Development and Validation of a Tool to Predict Onset of Mild Cognitive Impairment and Alzheimer Dementia

The ability to predict the onset of mild cognitive impairment (MCI) and Alzheimer dementia (AD) could allow older adults and clinicians to make informed decisions about dementia care.

To assess whether the age at onset of MCI and AD can be predicted using a statistical modeling approach.

This prognostic study used data from 2 aging and dementia cohort studies-the Australian Imaging, Biomarker and Lifestyle (AIBL) study and the Alzheimer's Disease Neuroimaging Initiative (ADNI)-for model development and validation of the Florey Dementia Index (FDI), a tool used to predict the age at onset of MCI and AD in older adults. Data from the Anti-Amyloid Treatment in Asymptomatic Alzheimer (A4) study were used for a simulated trial. Data were collected from 1665 AIBL participants, 2029 ADNI participants, and 93 A4 participants from October 1, 2004, to March 1, 2023. The data analysis was conducted between January and August 2024.

Predicted age at onset compared with clinically observed age at onset.

Among the 1665 AIBL participants (741 [44.5%] female) and 2029 ADNI participants (925 [45.6%] female), the mean (SD) age at first evaluation was 71.8 (7.1) years and 74.5 (6.7) years, respectively. The FDI achieved mean absolute errors of 2.78 (95% CI, 2.63-2.93) years for predicting MCI onset and 1.48 (95% CI, 1.32-1.65) years for predicting AD onset. In the simulated trial with 93 A4 participants (48 [51.6%] female; mean [SD] age at baseline, 73.4 [5.1] years), the FDI achieved mean absolute errors of 1.57 (95% CI, 1.41-1.71) years for predicting MCI onset and 0.70 (95% CI, 0.53-0.88) years for predicting AD onset.

In this prognostic study, the FDI was developed and validated to predict the onset age of MCI and AD. This tool may be useful in organizing health care for older adults with cognitive decline or dementia and in the future may help prioritize patients for the use of disease-modifying monoclonal antibody drugs.