Endocrinol 7(10):754-766. 10.1016/S2213-8587(19)30189-5. PMID 31353194.

Safety and hematology

11. Lincoff AM, Bhasin S, Flevaris P, et al. (2023). Cardiovascular Safety of Testosterone-Replacement Therapy. N Engl J Med 389(2):107-117. 10.1056/NEJMoa2215025. PMID 37326322.
12. Madsen MC, van Dijk D, Wiepjes CM, Conemans EB, Thijs A, den Heijer M (2021). Erythrocytosis in a Large Cohort of Trans Men Using Testosterone. J Clin Endocrinol Metab 106(6):1710-1717. 10.1210/clinem/dgab089. PMID 33599731.
13. Bachman E, Feng R, Travison T, et al. (2010). Testosterone suppresses hepcidin in men. J Clin Endocrinol Metab 95(10):4743-4747. 10.1210/jc.2010-0864. PMID 20660052.
14. Bachman E, Travison TG, Basaria S, et al. (2014). Testosterone induces erythrocytosis via increased erythropoietin and suppressed hepcidin. J Gerontol A Biol Sci Med Sci 69(6):725-735. 10.1093/gerona/glt154. PMID 24158761.
15. Guo W, Schmidt PJ, Fleming MD, Bhasin S (2019). Hepcidin is not essential for mediating testosterone's effects on erythropoiesis. Andrology 8(1):82-90. 10.1111/andr.12622. PMID 31001931.
16. Hennigar SR, Berryman CE, Harris MN, et al. (2020). Testosterone Administration During Energy Deficit Suppresses Hepcidin and Increases Iron Availability for Erythropoiesis. J Clin Endocrinol Metab 105(4). 10.1210/clinem/dgz316. PMID 31894236.
17. McManus JF, Nguyen NN, Davey RA, et al. (2020). Androgens stimulate erythropoiesis through the DNA-binding activity of the androgen receptor in non-hematopoietic cells. Eur J Haematol 105(3):247-254. 10.1111/ejh.13431. PMID 32311143.
18. Davis SR, Baber R, Panay N, et al. (2019). Global Consensus Position Statement on the Use of Testosterone Therapy for Women. J Clin Endocrinol Metab 104(10):4660-4666. 10.1210/jc.2019-01603. PMID 31498871.
19. Bhasin S, Brito JP, Cunningham GR, et al. (2018). Testosterone Therapy in Men With Hypogonadism: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab 103(5):1715-1744. 10.1210/jc.2018-00229. PMID 29562364.

Biomarkers

20. Rosner W, Auchus RJ, Azziz R, Sluss PM, Raff H (2007). Position statement: Utility, limitations, and pitfalls in measuring testosterone. J Clin Endocrinol Metab 92(2):405-413. 10.1210/jc.2006-1864. PMID 17090633.
21. Handelsman DJ, Wartofsky L (2013). Requirement for mass spectrometry sex steroid assays in JCEM. J Clin Endocrinol Metab 98(10):3971-3973. 10.1210/jc.2013-3375. PMID 24098015.
22. Vermeulen A, Verdonck L, Kaufman JM (1999). A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 84(10):3666-3672. 10.1210/jcem.84.10.6079. PMID 10523012.
23. Sniderman AD, Dufresne L, Pencina KM, et al. (2024). Individual Variation in the Distribution of Apolipoprotein B Levels Across the Spectrum of LDL-C or Non-HDL-C Levels. JAMA Cardiol. 10.1001/jamacardio.2024.1310. PMID 38865115.
24. Kronenberg F, Mora S, Stroes ESG, et al. (2022). Lipoprotein(a) in atherosclerotic cardiovascular disease and aortic stenosis: a European Atherosclerosis Society consensus statement. Eur Heart J 43(39):3925-3946. 10.1093/eurheartj/ehac361. PMID 36036785.
25. Koschinsky ML, Bajaj A, Boffa MB, et al. (2024). A focused update to the 2019 NLA scientific statement on use of lipoprotein(a) in clinical practice. J Clin Lipidol 18(3):e308-e319. 10.1016/j.jacl.2024.03.001. PMID 38565461.
26. Ridker PM, Bhatt DL, Pradhan AD, et al. (2023). Inflammation and cholesterol as predictors of cardiovascular events among patients receiving statin therapy. Lancet 401(10384):1293-1301. 10.1016/S0140-6736(23)00215-5. PMID 36893777.
27. Jia J, Dou P, Gao M, et al. (2019). Assessment of Causal Direction Between Gut Microbiota-Dependent Metabolites and Cardiometabolic Health: A Bidirectional Mendelian Randomization Analysis. Diabetes 68(9):1747-1755. 10.2337/db19-0153. PMID 31167879.
28. Tang WH, Wang Z, Levison BS, et al. (2013). Intestinal microbial metabolism of phosphatidylcholine and cardiovascular risk. N Engl J Med 368(17):1575-1584. 10.1056/NEJMoa1109400. PMID 23614584.
29. Harris WS, von Schacky C (2004). The Omega-3 Index: a new risk factor for death from coronary heart disease? Prev Med 39(1):212-220. 10.1016/j.ypmed.2004.02.030. PMID 15208005.
30. Nicholls SJ, Lincoff AM, Garcia M, et al. (2020). Effect of High-Dose Omega-3 Fatty Acids vs Corn Oil on Major Adverse Cardiovascular Events: The STRENGTH Randomized Clinical Trial. JAMA 324(22):2268-2280. 10.1001/jama.2020.22258. PMID 33190147.
31. Bhatt DL, Steg PG, Miller M, et al. (2019). Cardiovascular Risk Reduction with Icosapent Ethyl for Hypertriglyceridemia. N Engl J Med 380(1):11-22. 10.1056/NEJMoa1812792. PMID 30415628.
32. Manson JE, Cook NR, Lee IM, et al. (2019). Marine n-3 Fatty Acids and Prevention of Cardiovascular Disease and Cancer. N Engl J Med 380(1):23-32. 10.1056/NEJMoa1811403. PMID 30415637.
33. Prati D, Taioli E, Zanella A, et al. (2002). Updated definitions of healthy ranges for serum alanine aminotransferase levels. Ann Intern Med 137(1):1-10. 10.7326/0003-4819-137-1-200207020-00006. PMID 12093239.
34. Kwo PY, Cohen SM, Lim JK (2017). ACG Clinical Guideline: Evaluation of Abnormal Liver Chemistries. Am J Gastroenterol 112(1):18-35. 10.1038/ajg.2016.517. PMID 27995906.
35. Loomba R, Hartman ML, Lawitz EJ, et al. (2024). Tirzepatide for Metabolic Dysfunction-Associated Steatohepatitis with Liver Fibrosis. N Engl J Med 391(4):299-310. 10.1056/NEJMoa2401943. PMID 38856224.
36. Kowdley KV, Brown KE, Ahn J, Sundaram V (2019). ACG Clinical Guideline: Hereditary Hemochromatosis. Am J Gastroenterol 114(8):1202-1218. 10.14309/ajg.0000000000000315. PMID 31335359.
37. Lai S, et al. (2019). Effect of Testosterone on Natriuretic Peptide Levels. J Am Coll Cardiol 73(11):1288-1296. 10.1016/j.jacc.2018.12.062. PMID 30898204.
38. Suthahar N, Meijers WC, Ho JE, et al. (2018). Sex-specific associations of obesity and N-terminal pro-B-type natriuretic peptide levels in the general population. Eur J Heart Fail 20(8):1205-1214. 10.1002/ejhf.1209. PMID 29855124.
39. Heidenreich PA, Bozkurt B, Aguilar D, et al. (2022). 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure. J Am Coll Cardiol 79(17):e263-e421. 10.1016/j.jacc.2021.12.012. PMID 35379503.
40. KDIGO CKD Work Group (2024). Executive summary of the KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Int 105(4):684-701. 10.1016/j.kint.2023.10.016. PMID 38519239.
41. Surks MI, Hollowell JG (2007). Age-specific distribution of serum thyrotropin and antithyroid antibodies in the US population. J Clin Endocrinol Metab 92(12):4575-4582. 10.1210/jc.2007-1499. PMID 17911171.
42. US Preventive Services Task Force (2021). Screening for Vitamin D Deficiency in Adults. JAMA 325(14):1436-1442. 10.1001/jama.2021.3069. PMID 33847711.
43. Quest Diagnostics. Iron, TIBC and Ferritin Panel, test code 5616. Components: ferritin, iron binding capacity, iron total, percent saturation. Morning collection preferred, 8 to 12 hours fasting.

Neurobiology (from this project's companion research, published separately)

44. Daubner SC, Le T, Wang S (2011). Tyrosine hydroxylase and regulation of dopamine synthesis. Arch Biochem Biophys 508(1):1-12. 10.1016/j.abb.2010.12.017. PMID 21176768.
45. Purves-Tyson TD, Handelsman DJ, Double KL, et al. (2012). Testosterone regulation of sex steroid-related mRNAs and dopamine-related mRNAs in adolescent male rat substantia nigra. BMC Neurosci 13:95. 10.1186/1471-2202-13-95. PMID 22867132.
46. Haycock JW (2002). Species differences in the expression of multiple tyrosine hydroxylase protein isoforms. J Neurochem 81(5):947-953. 10.1046/j.1471-4159.2002.00881.x. PMID 12065606.
47. Levesque D, Di Paolo T (1988). Rapid conversion of high into low striatal D2-dopamine receptor agonist binding states after an acute physiological dose of 17 beta-estradiol. Neurosci Lett 88(1):113-118. 10.1016/0304-3940(88)90324-2. PMID 2969467.
48. Wallin KG, Alves JM, Wood RI (2015). Anabolic-androgenic steroids and decision making: probability and effort discounting in male rats. Psychoneuroendocrinology 57:84-92. 10.1016/j.psyneuen.2015.03.023. PMID 25900595.
49. Wallin-Miller KG, Kreutz F, Li G, Wood RI (2018). Anabolic-androgenic steroids increase sensitivity to uncertainty by inhibition of dopamine D1 and D2 receptors. Psychopharmacology 235(4):959-969. 10.1007/s00213-017-4810-7. PMID 29242988.
50. van Honk J, Schutter DJ, Hermans EJ, et al. (2004). Testosterone shifts the balance between sensitivity for punishment and reward in healthy young women. Psychoneuroendocrinology 29(7):937-943. 10.1016/j.psyneuen.2003.08.007. PMID 15177710.