Alzheimer disease prevention: what works, what doesn’t?

  • Yu JT & al.
  • J Neurol Neurosurg Psychiatry
  • 20 Jul 2020

  • curated by Susan London
  • Clinical Essentials
Access to the full content of this site is available only to registered healthcare professionals. Access to the full content of this site is available only to registered healthcare professionals.

Takeaway

Why this matters

  • Effective interventions for prevention and treatment are lacking.

Key results

  • A graphic showing which factors might offer protection vs harm at different lifespan points is given in figure 5 in the paper.
  • 19 factors were class I (intervention strongly recommended):
    • 10 had level A (strong) evidence:
      • Education.
      • Cognitive activity.
      • Late-life increased BMI.
      • Hyperhomocysteinemia.
      • Depression.
      • Stress.
      • Diabetes.
      • Head trauma.
      • Midlife hypertension.
      • Orthostatic hypotension.
    • 9 had level B (intermediate) evidence:
      • Midlife obesity.
      • Late-life weight loss.
      • Physical exercise.
      • Smoking.
      • Sleep.
      • Cerebrovascular disease.
      • Frailty.
      • Atrial fibrillation.
      • Vitamin C.
  • 2 interventions were class III (intervention not recommended):
    • Estrogen replacement therapy (level A2 evidence).
    • Acetylcholinesterase inhibitors (level B evidence).
  • 6 factors had level C (weak) evidence, prompting recommendation for further study:
    • Diastolic BP management.
    • NSAID use.
    • Social activity.
    • Osteoporosis.
    • Pesticide exposure.
    • Silicon from drinking water.

Study design

  • Systematic review, meta-analysis of 243 prospective observational cohort studies and 153 randomized controlled trials.
  • Meta-analysis included 134 risk factors and 11 interventions.
  • Main outcome: Alzheimer disease.
  • Funding: National Key R&D Program of China; Shanghai Municipal Science and Technology Major Project; Zhangjiang Laboratory.

Limitations

  • Residual and unmeasured confounders likely at play.
  • Causality was not established.
  • Generalizability is uncertain.