Alzheimer's disease has consumed more failed clinical trials than any other condition in modern medicine. Since 2002, over 99% of Alzheimer's drug candidates have failed in phase II or III — nearly all of them targeting amyloid-beta plaques. The FDA's controversial approval of lecanemab and donanemab in 2023–24, despite modest clinical benefit, marks the peak of a decades-long bet on the amyloid hypothesis. Yet amyloid clearance alone produces little cognitive improvement in most patients. Something upstream is being missed.

Two neglected mechanisms stand out. First, the glymphatic system — a brain-wide waste clearance network that operates almost exclusively during slow-wave sleep — is responsible for flushing amyloid-beta and tau from interstitial spaces. Studies show that even one night of sleep deprivation causes a 50% increase in amyloid accumulation in cerebrospinal fluid. Chronic sleep fragmentation, endemic in modern populations, creates a slow drip of amyloid that the brain cannot clear fast enough. Lateral (side) sleep position has been shown to optimise glymphatic flow compared to supine or prone positions. This is not a drug target — it is a lifestyle variable that clinical trials routinely fail to control for, let alone treat.

Second, emerging evidence from post-mortem brain studies and PET imaging consistently shows that Alzheimer's brains exhibit the same hallmarks of insulin resistance seen in type 2 diabetes — reduced glucose uptake, impaired insulin receptor signalling, and dysregulated tau phosphorylation downstream of PI3K/Akt pathway dysfunction. APOE4 carriers, who face a 3–12× increased Alzheimer's risk, show accelerated cerebral glucose hypometabolism decades before symptom onset. Ketones provide an alternative fuel source that bypasses this impairment — which partly explains why medium-chain triglyceride (MCT) supplementation and ketogenic diets show promising results in early cognitive decline. Metformin, an AMPK activator widely used for type 2 diabetes, has demonstrated neuroprotective effects in observational studies and is now in randomised trials (TAME study) for age-related disease.

The hypothesis: a combined dual-axis protocol could achieve meaningful prevention and early-stage reversal in high-risk individuals, particularly APOE4 carriers. Protocol: (1) Sleep axis — structured sleep hygiene targeting 7–9 hours with lateral sleep position, combined with short-duration morning bright light therapy to normalise circadian rhythm; (2) Metabolic axis — metformin 500mg/day (or MCT oil 20–30ml/day for those avoiding pharmacotherapy) plus time-restricted eating (16:8). Both arms are safe, inexpensive, accessible in LMIC settings, and could be trialled in a 12-month RCT in APOE4-positive adults aged 50–65, using FDG-PET glucose metabolism and CSF amyloid/tau ratios as primary endpoints. The prediction: combined intervention reduces amyloid accumulation rate by ≥30% and improves cerebral glucose uptake measurably within 12 months — a signal no single-target drug trial has yet achieved cleanly.

The broader implication is that Alzheimer's may be less a protein aggregation disease than a metabolic and clearance disease with protein aggregation as its visible symptom. Treating the plaques without fixing the upstream drivers is like mopping the floor while the tap runs. The interventions already exist. The trial design is straightforward. What's missing is the will to test a hypothesis that doesn't require a billion-dollar drug.