Method comparison of plasma and CSF GFAP immunoassays across multiple platforms.

Objectives: Glial fibrillary acidic protein (GFAP) is a biomarker of astrocytic activation associated with neurodegenerative diseases, neuroinflammatory disorders, and traumatic brain injury. However, the lack of standardized methods for quantifying GFAP across different immunoassay platforms poses challenges for its clinical implementation. This study aimed to compare the analytical performance of multiple commercially available and in-house immunoassays for GFAP quantification in plasma and cerebrospinal fluid (CSF) to assess their agreement and potential interchangeability.

Methods: We conducted a method comparison using four plasma GFAP immunoassays (Simoa, Ella, Alinity, and MSD) and four CSF GFAP assays (ELISA, Ella, Alinity and MSD). Anonymized leftover plasma and CSF samples were analyzed across platforms. Sample sizes for the pairwise comparisons ranged from 23 to 52 for plasma and 34 to 51 for CSF. Pairwise comparisons were performed using Spearman correlation, Bland-Altman analysis, and Passing-Bablok regression to assess systematic and proportional biases. Outliers were identified and excluded to ensure robust statistical evaluation.

Results: Strong correlations were observed across all platforms (Spearman's r=0.827-0.927 for plasma; r=0.937-0.958 for CSF). However, significant systematic and proportional biases were present in several comparisons, preventing direct interchangeability of results. In plasma, Simoa consistently reported higher GFAP concentrations compared with Ella and Alinity, while Alinity overestimated levels relative to Ella. Similarly, in CSF, ELISA tended to underestimate GFAP concentrations compared with Alinity, MSD, and Ella, with the largest discrepancy observed between ELISA and MSD.

Conclusions: Despite strong correlations, substantial method-dependent biases indicate that GFAP measurements across different immunoassay platforms need to be standardized to ensure harmonization and reliable clinical application of GFAP as a biomarker.