3.1 The UK Biobank aging population
From the entire UK BioBank cohort of more than 500,000 individuals, 252,327 who were aged 55 or over at recruitment (baseline) had a determined APOE genotype and were selected for our investigations (Table 1). Of these 14,523 (4.717%) had data available from their first follow-up visit and 2,677 (0.87%) had data available from baseline, first and second follow-up visits.
We found no differences in the population distribution in the four main classes (defined by sex and APOE genotypes), nor where there any difference when stratified by ethnicity.
A comparison of females (n=136,665) and males (n=115,662) revealed that the two groups differ in terms of age, education level, cognitive measures, disease diagnoses, and statin use, but not for the Townsend deprivation index or the incidence of Alzheimer Diseases.
The data illustrates that in the selected population, males are older, generally have a higher level of education, better cognitive scores, higher incidence of all the cardiovascular and metabolic diseases included in the analysis, and they have a higher rate of statins use. We further compared between males and females stratified by APOE genotypes (carriers vs. non-carriers of the APOE4 allele). In both females and males, statistically significant differences were found for several disease diagnoses, including Alzheimer and Dementia, and in use of different statins, excluding pravastatin in males. As for cognitive measures at baseline, only RT was found statistically significant different in both females and males when comparing between APOE4 carriers and non-carriers.
3.2 Drug exposure in the ageing population
To assess drug exposure in the aging population, datasets matched via Propensity Score (PS) for each drug were created (Table 2). We observed significant differences in drug exposure between females and males, as also depicted in Figure 2. Females are less likely to be treated with antidepressants, asthma, diabetes drugs, and non-statins lipid lowering drugs; they are more likely to be treated with non-steroidal anti-inflammatory drugs and Omega 3.
We were specifically interested in examining statin exposure differences, and so applied a logistic regression model and conditional inference tree to assess the exposure in the matched dataset (Supplementary Table 1) on the basis of features not included in the propensity score analyses (i.e. sex, AD and dementia diagnoses, and APOE4 genotype – indicated in Supplementary Table 1 as non-matched).
Based on the regression model, males (z-value=51.2, p-value < 2e-16) and participants with an APOE4 positive genotype (z-value=10.6, p-value < 2e-16), have a higher probability of being treated with statins. Interestingly, in this population, participants with an AD diagnosis were slightly less likely to be treated with statins (z-value=-3.0, p-value = 0.00246). Models output are reported in Supplementary Table 2.
Taking a second approach to better illustrate statin exposure differences by stratifiers, we apply recursive-partitioning and present the results in the form of a logical tree structures (Figure 3). Treatment with statins is stratified on the basis of sex, APOE genotype and degenerative diseases. However, the model suggests that treatment is stratified on the basis of APOE genotype in males (nodes 14 and 15), but not in female participants. Tree models also provide lists of rules, which summarize the branch path to each final node and its predicted probability. Within our model, the rule associated with the lowest probability of being treated (0.21) is the one including females without a diagnosis of AD (node 4); while the one with the highest probability of being treated (0.63) is that which includes males diagnosed with AD or dementia and who have an APOE genotype (node 31).
3.3 Effects of exposure to statins
To examine the effect of statin use on survival in the aging population, death records captured by UK BioBank were used for this analysis. We performed the following analysis on the dataset matched on the statin propensity score, thus including as covariates sex, APOE genotype, Alzheimer’s, and dementia diagnoses, as well as their interactions with statin treatment.
The matched data set included a total of 6622 death events (3170 in statin users and 3452 in non-users). The multivariate cox regression analysis (Table 3) revealed that use of statins was not significantly associated with overall higher rates of survival (p-value = 0.206). On the other hand, when considering the interaction of statin use with sex, the results suggest higher survival rates in males treated with statins.
As suggested by our analyses of statins exposure, individuals differ in prevalence of statin use on the basis of strata defined by sex and APOE genotype. Here we examined whether differences in use of statins have an effect on changes in RT, where higher RT scores indicate worse cognitive function.
To assess changes in cognitive patterns, as measured by RT, in relation to statin use, we included individuals who had at least two measurements (from 2 visits) following baseline assessment. The average length of time (days) between baseline and first follow-up was 1565.64 ±343.2, and 962.66 ±288.6 between first and second follow-up visits, respectively. A total of 3,877 individuals from the matched cohort had available RT measures (milliseconds) data at least two visits (Supplementary Figure 1)
A linear mixed effects model was used to test for differences in the rate of change in the RT measures over the entire follow-up period (3 time points) in the statins matched dataset. The model includes a random effect term indicating variation over time in each subject (Time from baseline | Subject), and adjusted for sex, APOE genotype and their interactions with statins treatments (Table 4). Changes in reaction time measures were significantly associated with time from baseline (scores worsened in time, as previously described (25)) as well as sex; males had worse performance over time. Statistically significant differences (p=0.03) were found in the change in reaction time between statin users and non-users when stratified by APOE genotype.
Figure 4A illustrates RT scores at the each of the available time points (0=Baseline, 1=First Visit, 2=Second Visit) in statin users (red) and non-users (grey) in each of the strata suggested by the model (male and female, and APOE4 carriers and non-carriers). As suggested by the mixed effect model (Table 4), significant differences are observed only when the interaction between treatment and APOE4 genotype is considered. In general, statin users have worst RT score during the whole observation period, but these differences are reduced in APOE4 carriers. Specifically, in male APOE4 carriers, statin users and non-users demonstrate substantial overlap of RT scores in time and, while not statistically significant, they are the only strata where RT is higher in non-users (mean=6.33, sd=0.1) than in user (mean=6.32, sd=0.1) at baseline.
We tested the differences in RT Slope.yrs between statin users and non-users in each of the strata (see Figure 4B). Larger slopes indicate faster deterioration of cognitive function in time. No significant differences were seen. However, as already suggested by Figure 4B, different behaviors in deterioration can be seen: in females with an APOE4 carrier genotype, statin non-users deteriorate faster (mean RT slope=6.24 (mmsec)/years) than statin users (mean RT slope=6.02 (mmsec)/years); this is unlike males without an APOE4 carrier genotype, where statin non-users deteriorate slower (mean RT slope=4.70 (mmsec)/years) than users (mean RT slope=4.77 (mmsec)/years).
Our results therefore indicate that statins may have a beneficial effect on cognitive functions, however this may be limited to specific combinations of sex strata and APOE4 genotypes.
3.4 Statin use and Alzheimer’s disease
To examine the potential effect of statin use on prevalence of Alzheimer’s disease, a multivariate logistic regression model including statin use, APOE genotype, and sex as interaction terms found that, as expected, APOE4 carriers demonstrate an increased risk for AD (z-value = 11.05, p = <2e-16). More interestingly, while statin users have increased risk of AD (z-value = 3.76, p = 0.00017), APOE4 carriers, reported to be using statins, demonstrate a decreased risk for AD (z-value = -1.77, p = 0.07), though marginally significant. The full models’ outputs are reported in Supplementary Table 3.