Dementia Does Not Begin with a Diagnosis. It Begins Years Earlier.

By the time a patient receives a diagnosis of dementia, the underlying biology has been active for years, sometimes decades.

A longitudinal cohort study by Ward and colleagues from the University of Queensland, published in JAMA Neurology in 2025, analysed 29,849 participants from four prospective cohort studies. Frailty trajectories accelerated 4 to 9 years before incident dementia. Frailty was positively associated with dementia risk across all cohorts, with adjusted hazard ratios from 1.18 to 1.73. Physical and biological vulnerability signals appear long before a memory clinic visit.

A meta-analysis by Malek-Ahmadi and colleagues from Banner Alzheimer's Institute, published in JAMA Neurology in 2026, synthesised data from 7,834 participants and found plasma p-tau217 achieved an AUC of 0.87 in identifying amyloid-positive cognitively unimpaired individuals. The authors concluded that these findings support the clinical utility of p-tau217 as a scalable, minimally invasive tool for early identification.

An earlier diagnostic accuracy study by Ashton and colleagues from the University of Gothenburg, published in JAMA Neurology in 2024, found that the p-tau217 immunoassay identified elevated amyloid and tau pathology with AUC values of 0.92 to 0.96, and reduced confirmatory testing by approximately 80%.

The DETeCD-ADRD clinical practice guideline by Atri and colleagues, published in 2025, reviewed 7,374 publications to empower primary care clinicians to conduct structured cognitive evaluations outside specialist settings. ReCOGnAIze was designed precisely for this context. If you are over 50 with one or more vascular risk factors and have not had a cognitive screening, speak to your GP. The window to act is open.