Differential Effect of APOE ɛ 4 Status and Elevated Pulse Pressure on Functional Decline in Cognitively Normal Older Adults

INTRODUCTION

Alzheimer’s disease (AD) is a chronic neurodegenerative disease and a growing global health concern, with its prevalence expected to triple over the next 50 years due to demographic increases in older populations and longer life expectancies [1]. In addition to cognitive decline, a central component of AD—both with respect to diagnosis and its devastating individual and societal effects—is the progressive loss of everyday functioning skills. Functional skills are normally classified into basic activities of daily living (ADL) which comprise various self-maintenance skills (e.g., bathing, dressing, grooming, and eating) and instrumental ADL (IADL) which involve more complex activities (e.g., managing finances, preparing meals, and using public transportation). Impairment in functional abilities significantly reduces independence and quality of life, while increasing caregiver burden and healthcare costs [2–4]). Moreover, the loss of functional independence due to cognitive impairment is a central criterion for AD diagnosis [5]. As such, identification of objective factors that predict functional decline among older adults has become a clinical and public health imperative.

More recently, arterial stiffness has been established as an important risk factor for cardiovascular events and maladaptive brain aging [6–8]. Arterial stiffness is marked by the degeneration of compliant elastin fibers and concomitant deposition of collagens, and thus increases steadily with advanced age [9]. The resulting decrease in arterial elasticity reduces the ability of arteries to accommodate the blood ejected from the heart during systole, and therefore increases the propagation of pulsatile pressure into the vascular structure of peripheral organs. These changes to arterial pulsatility ultimately result in damage to vulnerable microvascular tissue due to hypoperfusion and ischemia [10, 11]. As the brain has an especially high metabolic demand, it is particularly susceptible to the detrimental effects of these vascular alterations [12]. Accordingly, several studies have linked elevated pulse pressure (PP)—a proxy for arterial stiffness—to the presence of AD biomarkers [6, 13], cognitive decline [8], and progression to dementia [7, 14]. It remains unclear, however, how PP relates to functional difficultly and decline in cognitively normal older adults.

It is well known that the presence of the ɛ4 allele of the apolipoprotein E (APOE) gene—the most prominent susceptibility gene for late onset AD—is associated with cognitive impairment and decline over time [15, 16]. Moreover, brain volumetric changes and white matter decrements associated with APOE ɛ4 status have been noted across the aging spectrum, often in the presence of one or more vascular risk factors [17–19]. Given established associations between APOE and neurocognitive impairment [20, 21], several studies have attempted to explore the relationship between APOE ɛ4 status and functional decline. Findings are mixed, with some studies showing direct links between APOE genotype and functional outcomes [22], while others have reported that the effect of APOE ɛ4 status exists only in the presence of other risk factors for decline such as female sex or baseline functional difficulties [23]. Still, others have found no effect of APOE ɛ4 on functional outcome [24]. These varied findings are likely explained by wide-ranging methodologies and samples across studies. For example, the extant literature has focused on several different populations (e.g., cognitively normal, MCI, and AD; high versus low “AD risk”), employed various different study designs, and has used functional assessment methods with ranging reliabilities.

Although PP and the APOE ɛ4 allele have both been independently linked to cognitive impairment and pathological brain changes across the aging spectrum, it remains unclear how these risk factors may interact to affect functional decline. Thus, we investigated the longitudinal relationship between APOE ɛ4 genotype and elevated PP on changes in functional abilities in a well-characterized sample of cognitively normal older adults from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). We hypothesized that individuals with elevated PP would show poorer everyday functioning over time, and that this rate of functional decline would be more pronounced in APOE ɛ4 carriers than non-carriers.

MATERIALS AND METHODS

Participants

The sample comprised 738 cognitively normal older adults from ADNI. All participants were between the ages of 55 and 90 years old, had completed at least six years of education, were fluent in Spanish or English, and were free of any significant neurological or psychiatric disease. Additionally, all participants completed a neuropsychological assessment, APOE genotyping, and blood pressure assessment at baseline. Full criteria for ADNI eligibility and diagnostic classifications are described in detail at http://www.adni-info.org/Scientists/ADNIStudyProcedures.html.

Figure 1 shows participant selection for the present study sample. Of the 1,109 ADNI participants who completed baseline neuropsychological assessment, 749 participants were identified as cognitively normal (CN) since they were not classified by ADNI as having dementia and they did not meet criteria for MCI based on comprehensive neuropsychological criteria 2 [25, 26]. All individuals with the APOE ɛ2/ɛ4 genotype were excluded due to the unknown paired effects of the potential protective role of the ɛ2 allele combined with the known risk of the ɛ4 allele. Consistent with previous work, individuals with functional dependence at baseline were also excluded (i.e., Functional Activities Questionnaire (FAQ) score >5), as to ensure that that the final study sample consisted of older adults who could independently complete their activities of daily living at baseline [27]. Participants were not excluded for functional dependence at any subsequent visits. This yielded a final analytic sample of 738 cognitively normal older adults (ɛ4 carriers = 236; ɛ4 non-carrier = 502). Within the final analytic sample, 357 participants had complete data at the 48-month follow-up visit.

OPEN IN VIEWER

Fig.1

Participant selection.

RESULTS

Longitudinal effects of PP and APOE ɛ4 genotype on functional decline over four-year follow-up

A MLM, adjusted for age, education, sex, depressive symptoms, and total number of antihypertensive medications was used to determine whether APOE ɛ4 status and baseline PP predicted change in FAQ scores across the 4-year follow-up period. Table 2 includes the MLM fixed and random effects. Of the baseline covariates that were included in the model to control for confounding effects on FAQ, age and sex were significant (p < 0.05). Specifically, women (r = 0.227) and participants who were older at baseline (r = 0.132) had higher FAQ scores (indicating worse performance) across all visits (p-values <0.05).

OPEN IN VIEWER

Table 2

Estimates for the full longitudinal model of the effect of genetic and vascular risk on functional decline

	Estimate	S.E.	df	F-statistic	p-value	r-value
Level 2 (Between Subjects Effects)
Age	0.020	0.009	300.65	2.32	0.021	0.132
Education	–0.035	0.022	301.22	–1.60	0.110	0.092
Gender	–0.476	0.118	300.48	–4.04	<0.001	0.227
GDS total score	–0.002	0.005	283.60	–0.31	0.756	0.018
Number of antihypertensive medications	–0.027	0.056	302.15	–0.48	0.635	0.028
APOE ɛ4 status	0.175	0.516	283.88	0.34	0.735	0.020
Pulse Pressure	0.313	0.043	299.19	7.26	<0.001	0.386
Pulse Pressure×APOE ɛ4	–0.002	0.008	284.11	–0.27	0.788	0.016
Intercept	–0.359	0.792	297.55	–0.45	0.651	0.026
Level 1 (Within Subjects Effects)
Visit	–0.023	0.015	1668.95	–1.61	0.108	0.039
Pulse Pressure×Visit	0.001	0.000	1670.87	2.96	0.003	0.072
APOE ɛ4×Visit	0.087	0.024	1678.11	3.64	<0.001	0.088
Pulse Pressure×APOE ɛ4×Visit	–0.001	0.000	1683.53	–2.16	0.031	0.053

GDS, Geriatric Depression Scale; APOE ɛ4 status, presence of the ɛ4 allele; S.E., standard error of the estimate.

Fixed effects

There was a significant main effect of baseline PP (p < 0.001), such that higher baseline PP was associated with higher FAQ scores (more functional difficulty) across all time points and participants. There were no significant main effects of APOE ɛ4 status or visit. The two-way PP×APOE ɛ4 status interaction was also not significant (ps > 0.05). However, the two-way visit×PP interaction was significant (p = 0.003), such that elevated baseline PP was associated with greater decline in everyday functioning, as indexed by increasing FAQ scores, across follow-up visits. A significant two-way visit×APOE ɛ4 status interaction was also observed (p < 0.001), such that ɛ4 carriers had greater increases in FAQ scores (more functional difficulty) across follow-up visits relative to non-carriers.

The significant two-way interactions are qualified by the significant three-way PP x APOE ɛ4 status×visit interaction, such that the effect of baseline PP on FAQ score trajectories over time differed between APOE ɛ4 groups (p = 0.031). Figure 2 shows the FAQ score trajectories of ɛ4 carriers and non-carriers by baseline PP (High versus Low). Findings suggest that, while ɛ4 carriers had the greatest increase in FAQ scores across the 4-year follow-up period, FAQ scores trajectories did not differ depending on baseline PP level. Conversely, while non-carriers were observed to have less decline in functional ability, the rate of decline was greater for non-carriers with elevated baseline PP.

OPEN IN VIEWER

Fig.2

APOE ɛ4 status modifies the effect of pulse pressure on everyday functioning over time. FAQ, Functional Activities Questionnaire; PP, pulse pressure. The model predicted values, controlling for age, education, sex, and depressive symptoms, are shown. Pulse pressure levels were determined by a median split of the values in the analytic sample (Low = PP ≤58; High = PP >58). Higher FAQ scores indicate greater functional difficulty. Error bars represent the standard error of the mean.

Random effects

The model examined between-subject variability in FAQ score (random intercept). The intercept accounted for significant random variance (p < 0.001), suggesting that there was inter-individual variability in the initial FAQ score that was not accounted for in the model.

DISCUSSION

We examined the independent and interactive effects of brachial PP and APOE genotype on decline in everyday functioning over time in a sample of well-characterized, healthy, initially cognitively normal older adults. Results showed that elevated PP predicted greater functional difficulty trajectories across all participants, such that individuals with higher baseline PP showed greater declines in everyday functioning (as indexed by increasing FAQ scores) across four years of follow-up relative to individuals with lower baseline PP levels. Additionally, APOE ɛ4 status also significantly predicted functional difficulty trajectories, such that APOE ɛ4 carriers demonstrated the greatest increases in FAQ scores over the follow-up period (relative to non-carriers). Interestingly, although non-carriers showed less dramatic declines in functional ability across time relative to ɛ4 carriers, the rate of decline was increased for non-carriers with higher baseline PP values. Such findings suggest that, while APOE ɛ4 status is the most prominent predictor of functional difficulty for APOE ɛ4 carriers, the relatively smaller effect of arterial stiffening on functional difficulty is still observed in non-carriers. Additionally, results suggest that APOE ɛ4 status and PP do not have an additive effect on risk of functional difficulty in relatively healthy older adults. Such results add to the growing literature attempting to disentangle the factors that contribute to the development and progression of prevalent degenerative brain aging conditions that result in functional decline, such as AD and vascular cognitive impairment.

The present study converges with a burgeoning body of research that confirms the deleterious role of APOE ɛ4 in aging. APOE ɛ4 remains the greatest known genetic risk factor for AD [35, 36], and has been repeatedly linked to AD neuropathology [37, 38], cerebrovascular disease [39], as well as cognitive impairment [15, 16]. There are several possible mechanisms through which the presence of an APOE ɛ4 allele may exert its effect of increasing risk for AD, such as altering the amyloid-β (Aβ) clearance and deposition in vulnerable brain regions, exacerbating tau-mediated neurodegeneration in AD [40], as well as influencing the extent of cognitive reserve available to compensate for accumulating neuropathology [41–43]. Alternatively, it may be that the presence of the APOE ɛ4 allele exacerbates Aβ accumulation in the brain secondary to an initial vascular damage to the brain (e.g., hypoxia, perfusion stress, blood-brain-barrier disruption; see [44] for a review). The exact mechanism, however, by which APOE exerts its effect on specific neurodegenerative processes remains unclear. Moreover, while the robust effects of the APOE ɛ4 genotype on pathological brain aging makes it a prime candidate for the early detection of individuals at risk for functional decline, there is a lack of effective risk-modifying treatments based on genetic susceptibility. Thus, a focus on targeting modifiable risk factors—such as elevated PP—to prevent or slow functional decline and pathological brain aging may prove to be more fruitful. Indeed, the present study confirms the importance of considering vascular factors, even in asymptomatic older adults, when determining risk for future functional decline.

There exists a large body of evidence implicating vascular risk factors (e.g., hypertension, atherosclerosis, hyperlipidemia, metabolic syndrome, diabetes mellitus) in pathological brain aging, as evidenced by the development and progression of cerebrovascular disease, AD brain pathology, cognitive impairment, and functional decline [45]. Recent studies have revealed a particularly robust association between arterial stiffness with various aspects of brain disease in old age, including gray and white matter degradations, accumulation of AD neuropathology, and cognitive decline [7, 46]. As previously stated, arterial stiffness leads to a deleterious hypertensive state that can cause further damage to vulnerable microvasculature and associated tissue [46, 47]. The brain is especially vulnerable to these effects due to the high density and low impedance of its microvasculature, allowing for increased pulsatile loads to deeply penetrate the delicate microvascular beds [10]. Hypertension resulting from increased arterial stiffness is thus thought to cause chronic hypoperfusion and ischemia of brain tissue [11], ultimately resulting in the development of tissue damage [7, 48]. Such deleterious effects of arterial stiffening on brain structure and function have been demonstrated as early as young and mid-adulthood [13, 50], highlighting the chronic, insidious contribution of vascular disease to cognitive aging. The findings of the present study extend this literature by demonstrating that PP predicts 4-year functional decline in cognitively normal, asymptomatic older adults.

While APOE ɛ4 and PP both predicted functional difficulty, it should be noted that the sample was, on average, still functionally independent at the 48-month visit. Prior work has shown that a cutoff of ≥6 on the FAQ best discriminated MCI from AD [33]. While this suggests that the change in functional ability captured by the present study may reach the level of dependence on IADLs, previous studies have demonstrated that increased functional difficulty is a significant risk factor for future functional disability and cognitive decline [51–53]. Thus, the present findings have implications for the early identification of individuals that are increased risk for functional decline, and they add to a promising body of literature highlighting the utility of employing cardiovascular-based intervention strategies as a method for reducing the prevalence of functional decline due to cognitive impairment in the aging population (see [54] or [55] for a review).

An unexpected finding of the present study was the nature of the interaction between genetic and vascular risk factors on functional ability trajectories. We hypothesized that elevated vascular risk would modify the effect of APOE ɛ4 status on longitudinal FAQ scores, such that ɛ4 carriers with elevated baseline PP would have the greatest amount of observed functional decline in our sample. We predicted this, given the growing body of literature to suggest that the presence of one or more ɛ4 allele may exacerbate the effect of certain vascular risk factors on pathological brain aging [15, 56]. While we did observe that the presence of at least one ɛ4 allele was, indeed, associated with the greatest risk for declines in functional ability, the absence or presence of elevated PP at baseline did not significantly alter the functional decline trajectories of ɛ4 carriers. Conversely, while non-carriers were observed to generally have less dramatic declines in functional ability across time relative to their ɛ4 carrier counterparts, the rate of decline was increased for non-carriers with elevated PP at baseline. Differently stated, elevated PP only appeared to increase the risk of functional decline in individuals who were not genetically susceptible to developing AD.

There are several possible explanations for this pattern of observed findings. As previously noted, APOE ɛ4 carriers were observed to have much steeper rates of functional decline relative to non-carriers. Thus, it is possible that this relatively larger effect of APOE ɛ4 status may have subsumed the potentially smaller effect of arterial stiffening on FAQ trajectories in APOE ɛ4 carriers. The characteristics of the ADNI sample, combined with the method used to assess functional activity, may furthermore contribute to such potential masking effects of ɛ4 status. Specifically, the ADNI sample is more highly educated and exhibits fewer vascular health related issues than would be expected for the community-dwelling population. Thus, it is possible that FAQ was not sensitive enough to capture slight PP effects on functional decline in ɛ4 carriers with relatively low vascular illness at baseline, especially given their comparatively faster rate decline. Moreover, the effects of PP could have further been masked by compensation in this highly-educated group of participants. Alternatively, the present findings might also be attributable to differential age effects across the APOE ɛ4 groups. That is, APOE ɛ4 carriers were significantly younger than non-carriers in this sample. Thus, it is possible that pulse pressure did not predict functional decline in ɛ4 carriers because of their younger age (and potentially greater cerebral resilience against the deleterious effect of increased pulse pressure).

A recent meta-analysis of studies assessing the association between arterial stiffness, cerebral small vessel disease, and cognitive impairment concluded that, while the literature supports an association between arterial stiffness and microvascular brain disease, the link with cognition is much weaker [57]. Similarly, there is evidence to suggest that the association between arterial stiffness and cognition is mediated by pathways that include cerebral microvascular remodeling and microvascular parenchymal damage [58]. Such findings suggest that the relationship between arterial stiffness and cognition is complex, and that it is critical to consider additional factors that may affect whether chronic vascular disease results in subsequent brain damage that gives rise to changes in cognitive performance. Given that cognitive impairment is largely thought to drive functional decline in populations with AD or related dementias [5], we speculate that similarly complex relationships exist between PP, ɛ4 status, and functional decline. Future studies are warranted that attempt to model such multifarious relationships, with careful consideration of potential factors that may mediate or moderate the effects of arterial stiffening at at-risk populations, Furthermore, there is an increasing focus in the fields of aging and dementia on understanding the time-signatures of factors initiating or promoting brain pathology and clinical dementia syndromes [59]. Thus, consideration of potentially different timing-related effects of PP and APOE ɛ4 status on functional trajectories should be considered in future studies.

There are several strengths to the present study. Critically, the study employed a prospective study design to investigate risk factors for functional decline over time in older adults that were asymptomatic at baseline. The longitudinal clinical follow-up allowed for the investigation of the progression of functional ability, rather than only characterizing cross-sectional associations with everyday function. As demonstrated by the findings of our study, such an approach is imperative when exploring such associations in asymptomatic older adults. Indeed, we found that neither PP nor APOE ɛ4 status was associated with functional difficulty cross-sectionally; however, both interacted with time (visit) to predict future decline in functional ability. Additionally, the large sample size and multilevel modeling approach allowed the power to test for both independent and interactive associations between PP and APOE genotype on longitudinal FAQ scores, and afforded us the ability to statistically control for a variety of potential confounds. Finally, we identified cognitively normal older adults within the ADNI dataset using highly sensitive diagnostic criteria [25]. These criteria have been shown to be more reliable and stable than the conventional diagnostic criteria for MCI, and they yield far fewer ( 33%) false positives diagnostic errors [26, 60].

There are some important limitations to the present study that must also be considered. For example, there was a notable amount of attrition observed between baseline and the 48-month follow up visit. That is, while the modeling design employed in this study (i.e., mixed-effects design, full-information maximum likelihood, and the inclusion of attrition covariates) was selected because it is considered the best practice for estimating longitudinal relationships in a sample with selective attrition [28], bias due to attrition should still be considered as a potential limitation when interpreting the present findings. Additionally, while the present study did control for antihypertensive use at baseline, longitudinal antihypertensive use was not incorporated due to modeling constraints. Additionally, the ADNI exclusion and inclusion criteria yielded a sample with a relatively low frequency of health issues, including the presence of vascular disease, compared to the general population. Thus, the direction of the interaction between vascular and genetic risk on functional ability trajectories may apply only to populations with a relatively restricted range of vascular disease, thus limiting the generalizability of the present study findings. Moreover, there was limited information available for this sample regarding the presence of other risk factors to arterial stiffness and vascular disease (e.g., fasting blood glucose, A1c levels, dietary factors, and physical exercise). Thus, replication of these results in a more representative community sample that is well characterized with respect to vascular risk is recommended to further elucidate the role of arterial stiffness as a predictor of functional decline in late life. Indeed, future research that employs multivariate models to explore more complex associations between longitudinal PP measurements, vascular risk management (e.g., antihypertensive medication use and compliance, lifestyle modifications), and functional decline is needed.