Door 1

Do I need testosterone?

For a lot of women reading this, the answer is no, or at least not yet, and I'd rather say that on the first page than bury it.

I'm on testosterone. I think the subject is badly covered and that women get dismissed when they ask about it. Both of those can be true while the honest evidence summary is still narrower than you'd hope. So here's what the research actually supports, what it doesn't, and where I think the interesting question sits.

What the evidence supports

In 2019, eleven professional bodies published a joint position statement on testosterone therapy for women. The Endocrine Society, ACOG, the North American Menopause Society, the Royal College of Obstetricians and Gynaecologists, and seven others.

Their conclusion was narrow. The only evidence-based indication for testosterone in women is hypoactive sexual desire disorder, meaning distressing low sexual desire. Not low desire on its own. Low desire that bothers you.

For everything else, they found the data insufficient to recommend it.

(Davis et al., Journal of Clinical Endocrinology and Metabolism, 2019.)

What "insufficient data" does and doesn't mean

This is where a lot of writing on this subject goes wrong in both directions.

"Insufficient data" does not mean testosterone was tested for fatigue, mood, cognition and bone and found useless. It mostly means the trials that would answer those questions haven't been run properly in women, at the right doses, measuring the right things.

It also doesn't mean the effect is there and being suppressed. Sometimes when the trial finally gets run, the answer is no.

Here's a recent example of exactly that.

A trial that found nothing

STEP-HI, published in JAMA Network Open in 2025, enrolled 129 women aged 65 and over who had recently had surgical repair of a hip fracture and still had measurable mobility problems. Eight US sites, phase 3, double-blind, placebo-controlled. For 24 weeks, participants did a supervised high-intensity exercise program including progressive resistance training, and were randomized to topical testosterone gel or placebo gel on top of it.

The result: adding testosterone did not significantly improve six-minute walking distance compared with exercise plus placebo.

That's a well-designed trial in exactly the population where you'd expect a benefit, and it came back negative for that outcome. I'm including it because a site that only reports the encouraging studies isn't worth reading.

(Binder et al., STEP-HI Randomized Clinical Trial, JAMA Network Open, 2025, n=129. DOI 10.1001/jamanetworkopen.2025.10512.)

The muscle question, with the caveat that matters

The strongest causal evidence that testosterone builds muscle in women comes from a dose-response trial published in Menopause in 2014. Hysterectomized women, standardized on estradiol, were randomized to weekly intramuscular testosterone at four escalating doses (3, 6.25, 12.5 and 25 mg) or placebo, for 24 weeks, with DXA body composition and muscle power measured.

Testosterone increased lean body mass and muscle power in a dose-dependent way. Multiple doses tested, with a gradient across them, which is more persuasive than a single-dose trial.

The caveat is the whole point though. The clear gains showed up at the higher doses, producing testosterone levels above the female physiological range. Guideline-concordant therapy deliberately targets concentrations within the normal premenopausal range. The trial also used intramuscular injection, not the transdermal route typical in current practice.

So it demonstrates that testosterone can build muscle in women. It does not demonstrate that the doses a guideline-following clinician would actually prescribe will do it. Those are different claims, and they get conflated constantly.

(Huang et al., Menopause, 2014.)

The flatness, which is what I actually care about

Testosterone gets filed under libido. That's the approved indication, that's what the trials measured, and that's what a clinician hears when you raise it. So if your complaint isn't desire, you don't get offered it, and the research that would settle whether you should hasn't been done.

The symptom I care about is the flatness. Not sadness, and not tiredness. The thing where you do what you used to love and notice, somewhere in the middle, that nothing is reaching back. Clinically it's anhedonia, a reduced capacity to feel pleasure and, more to the point, reduced pull toward things.

I had it for thirty years while being treated for depression the entire time. That's not a coincidence or a failure of my medication. SSRIs work on serotonin. Anhedonia of this kind runs on dopamine, and testosterone acts on those pathways. They're different systems, and treating one does nothing for the other. On testosterone, mine lifted about ninety percent. Not all of it. About ninety percent.

So when I say this is the most important unanswered question in the subject, I'm not being coy about where I stand.

The one trial people cite against this, and why it doesn't settle it

If you raise anhedonia with a well-read clinician, you may get pointed at Dichtel 2020 in the American Journal of Psychiatry. It's the study that looks like it closes the question. It doesn't, and the authors say so themselves.

101 women, eight weeks, double-blind, two sites, low-dose transdermal testosterone added on top of antidepressants. The result was flatly null: depression scores fell from 26.8 to 15.3 on testosterone and from 26.3 to 14.4 on placebo. A p-value of 0.91. No difference in fatigue, sexual function, or the brain imaging measure either.

I'll give it real credit first, because this is where most critiques cheat. The measurement was impeccable. Mass spectrometry at Mayo, free testosterone by equilibrium dialysis. The assay complaint I make everywhere else on this site does not apply here. This was a well-run trial.

Here's why it still doesn't answer the question:

• The placebo response was 49%, and the authors flag it themselves. In their own words, the placebo response rate "was high (49%), which might have accounted in part for the lack of observed treatment effect." When the people who ran the trial tell you it may not have been able to detect an effect, believe them. A trial where half the placebo arm improves is a hard place to find a drug signal.
• It measured the wrong thing. The outcome was MADRS, a general depression scale. There was no anhedonia instrument. If your claim is about motivational flatness specifically, measuring overall depression tests a different question.
• It was the wrong women. Participants had active antidepressant-resistant major depression, with baseline scores around 26. That is not the same population as women whose depression is controlled and who are left with residual flatness. Not the same people, not the same question.
• It was powered to detect a 5-point difference. Anything smaller was invisible by design.
• Eight weeks. Androgen effects commonly take three to six months.

What that trial genuinely refutes is the broad claim that testosterone augments antidepressants for general treatment-resistant depression in women. That claim is dead, and it deserved to be. It does not touch the narrower question about anhedonia, which nobody has properly tested.

Where that leaves it, said plainly

I'm not going to tell you testosterone is a proven treatment for anhedonia in women, because the trial that would prove it has not been run and I won't pretend otherwise. That's the line I won't cross, and you should distrust anyone selling you across it.

What I will tell you is this. The mechanism is real, the one null trial was aimed at a different question and its own authors flagged why it couldn't see an effect, and in the meantime women describing exactly this symptom are being sorted into "depression," medicated on the serotonin system, and never asked about the dopamine one. I was one of them for thirty years.

That's worth raising with your doctor. Go in knowing it's an off-label conversation about an unsettled question, not a proven indication, and you'll have a better one.

So who might reasonably raise it

Not a diagnostic checklist. A rough sense of who has grounds to have the conversation:

• Distressing low sexual desire, where other contributors have been looked at. This is the indication with actual evidence behind it.
• Symptoms that persisted after estrogen and progesterone were optimized. If hormone therapy helped some things and left others untouched, that's a reasonable thing to bring up.
• Surgical menopause, where the drop is abrupt rather than gradual.
• The flatness described above, with the honest caveat that you'd be asking about something the evidence hasn't settled.

And who probably shouldn't yet:

• Anyone who hasn't had thyroid, iron, B12 and sleep looked at. Those produce overlapping symptoms and are easier to fix.
• Anyone hoping it will do what a resistance training program would do. See STEP-HI above, where testosterone was added on top of supervised high-intensity training and didn't improve the outcome measured.
• Anyone whose estrogen and progesterone haven't been addressed. Testosterone used alongside them is the subject of this site, and the sequence matters.

What I'd want you to take away

The case for testosterone in women isn't that it's a proven fix for a long list of midlife symptoms. It isn't, yet, and the honest evidence base is narrower than the enthusiastic version.

The case is that the question was never properly asked for most of that list, that the one approved indication is real, and that women get turned away using a lab number the guidelines say shouldn't be used that way. That's a different and more defensible argument, and it's the one this site makes.

What I am and am not. I'm a medical technologist. I read primary research and I explain study design, including who was in the trial and what was actually measured. I'm not a physician, I'm not diagnosing you, and none of this is a treatment recommendation. Take it to a clinician who knows your history.

The six-part email series goes through the rest: what testosterone does, why the trials missed it, why your labs read "normal," the side effects nobody quantifies honestly, and how women are getting prescriptions.

Source notes (internal, strip before publishing)

| Dichtel 2020: N=101 women aged 21-70, 8wk, two-site, double-blind, adjunctive transdermal T cream vs placebo in antidepressant-resistant MDD; MADRS 26.8→15.3 vs 26.3→14.4, p=0.91; no difference in fatigue, sexual function or side effects; fMRI substudy n=20, no ACC change; LC-MS/MS at Mayo with free T by equilibrium dialysis; placebo response "high (49%), which might have accounted in part for the lack of observed treatment effect" | 10-1176-appi-ajp-2020-19080844 | fulltext ✅, every figure checked against the paper 2026-07-25 | | Annette's own history: 30 years on an antidepressant, pre-testosterone anhedonia, ~90% resolution with ~10% residue | ~/memory/topics/health-hormones.md, annette.md (her own words, 2026-07-16) | her standing choice to publish |

⚠️ Deliberately not claimed: that testosterone treats anhedonia, depression, fatigue or cognition in women. The consensus found insufficient evidence and no trial has tested the narrow anhedonia question. The page argues forcefully that Dichtel 2020 does not close it, which is a different and defensible claim. Do not let a later edit turn "the null doesn't settle it" into "testosterone works." That is the line, and it is the line Annette set herself.

⚠️ The 90/10 number is load-bearing. Her instruction is to use "about 90% resolved, ~10% residue" and never "it fixed it." Do not round it up.

⚠️ Dropped as unverifiable: an earlier draft said the authors cite a literature threshold of >40% placebo response making effects undetectable. That framing is in the project's memory notes but I could not confirm the threshold citation in the paper's text, so the page now quotes only what the authors actually wrote.

Author line verified 2026-07-25. First author is Binder, Ellen F. (Washington University School of Medicine in St Louis), senior author Magaziner, Jay S. The earlier draft said Bhasin, which was wrong and has been corrected. DOI 10.1001/jamanetworkopen.2025.10512.

Door 2

Why are my labs "normal"?

You asked for the test. That already puts you ahead of most women, because usually nobody offers it. The result came back, someone told you it was normal, and that was the end of the conversation.

I want to explain what that number actually is, because "normal" is doing a lot of work in that sentence, and most of it isn't the work you think.

What happened to me

I've been on an antidepressant for thirty years. It works, in the sense that it does the job serotonin drugs do. What it never touched was the flatness.

I didn't want to get out of bed in the morning. Not sadness, and not exhaustion either. I couldn't feel any pull toward the things I used to care about. I'd look at work I'd chosen, that I'd built myself, and feel nothing reaching back. Thirty years of treated depression and that part never moved, because SSRIs work on serotonin and what I'm describing runs on dopamine. Those are different systems. You can be genuinely well treated for depression and still have this.

Nobody ever suggested my hormones. I'm a medical technologist. I've drawn blood for years and I can read my own results, and it still didn't occur to anyone to look, including for a long time me.

Here's the part that changed how I think about lab reports. It wasn't even testosterone that taught me, it was estrogen. I was on hormone therapy, and my estradiol was too low, and nobody ever told me. Not once. I only found out because I'm a medical technologist and I know what estradiol looks like in a twenty-year-old, so when I read my own number I could see it was nowhere near that. My doctor was looking at the same result and calling it fine.

That's the whole problem, and I got lucky in a way most women don't. I had the background to see it. If you don't, you get the word "normal" and no reason to push.

On testosterone, the flatness lifted by about ninety percent. I want to be exact about that, because "it fixed me" would be a lie and this subject has enough of those. Roughly ten percent is still there and I still notice it. But I wake up wanting the day now. I finish things because I want to, not because I've talked myself into it. I hate, hate, hate anhedonia, and I got most of the way out of it, and no lab number was ever going to be the thing that told me so.

The number is less solid than it looks

In 2017, researchers at the CDC and the New York State Department of Health ran a proficiency test. They took unaltered blood serum from four donors, two men and two women, split each sample, and sent it to 142 certified clinical laboratories. Those labs used 16 different testosterone assays. Everyone was measuring the exact same blood.

Then the researchers compared every lab's answer against a reference method, the closest thing there is to a true value.

Here's what came back. For the sample with the lowest testosterone concentration, 15.5 ng/dL, which is squarely in the female range, the labs' results ranged from 17.8% below the true value to 73.1% above it. Same blood. Same day. Different labs.

For the two samples in the male range, 402 and 498 ng/dL, the spread was much tighter, roughly 25% below to 9% above.

Read those two paragraphs again, because that gap is the whole story. The assays do a reasonable job at male concentrations. At the concentrations found in women's blood, they scatter.

(Cao, Botelho, Rej and Vesper, Clinica Chimica Acta, 2017. 142 laboratories, 16 assays, four single-donor specimens, reference-method target values.)

This isn't a fringe complaint

I want to be careful here, because "the tests are wrong" is exactly the kind of thing that shows up on supplement websites, and I don't want to be mistaken for one.

So: in 2019, eleven professional bodies published a joint position statement on testosterone therapy for women. The list includes the Endocrine Society, the American College of Obstetricians and Gynecologists, the North American Menopause Society, and the Royal College of Obstetricians and Gynaecologists. This is the establishment, not the fringe.

Their statement grades each recommendation by evidence strength. On measurement, they wrote that direct assays for total and free testosterone are highly unreliable in the female range, and they graded that finding A, their highest level of evidence.

The same document says total testosterone can be measured accurately using chromatography with tandem mass spectrometry, which is graded B.

So the professional consensus and the proficiency data agree. The common test is not reliable at your concentrations, and a better method exists.

(Davis et al., Journal of Clinical Endocrinology and Metabolism, 2019. Global Consensus Position Statement, endorsed by eleven societies.)

What to ask for

If a testosterone level is going to be measured at all, ask whether the lab is running LC-MS/MS, liquid chromatography with tandem mass spectrometry. That is the method the consensus statement grades as accurate.

The alternative, usually called a direct or immunoassay method, is the one that produced the scatter above. It is cheaper and it is what most labs run by default.

This is a reasonable question to ask, and any lab can tell you which method they use.

The part that complicates the popular version

Here is where I have to disagree with a lot of what's written about this, including things I've seen repeated confidently.

The usual advice is that total testosterone is meaningless and free testosterone is what matters, because free testosterone is the fraction not bound to sex-hormone-binding globulin and therefore the biologically active part.

The 2019 consensus statement does not endorse that. It says research should focus on total testosterone as the main biomarker, specifically because the evidence that free testosterone is the biologically active fraction is weaker than people assume. It also notes that direct assays for free testosterone are unreliable in the female range, which is the same problem again.

I'm flagging this because I'd rather tell you the evidence is unsettled than hand you a clean story that isn't true. Free testosterone measured properly, by equilibrium dialysis, is a real measurement. But "get your free T checked" is not the settled answer it's often presented as, and a free testosterone number from a direct assay carries the same reliability problem as the total.

The androgens nobody put on your panel

There's a second issue, and it isn't about accuracy. It's about what gets measured at all.

Your adrenal glands make a family of androgens called 11-oxygenated androgens. The best studied is 11-ketotestosterone. They are not on a standard hormone panel, so if you've had testosterone tested, you almost certainly have no idea what yours are.

In a 2020 study measuring hormones by mass spectrometry in 590 adults aged 18 to 97, 11-ketotestosterone stayed statistically flat across adult life in women, while classic testosterone declined with age.

That matters because it means the hormone most panels measure is falling while another meaningful androgen pool isn't, and only one of them is being looked at.

Where do they come from? A 2022 study sampled blood directly from the adrenal and ovarian veins and found these androgens come from the adrenal glands rather than the ovaries. That study looked at only four patients, all of them cases of hyperandrogenism rather than typical women, so treat it as suggestive rather than settled. I'm telling you the sample size because a four-person study and a 590-person study are not the same kind of evidence, and papers that don't tell you which one they're citing are hiding something.

The practical implication, if it holds: menopause doesn't touch this pool, and your standard panel never showed it to you.

(Davio et al., Journal of Clinical Endocrinology and Metabolism, 2020, n=590. Auer et al., European Journal of Endocrinology, 2022, n=4.)

So what does "normal" mean

Putting it together, when you were told your testosterone was normal, that sentence rested on:

• a measurement method the consensus grades as highly unreliable at your concentrations, unless the lab happened to run mass spectrometry
• a reference range built from a population, which tells you where you sit relative to other women, not whether the level is right for you
• one androgen, measured once, while a second adrenal androgen pool went unmeasured

None of that means your level is actually low. It might genuinely be fine. What it means is that "normal" was a weaker statement than it sounded, and you were not given enough to know either way.

What I'd actually do with this

Not medical advice, and I'll say why at the end. But practically:

1. Find out which assay your result came from. If it was a direct immunoassay, you know how much confidence to place in the specific number.
2. If testing again, ask for LC-MS/MS.
3. Test in the morning, since testosterone follows a daily rhythm, and be consistent about timing between tests.
4. Ask for SHBG alongside it. It changes how much testosterone is available, and oral estrogen raises it substantially while transdermal estrogen does not.
5. Hold the number loosely. The consensus statement is explicit that a blood testosterone level should not be used to diagnose the one condition testosterone therapy is actually evidence-based for.

That last point surprises people, so I'll repeat it. The eleven societies say the diagnosis comes from clinical assessment, not from the lab value. If a number sent you away, the number was being asked to do something the evidence says it can't do.

A word about what I am and am not. I'm a medical technologist. I've spent a long time reading primary research and I run labs for a living, which is exactly why the proficiency data above bothers me. I'm not a physician, I'm not diagnosing you, and nothing here is a treatment recommendation. Take it to a clinician who knows your history, and ask them the assay question directly.

If this was useful, the six-part email series goes through the rest of it: what testosterone actually does, why the trials missed it, the side effects nobody quantifies honestly, and how women are getting prescriptions.

Source notes (internal, strip before publishing)

Open item: the SHBG route-of-administration claim in step 4 traces to the genetics synthesis, but shifren-2007-menopause-oral-vs-transdermal-shbg did not come back from the fulltext fetch. Either soften that line or source it to a retrievable paper before this page goes live.

Door 3

Who will prescribe it?

This is the question that actually gets typed into search boxes, and it's the one with the least honest writing behind it. Most of what exists is either a clinic advertising itself or a forum thread from 2019.

I can't tell you who will prescribe for you, and I'll explain later why I won't name clinics. I can tell you how the system works, which turns out to be more useful, because once you understand why doctors hesitate you can ask better questions.

First, the thing nobody says out loud: more women are getting it

Prescribing to US women has climbed sharply. A 2026 analysis using Epic Cosmos, a database drawing on more than 300 million patient records, found testosterone prescribing to women rose 2.6-fold between 2016 and 2025, from 50.0 to 130.8 prescriptions per 100,000. Nearly a 59% jump in 2025 alone.

So you're not asking for something exotic. You're asking for something that a lot of clinicians have quietly started doing, faster every year.

(Avivi, Stuenkel, Sampath-Kumar and Ben-Yehuda, JACC: Advances, 2026.)

And the thing that paper says next, which you should hear

The same authors are cardiologists, and their point wasn't to celebrate the trend. It was to note that the safety evidence hasn't kept up.

They measured zero cardiovascular outcomes in women. There were none to measure. Everything known about testosterone and cardiovascular safety in this context is extrapolated from trials in men, principally TRAVERSE.

There was one cardiovascular trial in women, BLISS, run on the LibiGel product. It completed. It was never published.

I'm putting this on the page that's supposed to help you get a prescription because I'd rather you walk into that appointment knowing the honest state of the evidence. Prescribing is rising fast, and the long-term cardiovascular safety data in women does not exist. Both of those are true at once. A clinician who hesitates is not necessarily being obstructive.

Why it's harder than getting estrogen

Four separate frictions, and they stack.

1. It's entirely off-label. There's no FDA-approved testosterone product dosed for women in the United States. I checked the FDA's own database in July 2026: ten marketed brand-name testosterone products, none carrying a female indication. Prescribers therefore use roughly a tenth of a male product's dose, or a compounded preparation. Off-label prescribing is legal and routine, but it puts the decision entirely on the individual clinician, with no manufacturer's label to point to.

2. It's a Schedule III controlled substance. Estradiol and progesterone are ordinary prescriptions. Testosterone isn't, and hasn't been since 1990. That means DEA registration, prescription limits, and every dispensing recorded in the state prescription monitoring database. Some pharmacy chains apply extra scrutiny to off-label testosterone for women as a matter of policy.

3. Telehealth is genuinely unsettled. Tele-prescribing controlled substances has been running on temporary federal flexibilities for years. The permanent rule covering it was proposed in January 2025 and still isn't finalized. Some practices simply won't tele-prescribe testosterone, or require an in-person visit first. This is why a woman can be an established patient at an online menopause clinic, be prescribed estrogen without friction, and be told no on testosterone. It usually isn't the clinician's opinion of testosterone. It's the schedule.

4. Compounding varies. Because no female-dosed product exists, many prescriptions are compounded creams. Pharmacies differ in what they'll prepare and what they'll ship across state lines.

(Project regulatory research: research/regulatory/np-scope-and-testosterone-regulation.md. FDA counts: fda-orangebook-2026-07-testosterone-approved-products, queried 2026-07-25.)

What the professional guidelines actually authorize

This matters because it shapes what a cautious clinician believes they're permitted to do.

The 2019 global consensus statement, endorsed by eleven professional bodies including the Endocrine Society and ACOG, concluded that the only evidence-based indication for testosterone in women is hypoactive sexual desire disorder. For everything else, including fatigue, mood, cognition and bone, they found the data insufficient to recommend it.

They also said that since no approved female product exists, male formulations can be used judiciously, at doses producing concentrations in the normal premenopausal range, and that the diagnosis should come from clinical assessment rather than a blood level.

So a clinician following the guidelines to the letter has a narrow authorized indication, explicit permission to use male products at reduced doses, and instruction not to lean on your lab number. That combination explains a lot of appointments.

(Davis et al., Journal of Clinical Endocrinology and Metabolism, 2019.)

How to make the appointment go better

Not medical advice. Practical framing.

• Ask about the schedule, not the drug. "Does this practice prescribe Schedule III medications, and can that be done via telehealth?" A no here is administrative, and it saves you from concluding your doctor doesn't believe you.
• Know the guideline exists. Naming the 2019 global consensus statement signals you're not asking for something off the internet. It also gives a cautious clinician the cover of an endorsed document.
• Don't lead with your lab result. The consensus says the level shouldn't drive the diagnosis. If you open with a number, you've handed over the easiest way to end the conversation.
• Ask what monitoring they'd want. A clinician willing to prescribe usually has a monitoring plan. Asking for it reads as seriousness, and it tells you whether they've done this before.
• Ask directly whether they've prescribed it to women. Experience varies enormously, and there's no polite way to find this out other than asking.

The directory, and how I'm building it

I am building a list of named clinicians. Here is the rule it runs on: nobody appears on it who hasn't said yes, in writing, with a date.

Not scraped from practice websites. Not collected from forum posts about who prescribed for someone in 2023. Not inferred from a clinic's marketing copy. I write to the clinician, tell them exactly what would be listed, and publish only if they agree.

What that means for you as a reader:

• Every entry shows the date it was confirmed, so you can see how fresh it is rather than trusting that it's current.
• I re-check with every listed clinician every six months. Anything not re-confirmed within nine months gets hidden rather than quietly left up to rot.
• Entries say what the clinician told me about their practice. No entry claims anyone will prescribe for you specifically, because that isn't mine to promise and it isn't theirs to promise sight-unseen.
• Nothing about pricing unless they volunteered it in writing, and nothing at all about any individual patient.
• Any clinician can be removed the same day, by replying to me. No questions.

I researched prescribers in Colorado and watched entries go out of date within weeks. That's the real failure mode for a page like this, and it's why the dates and the re-checks matter more than the length of the list. A short, current, consented list beats a long stale one, and it isn't close.

I take no referral fees and sell nothing. If that ever changes it will be written at the top of the page in the same size type as everything else.

What I am and am not. I'm a medical technologist who reads primary research. I'm not a physician, I'm not diagnosing anyone, and none of this is a treatment recommendation. The decision belongs with a clinician who knows your history.

The six-part email series covers the rest: what testosterone does, why the trials missed it, why your labs read "normal," and the side effects nobody quantifies honestly.

Source notes (internal, strip before publishing)

⚠️ Do not add these figures. The JACC note records that the "8.2% HSDD prevalence" and "1 in 280 women" figures both failed verification in the adversarial check. They are not in this draft and must not be added.

⚠️ Re-check before publishing: the telehealth flexibility status was current as of 2026-07-11 project research. Controlled-substance telehealth rules have moved repeatedly. Verify against the Federal Register at publish time.