Lower mortality risk in APOE4 carriers with normal cognitive ageing

Figures

• 

  Figure 1. The cumulative incidence function (CIF) plot for the competing risk groups is shown for APOE4 carriers in panel (a), and non-carriers in panel (b). The cumulative incidence (as a proportion) is on the y-axis and the age at death (in years) is on the x-axis. APOE4 carriers who died with high amounts of Alzheimer’s lesions (DeadHighADnp APOE4 carriers, red solid line) had the highest incidence of death, followed by non-carriers (DeadHighADnp non-carriers and DeadLowADnp non-carriers, red and blue dashed lines, respectively). The lowest incidence of death was among APOE4 carriers who died with low amounts of AD neuropathology (DeadLowADnp APOE4 carriers, blue solid line). The 95% confidence intervals (shaded areas) of the CIF—calculated by the Fine and Gray method—show no overlap in the cumulative incidences of death for the competing risks of death among APOE4 carriers and non-carriers.

• 

  Figure 2. Forest plots of univariate Fine and Gray competing risk survival analyses are shown in the 17 Alzheimer’s Disease Research Centers (ADRC) that had at least 165 autopsied people (representing > 80% of the autopsy sample). The colored horizontal bars represent the estimated 95% confidence interval of the subdistribution hazard ratio (SHR) for APOE4 on mortality risk, with the box in the middle at the estimated SHR scaled to the size of the sub-sample. The dashed vertical line is at exp(β = 0) = 1, with decreasing mortality risk on the left and increasing mortality risk on the right. The SHR estimates for the competing risk group of death with low amounts of AD neuropathology (DeadLowADnp, shown in blue) are all to the left of the dashed vertical line, indicating decreased mortality risk was associated with APOE4 for this competing risk group in all 17 ADRC. The SHR estimates for the competing risk group of death with high amounts of AD neuropathology (DeadHighADnp, shown in red) are all to the right of the dashed vertical line, indicating increased mortality risk was associated with APOE4 for this competing risk group in all 17 ADRC. The SHR for APOE4 on mortality risk from the…

• 

  Figure 3. Forest plots of adjusted hazard ratios from multivariable Fine and Gray competing risk survival analyses within each stratum as stratified by baseline median age for autopsied subjects in panel (a), sex in panel (b), baseline cognitive impairment in panel (c), race in panel (d), and ethnicity in panel (e). Each multivariate model included the main explanatory variable APOE4 and all covariates apart from the stratification variable. The models stratified by baseline median age, sex, and baseline cognitive impairment also included significant two-way interactions. The colored horizontal bars represent the estimated 95% confidence interval of the subdistribution hazard ratio (SHR) for APOE4 on mortality risk, with the box in the middle at the estimated SHR scaled to the size of the sub-sample in each stratum. The dashed vertical line is at exp(β = 0) = 1, with decreasing mortality risk on the left and increasing mortality risk on the right. The SHR estimates for the competing risk group of death with low amounts of AD neuropathology (DeadLowADnp, shown in blue) are all to the left of the dashed vertical line, indicating decreased mortality risk was associated with APOE4 for this…