Sex hormones

The loop

The hypothalamus signals the pituitary, the pituitary releases LH and FSH, and those instruct the testes or ovaries to produce hormone. The finished hormone then feeds back and quiets the signal. Measuring testosterone alone tells you the output. Measuring LH and FSH alongside it tells you whether the instruction is the problem or the factory is.

The signal is delivered in pulses, not as a steady stream. In men the hypothalamus fires roughly every 2 hours. In women it speeds up to about once an hour before ovulation and slows to every 3 to 4 hours by the end of the cycle. LH follows it in bursts. A single LH measurement is therefore one point on a wave rather than a level, which is why it is read as a broad picture rather than a precise figure. Clearly high, clearly low, or somewhere in the middle.

Binding, and the numbers that come out of it

Most of the testosterone in your blood is not available to your tissues. The bulk is bound tightly to sex hormone binding globulin, a good deal more is loosely attached to albumin, and only a small fraction circulates free. Since SHBG changes with age, thyroid status, insulin, liver function, alcohol and body composition, two people with identical total testosterone can have quite different amounts available. That is why SHBG sits on the profile rather than being an optional extra.

Two of the numbers on your report are worked out from that arrangement rather than measured.

Free testosterone is calculated from total testosterone, SHBG and albumin. Free androgen index is simpler still, at total testosterone divided by SHBG. Neither of those figures involves an assay of its own. Both are attempts at the same question, how much of the hormone is actually available, arrived at by different routes. Free testosterone can also be measured directly, by equilibrium dialysis with mass spectrometry.

They count as markers on the same basis as eGFR, or the absolute white cell counts on a full blood count: calculated from measured inputs, reported as a result, and useful.

But the two are not equally good, and the difference is not a technicality. Leaving albumin out matters, because albumin carries a substantial share of the loosely bound hormone. The index over-estimates how much is available when SHBG is low, and increasingly so the lower SHBG goes. And SHBG falls with obesity, insulin resistance and an underactive thyroid. So it is least reliable in exactly the people most likely to be testing it. UK guidance on androgens in men, published in 2010, describes it as of limited value in men, and the Biomarker Guide sets out why in full, along with the calculated-versus-measured distinction.

The free androgen index is not sold on its own, and the reason is simple. A standalone version would contain no assay at all. There would be nothing to run. It appears where testosterone and SHBG are both already being measured, which is the only place it can exist.

Albumin turns up in the free testosterone calculation as a carrier, which is the third job it does on a report. The Minerals and bone section pulls all three together.

The rest of the profile

Prolactin can suppress the whole axis, which is why it is measured alongside rather than separately.

Oestradiol and progesterone complete the female profile. The menstrual cycle has distinct phases and both hormones move substantially through them, so the expected picture depends on where in the cycle a sample sits. The physiology below explains why, and the Biomarker Guide covers what that means for a result.

Testosterone is also a raw material

Almost nobody explains this part, and it is why one testosterone number cannot describe what is happening everywhere in the body.

Testosterone is an active hormone in its own right. It is also a starting material, and two enzymes take it in different directions.

Aromatase converts testosterone into oestradiol. It is not confined to the ovaries: it sits in fat tissue, bone, brain and the testes, so a man's oestradiol is made largely from his own testosterone, in his own tissues. More fat tissue means more aromatase, which is one of the ways body composition feeds back into the hormone picture.

5-alpha reductase converts testosterone into dihydrotestosterone, a considerably more potent androgen. Only a small percentage of testosterone takes this route, but in the tissues that use it, dihydrotestosterone is the androgen that matters rather than testosterone. There are two forms of the enzyme. One sits mainly in the prostate, the genital tract and hair follicles. The other sits mainly in liver and skin, and also in fat, muscle and bone. That includes the sebaceous glands of the scalp, so scalp skin carries both forms.

So which arm matters depends on the tissue. Muscle responds largely to testosterone itself. The prostate and the scalp respond to dihydrotestosterone. Bone and fat are strongly influenced by oestradiol.

That is the mechanism behind the medications described in the Prostate section. Finasteride blocks one form of 5-alpha reductase, which is why it is used for prostate enlargement and for scalp hair loss, and why it roughly halves PSA. Dutasteride blocks both forms. Neither touches aromatase, so the oestradiol arm carries on regardless.

And oestradiol in men does real work rather than being a by-product. Where the two have been told apart, lean mass, muscle size and strength depended on testosterone, while body fat and bone loss depended on oestradiol, and both contributed to libido and erectile function. An oestradiol in a man is therefore describing something the testosterone number does not.

The loop runs in both directions, and treatment closes it

This is the piece that explains what happens on testosterone treatment, and it depends on a number that is never on a blood test.

The testosterone concentration inside the testes is something like 40 to 100 times the concentration in blood. It is produced locally, and the cells that produce it are driven by LH. Sperm production depends on that local concentration, not on the blood level.

Testosterone given from outside raises the blood level, and the loop reacts exactly as it is designed to. The hypothalamus and pituitary see plenty, so the instruction is withdrawn. LH and FSH fall. Local production mostly stops, and the concentration inside the testes collapses towards the concentration in blood, which is a small fraction of what sperm production requires.

So a blood testosterone can look excellent while the number that governs fertility has fallen by an order of magnitude. A suppressed LH is the expected finding rather than a surprise, and it is the reason LH is worth measuring alongside testosterone rather than instead of it.

The test that actually answers the fertility question

No blood test tells you whether you are making sperm. Not testosterone, not LH, not FSH. They tell you what the signal is doing. The only thing that answers the question directly is looking at the sperm.

A proper semen analysis is a lab test. It measures volume, concentration, how many are moving, how many are alive, the shape of them, and the acidity of the sample. Your GP can arrange one. Privately it runs to somewhere between about £140 and £200 at a fertility clinic and needs no referral.

Home tests exist too, and they are cheap. Fourteen of them are on sale in the UK. Every one reports sperm concentration, about 6 also report motility, and not a single one can assess shape. Accuracy for the things they do measure runs from 93% to 99%, and they are at their best when the count is very low. At borderline or normal numbers they are much less reliable.

So a home test is a screen, not an answer. A clearly bad result is worth acting on. A reassuring one from a test that only counts sperm has told you about one of the 6 things a lab would look at.

If you are on testosterone, or thinking about it, and fertility matters to you at any point in the future, this is the test to do before you start rather than after. Human chorionic gonadotropin is used alongside testosterone precisely to keep that local production going, and sperm banking exists. Both work far better as a decision made in advance than as a rescue.

Ovarian reserve, and the one marker that reads inside the testis

Anti-Müllerian hormone is produced by the small growing follicles in an ovary, so the amount circulating tracks roughly how many of them are left. That is what ovarian reserve means, a count rather than an assessment of quality, and not a prediction of natural fertility. Unusually among female hormones it can be measured on any day of the cycle, because it does not swing with the phases the way oestradiol and progesterone do.

It is also only half of the assessment. The other half is an antral follicle count, a transvaginal ultrasound that literally counts the small follicles sitting in each ovary at the start of a cycle. The two get used together because they measure the same thing by different routes and they do not always agree. A blood test alone gives you one of the two. If the number matters to you, the scan is the other one to ask for.

Men produce it too, from the cells in the testis that nurse developing sperm, and here it connects directly to the paragraph above. What suppresses anti-Müllerian hormone in a man is not the testosterone in his blood. It is the testosterone concentration inside the testis, the number that never appears on any blood test. Which makes this one of very few markers on any panel that reads, indirectly, what is happening inside an organ rather than in the circulation.

What it is and is not established for, in both sexes, is in the Biomarker Guide, including the guideline positions on when not to use it. Those are unusually blunt and you should read them before ordering it.

What treatment actually does, and what it does not

The rest of this section describes the system. This part is about what happens when you intervene in it, because it is the question most people arrive with, and because until recently nobody knew the answer.

The question that hung over testosterone treatment for 40 years was whether it caused heart attacks. It has now been tested properly. A large randomised trial gave either testosterone gel or an identical dummy gel to middle-aged and older men who had low testosterone with symptoms, and who already had cardiovascular disease or a high risk of it, then followed them for around 3 years. There was no excess of heart attacks, strokes or cardiovascular deaths. That is the strongest evidence that exists on the question, and it points the same way as the smaller work before it.

The same trial answered several other questions at the same time, and not all of the answers were the expected ones.

What improved. Anaemia corrected more often on treatment than on placebo. Sexual desire and sexual activity both improved.

What did not. Erectile function did not improve. In men with prediabetes, treatment did not prevent progression to diabetes, which many people had expected to be one of its clearest benefits.

What went the wrong way. Fractures were more common on treatment, not less. That result deserves attention, because testosterone reliably improves bone density, and bone density is the number usually quoted to argue that it protects the skeleton. The only trial large enough to count actual fractures found more of them. A marker moving in the right direction is not the same thing as the outcome moving in the right direction, and this is one of the cleanest examples of that gap anywhere in medicine.

Prostate. There was no excess of aggressive prostate cancer. The caveat is a real one. The total number of such cancers across the entire trial was very small, and men who already had a raised PSA or significant urinary symptoms were excluded before it began. So the result rules out a large short-term hazard in men who have been screened first. It does not establish long-term prostate safety, and the Endocrine Society restated that in 2026.

Three other events occurred more often on treatment: atrial fibrillation, acute kidney injury and clots on the lung. These were secondary findings rather than what the trial was designed to measure, so they are a reason to watch rather than a settled harm.

Regulators moved on the back of all this. In February 2025 the American regulator removed the cardiovascular warning that had sat on testosterone products for a decade, and added a new one about raised blood pressure, having confirmed that rise across every product in the class. Which leaves 2 things that genuinely need monitoring on treatment: haematocrit, because testosterone thickens the blood, and blood pressure.

What is still unknown is the part that matters most to anyone reading this. The trial ran for about 2 years of treatment in older men who already had heart disease. Nobody has run the equivalent trial in healthy men in their thirties and forties taking testosterone for the following 30 years, and that is who most private prescribing actually serves. The safety evidence everyone quotes was not collected in people like them. The trial was also funded by the companies that make the drug. It was well conducted and that funding does not make it wrong, but it is a fact a reader is entitled to have before weighing it.

The thresholds that decide who gets offered treatment in the first place, and the disagreement between the two British guidelines about where those lines sit, are in the Testosterone entry in the Biomarker Guide.

What changes when these are low

Low testosterone tends to bring fatigue, reduced libido, low mood, loss of muscle and strength, poor concentration and disturbed sleep. Falling oestradiol produces hot flushes, disturbed sleep, mood changes, vaginal dryness and accelerated bone loss.

Both lists overlap heavily with thyroid problems, iron deficiency, poor sleep and simple overwork, which is why a hormone result read without the rest of the picture is a poor guide.