Serum GFAP Linked to Multiple Sclerosis Progression Independent of Relapse

Elevated serum glial fibrillary acidic protein (GFAP) was associated with a higher risk of progression independent of relapse activity (PIRA) in multiple sclerosis (MS), according to a study published in JAMA Neurology.¹ By enrolling only patients with high GFAP in clinical trials could cut required sample sizes by roughly one-fifth.

PIRA is disability worsening that occurs without relapses, and it significantly drives the long-term disability burden in MS. Disease-modifying therapies (DMTs) that suppress relapses and new lesions have shown limited effect on it. This creates a measurement problem for payers and drug developers. Phase 2 and 3 trials typically run 2 years or less, which the authors note is likely too short to capture PIRA as a clinical end point. Neurofilament light chain (NfL), a marker of neuroaxonal injury, is already established for tracking acute inflammatory activity. GFAP is released when astrocytes are activated or injured. It may reflect diffuse degenerative processes that NfL captures only partially.

The researchers analyzed 18,629 serum samples from 2329 people with MS in 2 cohorts. The Swiss MS Cohort (SMSC; n = 1709) was the main cohort. The University of California, San Francisco, EPIC study (n = 620) served as independent validation. Consistent with prior work, elevated NfL was associated with relapse risk within the next year, and GFAP was not. GFAP instead tracked progression. A GFAP z score above 1.0 was associated with a higher hazard of PIRA during the following visit interval, roughly 1 year. This held in both SMSC (HR, 1.45; 95% CI, 1.21-1.75) and EPIC (HR, 1.36; 95% CI, 1.07-1.71). Elevated NfL was not associated with short-term PIRA in SMSC.

GFAP Enrichment Could Reduce MS Clinical Trial Sample Sizes

The finding with the most direct cost implications came from a simulation of trial enrichment. Approximately 27.5% of SMSC participants had a GFAP z score above 1.0. By limiting enrollment to these patients reduced required sample sizes by 16.1% for time-to-first-PIRA analyses and 22.1% for recurrent-event analyses. A stricter threshold of 1.5 reduced them further. Enrichment raises event rates, so the authors suggested GFAP could make it feasible to show efficacy against progression within standard trial timelines.

GFAP may also be useful as a treatment-response marker. The researchers followed 241 patients starting fingolimod and 342 starting B-cell–depleting therapy. In each group, every yearly 1-unit decline in GFAP z score during the first 2 years of treatment was associated with a lower risk of later PIRA: 54% lower with fingolimod and 67% lower with B-cell–depleting therapy. NfL slope showed no association with later PIRA in either group.

Moving GFAP Testing From Research Labs to Clinical Practice

Wider use depends on whether GFAP can be measured reliably outside research labs. Most of the evidence, including the JAMA Neurology study, relies on the ultrasensitive Simoa platform. Simoa is approved for research use only and has had trouble with lot-to-lot consistency.

Recent study in Clinical Chemistry and Laboratory Medicine compared Simoa with the Roche Elecsys GFAP assay.² Elecsys runs on automated Cobas analyzers that are already widely used in clinical laboratories. Measurements from the 2 platforms correlated strongly (r = 0.94). Elecsys-derived z scores were associated with time to PIRA in both treatment cohorts, with HRs similar to those from Simoa. The investigators also updated their online GFAP z score calculator to accept Elecsys values.

Study Limitations

The study was observational. The 2 cohorts used different PIRA definitions: EPIC used a composite that also counted worsening on the 9-hole peg test and timed 25-ft walk. Findings may not generalize to more racially and ethnically diverse populations. Several authors reported ties to Roche and Novartis, which also helped fund the work.

References

1. Einsiedler M, Sandgren S, Schaedelin S, et al. Serum glial fibrillary acidic protein dynamics, disease progression, and therapy response in multiple sclerosis. JAMA Neurol. 2026;83(9):894-905. doi:10.1001/jamaneurol.2026.2500
2. Willemse EAJ, Sandgren S, Benkert P, et al. Serum GFAP as a biomarker for progression in multiple sclerosis: assay comparison and a large reference database of healthy controls. Clin Chem Lab Med. 2026;64(5):1046-1053. doi:10.1515/cclm-2025-1480