Amyloid-beta and APOE genotype predict memory decline in cognitively unimpaired older individuals independently of Alzheimer's disease polygenic risk score

Background: What combination of risk factors for Alzheimer's disease (AD) are most predictive of cognitive decline in cognitively unimpaired individuals remains largely unclear. We studied associations between APOE genotype, AD-Polygenic Risk Scores (AD-PRS), amyloid-beta pathology and decline in cognitive functioning over time in a large sample of cognitively unimpaired older individuals. Methods: We included 276 cognitively unimpaired older individuals (75 +/- 10 years, 63% female) from the EMIF-AD PreclinAD cohort. An AD-PRS was calculated including 83 genome-wide significant variants. The APOE gene was not included in the PRS and was analyzed separately. Baseline amyloid-beta status was assessed by visual read of [F-18]flutemetamol-PET standardized uptake value images. At baseline and follow-up (2.0 & PLUSMN; 0.4 years), the cognitive domains of memory, attention, executive function, and language were measured. We used generalized estimating equations corrected for age, sex and center to examine associations between APOE genotype and AD-PRS with amyloid-beta status. Linear mixed models corrected for age, sex, center and education were used to examine associations between APOE genotype, AD-PRS and amyloid-beta status, and their interaction on changes in cognitive functioning over time. Results: Fifty-two participants (19%) had abnormal amyloid-beta, and 84 participants (31%) carried at least one APOE epsilon 4 allele. APOE genotype and AD-PRS were both associated with abnormal amyloid-beta status. Increasingly more risk-full APOE genotype, a high AD-PRS and an abnormal amyloid-beta status were associated with steeper decline in memory functioning in separate models (all p & LE; 0.02). A model including 4-way interaction term (APOExAD-PRSxamyloid-beta xtime) was not significant. When modelled together, both APOE genotype and AD-PRS predicted steeper decline in memory functioning (APOE beta(SE)=-0.05(0.02); AD-PRS beta(SE)=-0.04(0.01)). Additionally, when modelled together, both amyloid-beta status and AD-PRS predicted a steeper decline in memory functioning (amyloid-beta beta(SE)=-0.07(0.04); AD-PRS beta(SE)=-0.04(0.01)). Modelling both APOE genotype and amyloid-beta status, we observed an interaction, in which APOE genotype was related to steeper decline in memory and language functioning in amyloid-beta abnormal individuals only (beta(SE)=-0.13(0.06); beta(SE)=-0.22(0.07), respectively). Conclusion: Our results suggest that APOE genotype is related to steeper decline in memory and language functioning in individuals with abnormal amyloid-beta only. Furthermore, independent of amyloid-beta status other genetic risk variants contribute to memory decline in initially cognitively unimpaired older individuals.