Serum GFAP levels in optic neuropathies

BACKGROUND Complement mediated autoimmunity against aquaporin-4 results in astrocytic damage in neuromyelitis optica (NMO). There is evidence for increased CSF glial fibrillary acidic protein (GFAP) and S100B levels in acute NMO. Here we tested whether the CSF finding also holds true for the diagnostic value of serum GFAP and S100B levels in NMO. METHODS A multicentre study included 322 patients from London (n=160), Nijmegen (n=95), Pecs (n=44), and Lyon (n=24). Patients were classified into the following diagnostic categories: neurological control patients (n=45), MS optic neuritis (MSON, n=38), isolated optic neuritis (ION, n=11), relapsing isolated optic neuritis (RION, n=48), chronic relapsing isolated optic neuropathy (CRION, n=18), unclassified optic neuritis (UCON, n=39), NMO (n=77) and relapsing remitting multiple sclerosis (RRMS, n=47). Serum GFAP and S100B levels were quantified using ELISA. RESULTS Median serum GFAP but not S100B levels were significantly higher (p<0.0001, general linear model) in patients with NMO (4.83 pg/mL) if compared to MSON (1.5 pg/mL, p=0.0001), UCON (1.92 pg/mL, p<0.01), ION (0.0 ng/mL, p<0.05), RION (1.3 pg/mL, p<0.0001) and CRION (2.2 pg/mL, p=0.01). Serum GFAP levels in the control cohort (3.6 pg/mL) were not significantly different to NMO. There was no relationship between serum GFAP levels and any other clinical or demographic parameter. Serum S100B concentrations correlated with the number of relapses in MSON (R=0.83, p=0.005). CONCLUSION In contrast to the CSF, neither serum GFAP nor S100B levels were of major diagnostic value for the laboratory supported differential diagnosis between optic neuritis in the context of NMO and other optic neuropathies.