Plasma glial fibrillary acidic protein (GFAP) is a biomarker for central nervous system involvement in infantile-onset Pompe disease

BACKGROUND: Pompe disease (PD), caused by acid α-glucosidase (GAA) deficiency, leads to glycogen accumulation in various tissues including the central nervous system. While enzyme replacement therapy (ERT) is lifesaving, it does not cross the blood-brain barrier. CNS manifestations including sensorineural hearing loss, dysarthria, and cognitive delay-persist despite ERT. Hence, CNS-specific biomarkers are needed to identify at-risk patients and guide monitoring and treatment. This clinical, retrospective study evaluated the utility of plasma glial fibrillary acidic protein (GFAP) as a potential biomarker of CNS involvement in infantile-onset Pompe disease (IOPD) compared to plasma neurofilament light chain (NfL). METHODS: Plasma GFAP and NfL levels were measured longitudinally in 37 patients with PD (180 samples, ages 0.1-21 years) and 54 age- and sex-matched controls. Patients were grouped as (1) IOPD with severe neurologic involvement (n = 7), (2) IOPD with attenuated neurologic involvement (n = 13), and (3) late-onset PD without neurologic involvement (LOPD; n = 17). Neurologic status was determined via clinical examination and/or brain MRI white matter grading via Modified Fazekas Scores (MFS). FINDINGS: GFAP levels were highest in IOPD patients with severe neurologic involvement. Patients with IOPD in the attenuated neurologic involvement group had, on average, lower plasma GFAP concentrations than those with severe neurologic involvement but greater GFAP levels than patients with LOPD, which resembled controls. GFAP outperformed NfL in differentiating patients with IOPD from controls (AUC = 0.886 vs 0.705) and identifying severe from attenuated neurologic involvement group (AUC = 0.801 vs 0.745). NfL showed high variability between controls and PD subgroups. INTERPRETATION: Plasma GFAP levels reliably reflect CNS disease burden in IOPD and outperform NfL in diagnostic performance. GFAP may be a useful biomarker for detecting and monitoring CNS involvement in PD. FUNDING: Funding was provided in part by Sanofi (Cambridge, Massachusetts, USA). Philanthropic support was provided in part by Judy and Monty Frost, Abigail and JB Spaulding, AG Cox Charity Trust and Macie's Mission.