We aimed to determine the amyloid (Aβ) and tau biomarker levels associated with imminent Alzheimer’s disease (AD) - related metabolic decline in cognitively normal individuals.

A threshold analysis was performed in 120 cognitively normal elderly individuals by modelling 2-year declines in brain glucose metabolism measured with [¹⁸F]fluorodeoxyglucose ([¹⁸F]FDG) as a function of [¹⁸F]florbetapir Aβ positron emission tomography (PET) and cerebrospinal fluid phosphorylated tau biomarker thresholds. Additionally, using a novel voxel-wise analytical framework, we determined the sample sizes needed to test an estimated 25% drugeffect with 80% of power on changes in FDG uptake over 2 years at every brain voxel.

The combination of [¹⁸F]florbetapir standardized uptake value ratios and phosphorylated-tau levels more than one standard deviation higher than their respective thresholds for biomarker abnormality was the best predictor of metabolic decline in individuals with preclinical AD. We also found that a clinical trial using these thresholds would require as few as 100 individuals to test a 25% drug effect on AD-related metabolic decline over 2 years.

These results highlight the new concept that combined Aβ and tau thresholds can predict imminent neurodegeneration as an alternative framework with a high statistical power for testing the effect of disease-modifying therapies on [¹⁸F]FDG uptake decline over a typical 2-year clinical trial period in individuals with preclinical AD.