Cardiovascular

Circulating Serum AGRN and ADAMTS8 With Related Candidates as Potential Biomarkers for Carotid Plaque Presence and Vulnerability.

TL;DR

Circulating AGRN and ADAMTS8 are associated with carotid plaque presence versus healthy controls but do not independently discriminate plaque presence or instability after multivariable adjustment, and clinical covariates alone provided near-perfect discrimination, indicating substantial spectrum bias in the study design.

Key Findings

Serum AGRN concentrations were significantly elevated in patients with carotid plaques compared to healthy controls.

  • Mean ± SD: Stable plaque group: 2222.42 ± 508.30 pg/mL; Unstable plaque group: 2109.59 ± 513.51 pg/mL; Healthy controls: 1800.36 ± 375.89 pg/mL
  • Overall p < 0.001 across the three groups
  • 128 participants total: stable plaques (n=41), unstable plaques (n=49), healthy controls (n=38)
  • Plaque stability was determined by carotid ultrasonography
  • Proteins were quantified by enzyme-linked immunosorbent assay (ELISA)

Serum ADAMTS8 levels were comparable across the stable plaque, unstable plaque, and healthy control groups.

  • Overall p = 0.110 across the three groups
  • No statistically significant difference was observed between any group pair for ADAMTS8
  • Measured alongside AGRN, ADAMTS9, COL18A1, ITGB4, LOXL4, and SLIT3 as a prespecified seven-protein panel

Neither AGRN nor ADAMTS8 was independently associated with plaque presence or plaque instability after multivariable adjustment.

  • Multivariable Firth penalized logistic regression was used with log2-transformed biomarkers
  • Covariates adjusted for: age, sex, hypertension, alcohol use, dyslipidemia, smoking, diabetes mellitus, and coronary heart disease
  • All adjusted p > 0.05 for both AGRN and ADAMTS8 for both outcomes
  • Firth penalization was applied, likely due to small sample sizes and potential separation issues

The combined AGRN+ADAMTS8 model achieved moderate discrimination for carotid plaque presence versus healthy controls.

  • Apparent AUC of 0.771 (bootstrap 95% CI, 0.678–0.865)
  • Optimism-corrected AUC: 0.764
  • Repeated stratified 5-fold cross-validated AUC: 0.749
  • ROC analysis was used to evaluate individual biomarkers and the combined prediction probability

The combined AGRN+ADAMTS8 model showed no discriminatory capacity for plaque instability.

  • Apparent AUC: 0.549
  • Optimism-corrected AUC: 0.498
  • Repeated 5-fold cross-validated AUC: 0.471
  • These values are near or below chance (0.5), indicating the model cannot distinguish stable from unstable plaques

Clinical covariates alone provided near-perfect discrimination of plaque presence, indicating substantial spectrum bias from the healthy-control comparator design.

  • Age, sex, and vascular risk factors (hypertension, alcohol use, dyslipidemia, smoking, diabetes mellitus, coronary heart disease) alone nearly perfectly separated plaque patients from healthy controls
  • This finding demonstrates that the healthy-control comparator design introduces spectrum bias
  • Authors note this underscores the need for age- and risk-matched comparator populations in future biomarker studies
  • The study design is described as inconsistent with STARD 2015 recommendations regarding appropriate comparator selection

Baseline characteristics differed substantially between plaque groups and healthy controls.

  • Study enrolled 128 participants: stable plaques (n=41), unstable plaques (n=49), healthy controls (n=38)
  • Differences in baseline characteristics included established cardiovascular risk factors
  • These differences confound interpretation of biomarker associations, as the groups were not matched on age or risk factor burden

What This Means

This research suggests that two blood proteins — AGRN (agrin) and ADAMTS8 — along with five related proteins were measured in the blood of 128 people to see if they could help identify whether someone has carotid artery plaque (fatty buildup in neck arteries) or whether that plaque is unstable and dangerous. People with carotid plaque had higher levels of AGRN compared to healthy individuals, but ADAMTS8 levels were similar across all groups. When these proteins were combined into a prediction model, they showed moderate ability to distinguish plaque patients from healthy controls, but essentially no ability to tell whether a plaque was stable or unstable — a clinically more important question since unstable plaques are more likely to cause strokes. However, a critical finding was that simply knowing a person's age and standard risk factors (like high blood pressure, diabetes, and smoking history) was nearly enough on its own to predict who had carotid plaque — meaning the healthy control group was so different from the plaque patients that the biomarkers' apparent usefulness was likely inflated. This is a form of bias called "spectrum bias," where comparing very different groups makes a test look more useful than it would be in real-world screening. This research suggests that future studies testing blood biomarkers for carotid plaque should compare plaque patients against people of similar age who also have cardiovascular risk factors but no plaque, rather than against young, healthy individuals. Without this type of matched comparison, it is difficult to know whether biomarkers like AGRN truly add value beyond what doctors can already determine from a patient's medical history and standard risk assessment.

Have a question about this study?

Citation

Guo S, Cui S, Zhang X, Xing Z, Sun Y, Qin M, et al.. (2026). Circulating Serum AGRN and ADAMTS8 With Related Candidates as Potential Biomarkers for Carotid Plaque Presence and Vulnerability.. CNS neuroscience &amp; therapeutics. https://doi.org/10.1002/cns.71130