Association Between Chronic Inflammation Status and Serum Testosterone and Free Testosterone in Adult Males

In 7,576 men from NHANES, higher composite inflammation scores independently predicted lower total and free testosterone — but the age-stratified FT findings raise questions

Journal: The Aging Male | Published: 2026-02-18 | Type: Cross-sectional | PMID: 41708584 Authors: Zhang Si-Zheng et al. (Guang'anmen Hospital, China Academy of Chinese Medical Sciences, Beijing) Funding/COI: Not disclosed

Summary

Using NHANES data from 2011–2016 and 2021–2023, this study tested whether two composite inflammation indices — SIRI (systemic inflammation response index) and AISI (aggregate index of systemic inflammation), both calculated from routine blood differentials — track with testosterone levels in adult men. They do, inversely and significantly. The association with free testosterone (FT) is messier, appearing only in men aged 20–39 and ≥60, skipping middle age entirely, with no clear mechanistic explanation offered.

Claims

• Higher SIRI and AISI were significantly associated with lower serum testosterone and FT (P < 0.0001) across all adjusted models
• Men in the highest quartiles of SIRI and AISI had the lowest testosterone levels
• Both indices showed nonlinear inverse associations with serum testosterone
• AISI also showed a nonlinear association with FT; SIRI's association with FT was linear
• The link to biochemical hypogonadism was significant only when defined by total serum testosterone — not when defined by FT
• Age-stratified analysis: FT associations were significant in men aged 20–39 and ≥60, but not in men aged 40–59

Study Quality

NHANES is a nationally representative US survey with rigorous sampling methodology, making the 7,576-man testosterone cohort one of the more credible cross-sectional datasets available for this kind of analysis. The authors applied multivariable regression with threshold effect and smoothing curve analyses, which appropriately tests for the nonlinearity they found.

That said, this is fundamentally a cross-sectional snapshot. Both inflammatory markers and testosterone were measured at a single timepoint. Testosterone has well-documented diurnal and day-to-day variability; a single draw introduces substantial measurement noise. The study cannot determine whether inflammation precedes testosterone decline, whether low testosterone promotes inflammatory states, or whether a third variable (obesity, metabolic syndrome, chronic disease) drives both.

Red Flags

• Cross-sectional design makes causal inference impossible — correlation is all that's on the table
• Funding source and conflicts of interest are not disclosed; this is a transparency failure regardless of what the answer would have been
• SIRI and AISI are composite markers built from CBC differentials (neutrophils, monocytes, lymphocytes, platelets) — they are not universally validated biomarkers of chronic inflammation and their clinical utility is debated
• The age-stratified FT result — significant in men 20–39 and ≥60 but not 40–59 — is biologically unexplained and pattern-consistent with multiple testing artifact
• No data on exogenous testosterone use, which is common in this age range and would directly confound results
• Single blood draw for testosterone lacks time-of-day standardization details in the abstract

Strengths

• Large, nationally representative US sample (7,576 for total testosterone; 3,590 for FT)
• Draws from multiple NHANES cycles spanning over a decade
• Age-stratified analysis is methodologically sound even if the FT results are unexplained
• Tests for nonlinearity rather than assuming linear dose-response
• Used multivariable adjustment to account for confounders

Verdict

This paper adds a reasonably large data point to the existing literature linking systemic inflammation to testosterone suppression — a hypothesis with plausible biological grounding (inflammatory cytokines suppressing Leydig cell function is established in animal models). The NHANES sample gives it more weight than most single-site studies. But the missing COI disclosure is a problem, the single-timepoint design is the hard ceiling on interpretation, and the age-stratified FT findings look more like noise than signal without a mechanism. File this under "consistent with prior evidence, does not advance the field" rather than anything requiring a change in how you read the literature.