Plasma FGF21 is associated with intracranial atherosclerotic stenosis in patients with type 2 diabetes mellitus

Front Neurol. 2026 Sep 7;17:1904036. doi: 10.3389/fneur.2026.1904036. eCollection 2026.

ABSTRACT

INTRODUCTION: FGF21 is associated with atherosclerosis and the anatomical severity of coronary artery disease. No study has explored the relationship between FGF21 and intracranial atherosclerotic stenosis. The purpose of this study is to explore the association between FGF21 and intracranial atherosclerotic stenosis in patients with type 2 diabetes mellitus.

METHODS: One hundred ten patients with type 2 diabetes were recruited. The levels of plasma FGF21 were measured by enzyme-linked immunosorbent assays. The severity of intracranial atherosclerotic stenosis was evaluated in strict accordance with the internationally recognized modified Warfarin-Aspirin Symptomatic Intracranial Disease criteria. Patients were categorized into three groups based on percent stenosis: no stenosis (0%), mild stenosis (1-49%), and moderate to severe stenosis (≥50%).

RESULTS: Plasma FGF21 level was different among no stenosis group, mild stenosis group and moderate to severe stenosis group (114.35 (105.35-124.04) vs. 124.81 (118.84-134.47) vs. 132.18 (120.43-150.05), H = 17.327, p < 0.001). FGF21 levels in the moderate-to-severe stenosis group were higher than those in the no stenosis + mild stenosis group (132.18 (120.43-150.05) vs. 122.44 (109.88-129.19), U = -3.138, p = 0.002). The highest tertile (T3) of FGF21 level (>129.55 ng/L) and lnFGF21 was associated with moderate to severe intracranial atherosclerotic stenosis in patients with T2DM after adjusting for age andα-glucosidase inhibitor use (OR = 6.453, 95% CI: 1.944-21.415, p = 0.002; OR = 2.191, 95% CI: 1.277-3.759, p = 0.004). Based on the receiver-operating characteristic (ROC) curve, the combination of lnFGF21, age, and α-glucosidase inhibitor use predicted moderate to severe intracranial atherosclerotic stenosis in patients with type 2 diabetes mellitus, with an area under the curve (AUC) of 0.768 (95% CI, 0.670-0.865). Bootstrap resampling with 1,000 replicates was performed for internal validation to assess model stability and calibration (bootstrap B = 1,000; MAE = 0.021, MSE = 0.00064, 90th percentile of absolute error = 0.038).

CONCLUSION: We observed a significant cross-sectional association between plasma FGF21 levels and intracranial atherosclerotic stenosis among patients with T2DM.

PMID:42769059 | PMC:PMC13590385 | DOI:10.3389/fneur.2026.1904036