You have been trying for over a year. The semen analysis came back abnormal. Then came the repeat sample, the genetic screen, the scrotal ultrasound, the hormone panel, and at the end of all of it, a word that explains nothing. Idiopathic. No cause found. About 40 percent of male infertility ends in that word.
Those figures come from an Italian andrology unit that went back through its own case records from 2016 to 2023 and published the analysis in 2024. They deserve your attention for one uncomfortable reason: under the rules of the diagnosis itself, nobody was looking.
Who was in the study
214 men carrying the label of idiopathic infertility, mean age 38.2 years. 190 of them (88.8 percent) had some degree of abnormality across semen parameters, and 24 (11.2 percent) had no sperm in the ejaculate at all. Every one of them had been investigated until the known causes were gone:
- chromosomal abnormalities and Y chromosome microdeletions
- CFTR gene mutations
- varicocele
- urogenital infection
- obstructive causes
- other endocrine disease affecting the testis, including Cushing syndrome, pituitary tumors and adrenal dysfunction
The comparison group was 224 men with entirely normal semen analysis, mean age 33.7 years. In both groups the blood was drawn fasting, at 8 in the morning. That detail matters more than it sounds, because testosterone measured at any other hour is not the same test.
Why the search stopped where it did
Idiopathic infertility is defined partly by things being normal. FSH inside 1 to 12 IU/L. LH inside 1 to 9 IU/L. Once those two pituitary hormones read normal, the pituitary is declared innocent and the endocrine part of the investigation closes. Guidelines do tell the doctor to measure testosterone at least once in an infertile man. But inside a framework where the label is already fixed, the result has nowhere to go.
The shape of the curve, not the average
Mean total testosterone was 5.2 ng/mL in the infertile men and 5.7 ng/mL in the controls. Statistically significant (p = 0.002), and close to useless for any single man. The two averages sit near each other and both look unremarkable on a lab report.
The difference lives at the bottom of the distribution. In the infertile men the curve was skewed and pushed to the left; in the controls it sat closer to a normal bell. 51 of the 214 infertile men, 23.8 percent, came in below 3.5 ng/mL. Among the men with normal semen, 10 out of 224, 4.5 percent (p < 0.001). Same threshold, very different rates.
The gray zone
These were not obviously hypogonadal men. Anyone below 2.1 ng/mL had been excluded before the study began, because at that level the diagnosis is not in question. What is left is the band between 2.1 and 3.5 ng/mL, roughly 7.3 to 12.1 nmol/L if your lab reports in those units. Not frankly pathological. Not reassuring either, in a group whose mean age was 38.2 years. The authors' own phrasing is deliberately restrained: approximately a quarter of these men, they write, present "some sort of functional hypogonadism."
One signal, in one group only
Split the infertile men at 3.5 ng/mL and the two halves stop behaving alike. Above the line, testosterone correlated with nothing in the semen analysis. Below the line, it correlated with one thing: the percentage of sperm with normal morphology. That held in multivariate regression (R = 0.430, p = 0.020), and held again after adjustment for cryptorchidism, previous varicocelectomy, comorbidities, medication use, smoking and alcohol (R = 0.390, p = 0.025).
On almost everything else the two subgroups were indistinguishable. Same LH, same FSH, same testicular volume, same sperm concentration, same motility, no significant difference in any of it. The one hormonal measure that separated them was the testosterone to LH ratio, 0.9 in the low group against 1.7 in the rest (p < 0.001). That is a pituitary signaling normally and a testis answering weakly.
Nothing else in the file predicted it
The authors went looking for a shortcut. One statistical model built from clinical history, another built from hormones and semen parameters. Neither could predict which men would fall below 3.5 ng/mL (p = 0.588 and p = 0.063). There is no way to infer this number from the rest of the chart. It has to be measured, properly, in the morning, fasting.
What this study does not show
The limits are real. This is retrospective, real world data, not a trial. Testosterone was measured by immunoassay, which the authors themselves note is less accurate at low values than mass spectrometry. The 3.5 ng/mL threshold is a guideline convention, not a proven biological cutoff. The control group had normal semen analysis, but their fertility was never actually established. And above all, nothing here demonstrates that correcting a low testosterone restores fertility. The study did not test that, and it does not claim it.
If you are the man being investigated
A quarter is not a rounding error. It is the difference between a diagnosis that has genuinely run out of road and one that stopped at the second hormone. If your workup ended in the word idiopathic, and there is no morning fasting total testosterone in the file, or there is one and nobody ever discussed where it sat, then one of the two subgroups in this study has not been ruled out in your case. That is a question for a specialist who reads the whole endocrine picture rather than the two gonadotropins that closed the file.
What the data support is narrow and worth stating plainly. Men labeled idiopathic infertile are not one population. They are at least two, they look identical on paper, and one blood sample taken at the right hour is what tells them apart.