Serum interleukin-6 as a neuroinflammatory biomarker across the spectrum of neurological disorders: a large-scale retrospective cohort study of 6,465 individuals.

BACKGROUND: Neurological disorders have overlapping clinical manifestations, creating an urgent need for accessible biomarkers to aid differential diagnosis. Interleukin-6 (IL-6) is a core neuroinflammatory mediator, yet its expression profile across the full spectrum of neurological disorders remains poorly characterized in large-scale cohorts using a uniform detection platform. We aimed to map serum IL-6 levels across 10 categories of neurological disorders, with healthy individuals as controls, identify its independent predictors, and evaluate its multi-scenario diagnostic performance.

METHODS: We conducted a retrospective observational non-interventional cohort study at Renmin Hospital of Wuhan University (a tertiary academic medical center in Wuhan, China) enrolling 6,465 individuals (515 healthy controls, 5,950 patients across 10 neurological disease categories) who underwent serum IL-6 testing via cytometric bead array (CBA) between January 2018 and September 2025. Statistical analyses included Kruskal-Wallis test, multivariable linear regression, receiver operating characteristic (ROC) curve analysis, and decision curve analysis (DCA). Nested logistic regression models were constructed to evaluate the incremental diagnostic value of IL-6 beyond age and sex.

RESULTS: Serum IL-6 levels showed significant heterogeneity across disease categories (p < 0.001). The highest elevations were observed in traumatic brain injury (TBI), metabolic/toxic encephalopathy, and hemorrhagic stroke (median: 59.93, 28.41, 24.09 pg/mL), with only mild increases in neurodegenerative diseases and intracranial tumors. Disease category, age (U-shaped association), and male sex were independent predictors of IL-6 levels (all p < 0.001). IL-6 achieved good diagnostic performance for distinguishing TBI (AUC 0.974), metabolic/toxic encephalopathy (AUC 0.940), and hemorrhagic stroke (AUC 0.925) from healthy status; notably, IL-6 showed limited discriminatory power for differential diagnosis between TBI and other neurological diseases (AUC = 0.741). Incorporating age and sex significantly improved discrimination for CNS infections (ΔAUC = 0.129, p < 0.0001), with the age+sex-only model achieving an AUC of 0.785 (95% CI 0.75-0.82).

CONCLUSION: In this large-scale cohort study, we demonstrate that serum IL-6 is a neuroinflammatory biomarker with robust efficacy for screening acute brain injury from healthy status. Its clinical interpretation requires age- and sex-stratified reference intervals, and it holds potential as a preliminary screening indicator for acute neurological conditions in specific clinical settings. The diagnostic utility of IL-6 in unselected emergency triage settings, as well as the proposed clinical interpretation framework, require prospective validation in multi-center, real-world cohorts.