Abstract
Genetics and lifestyle independently determine dementia risk, but the interaction is unclear. We assessed the interactive relationship of apolipoprotein E (APOE) genotype and physical exercise on dementia risk over a 5-year period in 1,646 older adults from the Canadian Study of Health and Aging who were dementia-free at baseline. Physical exercise moderated the relationship between genotype and dementia (p < 0.01). Specifically, for APOE ɛ4 non-carriers, the odds of developing dementia were higher in non-exercisers than exercisers (OR = 1.98, 95% CI = 1.44, 2.71, p < 0.001), whereas, for APOE ɛ4 carriers, the odds of developing dementia were not significantly different between non-exercisers and exercisers (OR = 0.71, 95% CI = 0.46, 1.31, p = 0.34). Given that most individuals are not at genetic risk, physical exercise may be an effective strategy for preventing dementia.
INTRODUCTION
Neurodegenerative diseases that cause dementia syndromes are characterized by gradual neural decline resulting in severe cognitive impairment that interferes with daily life [1]. Currently, approximately 47.5 million individuals are living with dementia worldwide [2]. This number is expected to surge to 115.4 million by 2050. With no known cure, an urgent need exists to uncover modifiable lifestyle factors that can reduce dementia risk.
Although chronological age is the greatest risk factor for dementia, both genetics and lifestyle factors also independently predict an individual’s outcome [3]. Various genetic mutations increase dementia risk [4, 5]; however, the apolipoprotein E (APOE) ɛ4 allele is the strongest genetic risk factor for non-pathological cognitive decline [6], vascular dementia [7, 8], dementias with synucleinopathy (dementia with Lewy bodies and Parkinson’s [9]), and especially Alzheimer’s disease [10–12]. Carriers of a single APOE ɛ4 allele are estimated to have a three to four fold increased risk of disease over age-matched non-carriers [13]. With respect to lifestyle, physical inactivity is the greatest modifiable risk factor for dementia [14], and individuals who engage in more physical exercise earlier in life have a reduced risk of developing dementia later in life [15]. However, it is unclear whether the benefits of physical exercise are similar for those with and without a genetic risk, and this knowledge is imperative when designing interventions to mitigate risk at the individual level.
Several population-based studies have examined the interaction between APOE genotype, physical exercise and dementia risk [16, 17], and yield equivocal results. Podewils et al. [16] reported that physical exercise reduced dementia risk for APOE ɛ4 allele non-carriers but not for carriers. In contrast, research using the Cardiovascular Risk Factors, Aging and Dementia (CAIDE) dataset [17–19] reported that physical exercise reduced dementia risk for APOE ɛ4 allele carriers but not for non-carriers. Yet others have demonstrated no interactive effect of physical exercise on dementia risk and genotype [20, 21]. The aim of the current study was to clarify the moderating effect of physical exercise on the relationship between APOE genotype and dementia risk, and to propose a hypothesis that may explain prior contradictory findings.
MATERIALS AND METHODS
Dataset
Detailed methods of the Canadian Study of Health and Aging (CSHA) have been described elsewhere [22]. Briefly, in the baseline phase (CSHA-1 : 1991-1992), individuals aged ≥65 years were drawn from 36 urban and rural areas from all 10 Canadian provinces. At CSHA-1, a nurse assessed cognitive impairment using the Modified Mini-Mental State Examination (3MS). Those who screened positive (3MS <78) and a random sample of participants who screened negative (3MS ≥78) were invited for complete clinical assessment by a nurse, which included the administration of the 3MS and screening for hearing and vision problems (CSHA-1: N = 2,914; CSHA-2, N = 2,305). A physician obtained medical history and performed a neurological examination. A psychometrist, who was blind to the participant’s medical history or current clinical findings, administered neuropsychological tests to participants who scored between 50 and 78 on the nurses’ 3MS. The study physician and neuropsychologist met together to reach a consensus diagnosis [22]. Physical exercise and self-reported health-related risk factors were collected at CSHA-1; participants who were found to be cognitively normal completed the questionnaires themselves.
In 1996–1997, the same diagnostic criteria was applied as in CSHA-1, with the addition of DSM-IV criteria for Alzheimer’s disease and new criteria for vascular dementia [23, 24]. The mean length of follow-up was 5 years. Blood samples were also collected and APOE allele status was determined in a subsample of clinically examined participants [25]. For those who passed away, their cognitive status three months prior to death was provided by a relative or other informant. The ethics review committees at all participating study centers and at the coordinating center approved both phases of the CSHA. Participants or their proxies provided informed consent for each component of the study.
Sample for analysis
For the current study, we analyzed data from respondents who provided blood samples for APOE allele status evaluation (n = 2,146). A total of 286 individuals (13%) were removed due to missing values for the physical exercise variable; a total of 167 individuals (8%) were removed due to a positive dementia diagnosis at CSHA-1, and a total of 47 individuals (2.8%) were removed due to a diagnosis of cognitive impairment no dementia. Thus, we included 1,646 persons in the analyses. Methods for APOE genotyping in CSHA have been described elsewhere [25]. To be included, participants’ initial screening results had to be negative or were clinically diagnosed without dementia at CSHA-1. Cases diagnosed with dementia at CSHA-2 (n = 331) included probable Alzheimer’s disease (47.7%), possible Alzheimer’s disease atypical (9.7%), possible Alzheimer’s disease vascular (8.5%), possible Alzheimer’s disease Parkinson’s (0.01%), possible Alzheimer’s disease with co-existing condition (3%), vascular dementia (22.1%), other specified dementia (2.7%), or unclassified dementia (5.7%). The remaining participants were dementia free at the follow-up and formed the control group (n = 1,315).
Covariates
Demographic factors of sex (0, male; 1, female), age, and years of education are important risk factors for cognitive changes, and were included as covariates in the regression analyses described below. Self-reported health-related risk factors for dementia were also included as covariates: prior heart attack (1, yes; 0, no), prior stroke (1, yes; 0, no), diabetes (1, yes; 0, no), high blood pressure (1, yes; 0, no), depression (1, yes; 0, no), and smoking (1, yes; 0, no).
Statistical analysis
Multivariable logistic regression analysis was conducted using SPSS 23. Dementia status at CSHA-2 was categorically coded as diagnosed with dementia (1) or not diagnosed with dementia (0). APOE genotype was binary coded according to presence or absence of ɛ4 alleles (0, non-carrier with no ɛ4 alleles; 1, carrier with at least one ɛ4 allele). To assess engagement in regular physical exercise, participants were asked, “Do you engage in regular exercise?” and provided a yes/no response. Thus, exercise was categorically coded as exercisers (0) or non-exercisers (1). Exercisers were asked two additional questions regarding the type and frequency of their exercise; the majority of exercisers walked (67%) or participated in exercises more strenuous than walking (26%) at a frequency of three times per week (75%). These physical exercise questions have been previously determined to be reliable and valid [26].
A multivariable logistic regression was conducted to evaluate whether APOE genotype and physical exercise predicted dementia risk, and to evaluate whether engagement in physical exercise at CSHA-1 moderated the relationship between APOE genotype and dementia outcome at CSHA-2. We also stratified by APOE genotype and conducted a separate multivariable logistic regression to further examine the interaction between exercise, genotype and dementia risk.
RESULTS
Table 1 displays participant characteristics. Table 2 reports dementia prevalence for APOE ɛ4 non-carriers and carriers by their exercise status. Table 3 shows the results of the regression analyses. The logistic regression revealed that the odds of developing dementia at CSHA-2 were significantly higher for non-exercisers than exercisers (OR = 1.96, 95% CI = 1.43, 2.67, p < 0.001). The odds of developing dementia were also significantly greater in APOE ɛ4 allele carriers than non-carriers (OR = 2.02, 95% CI = 1.26, 3.23, p < 0.01), and there was a significant interaction between exercise status and APOE genotype (p < 0.01). To understand this interaction, we stratified by APOE ɛ4 genotype and reran the logistic regression. This revealed that for APOE ɛ4 non-carriers, the odds of developing dementia were higher in non-exercisers than exercisers (OR = 1.98, 95% CI = 1.44, 2.71, p < 0.001), whereas, for APOE ɛ4 carriers, the odds of developing dementia were not significantly different between non-exercisers and exercisers (OR = 0.71, 95% CI = 0.46, 1.31, p = 0.34).
DISCUSSION
The present study examined the interactive effects of genetic risk and physical exercise on predicting dementia outcome. Overall, physical exercise moderated the relationship between APOE genotype and dementia risk after controlling for the effects of covariates. Exercisers were less likely to develop dementia than non-exercisers but only if they did not carry the APOE ɛ4 allele. The results point to the benefits of prescribing physical exercise to reduce the risk of dementia in older adults without a preexisting genetic risk.
When we assessed the entire sample, we observed the expected association of APOE ɛ4 allele carriers at an elevated dementia risk compared to APOE ɛ4 allele non-carriers. Surprisingly, non-exercisers without this genetic risk were similarly likely to develop dementia as those with it (Table 2), suggesting that physical inactivity may negate genetic protection from dementia, at least for APOE ɛ4 non-carriers. This is important because the majority of the population does not carry the APOE ɛ4 allele. Thus, increasing physical exercise among all older adults may represent an economical strategy for reducing dementia risk. Although more research is needed to understand the appropriate dose of exercise needed for protection, most of our active participants walked three times per week suggesting that this form of exercise may be particularly beneficial. Indeed, this is consistent with prior neuroimaging studies that report the slowing of age-related neural atrophy as a result of regular walking [27].
Unfortunately, individuals with one or two APOE ɛ4 alleles were not associated with dementia protection from physical exercise. This finding is consistent with results from Podewils et al. [8], who used a similarly aged cohort (Mean = 70 y) as the present study and also reported a moderating effect of physical exercise on dementia for APOE ɛ4 non-carriers but not for carriers. In contrast, studies using the CAIDE dataset [17–19] assessed a substantially younger cohort (Mean = 50 y) and yielded the opposite result, with APOE ɛ4 carriers benefitting the most from physical exercise. The large age differences across studies, confounded by the corresponding age-related accumulation of neural pathology associated with dementia, may account for these divergent findings [28]. In other words, there may be a potential threshold in disease progression above which the neural damage accumulation in APOE ɛ4 carriers may be too extensive to benefit from the therapeutic effects of physical exercise. Indeed, carriers of at least one APOE ɛ4 allele show metabolic and structural brain changes characteristic of dementia prior to the clinical manifestation of the disease [29], and those with two APOE ɛ4 alleles tend to experience greater accumulation of neural plaques and neurofibrillary tangles [30]. Although it remains unclear how physical exercise is changing the human brain to reduce dementia risk, rodent models demonstrate reduced plaque burden with regular aerobic exercise [31]. This may mean that there is a sensitive period during which physical exercise can benefit homozygous APOE ɛ4 carriers that occurs at an earlier stage of the disease before neural plaques and tangles can amass. While our results suggest there is no associated benefit of engaging in physical exercise to reduce dementia risk in APOE ɛ4 carriers after the age of 65, physical exercise has tremendous physical health benefits [32, 33] and therefore APOE ɛ4 carriers should still be encouraged to exercise. Importantly, future research is needed to directly examine how dementia risk and severity change as a function of age and physical exercise among different APOE genotypes.
These important findings are not without limitations. Physical exercise was assessed using validated but simple self-report questions that were only posed at CSHA-1. Although we cannot confirm whether activity level was maintained over the five-year period, prior longitudinal evidence suggests that the walking behavior of older adults’ (aged 65+) remains stable for up to ten years [34–36]. Critically, walking was the most common form of physical activity reported by CSHA participants and the follow-up time for CSHA-2 was five years. Longitudinal studies with objective measures of physical activity are needed to further test the interaction between physical exercise and genetic risk along with randomized-controlled trials to directly assess the causal relations.
In conclusion, the present study demonstrates a moderating effect of physical exercise on the relationship between the APOE genotype and dementia risk. Given that the majority of the population are without genetic risk, these results highlight the need for interventions to determine whether a dementia strategy should focus on increasing physical exercise among all older adults.
Footnotes
ACKNOWLEDGMENTS
The CSHA data presented in this paper were primarily funded by the Seniors’ Independence Research Program through the National Health Research and Development Program (NHRDP) of Health Canada (project no. 6606-3954-MC (S)). Additional funding was provided by Pfizer Canada Incorporated through the Medical Research Council/Pharmaceutical Manufacturers Association of Canada Health Activity Program, NHRDP (project no. 6603-1417-302 (R)), Bayer Incorporated, and the British Columbia Health Research Foundation (project no. 38 (93-2) and no. 34 (96-1)). The study was coordinated through the University of Ottawa and the Division of Aging and Seniors, Health Canada.
The Banting Research Foundation research grant awarded to JJH provided funding to support the data analysis. PR holds a Tier 1 Canada Research Chair in Geroscience and the Raymond and Margaret Labarge Chair in Research and Knowledge Application for Optimal Aging. AK holds a NSERC CGS-M scholarship.
