Modifiable Lifestyle Factors and Cognitive Function: A Population-Based Study Amongst Nondemented Men
Abstract
Background: As cognitive decline and dementia-related diseases are currently incurable, identifying modifiable and preventative factors is imperative. Thus, we aimed to investigate the associations between modifiable lifestyle factors and cognitive function in a population-based sample of nondemented men.
Methods: Data for 537 men were drawn from the ongoing, population-based, Geelong Osteoporosis Study (GOS). The CogState Brief Battery (CBB) was utilised to determine cognitive function (overall cognitive function [OCF], psychomotor function, attention, working memory, and visual learning). Demographic and lifestyle factors (diet, smoking, physical activity [sport, work/home, and leisure time], alcohol use, and body mass index [BMI]) were self-reported and documented across three waves of data collection. Generalised estimating equation (GEE) models were used to determine associations between the time-updating exposures of interest and each of the cognitive domains.
Results: Age was associated with OCF (η2 = 0.298), psychomotor function (η2 = 0.163), attention (η2 = 0.210), working memory (η2 = 0.215), and visual learning (η2 = 0.097). Playing sport was associated with all aspects of cognition: OCF (η2 = 0.014), psychomotor function (η2 = 0.007), attention (η2 = 0.004), working memory (η2 = 0.013), and visual learning (η2 = 0.004), and being physically active during leisure time was associated with visual learning (η2 = 0.008). Being physically active at work and home, smoking, diet quality, energy consumed per day, alcohol intake, and BMI were not associated with any of the cognitive domains.
Conclusion: In this study of nondemented men, elements of a healthy lifestyle such as playing sport and being physically active during leisure time were associated with better cognition. Further studies into whether these factors, amongst others, prevent the onset and progression of dementia are needed.
1. Background
Sustaining optimal cognitive function is essential for successful performance and independence in later life [1, 2]. The spectrum of cognitive ageing includes normal cognitive ageing and mild cognitive impairment (MCI) through to dementia. Normal cognitive ageing has been defined as ageing changes that occur in individuals free of overt diseases of the nervous system [3]. In contrast, MCI is defined as an early stage of neurodegenerative pathology [4] and dementia as a loss of cognitive abilities, being an umbrella term for a number of neurological conditions, with the most commonly known as Alzheimer’s disease (AD). In 2006, 26.6 million people worldwide were suffering from AD, with the number expected to increase to 107.8 million by 2050 [5]. Currently, there are no medications or treatments for cognitive decline. Therefore, it is imperative to identify risk factors that can better evidence-based preventative strategies to mitigate the impact of poor cognitive health.
Over the past decades, lifestyle behaviours have received increasing attention as factors that could be modified to improve cognitive ageing, including regular physical activity, low body mass index (BMI), healthy diet, not smoking, and moderate alcohol consumption [6–10]. A previous cross-sectional study of elderly men and women found certain lifestyle behaviours, such as adherence to a Mediterranean diet, quality sleep, physical activity, and engagement in activities of daily living to be associated with better cognitive function [7]. Moreover, prospective studies in the elderly have found similarities, indicating that a healthy lifestyle is associated with reduced dementia risk [8–10]. However, few studies have focused on the effect of lifestyle on specific cognitive domains in populations spanning the full adult age range.
Assessment of modifiable risk and protective factors from an earlier age is important in regards to dementia prevention, as cognitive decline often appears many years before a dementia diagnosis, with the importance of early intervention having been emphasized [11]. Furthermore, examining the effects of lifestyle on cognitive health across the lifespan, especially in cognitively unimpaired community-dwelling adults, may aid in the identification of subpopulations requiring intervention. Targeting many lifestyle behaviours simultaneously may help understand the combined effects of healthy lifestyle on cognitive health.
To fill this gap, we examined a number of modifiable lifestyle factors including diet, smoking, physical activity, alcohol use, and BMI and their associations with both overall and specific domains of cognitive function in a population-based sample of nondemented men. We hypothesized that a quality diet, nonsmoker, being physically active, low alcohol use, and BMI within the normal range is associated with better cognitive function.
2. Methods
2.1. Participants
Data were collected from an age-stratified, population-based sample of men enrolled in the Geelong Osteoporosis Study (GOS). In 2001, a total of 1540 men, aged between 20 and 97 years, were recruited via random selection from the electoral rolls for the Barwon Statistical Division in south-eastern Australia [12]. Participants have since returned for assessment 5-year (2006–2007, η = 978) and 15-year (2015–2019, η = 624) postbaseline. Those who completed a cognitive assessment at the 15-year follow-up were eligible for inclusion in this study, resulting in a sample of 537 men. This study was approved by the Barwon Health Human Research Ethics Committee (00/56), and all participants gave written informed consent.
2.2. Measurements
2.2.1. Outcome
Cognitive function was assessed at the 15-year follow-up using the CogState Brief Battery (CBB), a computer-based neuropsychology assessment. The CBB involves responding to stimuli cards as part of a detection task (DET), an identification task (IDN), a one-card learning task (OCL), and a one-back task (OBK). These four tasks assess the cognitive domains of psychomotor function, visual attention, recognition memory/learning, and working memory, respectively. Reaction time was the primary outcome measure for the DET, IDN, and OBK tasks, measured in log10 million seconds (lmn) and calculated according to the speed of performance, with lower scores indicating greater performance. A combination of reaction time and accuracy was the primary outcome measure for the OCL task, labelled arcsine square root proportion correct (acc) and calculated according to the accuracy of performance and reaction time, with higher scores indicating greater performance. Overall cognitive function (OCF) was generated by combining the four cognitive domains; higher scores indicated better performance.
2.2.2. Exposures
Data on physical activity, smoking habits, diet quality, alcohol intake, and BMI were collected across three waves (baseline and 5-year and 15-year follow-up assessments).
Physical activity was derived from two questionnaires, the Baecke questionnaire, designed for young adults (18–59 years), and the Voorrips questionnaire, designed for older adults (60+ years) [13, 14]. The Baecke questionnaire includes 16 items and assesses physical activity across three domains (sport, work, and leisure time) [13]. The Voorrips questionnaire also includes three domains: household activities (10 items), sporting activities (two items), and leisure time activities (two items) [14]. The median score of each domain was calculated and then dichotomised at the median for incorporation into statistical models.
Current and past patterns of tobacco use were self-reported (yes/no). Smoking included manufactured and “hand-rolled” cigarettes, cigars, and pipes.
Diet quality was determined using the Australian Recommended Food Score (ARFS). Each participant completed the Australian Cancer Council’s Dietary Questionnaire for Epidemiological Studies (DQESs) [15] with information collected used to determine an ARFS for each participant. The ARFS represents the degree of adherence to the Australian Dietary Guidelines, which is generated by allocating a score for the consumption of (i) vegetables, nuts, and beans; (ii) fruit; (iii) protein-containing foods; (iv) grains; (v) dairy; (vi) fats; and (vii) alcohol. Scores are totalled resulting in an overall score ranging from 0 to 74 [16]. A high score indicates better adherence to a quality diet. Mean dietary energy intake (kJ/day) and alcohol intake (g/day) were also derived from the DQES [15].
Weight was measured using electronic scales to the nearest 0.1 kg, and height was measured using a Harpenden stadiometer measured to the nearest 0.001 m. BMI was calculated as weight/height2 (kg/m2).
2.3. Statistical Analysis
Characteristics were summarized as percentage for categorical variables, mean (standard deviation [SD]), or median (interquartile range [IQR]) for continuous variables. Generalised estimating equation (GEE) models were utilised to determine associations between the time-updating exposures of interest (diet, smoking, physical activity [sport, work/home, and leisure time], alcohol use, and BMI) and each of the cognitive domains (OCF), psychomotor function (DET), attention (IDN), working memory (OBK), and visual learning (OCL)]. The three exercise domains (sport, work/household maintenance, and leisure time) were run in the models separately due to collinearity. A Bonferroni correction was applied, with an adjusted p value significant at a 0.01 level to correct for multiple comparisons. Eta-squared (partial) effect sizes were calculated; 0.01 indicates a small effect; 0.06 indicates a medium effect; 0.014 indicates a large effect [17]. Analyses were undertaken using SPSS Statistics (Version 28) [18].
3. Results
A total of 537 participants were eligible for inclusion (Table 1). At baseline, the median age of participants was 49.7 years (IQR 38.4–58.8). Four hundred and sixty-four (86.6%) participants were identified as nonsmokers at the time of assessment. The median BMI was 26.5 kg/m2 (24.3–29.1), alcohol intake was 16.8 g/day (4.7–29.8), diet quality score was 30 [19–31], and energy intake was 8658.4 kJ/day (6968.3–10612.2). Two hundred and sixty-one (261) (49.8%) participants were identified as playing physical sport, two hundred and thirteen (213) (59.4%) participants were identified as being physically active during leisure time, and two hundred and seventy-three (273) (51.8%) were identified as being physically active at work or during home maintenance.
| Age (year) | 49.7 (38.4–58.8) |
| BMI (kg/m2) | 26.5 (24.3–29.1) |
| Smoking (nonsmoker) | 464 (86.6%) |
| Alcohol intake (g/day) | 16.8 (4.7–29.8) |
| Healthy diet | 30 (25–37) |
| Energy (kJ/day) | 8658.4 (6968.3–10612.2) |
| Physically active (sport) | 261 (49.8%) |
| Physically active (work/household maintenance) | 273 (51.8%) |
| Physically active (leisure time) | 313 (59.4%) |
Age was associated with lower OCF (B − 0.030 ± 0.002, η2 = 0.298, p < 0.001), psychomotor function (B 0.003 ± 0.0003, η2 = 0.163, p < 0.001), attention (B 0.003 ± 0.0002, η2 = 0.210, p < 0.001), working memory (B 0.004 ± 0.0003, η2 = 0.215, p < 0.001), and visual learning (B − 0.002 ± 0.0003, η2 = 0.097, p < 0.001). The effect size for all relationships ranged from 0.097 to 0.298.
Playing sport was associated with better OCF (B 0.004 ± 0.001, η2 = 0.014, p < 0.001), psychomotor function (B < 0.0002 ± <0.0001, η2 = 0.007, p = 0.003), attention (B < 0.0002 ± <0.0001, n2 = 0.004, p < 0.001), working memory (B < 0.0001 ± <0.0001, n2 = 0.013, p = 0.001), and visual learning (B < 0.0001 ± <0.0001, η2 = 0.004, p = 0.006). The effect size for all relationships ranged from 0.004 to 0.014. Physical activity during leisure time was associated with better visual learning (B < 0.0001 ± <0.0001, η2 = 0.008, p = 0.003). No differences were detected in OCF, psychomotor function, attention, and working memory between those who were and were not physically active during leisure time.
Being physically active at work or at home was not associated with any of the cognitive domains (OCF, psychomotor function, attention, working memory, and visual learning) (all p > 0.01; Table 2).
| Overall cognitive function | Psychomotor function | Attention | Working memory | Visual learning | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| b | SE | η2 | p | b | SE | η2 | p | b | SE | η2 | p | b | SE | η2 | p | b | SE | η2 | p | |
| Age | −0.030 | 0.002 | 0.298 | < 0.001 | 0.003 | 0.0003 | 0.163 | < 0.001 | 0.003 | 0.0002 | 0.210 | < 0.001 | 0.004 | 0.0003 | 0.215 | < 0.001 | −0.002 | 0.0003 | 0.097 | < 0.001 |
| Diet quality | 0.0002 | 0.0001 | 0.002 | 0.106 | <− 0.0001 | < 0.0001 | 0.001 | 0.243 | <− 0.0001 | < 0.0001 | < 0.001 | 0.519 | < 0.0001 | < 0.0001 | 0.002 | 0.134 | < 0.0001 | <0.0001 | 0.001 | 0.209 |
| Energy (kJ/day) | <− 0.0001 | < 0.0001 | 0.002 | 0.200 | < 0.0001 | < 0.0001 | 0.002 | 0.137 | <0.0001 | < 0.0001 | < 0.001 | 0.493 | < 0.0001 | < 0.0001 | 0.003 | 0.177 | < 0.0001 | < 0.0001 | < 0.001 | 0.975 |
| BMI (kg/m2) | 0.001 | 0.001 | 0.001 | 0.833 | <− 0.0001 | < 0.0001 | 0.001 | 0.648 | <− 0.0001 | < 0.0001 | 0.004 | 0.236 | < 0.0001 | < 0.0001 | 0.007 | 0.378 | <− 0.0001 | < 0.0001 | 0.006 | 0.059 |
| Alcohol (g/day) | < 0.0001 | < 0.0001 | 0.001 | 0.321 | <− 0.0001 | < 0.0001 | 0.002 | 0.255 | <− 0.0001 | < 0.0001 | 0.001 | 0.481 | < 0.0001 | < 0.0001 | 0.001 | 0.288 | < 0.0001 | < 0.0001 | < 0.001 | 0.928 |
| Nonsmoking | 0.003 | 0.003 | 0.001 | 0.288 | −0.0002 | 0.0002 | 0.001 | 0.317 | <−0.0001 | 0.0003 | < 0.001 | 0.771 | < 0.0001 | < 0.0001 | 0.001 | 0.418 | < 0.0001 | < 0.0001 | 0.005 | 0.023 |
| Physical activity | ||||||||||||||||||||
| Sport | 0.004 | 0.001 | 0.014 | < 0.001 | −0.0002 | < 0.0001 | 0.007 | 0.003 | −0.0002 | 0.0001 | 0.004 | < 0.001 | < 0.0001 | < 0.0001 | 0.013 | 0.001 | < 0.0001 | < 0.0001 | 0.004 | 0.006 |
| Work/home | −0.0001 | 0.0008 | < 0.001 | 0.885 | < 0.0001 | < 0.0001 | < 0.001 | 0.781 | < 0.0001 | 0.0001 | < 0.001 | 0.932 | < 0.0001 | < 0.0001 | 0.002 | 0.096 | < 0.0001 | < 0.0001 | 0.002 | 0.134 |
| Leisure time | −0.0001 | 0.0005 | 0.001 | 0.670 | < 0.0001 | < 0.0001 | < 0.001 | 0.373 | <− 0.0001 | < 0.0001 | < 0.001 | 0.903 | < 0.0001 | < 0.0001 | < 0.001 | 0.280 | < 0.0001 | < 0.0001 | 0.008 | 0.003 |
- Note: Eta-squared (partial) effect sizes were calculated as η2 = 0.01 (small effect), 0.06 (medium effect), and 0.014 (large effect). The full model presented includes physical activity—sport; equivalent multivariate models were run for leisure time and work/home maintenance. Bold entries mean significant (p < 0.01).
BMI, diet quality, smoking, energy, and alcohol intake were not associated with any of the cognitive domains (all p > 0.01; Table 2).
4. Discussion
In this longitudinal population-based study of men, age, playing sport, and being physically active during leisure time were associated with better cognitive function. Specifically, playing sport was associated with better OCF, psychomotor function, attention, working memory, and visual learning, while being physically active during leisure time was associated with better visual learning. Interestingly, there were no significant associations detected between BMI, diet quality, smoking, energy, alcohol intake, and being physically active at work or at home with any of the cognitive domains.
In support of our findings, a recent systematic review of 24 published systematic reviews and meta-analyses investigating the relationship between physical activity and physical and mental health outcomes in older adults found that physically active older adults are at considerably reduced risk of cognitive decline, dementia, and AD, amongst other health conditions [32]. They also found those who experience healthier ageing trajectories and better quality of life have better cognitive functioning than their counterparts. Previous evidence suggests physical activity preserves cognition via a range of neurological mechanisms [33–35]. Recently, the effects of physical activity were found to be associated with positive modulation of myokines, which improves memory and increases brain-derived neurotrophic factor expression [19, 36, 37]. Other studies have suggested that physical activity drives autophagy degradation to reverse the proteostasis dysfunction linked with ageing and neurodegeneration [20–22]. Additionally, a cross-sectional study by Kurita et al. [23] reported that older adults with longer daily sedentary time and low cognitive engagement were associated with decreased volumes of the rostral and caudal cingulate cortex.
However, few studies have gone on to evaluate what type of activity and when it is protective. Similar to our results, a recent longitudinal study by Sabia et al. reported a positive effect of sports and leisure activities on cognition [24]. Although, Dupré et al. [25] found a positive effect of domestic activities but not sports or leisure time activity on cognition over 8 years of observation in an elderly cohort of men and women. Thus, further studies on the type and duration of physical activity and cognitive function across the lifespan are warranted to assess whether our results are corroborated.
Interestingly, we did not detect an association between diet quality, smoking, alcohol consumption, or BMI with any of the cognitive domains measured. Similar to our findings, using data from a large 25-year population-based cohort study found that diet quality assessed during midlife was not associated with risk of dementia [25]. Moreover, other studies have similarly found alcohol consumption to have no association with cognitive function [28–30], whereas others have found that low to moderate alcohol consumption possibly reduces the risk of dementia compared to not drinking at all [31]. In regards to BMI, a review of 19 longitudinal studies published over 42 years reported that obesity (BMI ≥ 30 units?) but not being overweight (BMI 25–30 units?) or underweight was associated with late-life dementia [38]. Smokers too have been found to be at a higher risk of poor cognitive health compared to nonsmokers [39]; however, evidence exists showing that ex-smokers who had given up smoking at least 10 years prior have no adverse cognitive effects [40]. Taken together, the relationship between lifestyle factors and cognitive health is complex, and there is a need for further research to help clarify the associations between lifestyle factors and cognition.
Several mechanisms underlying the link between lifestyle attributes and cognition have been suggested, one of which is cognitive reserve. Cognitive reserve has been defined as the ability of the brain to optimize performance through differential recruitment of brain networks or use of alternative strategies [41–43]. Engagement in mental activity, for example, through undertaking education or working in occupations that involve complex demands, is a key determinant of cognitive reserve [41]. Cognitive reserve reflects the capacity to provide a buffer against the effects of dementia-related brain pathology so that a greater burden of pathology is needed before signs of cognitive decline or symptoms of dementia become evident [42]. It is possible that lifestyle factors may exert their effects on risk by increasing the efficiency of neural networks and hence enhancing cognitive reserve, resulting in greater resilience against the effects of developing neuropathology [41, 42]. Additionally, the concept of brain maintenance has also been suggested, referring to lifestyle factors, such as physical activity, preserving the brain by reducing overall brain changes, and improving neural repair processes [44–46]. Furthermore, it is possible that each lifestyle behaviour is interrelated, with previous evidence describing a life-course approach [27, 39]. For example, smokers may be more likely to drink high levels of alcohol, or physically inactive individuals may have a poorer diet. It might be a combination of these factors that exacerbate the relationship between lifestyle factors and cognition. Lifestyle attributes may also affect cognition indirectly, via their association with medical conditions such as cancer, metabolic syndrome, and/or cardiovascular disease [27].
4.1. Strengths and Limitations
This study has several strengths. First, we included a large population-based sample of men aged 35 years and older. Second, this study examined a broad range of lifestyle factors simultaneously. Third, the CBB was used to assess cognition, a comprehensive cognitive test battery which allowed the investigation of overall and specific cognitive domains. Our study also has limitations. As our study included only men, interpretation may not be generalisable to women, and therefore, future research may consider collecting comparable data for women. Additionally, the assessment of lifestyle factors was based on self-reported tools, which can be prone to differential recall bias. Lastly, longitudinal cognitive data were not available; therefore, further studies are warranted to assess whether our results are corroborated.
5. Conclusion
In this study of nondemented men, elements of a healthy lifestyle such as playing sport and being physically active during leisure time were associated with better cognition. Further awareness of the risk of cognitive decline in the presence of lifestyle behaviours is needed. Such evidence highlights the emphasis on promoting healthy lifestyle attributes to preserve cognitive health.
Disclosure
The funding sources have no role in the study design, data collection, data analysis, data interpretation, writing of the report, or the decision to submit the paper for publication.
Conflicts of Interest
The authors declare no conflicts of interest.
Author Contributions
K.B.C., L.J.W., and J.A.P. planned and designed the study. K.B.C. and M.M. performed the statistical analyses. K.B.C., A.L.S., L.J.W., and J.A.P. contributed to the interpretation of the data. K.B.C. took primary responsibility for writing the manuscript. L.J.W., J.A.P., A.L.S., B.E.K., M.M., and S.X.S. provided critical revisions to the article. All authors read and approved the final manuscript to be published.
Funding
K.B.C. was supported by the Australian Rotary Health/Bing Taylor PhD Scholarship in Dementia, S.X.S was supported by an Executive Dean Health Research Fellowship (Deakin University), and L.J.W was supported by the National Health and Medical Research Council (NHMRC) Investigator grant (1174060). The Geelong Osteoporosis Study was funded by the NHMRC (628582 and 299831). Open access publishing was facilitated by Deakin University, as part of the Wiley-Deakin University agreement via the Council of Australian University Librarians.
Acknowledgments
The authors thank Professor Graham Giles of the Cancer Epidemiology Centre of The Cancer Council Victoria, for permission to use the Dietary Questionnaire for Epidemiological Studies (Version 2), Melbourne: The Cancer Council Victoria 1996. We also thank Dr Jessica Davis for calculating and interpreting the Australian Recommended Food Score. Study data were collected and managed using the Redcap electronic data capture tool hosted and managed by Research Technology Services at UNSW Sydney [47]. Open access publishing was facilitated by Deakin University, as part of the Wiley-Deakin University agreement via the Council of Australian University Librarians.
Open Research
Data Availability Statement
The findings of the study will be published in a peer-reviewed scientific journal and presented at national and international conferences.
