Serum Antimüllerian Hormone as a Biomarker of Functional Testicular Reserve: A Comparative Analysis

AMH is lower in men with non-obstructive azoospermia, but doesn't independently predict sperm concentration once FSH and testicular volume are accounted for

Journal: The Journal of Clinical Endocrinology and Metabolism | Published: 2026-02-20 | Type: Cross-sectional comparative study | PMID: 40976239 Authors: Pozzi E et al. — Vita-Salute San Raffaele University and Foundation IRCCS Ca' Granda, Milan Funding/COI: Not listed

Summary

This cross-sectional study of 1,085 men across three fertility categories found that serum AMH was significantly lower in men with non-obstructive azoospermia (NOA) compared to fertile controls and non-azoospermic infertile men. The headline result is what AMH can't do: after adjusting for age, FSH, and testicular volume, AMH lost its independent association with sperm concentration. AMH tracks Sertoli cell activity and testicular status — it's a marker of the environment, not the output.

Claims

• AMH was significantly lower in NOA men vs. fertile controls and primary infertility: 3.8 (IQR 1.6–7.2) vs. 5.1 (3.6–7.0) vs. 4.9 (3.0–7.8) ng/mL (P < .001)
• AMH correlated negatively with age and FSH, positively with testicular volume and sperm concentration
• In multivariable regression, AMH was not independently associated with sperm concentration — FSH and testicular volume were
• Age, FSH, and testicular volume were the independent predictors of AMH level itself

Study Quality

The sample size (n = 1,085) is solid for this type of analysis and the three-group comparison — confirmed fertile, primary infertile non-azoospermic, and NOA — is clinically meaningful. Use of WHO 2010 criteria for semen analysis is appropriate if now slightly dated. Kruskal-Wallis and Spearman rank correlation are correct choices given likely non-normal hormone distributions, and multivariable regression to disentangle AMH's role from confounders is exactly what this kind of dataset needs.

The cross-sectional design is the primary methodological ceiling: it can show associations but cannot establish whether low AMH precedes or results from impaired spermatogenesis. The study is restricted to White European non-Finnish men, which limits generalizability to other ethnic groups where AMH reference ranges may differ.

Red Flags

• No funding or conflict of interest disclosed — notable absence for a biomarker study with potential commercial diagnostic implications
• All authors are from two Milan institutions; no external validation cohort
• Restricted to White European men by design — authors acknowledge limited ethnic generalizability
• "Confirmed fertility" group (n = 116) is small relative to infertility groups; selection criteria for this comparator aren't detailed in the abstract
• Cross-sectional design precludes any causal inference

Strengths

• Large, well-characterized cohort (n = 1,085) with comprehensive hormonal and semen data
• Three clinically distinct comparison groups, not just fertile vs. infertile
• Multivariable modeling appropriately tests AMH's independent contribution rather than stopping at correlation
• Honest conclusion: the paper doesn't oversell AMH; it correctly identifies where it adds and doesn't add information

Verdict

This paper does what a good comparative study should: it finds a signal (AMH distinguishes NOA from other groups) and then methodically tests whether that signal is redundant with what clinicians already measure (it mostly is, once you have FSH and testicular volume). The practical takeaway is that AMH may be a useful supplementary marker in the azoospermia workup, but it doesn't replace established parameters for predicting sperm concentration. Worth reading for anyone building a male fertility biomarker panel. The missing funding/COI disclosure is a frustrating gap in an otherwise tidy paper.