An analysis of the Swiss MS Cohort has found that routine monitoring of serum biomarkers can aid in predicting progression independent of relapse activity (PIRA) in patients during B cell-depleting therapy (Benkert et al. Ann Neurol 2025;97:104-115).

The longitudinal study examined levels of glial fibrillary acidic protein (GFAP) and serum neurofilament-light chain (sNfL) in treated MS patients (n=362) and healthy controls (n=2861). The MS phenotypes were RRMS (76%), SPMS (12%) and PPMS (12%). GFAP is a biomarker of astrogliosis and NfL is a marker of neuroaxonal damage; neither biomarker is specific to MS.

Serum samples collected after a median of one year after starting B cell-depleting therapy were stratified as High (sGFAP z-score >1) and Low (z-score <1). Elevated sGFAP z-scores were more common in progressive versus relapsing MS (41.4% vs. 24.8% of samples).

In the first three years of treatment, 29.5% of patients in the High group had a PIRA event compared to 14.5% in the Low group (hazard ratio 1.80). Age and EDSS score were also associated with a higher PIRA risk. PIRA risk was even higher in patients with elevated sGFAP + elevated sNfL (HR 2.66). High sGFAP levels were also predictive of progression independent of relapse and MRI activity (PIRMA) (HR 2.91).

Serum GFAP significantly increased over time by 0.49 z-score units/10 years during treatment. sNfL levels decreased by 0.92 z-score units/10 years in patients without PIRA but remained elevated in those with PIRA. It should be noted that in healthy controls, sGFAP levels increased by 2.0%/year, suggesting some of the change may be attributable to aging.

The authors noted that the gradual increase of sGFAP during treatment indicate that B cell depletion may have only a limited impact on progressive disease biology. There were similar findings in natalizumab studies (Jiang X, et al. Mult Scler 2023;29:1070-1079. Wessels et al. Mult Scler 2023;29:1229-1239). In contrast, there was a modest decrease in sGFAP in one study of siponimod (Kuhle et al. Neurology 2020; 94(15 suppl):1782).

A further observation was that sNfL levels were decreased during treatment but remained elevated in patients experiencing disease progression. This may suggest that there is an aspect of neuroaxonal damage that is part of progressive biology that may not be responsive to treatment.

The same group previously reported that sGFAP was a better predictor of disease progression than sNfL in a similar database analysis (Meier et al. JAMA Neurol 2023;80:287-297).