Predicting and Preventing Alzheimer's Disease (erictopol.substack.com)

Researchers and commentators have updated a roadmap for predicting Alzheimer’s years—often decades—before clinical symptoms appear, leveraging biomarkers, aging “clocks,” and AI. Key advances include highly predictive blood tests for phosphorylated tau at threonine‑217 (p‑tau217), which rises early in the disease course and matches or exceeds cerebrospinal‑fluid assays and PET tau scans. One FDA‑cleared Fujirebio blood test (p‑tau217 + Aβ42) showed ~92% positive and ~97% negative concordance with PET/CSF. Large proteomic screens (SomaScan of ~7,000 plasma proteins in ~1,300 AD and ~2,100 controls) with machine learning and smaller cohort studies (e.g., 400 middle‑aged, cognitively intact participants) validated p‑tau217 and p‑tau217/Aβ42 ratios. Complementary predictors include DNA‑methylation and multi‑organ “aging clocks,” polygenic risk scores, and AI analyses of retinal imaging; combining these modalities into multimodal AI models can both flag high risk and project a likely timeline of progression across the ~20‑year preclinical window. The clinical implication is a shift from late-stage treatment to proactive prevention. Interventions shown or associated with reduced biomarker levels and dementia risk include lifestyle modifications (exercise, Mediterranean‑like diet, sleep optimization), vascular risk control, vaccinations (observational signals for shingles and influenza), and emerging therapeutics—GLP‑1 agonists currently in randomized trials and brain‑penetrant drugs targeting inflammation (e.g., NLRP3). Ongoing work on the gut microbiome and multi‑marker surveillance strategies aims to enable aggressive, individualized prevention for those identified at high risk.