Amyloid‐ß aggregation, APOE‐e genotype, and the trajectory of cognitive decline in persons without dementia.

Background: The association of amyloid abnormality with cognitive decline in persons with normal cognition (NC) and mild cognitive impairment (MCI) is important to guide AD trial design and understand prognoses. We examined the relation between amyloid abnormality and decline over time in several cognitive domains, and tested whether apolipoprotein (APOE)‐e genotype modified this association. Method: We included 3,820 persons (2,599 NC, 1,221 all‐cause MCI) from 22 Amyloid Biomarker Study cohorts, with baseline PET (n = 846) or CSF (n = 2,974) amyloid measurements available. Amyloid abnormality on PET was based on visual read or cohort‐specific cutoffs, and for CSF on cohort‐specific or data‐driven cutoffs. Global cognition was assessed by Mini‐Mental State Examination (MMSE), functional performance by Clinical Dementia Rating global scale (CDR), and memory immediate and delayed recall by verbal learning tests. We examined baseline amyloid abnormality associations with cognitive performance over time and potential modification by APOE‐e (up to 3‐way interactions) using random‐slope/random‐intercept linear mixed‐models corrected for age, sex, education with random cohort effects. Result: Mean follow‐up duration was 2.1 (range 0.5‐8) years. Baseline MMSE was worse in persons with abnormal‐amyloid compared with normal‐amyloid (NC mean difference = 0.2, p = 0.04, MCI mean difference = 2.4, p<0.001). In NC, amyloid abnormality was not associated with decline in any domain (Table 1), nor was the association influenced by APOE‐e4. In MCI, amyloid abnormality was associated with lower baseline performance and steeper decline on all domains (Table 1). In APOE‐e4‐carriers with MCI, MMSE decline was steeper than in non‐carriers (p<0.001, 1.5x steeper in amyloid‐abnormal, 3x steeper in amyloid‐normal). The rate of decline on immediate/delayed recall in MCI was higher in APOE‐e4‐carriers than non‐carriers, independent of amyloid status (p<0.001). For both NC and MCI, decline was steepest in persons with e4e4, followed by e3e4, e3e3, e2e4, and e2e3 (p = 0.001, Figure 1), independent of amyloid status. Conclusion: Amyloid abnormality was not associated with cognitive decline in persons with NC after a short‐term follow‐up. In MCI, amyloid abnormality was closely associated with cognitive decline in several domains, with an additive effect of APOE‐e on decline. These results may be relevant for future trial recruitment strategies. Future studies should examine long‐term outcomes in NC. [ABSTRACT FROM AUTHOR]