FSH reports on sperm, not on testosterone.
A weaker marker of spermatogenesis than you have been told — and the claim that a high FSH predicts failed sperm retrieval is not merely unsupported, it is backwards.
FSH reports on a different part of the testis from the part that makes testosterone, and it is the test most often over-read.
What it measures
Follicle-stimulating hormone acts on the Sertoli cells, which support sperm production. LH acts on the Leydig cells, which make testosterone. The two hormones come from the same pituitary cells and are usually ordered together, but they are reporting on two different compartments.
A raised FSH means the Sertoli compartment is not returning its usual feedback signal — the classic pattern of impaired spermatogenesis. A low FSH alongside a low testosterone means the pituitary is not driving either compartment, which points upstream.
An unexpected fact about what controls it
Testosterone is usually described as the negative feedback signal for both gonadotropins. For FSH that is not correct.
In sixteen normal men, blocking aromatase raised testosterone from 563 to 817 ng/dL while dropping estradiol; medical castration dropped testosterone from 491 to 40 ng/dL and dropped estradiol by a similar amount. LH responded very differently to the two manipulations — up 275% after castration versus 96% after estradiol withdrawal alone. FSH responded almost identically to both — 91% versus 71%, no significant difference — despite testosterone moving in opposite directions.
The authors’ conclusion: “E2 is the predominant regulator of FSH secretion in the human male.”
That is worth carrying, because it means an FSH is not simply a second reading of the same axis that LH reports on.
Evidence: Established.
Guideline position — and an asymmetry worth knowing
The Endocrine Society pairs them: distinguish primary from secondary hypogonadism “by measuring serum luteinizing hormone (LH) and follicle-stimulating hormone (FSH) concentrations,” a strong recommendation on moderate-quality evidence. The EAU rates the pair Strong. The VA’s Recommendations for Use — formulary guidance rather than a clinical practice guideline — requires both once low testosterone is confirmed.
The AUA does not. Its numbered statement covers LH only — “In patients with low testosterone, clinicians should measure serum luteinizing hormone levels,” Strong Recommendation, Grade A — and FSH appears nowhere in its graded statements, only in the supporting discussion. That is a real difference between two American guidelines and it is rarely noted.
The variation problem is smaller than LH’s, and less settled than it looks
FSH is the steadier of the two. Unlike LH, discrete pulses are rarely detectable, and there is no diurnal variation to speak of — one study sampling twenty men every ten minutes found FSH pulses “rarely observed” and no daily rhythm.
Its biological variation is correspondingly lower, but the published figures do not agree. Two modern studies put the within-person coefficient of variation at about 8%. A 1993 study of 20 men over a year, using a different statistic and an older assay generation, gave 17.3% — and one of the modern studies’ own indirect method gave 19%. We would quote the range rather than the tidy 8%.
Its index of individuality was 0.14 in the study that measured it — the lowest of the five hormones that study covered. In practical terms that is the opposite of a reassuring number: it means your own FSH sits in a narrow band that the wide population reference interval cannot see. A value at the top of “normal” may be far from normal for you.
As with LH, there is no CDC standardization program for FSH. That program covers testosterone and estradiol only. WHO International Standards exist; comparability across platforms does not follow from them, and the director of the UK external quality scheme for these hormones has said so in print.
The correction that matters most: high FSH and sperm retrieval
For years men with non-obstructive azoospermia and a high FSH were told the outlook for surgical sperm retrieval was poor. That belief comes from the era of random testicular biopsy. It does not survive microdissection TESE, and the data run in the opposite direction.
In 792 men with non-obstructive azoospermia undergoing microdissection TESE:
| Serum FSH (IU/mL) | Sperm retrieval rate |
|---|---|
| Under 15 | 51% |
| 15 to 30 | 60% |
| 31 to 45 | 67% |
| Over 45 | 60% |
Retrieval was highest in the high-FSH strata, and pregnancy and live-birth rates among couples where sperm was retrieved were similar across all four. The authors’ own words: “High FSH is not a contraindication for micro-TESE.”
This is not an isolated finding. A separate series of 311 men reported that “spermatozoa recovery was independent of serum follicle-stimulating hormone.” A study selecting men with both small testes and an FSH above 20 found retrieval rates of 24% versus 32% in men with only one of those features, and concluded that “no successful pre-operative predictors of sperm recovery were identified.” In 191 microdissection procedures, FSH differed between success and failure on simple comparison but was not an independent predictor once biopsy score, testosterone, family history and Y-chromosome microdeletion were accounted for. And in 148 men with presumed Sertoli-cell-only syndrome, retrieval was 28.9% above an FSH of 15.25 and 11.8% below it — again favouring the higher FSH.
We should present the dissent too. One smaller series found a raised FSH did independently predict retrieval failure — with conventional TESE — while also finding that microdissection was independently predictive of success in exactly those men. That is consistent with the rest: the problem was the surgical technique, not the hormone.
The likely biology is that a low FSH in azoospermia often flags maturation arrest with globally poor tubules, while a high FSH flags Sertoli-cell-only histology with focal preserved areas that microdissection is designed to find.
Evidence: Established (that high FSH does not contraindicate microdissection TESE).
FSH is a weaker marker of spermatogenesis than you were told
Directionally it is right. As a discriminator it is second-best.
In 218 subfertile men, with testicular biopsy in a subset, the receiver-operating characteristic accuracy for distinguishing competent from impaired spermatogenesis was 80% for FSH and 95% for inhibin B. Inhibin B correlated with biopsy score at r = 0.76, with sperm count at 0.54, and with testicular volume at 0.63. The authors concluded that “inhibin B is the best available endocrine marker of spermatogenesis in subfertile men.”
The practical consequence: a normal FSH does not exclude impaired spermatogenesis. If fertility is the question, the answer comes from a semen analysis — which is what the Endocrine Society actually recommends, and it recommends at least two of them, separated by several weeks.
Evidence: Moderate.
What raises FSH
Anything that damages the seminiferous tubules: Klinefelter syndrome, prior orchitis, chemotherapy, testicular radiation, a history of undescended testes, and testicular surgery itself. After testicular sperm extraction, mean FSH rose from 22 to 30 IU/L in one series of 435 men, while LH rose non-significantly — a clean illustration of which compartment the procedure costs you.
Aging, in the same pattern as LH.
Rarely, a gonadotropin-secreting pituitary adenoma. In one surgical series of clinically non-functioning adenomas, five of 22 showed hormone hypersecretion in life — and all five were men.
What lowers FSH
Exogenous testosterone and anabolic steroids, profoundly and for a long time. In men currently using androgens, mean FSH was 0.5 IU/L against 4.9 in non-users.
Obesity, opioids, hyperprolactinemia, pituitary disease and congenital hypogonadotropic hypogonadism — the same central causes that lower LH.
Energy deficit is the exception. In healthy men fasted for five days with sampling every five minutes, mean LH fell 30% and FSH was unaffected entirely. The two hormones do not always move together.
Recovery is slower than for LH
This is the part most men on testosterone are not told, and it is the reason to settle the fertility question before the first dose rather than after.
In a series comparing current androgen users, past users and non-users, mean time to recovery was 10.7 months for LH but 19.6 months for FSH — and 31.7 months for inhibin B, the marker that actually tracks the Sertoli compartment. Sperm output recovered at a mean of 14.1 months. Longer duration of use was the only variable associated with slower recovery of sperm output.
The frequently quoted figure that everything returns within three or four months comes from a different study — a large analysis of recovery after hormonal male contraception, whose outcome was sperm concentration in young healthy men in supervised trials. Its median time to reach 20 million per millilitre was 3.4 months, with 100% recovery by 24 months. That is a real and reassuring finding about a different population, and it does not measure gonadotropin recovery at all.
Evidence: Moderate.
Assay artefacts
Macro-FSH — FSH bound into a large immunoglobulin complex, producing a spuriously high result — is documented, and the pattern is distinctive: an isolated high FSH with a normal LH, in someone whose gonadal function is clinically normal. In one reported case FSH fell from 41.1 to 6.54 IU/L after the complex was precipitated out. Fewer than ten cases have been published, and every one of them is in a female patient. The mechanism is not sex-specific, so it belongs on the differential for an unexplained isolated FSH elevation in a man with normal testes and normal semen parameters — but we should be clear that no male case is on record.
Heterophile and anti-animal antibodies produce the same picture. In several published cases the standard blocking tubes were negative, and in one the interference was exposed only by switching to a platform whose antibodies came from a different species.
Biotin interferes with the assays used for FSH and LH. Stop high-dose supplements before the draw.
The hook effect is documented for hCG and prolactin. We found no case for FSH and will not assert one.
Related
The Lab Library · LH · TRT and fertility · Total testosterone
Questions patients ask
My FSH is high, so sperm retrieval won't work.
Wrong, and backwards. With microdissection, retrieval was highest in the men with the highest FSH.
What the evidence showsIn 792 men with non-obstructive azoospermia, sperm retrieval was 51% at an FSH below 15 and 60 to 67% above it, with similar pregnancy and live-birth rates. The authors wrote that "high FSH is not a contraindication for micro-TESE." Four other series agree, including one in presumed Sertoli-cell-only syndrome where retrieval was 28.9% above an FSH of 15.25 versus 11.8% below.
What remains uncertainWhich men will succeed. The best independent predictors in one series were biopsy score, testosterone, family history and Y-chromosome microdeletion — not FSH.
Bottom lineA high FSH is a reason to see a fertility specialist, not a reason to give up.
Strong
My FSH is normal, so my sperm production is fine.
It makes it more likely, and it does not establish it.
What the evidence showsIn 218 subfertile men, FSH's accuracy for distinguishing competent from impaired spermatogenesis was 80%, against 95% for inhibin B. FSH's index of individuality is about 0.14 — meaning your own range is narrow and the population interval is a blunt tool for you.
What remains uncertainWhether measuring inhibin B routinely would change decisions. It is not in any guideline.
Bottom lineIf fertility is the question, the answer is a semen analysis. Two of them.
Moderate
FSH and LH tell you the same thing.
They report on different compartments, and different signals control them.
What the evidence showsIn sixteen men, medical castration and selective estradiol withdrawal produced very different LH responses — 275% versus 96% — but essentially identical FSH responses of 91% and 71%, despite testosterone moving in opposite directions. The conclusion was that estradiol, not testosterone, is the predominant regulator of FSH in men. A five-day fast lowered LH by 30% and left FSH untouched.
What remains uncertainNothing about the direction; the fine control of FSH in men is still being worked out.
Bottom lineLH reports on testosterone production. FSH reports on sperm production.
Moderate
My FSH will bounce back a few months after I stop testosterone.
FSH is roughly twice as slow as LH, and inhibin B slower still.
What the evidence showsIn men who had used androgens, mean time to recovery was 10.7 months for LH, 19.6 months for FSH, 14.1 months for sperm output and 31.7 months for inhibin B. The often-quoted "3.4 months" figure is the median time to recover a sperm concentration of 20 million per millilitre in young men in supervised contraceptive trials — a different question and a different population.
What remains uncertainHow closely those figures transfer to men on ordinary replacement doses rather than supraphysiological ones.
Bottom lineSettle fertility before the first dose. See TRT and fertility.
Moderate
An isolated high FSH with everything else normal means something is wrong with my testicles.
Usually. But that exact pattern is also the signature of an assay artefact.
What the evidence showsMacro-FSH — the hormone bound into a large complex with immunoglobulin — produces an isolated high FSH with a normal LH in a person whose gonadal function is clinically normal. In one case the value fell from 41.1 to 6.54 IU/L once the complex was removed. In several published cases the standard blocking tubes were negative, and in one the interference was exposed only by switching to a platform whose antibodies came from a different species.
What remains uncertainHow often this happens in men. Fewer than ten cases have been published and all of them are in women.
Bottom lineIf the FSH does not fit the man in front of you, ask the laboratory before you act on it.
Limited
My urologist checked LH but not FSH, so something was missed.
Not necessarily — the two American guidelines genuinely disagree here.
What the evidence showsThe Endocrine Society recommends measuring both, as a strong recommendation. The AUA's only graded statement covers LH; FSH appears nowhere in its numbered recommendations, only in the supporting text. The EAU and the VA's formulary guidance both call for the pair.
What remains uncertainWhether adding FSH changes management in a man who is not asking about fertility. Nobody has tested it.
Bottom lineIf fertility matters to you, ask for it — and ask for a semen analysis, which tells you more.
Moderate
Ready for testosterone care built on a diagnosis?
The first step is a full endocrine evaluation, not a prescription. We see patients across San Diego County and welcome referrals from other physicians.
