APOE2 is associated with longevity independent of Alzheimer’s disease

(a) Subject characteristics in the overall NACC cohort. The sample median (minimum, maximum) is given for continuous variables. Information was unavailable regarding hypertension (N = 28), transient ischemic attack (N = 116), pacemaker (N = 260), angioplasty/endarterectomy/stent (N = 19), heart attack/cardiac arrest (N = 42), cardiac bypass procedure (N = 9), atrial fibrillation (N = 48), hypercholesterolemia (N = 102), congestive heart failure (N = 25), and stroke (N = 45). (b) Subject characteristics for individuals in the NACC cohort with a neuropathological assessment. The sample median (minimum, maximum) is given for continuous variables. Information was unavailable regarding hypertension (N = 3), transient ischemic attack (N = 24), pacemaker (N = 1), heart attack/cardiac arrest (N = 4), atrial fibrillation (N = 3), hypercholesterolemia (N = 17), congestive heart failure (N = 4), stroke (N = 7),. density of neocortical neuritic plaques CERAD score (N = 12), density of diffuse plaques CERAD score (N = 328), Braak NFT stage (N = 55), presence of vascular pathology (N = 51), and minimal amyloid pathology. (c) Association between APOE genotype and survival using the NACC data. HR = hazard ratio; CI = confidence interval; AD = Alzheimer’s disease; CV = cardiovascular. HRs, 95% CIs, and p-values result from Cox proportional hazards regression models. ¹ These analyses were only performed when considering all subjects. ² Cognitive status at final visit and AD at final visit were adjusted for only in the model involving all subjects. ³ CV factors included hypertension, transient ischemic attack, pacemaker, angioplasty/endarterectomy/stent, heart attack/cardiac arrest, atrial fibrillation, hypercholesterolemia, congestive heart failure, and stroke. ⁴ For the neuropathological assessment cohort, all models were additionally adjusted for CERAD diffuse plaque score, CERAD neuritic plaque score, Braak NFT stage, and the presence of vascular pathology. p-values<0.025 are considered as statistically significant after applying a Bonferroni correction for multiple testing for the two primary comparisons of survival that were made (i.e. between APOE3 and APOE4 subjects, and between APOE3 and APOE2 subjects). (d) Summary of mouse cohorts. For continuous data, values are means ± SD [range]. P-values were calculated using one-way ANOVA (continuous data), the Pearson’s chi-square test (categorical value) or log-rank test (median age at death). (e) Effects of APOE on ‘Dropped activities and interests’ and other GDS items in cognitively normal subjects over 60 years old. ORs, 95% CIs, and p-values result from binary logistic regression models for analysis of individual GDS items, and from proportional odds logistic regression models for analysis of the ordinal total GDS score, which is described as the mean (25th percentile, 75th percentile). Information was unavailable for GDS items for a maximum of 23 subjects. (f) Associations between activity levels in the OFA and protein/lipid levels across APOE genotype groups at old age. Correlation coefficients (r) and p-values were calculated using the Pearson correlation test. P-values were corrected by Bonferroni test adjusted for the number of protein/lipids analyzed in the entire cohort. CX = cortical, HP = hippocampal. (g) List of NACC contributors.