Glial fibrillary acidic protein in progressive supranuclear palsy: A systematic review and meta-analysis

Astrocyte pathology is a defining feature of progressive supranuclear palsy (PSP), yet the relevance of fluid glial fibrillary acidic protein (GFAP) remains uncertain.

To determine whether mean GFAP concentrations differ between PSP and healthy controls or other parkinsonian disorders, particularly Parkinson's disease (PD) and multiple system atrophy (MSA).

PubMed and Scopus were searched from inception to 24 January 2026. Original human studies that measured GFAP in plasma, serum, or cerebrospinal fluid (CSF) and included a PSP group were eligible. Standardized mean differences were pooled using random-effects meta-analysis and interpreted as group-level separation rather than individual diagnostic accuracy.

Of 1917 records, 1590 were screened after deduplication; 19 studies were included qualitatively and 12 quantitatively. Seven blood studies contributed to the primary PSP versus healthy-control analysis (PSP n = 283; controls n = 312). Blood GFAP was moderately higher in PSP (Hedges' g = 0.63, 95% CI 0.30 to 0.97; p = 0.003; I² = 53.4%). The estimate persisted in leave-one-out and in a sensitivity analysis adding three studies with median-derived means (k = 10; g = 0.72, 95% CI 0.36 to 1.08). CSF GFAP was directionally concordant. Group-level separation between PSP and PD or MSA was limited and inconsistent.

Mean blood GFAP concentrations are higher in clinically diagnosed PSP than in healthy controls, although the biological basis and disease specificity of this finding remain uncertain. However, these analyses do not establish sensitivity, specificity, or individual-level diagnostic performance, and GFAP should not be considered a stand-alone diagnostic marker.