Serum Glial Fibrillary Acidic Protein Dynamics, Disease Progression, and Therapy Response in Multiple Sclerosis.

Importance Capturing individual multiple sclerosis (MS) progression is difficult; few studies have evaluated glial fibrillary acidic protein (GFAP) in large longitudinal cohorts with independent validation.Objective To investigate whether serum GFAP levels and treatment-related changes are associated with future progression independent of relapse activity (PIRA).Design, Setting, And Participants This was a prospective observational study using 2 large MS cohorts: the Swiss MS Cohort (SMSC; initiated in June 2012; data extraction September 22, 2025) and the Expression, Proteomics, Imaging, Clinical study (EPIC; initiated in July 2004; data extraction February 21, 2024). The study took place at tertiary MS centers (8 for SMSC and 1 for EPIC). A total of 2329 persons with MS from both cohorts with at least 1 available time point with neurofilament light chain (NfL) and GFAP measurements were included (overall 18 629 measurements).Exposures Clinical data and NfL and GFAP z scores, collected and calculated every 6 or 12 months.Main Outcomes And Measures Risk of future PIRA, defined as Expanded Disability Status Scale score worsening confirmed after 6 or more months without relapses (in SMSC), or a composite additionally including greater than 20% worsening in the 9-hole peg test or timed 25-ft walk test (in EPIC).Results The SMSC and EPIC cohorts consisted of 1709 (13 375 samples; median [IQR] follow-up, 6.9 [2.5-10.7] years and age, 40.6 [32.1-50.1] years; 1128 [66.0%] female) and 620 (5254 samples; median [IQR] follow-up, 13.1 [9.4-14.0] years and age, 42.0 [35.0-50.0] years; 432 [69.7%] female) persons with MS, respectively. Consistent with prior work, high NfL was associated with relapse risk within the next year, whereas high GFAP was associated with long-term PIRA risk. In addition, elevated GFAP (z score >1.0 [84th percentile]) was associated with an average 40% higher hazard of short-term PIRA in the subsequent visit interval (SMSC: median [IQR] 346 [190-375] days; hazard ratio [HR], 1.45, 95% CI, 1.21-1.75; P < .001; EPIC: 385 [355-518] days; HR, 1.36; 95% CI, 1.07-1.71; P = .01). GFAP-based cohort enrichment in clinical trials targeting PIRA as an end point could reduce sample size by approximately 20%. Further, in SMSC, every yearly GFAP z score unit reduction during the first 2 years receiving fingolimod or B-cell-depleting therapy was associated with a lower risk of subsequent PIRA (54% risk reduction; HR, 0.46; 95% CI, 0.26-0.84; P = .01 and 67% risk reduction; HR, 0.33; 95% CI, 0.18-0.61; P < .001), respectively.Conclusions And Relevance In this cohort study, elevated GFAP was associated with a higher risk of PIRA, while treatment-associated reductions were associated with a lower PIRA risk. Together, these results suggest that serum GFAP may serve as a biomarker for personalized risk stratification and treatment monitoring and as a screening tool to reduce cohort size in clinical trials targeting MS progression.