Serum GFAP was associated with progression independent of relapses in multiple sclerosis

Design

Prospective observational study in two independent cohorts

Population

2,329 people with MS, 18,629 paired NfL and GFAP measurements

Primary outcome

Progression independent of relapse activity (PIRA)

Effect

GFAP z >1: HR 1.45 (1.21 to 1.75) Swiss cohort; 1.36 (1.07 to 1.71) EPIC

A prospective observational study combined two MS cohorts, the Swiss MS Cohort (1,709 people) and the US EPIC study (620), with 18,629 paired measurements of serum neurofilament light (NfL) and glial fibrillary acidic protein (GFAP) over a median 7 and 13 years.

High NfL was associated with relapses in the next year; high GFAP with progression independent of relapse activity (PIRA). A GFAP z score above 1 was associated with a higher hazard of PIRA at the next interval (HR 1.45 in the Swiss cohort, 1.36 in EPIC). In the Swiss cohort, each yearly unit fall in GFAP z score during the first 2 years of fingolimod or B-cell depletion was associated with lower later PIRA risk (HR 0.46 and 0.33).

The two markers appear to track different processes: NfL for inflammatory activity, GFAP for smouldering progression. That makes GFAP a promising monitoring tool, but it is not yet validated for changing treatment in an individual patient.

• NfL and GFAP answer different questions: relapse activity versus progression
• Interpret any biomarker as an age-adjusted z score, not a raw value
• Keep EDSS, walk and hand-function tests as the basis for treatment decisions
• Watch for GFAP in future monitoring guidance, but do not order it routinely yet

Why it matters

Progression without relapses is the main unmet problem in MS, and it has had no blood marker to track it.

Don't overread it

This was observational; a falling GFAP on treatment is associated with, not proven to cause, lower progression risk.

The statistics, in plain English

A hazard ratio of 1.45 means the rate of progression events was about 45% higher over the next year when GFAP was high. The finding held in two independent cohorts, which is the strongest thing about it. The treatment-response results (HR 0.46 and 0.33) come from one cohort and have wider intervals, so they are less certain.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.