Plasma GFAP and Amyloid Pathology Predict Cognitive Response to Multidomain Interventions in MCI

There is limited evidence on which biological markers can predict the effectiveness of interventions in mild cognitive impairment (MCI) patients, particularly in relation to amyloid pathology. This study aims to investigate whether plasma glial fibrillary acidic protein (GFAP) levels can serve as a predictive biomarker for short-term cognitive response to multidomain interventions in elderly individuals with MCI, stratified by probable amyloid-β plaque deposition. In this 24-week multicenter randomized controlled trial (RCT; SUPERBRAIN-MEET), 300 elderly participants with MCI were enrolled. Probable amyloid status was determined using a plasma phosphorylated tau 181 cutoffs derived from a validated cohort. Multivariable linear regression analyses were employed to assess the association between plasma GFAP levels and percentage changes in Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) scores, stratified by amyloid deposition status and intervention group. Higher plasma GFAP levels at baseline and 6 months were independently and significantly associated with smaller percentage improvements in RBANS scores over 6 months. Among participants with probable amyloid positivity who underwent the multidomain intervention, increased baseline GFAP levels were significantly associated with reduced cognitive improvement compared to those with lower levels (β = -8.661, p = 0.040). This post hoc exploratory subanalysis, based on data from a multicenter RCT, suggests that plasma GFAP may serve as a biomarker for early cognitive stage transitions in elderly individuals with MCI. Baseline GFAP levels-particularly among those with probable amyloid pathology-may help predict cognitive responsiveness to multidomain interventions.