Relationship between physical and cognitive performance in community dwelling, ethnically diverse older adults: a cross-sectional study

Introduction

Approximately 5.4 million people, or ∼22% of the population aged 71 years (yrs.) and older in the United States are affected by cognitive impairment without dementia (Plassman et al., 2008). Age-related chronic diseases such as hypertension and cardiovascular disease (Ylikoski et al., 2000), chronic atrial fibrillation (Farina et al., 1997), and diabetes (Desmond et al., 1993) are associated with impaired cognitive function. Although encouraging evidence suggests that regular physical activity reduces the risk of age-related chronic disease and slows cognitive decline (Bherer, Erickson & Liu-Ambrose, 2013; Kirk-Sanchez & McGough, 2014; Tarumi et al., 2013; Barnes et al., 2003), evidence is lacking to prescribe physical activity to benefit cognitive function beyond ‘be active’. Since training-induced physiological adaptations responsible for improving physical performance measures are specific to the type (aerobic vs. strength), intensity, and duration of training, studies are needed to resolve the relationship between specific physical performance measures and cognitive performance in older adults. These results could be used to guide exercise prescriptions with the goal of improving cognitive performance. In addition, regular physical training improves health and the ease of performing daily physical activities.

In older adults, the ability to perform activities of daily living (ADL) correlates with performance in the 6MWT (Enright et al., 2003), a test generally well-tolerated by older adults (>68 yrs. old). The 6MWT is an exercise test in which individuals are encouraged to cover as much distance as possible at a self-selected pace, and is a measure of cardiovascular fitness and functional capacity (Rikli & Jones, 1998); functional capacity is the integrated responses of all the systems involved during exercise, including the pulmonary and cardiovascular systems, systemic circulation, peripheral circulation, blood, neuromuscular system, and muscle metabolism.

Positive associations between functional capacity and cognitive performance have been reported in cross-sectional studies of healthy older adults (Matthé, Roberson & Netz, 2015; Lord & Menz, 2002), but the physiological mechanisms responsible for aging-associated cognitive decline are not yet fully understood. Some studies reporting a relationship between functional capacity and cognitive performance have found that this relationship disappears after controlling for demographic covariates such as age (Lord & Menz, 2002) while other studies report that this relationship remains. In a cross-sectional study of 80 outpatients with stable chronic heart failure (72.4 ± 6.2) yrs.; mean ± SD), Baldasseroni et al. (2010) found a positive association between distance walked in the 6MWT and overall global cognitive function assessed in the Mini Mental State Examination even after adjusting for demographic covariates, indexes of chronic heart failure severity, comorbidities, level of disability, and quality of life.

Higher hand-grip strength, an assessment of muscular fitness (Roberts et al., 2011), has also been associated with better age-related cognitive function (Alfaro-Acha et al., 2006; Taekema et al., 2010) and self-reported measures of physical activity (Eggermont et al., 2009). A longitudinal study of cognitively healthy older adults found that a higher composite measure of muscular strength was associated with a slower rate of cognitive decline (Boyle et al., 2009) and more specifically, maximal hand-grip strength measures have been associated with better age-related cognitive function (Alfaro-Acha et al., 2006; Taekema et al., 2010). Intervention studies report that a 16 and 9 week strength training program (Moul, Goldman & Warren, 1995; Özkaya et al., 2005) resulted in improvements to cognition, proposed mechanism being earlier sensory processing and attention to external stimuli found in the strength as compared to endurance trained group (Özkaya et al., 2005).

In addition to the broader dimensional aspects of exercise programs, e.g., cardiovascular vs. strength training, exercise intensity is a key component of physical training programs (Cress et al., 2005). Physical activity (PA) performed at a higher intensity may benefit cognitive function (Angevaren et al., 2007) as has been demonstrated in longitudinal (Abbott et al., 2004; Laurin et al., 2001; Van Gelder et al., 2004) and cross-sectional works (Hogan, Mata & Carstensen, 2013). In 2,257 men aged 71–93 yrs., higher intensity walking (>3 m in less than 3 s) at a baseline measure was associated with lowered risk of dementia assessed three and six years later (Abbott et al., 2004), and high intensity PA was associated with less cognitive loss in women (>65 yrs.) over 5 years (Laurin et al., 2001). Also, low-moderate intensity PA in older men followed for 10 years was found to be associated with cognitive benefits, and cognitive decline was greater in those men whose PA intensity declined over the study period (Van Gelder et al., 2004). Previous studies have linked distance walked during the 6MWT to cognitive function (Matthé, Roberson & Netz, 2015; Lord & Menz, 2002; Baldasseroni et al., 2010), but the relationship between exercise intensity during the 6MWT and cognitive performance has not been studied.

Guidelines for monitoring exercise intensity in older adults often prescribe exercise as a percentage of maximal oxygen uptake (VO_2max), percentage of heart rate reserve, Borg Rating of Perceived Exertion (RPE), or Metabolic Equivalent of Task (MET) (Chodzko-Zajko et al., 2009; Nelson et al., 2007; Warburton, Nicol & Bredin, 2006). For ease of use in a field test, such as the 6MWT, RPE and MET stand to be good candidates although some problems exist with application of these tests in older populations. For example, although RPE is often used as a monitoring technique, a deconditioned state and a prescribed intensity of VO_2max may limit its accuracy (Grange et al., 2004; Dunbar & Kalinski, 2004). Additionally, direct measures of resting metabolic rates in older adults were 2.6 ml kg⁻¹ min⁻¹, 31.6% lower than the resting metabolic standard of 3.5 ml kg⁻¹ min⁻¹ normally used when calculating METs. This had the effect of elevating MET levels during walking conditions when VO₂ measures were being taken (Hall et al., 2013). Although distance covered is the primary recorded measure during the 6MWT, the additional use of heart rate measures provides an easily obtained and objective method to monitor exercise intensity during the test.

The Physical Activity Scale for the Elderly, a self-reported index of physical activity, has also been reported to be associated with cognitive function in a large sample of 600 older participants recruited from a longitudinal cohort even when regression models were further adjusted for cardiovascular disease risk factors (Eggermont et al., 2009). This activity scale is a short survey that has been validated in the older population, is easily scored, and represents a composite physical activity score representing household, occupational, and recreational/exercise activities over a one-week period (Washburn et al., 1993).

Experimental approach to the problem:

The rationale for this study is as follows: considering the relationship between PA and cognitive performance in older adults, and the unique relationship between performance on physical measures and the type, intensity, and duration of PA, our research objective was to assess the relationships between physical and cognitive performance in community dwelling, ethnically diverse, older adults. Since sociodemographic factors are likely to moderate these relationships, multivariate analyses were used to tease out these complex relationships.

Using a cross-sectional approach, stepwise multiple regression analyses were performed to determine which physical measure best explained performance in each of the 4 dependent variables (different model for each response variable): 3MS, Trailmaking A & B, and Animal Naming test. This study suggests that peak heart rate in the 6MWT may be used to predict cognitive status.

Materials & Methods

Participants were volunteers recruited via word-of mouth, flyers, and e-mail announcements from fitness/fall prevention programs at independent-living senior dwellings and local community senior centers, and from the general population at California State University, East Bay (CSUEB). To be included in this study, participants had to be between the ages of 60–95 yrs., nonsmokers, ambulatory (without the use of an assistive device), able to communicate (speak and read) English, and could not have uncontrolled cardiac, pulmonary, or metabolic illnesses that would contraindicate submaximal exercise testing according to guidelines established by the American College of Sports Medicine (ACSM) (Riebe et al., 2015). In this exploratory study, the primary objective was to determine which physical performance test best correlates with cognitive performance. Sample size was based on feasibility. Study methods and procedures were approved by the Institutional Review Board at CSUEB (#2014-231-F). Participants signed an informed consent for study participation and completed written questionnaires concerning their health, current medications, and PA level.

Although the participants studied here were asymptomatic for cardiovascular disease and otherwise physically and cognitively healthy, many reported taking prescribed medications to manage their cardiovascular health. We believe that healthy older adults are more likely to manage their health with medications so our choice to include these participants more accurately represents most older adults. Similar to our population, a 2003 study, found that 53.2% of adults over 65 yrs. old were taking medications to manage their cardiovascular health (Gurwitz et al., 2003). Eligible individuals participated in a single 120-minute session during which sociodemographic and anthropometric measures, distance walked in the 6MWT, maximal hand-grip strength, cognitive performance, and heart rate pre-, during, and each minute for up to 5 min post-6MWT were measured.

Results

Ninety-five participants ages 60–95 yrs. were recruited; however, 5 participants were excluded due to not meeting the inclusion criteria (n = 1), or missing test data (n = 4). Ninety participants (69 females) with a mean age of 75 ± 9.5 yrs. participated in this study. Nearly 58 percent of our population (n = 52) self-identified as white, 12 participants identified as African American, 18 identified as Asian, and 8 self-reported to be from other ethnic backgrounds. Demographic, clinical, and physical performance measures are presented in Table 1. Although males reported completing significantly more years of education (16.7 ± 3.9 vs. 14.5 ± 2.8 yrs.; p < 0.05) and generated significantly greater hand-grip strength (19.7 ± 7.9 vs. 11.9 ± 3.7 kg; p < 0.001) males vs. females, respectively), no other significant differences were found between males and females in our population. Thus, male and female data were pooled for analyses. The cognitive performance measures for all participants are shown in Table 2.

Discussion

Main findings of the current study were: (1) exercise intensity (6MWT HR_PEAK) recorded during the 6MWT was correlated with better performance on a range of cognitive tests (3MS and TMT A & B); this correlation remained after adjusting for sociodemographics, (2) higher PA levels reported on the PASE was correlated with better performance on verbal fluency (Animal Naming) and TMT A, (3) distance walked in the 6MWT was associated with better performance on cognitive tests; however, these relationships were no longer evident after controlling for sociodemographics (age, ethnicity, education), (4) no relationship was found between maximal hand-grip strength and performance on cognitive tests, and (5) a negative correlation was found with β-blocker use and verbal fluency (Animal Naming).

In the current study, higher HR_PEAK during the 6MWT was positively correlated with cognitive performance in the 3MS and TMT A and B. These findings are in agreement with evidence that has shown intensity of PA positively influences cognitive abilities (Van Gelder et al., 2004; Angevaren et al., 2007; Brown et al., 2012). Longitudinal studies found that higher self-reported PA intensity preserved cognitive function in middle aged and older adults tested over a four (Angevaren et al., 2007) and ten-year follow-up period (Van Gelder et al., 2004). In a cross-sectional study, direct measures of daily PA intensity were assessed with an Actigraph monitor and those in the higher intensity tertiles scored higher on cognitive tests (Brown et al., 2012). Our study findings are unique in that peak heart rate recorded during the 6MWT provided a direct and easily obtained measure of exercise intensity that was correlated with cognitive performance measures.

In the bivariate correlations performed here, distance walked in the 6MWT was correlated with cognitive performance (3MS, Animal Naming, and TMT A & B), but after adjusting for confounding variables such as age, distance walked was no longer associated with global cognitive functioning. These results are consistent with those of Lord & Menz (2002) in which the association between cognitive and physical performance disappeared after controlling for age. Matthé, Roberson & Netz (2015) included participants using walking aids, and found that those participants who did not use walking aids, covered more distance in the 6MWT and scored higher on the MMSE. However, sample size was much smaller (n = 27) and confounding variables such as age were not controlled. In addition, our population may have been more physically homogeneous than populations in previous studies, thus associations with PA may have been more difficult to resolve in this study.

Physical activity as self-reported in the PASE maintained a positive association with Animal Naming performance after controlling for age, ethnicity, and education level. This association might be partially explained by the positive influence of PA on psychomotor function in older adults (Rodríguez-Aranda et al., 2006; Rosano et al., 2010). Evidence of an age-associated slowing of psychomotor function has been demonstrated as the slowing of handwriting, reading speed and time required to say a particular word in older participants (Eggermont et al., 2009). Additionally, PA may be associated with the frequency of social interactions (e.g., group exercise, exercise with a friend) and a higher PASE score may reflect less social isolation. This is in part supported by McAuley et al. (2003) who found self-reported PA levels assessed by PASE to be influenced by perceptions of social support with higher perceived support associated with increased self-efficacy and improved exercise adherence.

Maximal hand-grip strength was not found to be associated with cognitive performance, a finding that remains equivocal in the literature. Systematic reviews support the use of hand-grip measures to predict numerous outcomes, including mortality, functional limitations, activity levels, and independence, these measures most likely reflecting multidimensional variables that contribute to future risk in older adults (Bohannon, 2001; Bohannon, 2008). Cross-sectional and longitudinal studies suggest that muscular strength measures are positively associated with cognitive performance (Enright et al., 2003; Lord & Menz, 2002; Alfaro-Acha et al., 2006; Taekema et al., 2010; Ling et al., 2010). Longitudinal studies point to hand-grip as a significant predictor of cognitive decline in 555 participants, 85 years of age, followed for 4 years (Taekema et al., 2010) and in 2,160 older Mexican Americans 65 years and older followed for a 7-year period (Alfaro-Acha et al., 2006), however this association was no longer significant after controlling for age (Lord & Menz, 2002). Also, some participants in our study complained of limited ability to perform the hand-grip strength test due to arthritic conditions, which may have limited our ability to detect an association between grip strength and cognitive performance. Perhaps with a larger sample and/or a longitudinal study, a temporal association would emerge.

The negative correlation between β-blocker use and verbal fluency is an interesting one and warrants further investigation. Similar to results from Qato et al. (2008), ∼83% (75 of 90 participants) reported regular use of medications to manage chronic health conditions, and ∼21% of these participants (16 of 75 participants) reported using β- blockers. Beta blockade has been found to abolish the stress induced rise in heart rate (Benschop et al., 1994) and when administered to students with paralyzing test anxiety, improved Scholastic Aptitude Test scores (Faigel, 1991). Further benefit has been attributed to tests of cognitive flexibility (Alexander et al., 2007), but the authors are not aware of research links to verbal fluency.

Although the current study was not designed to test a specific mechanism, possible explanations exist and justify further research. Regular exercise at higher intensities improves cardiovascular health, cardiorespiratory fitness, lipid profiles, and is believed to benefit cerebral blood flow (Van Gelder et al., 2004; Kraus et al., 2002; Rogers, Meyer & Mortel, 1990), a theory in part supported by a cross-sectional study showing that cerebral blood flow reductions in sedentary older adults are associated with impaired cognitive performance (Rogers, Meyer & Mortel, 1990). Additional mechanisms include the exercise-stimulated rise in growth factors believed to promote hippocampal synaptic plasticity and neurogenesis (Cotman, Berchtold & Christie, 2007; Gómez-Pinilla, So & Kesslak, 1998; Van Praag et al., 1999).

In our study, a higher peak exercise intensity during the 6MWT was associated with better cognitive performance. The 6MWT is a test of functional capacity, and as such depends on the health and functioning of multiple systems (Stewart et al., 2000) including the cardiovascular and muscular systems. Our results show that a higher exercise intensity during an acute exercise test (i.e., the 6MWT) is correlated with better cognitive performance. We assume that higher heart rate during the 6MWT is indicative of chronic physical training effects on the heart. However, additional studies are needed to test whether chronically training at higher intensities benefits cognitive performance in older adults.

In this population of ethnically diverse older adults, ethnicity correlated with performance in the 3MS and Animal Naming even after controlling for self-reported education level. Work by Manly et al. (2003) suggests that in multi-ethnic older adults, cognitive test performance should be interpreted relative to education quality, not years of formal education. In a population of ethnically diverse community dwelling older adults, Manly et al. (2002) found that African American older adults obtained significantly lower scores than whites on measures of word list learning and memory, figure memory, abstract reasoning, fluency, and visuospatial skill even though the groups were matched on years of education. However, after adjusting for education quality, the overall effect of ethnicity was greatly reduced and ethnicity differences on tests became nonsignificant. Culture, ethnicity, and poverty affect education quality so that years of education and quality of education are not comparable in a population of multi-ethnic older adults (Manly et al., 2005).

It is worth noting that despite reaching significance, the standardized coefficients (β-weights) indicated a moderate ability to predict cognition, a finding that is consistent with other cross-sectional (Sprague, Watts & Burns, 2016; Hultsch, Hammer & Small, 1993; Baldasseroni et al., 2010; Angevaren et al., 2007) and longitudinal studies (Aichberger et al., 2010). Distance walked during the 6MWT was found to be a significant predictor of MMSE in patients with chronic heart failure (β = 0.20) (Baldasseroni et al., 2010), and selected PASE items predicted cognition (β-weights ranging from 0.30–0.31) (Sprague, Watts & Burns, 2016) as well as self-reported measures of activity frequency (β-weight of 0.019–0.262 (Hultsch, Hammer & Small, 1993). Angevaren et al. (2007) and Aichberger et al. (2010) found intensity to be predictive of cognition with β-weights that ranged from 0.16–0.46. These findings are consistent with ours and suggest that the relationship between physical activity or exercise intensity, and cognitive performance is complex and not fully understood, with ethnicity in the current study modifying this relationship. In conclusion, although the 6MWT is a practical measure, there are inherent limits as to what can be assessed and additional studies are needed to elucidate other contributors.

Study limitations include the cross-sectional study design and homogeneity of our population. It is possible that physical exercise is a marker for a healthy lifestyle; better overall health linked to cognitive health, however our data are unable to ascribe direction of causality. Additionally, many of our participants were recruited from local fall prevention/fitness programs and thus likely to represent a segment of more active persons aged 60 to 95 years. Other limitations include that we did not test for educational quality in our sample of multi-ethnic older adults, and facility and participant scheduling constraints precluded us from offering an acclimation period for the 6MWT. Finally, while our findings show that a higher intensity (HR _PEAK) during the 6MWT was associated with better performance on a range of cognitive tests, we did not control the intensity at which participants walked but allowed them to self-pace during the 6MWT.

Conclusion

Present study found a lack of association between grip strength and distance walked, and cognition, and β-blocker use negatively correlated to verbal fluency. Authors speculate that arthritis in participants’ hands limited grip strength, and our homogeneous sample prevented distance walked from being isolated as a contributor to cognitive performance. Additionally, authors are not aware of previous research linking β-blocker use to verbal fluency, although Faigel (1991) reports a reduction in test anxiety and improved Scholastic Aptitude Scores. These findings warrant future testing that controls for these variables. Self-reported PASE measures were associated with Animal Naming perhaps reflecting reduced social isolation and slowing of psychomotor function associated with physical activity.

Key to this study, was to identify safe, inexpensive and field-amenable methods to monitor cognition in community dwelling older adults. A submaximal field test, such as the 6MWT, may be part of that solution. Peak exercise intensity during the 6MWT is a simple and reliable measure of functional capacity, and is related to cognitive performance in multi-ethnic, community dwelling older adults (ages 60–95 yrs.). This is promising because it suggests that a submaximal field test may be used to monitor age-related changes in executive function and higher order processing. Thus, physicians, exercise professionals, and/or fitness/fall prevention programs could use this short, physical test to monitor cognition and motivate people to remain active with age.

Supplemental Information