Revisiting an overlooked biomarker: CSF-IgA demonstrates significant diagnostic value for viral encephalitis in a case-control study
TL;DR
A diagnostic model incorporating CSF-IgA showed good performance, even in atypical cases with normal routine CSF profiles, indicating its potential utility as an auxiliary tool for early VE diagnosis.
Abstract
Background Viral encephalitis (VE) is a serious central nervous system infection, yet its diagnosis is challenged by non-specific clinical manifestations, low pathogen detection rates, and often normal routine cerebrospinal fluid (CSF) findings. The diagnostic potential of CSF immunoglobulin A (CSF-IgA) has been largely overlooked. This study aimed to evaluate the diagnostic value of CSF-IgA for VE. Methods This single-center, retrospective case-control study analyzed data from 233 patients with VE and 80 symptomatic controls, collected over a 15.5-year period (January 1, 2010 to June 30 2025). A multivariable logistic regression model was constructed using variables that were significant in univariate analysis (p < 0.05). Model performance was rigorously evaluated in the overall cohort for discrimination (area under the receiver operating characteristic curve, AUC), calibration (using calibration plots and the Brier score), and clinical utility (via decision curve analysis). The model was further validated in a predefined subgroup of VE patients (n=50) with normal routine CSF findings, defined as normal CSF pressure, cell count, total protein, albumin, glucose, and chloride levels. Results The final parsimonious diagnostic model retained three robust predictors: CSF-IgA (OR 2.56, 95% CI: 1.73–4.06), CSF white blood cell count (OR 1.45, 95% CI: 1.25–1.76), and serum albumin (OR 0.81, 95% CI: 0.71–0.91). It demonstrated excellent discrimination in the overall cohort (AUC 0.933, 95% CI: 0.913–0.955), was well-calibrated (Brier score = 0.091), and provided superior net benefit across a wide threshold probability range. Importantly, robust performance was maintained in the clinically challenging subgroup with normal routine CSF parameters (AUC 0.812, 95% CI: 0.730–0.873; Brier score = 0.171). Conclusions CSF-IgA may serve as a promising adjunctive biomarker for viral encephalitis. A diagnostic model incorporating CSF-IgA showed good performance, even in atypical cases with normal routine CSF profiles, indicating its potential utility as an auxiliary tool for early VE diagnosis. However, given the single-center retrospective design, prospective external validation is warranted before routine clinical implementation.