Inverse relation between serum neurofilament light chain and cognitive function in chronic inflammatory demyelinating polyneuropathy

Abstract

Background: Chronic Inflammatory Demyelinating Polyneuropathy (CIDP) is a dysimmune disease primarily targeting the Schwann cell myelin sheath in the peripheral nervous system (PNS), resulting in sensorimotor deficits. Surprisingly, subtle cognitive impairments as well as axonal damage, indicated by elevated serum neurofilament light chain (sNfL), prevail in CIDP. This study investigated whether elevated sNfL is associated with lower cognitive performance in CIDP.

Methods: Thirty-five CIDP patients underwent digital cognitive testing across multiple domains, alongside assessments of sociodemographic, clinical, and sNfL measures. Patients were stratified into low- and high-sNfL groups based on the median value, and clinical variables were compared. Further, general linear models, controlled for clinical and sociodemographic factors, were employed to evaluate the predictive value of sNfL for global and domain-specific cognitive functioning.

Results: Higher sNfL values were associated with worse general cognitive performance (β = -0.31, p = 0.016) and reduced processing speed (β = -0.40, p = 0.008). Patients with increased sNfL levels had a longer disease duration (p = 0.016), also linked to poorer cognitive outcome (β = -0.26, p = 0.045).

Conclusions: In this CIDP cohort, high sNfL levels were associated with reduced cognitive performance and longer disease duration. The findings suggest that sNfL is a clinical meaningful biomarker for the detection and monitoring of central involvement in the course of CIDP, a condition traditionally viewed as purely peripheral.

Keywords: CNS manifestations; Chronic inflammatory demyelinating polyneuropathy; Cognition; Neurodegeneration; SNFL.

Fig. 1

Relationship between sNfL scores and A cognitive performance (expressed as z-scores) and B various clinical parameters. For all cognitive scores, the presented β and p values are derived from GLM models including age, depression, fatigue, disease duration, and motor symptoms as covariates. R-ODS Rasch-built overall disability scale, GLM generalized linear model, sNfL serum neurofilament light chain corrected for age and BMI, PIP posterior inclusion probability, BMA Bayesian model averaging

Fig. 2

Comparison of sNfL z-scores across DSM-5–based cognitive impairment categories in CIDP patients. Categories include no/mild/mild-to-moderate/moderate neurocognitive disorder. sNfL values are presented as age- and BMI-adjusted z-scores. Error bars represent the standard error. *p ≤ 0.05, based on one-way ANOVA with Tukey’s post hoc test