Introduction
In current gerontological research, a major goal is to understand how interindividual differences in cognitive functioning in old age emerge (Hultsch et al., Reference Hultsch, Hertzog, Small and Dixon1999). To address this issue, the cognitive reserve concept postulates that interindividual differences in the effective recruitment of neural networks and cognitive processes explain differences in individuals’ capacity to cope with or compensate for age-related decline (Stern, Reference Stern2009; Reference Stern2012). In individuals with healthy cognitive functioning, these mechanisms contribute to the adaptation of brain activity when task difficulty level is increased and thereby enhance cognitive performance (Stern, Reference Stern2012). Empirically corroborating the predictions of the cognitive reserve concept, evidence showed that longer education in early life, cognitively demanding occupation in midlife, and cognitive leisure activity in midlife and old age contribute to the build-up of cognitive reserve during the life course and are related to better cognitive functioning such as memory and executive functioning in old age (Gatz et al., Reference Gatz, Svedberg, Pedersen, Mortimer, Berg and Johansson2001; Karp et al., Reference Karp, Paillard-Borg, Wang, Silverstein, Winblad and Fratiglioni2006; McDowell et al., Reference McDowell, Xi, Lindsay and Tierney2007; see Hertzog et al., Reference Hertzog, Kramer, Wilson and Lindenberger2008; Adam et al., Reference Adam, Bonsang, Grotz and Perelman2013; Opdebeeck et al., Reference Opdebeeck, Martyr and Clare2016, for overviews).
Recently, a focus in research on this topic became to investigate the role of cognitive reserve in certain vulnerable (compared to less vulnerable) individuals. From a conceptual perspective following theoretical models on vulnerability and successful aging, differences between vulnerable (compared to less vulnerable) individuals seem reasonable. For instance, in the most recent overarching framework of vulnerability, Spini et al. (Reference Spini, Bernardi and Oris2017) defined vulnerability as a weakening process and lack of resources that expose individuals to greater difficulty to cope with for example health restrictions in old age. Moreover, as postulated in the Selective Optimization with Compensation (SOC) model (Baltes and Baltes, Reference Baltes and Baltes1989), a lack of resources in a given domain can (at least partly) be compensated with the help of resources in another domain. In line with the SOC model (Baltes and Baltes, Reference Baltes and Baltes1989) and the cognitive reserve concept (Stern, Reference Stern2009; Reference Stern2012), such cross-domain compensation mechanisms allow the compensation of a lack of health resources by buffering resources in terms of cognitive reserve. As postulated in the vulnerability framework of Spini et al. (Reference Spini, Bernardi and Oris2017), vulnerable individuals have greater difficulty (than less vulnerable individuals) to cope with health restrictions. Therefore, the need for cross-domain compensating resources in terms of cognitive reserve would be disproportionately higher in vulnerable (compared to less vulnerable) individuals. Empirically corroborating these conceptual assumptions, recent research suggests that cognitive functioning may more strongly depend on individuals’ cognitive reserve in older individuals with chronic diseases and in obese individuals (Robertson, Reference Robertson2013; Arenaza-Urquijo et al., Reference Arenaza-Urquijo, Wirth and Chetelat2015; Ihle et al., Reference Ihle2016; Reference Ihle2017).
According to the vulnerability framework of Spini et al. (Reference Spini, Bernardi and Oris2017), a group of particularly vulnerable individuals are physically frail older adults. A common measure to assess physical frailty is grip strength as it is indicative of overall muscle strength, loss of muscle mass (sarcopenia), functional decline, morbidity, and mortality (Syddall et al., Reference Syddall, Cooper, Martin, Briggs and Sayer2003; Chainani et al., Reference Chainani2016; Velghe et al., Reference Velghe, De Buyser, Noens, Demuynck and Petrovic2016). Evidence showed that physical frailty (in terms of e.g. lower grip strength) is related to lower cognitive status, faster cognitive decline, and greater cognitive impairment in old age (Raji et al., Reference Raji, Kuo, Al Snih, Markides, Peek and Ottenbacher2005; Deary et al., Reference Deary, Whalley, Batty and Starr2006; Buchman et al., Reference Buchman, Wilson, Boyle, Bienias and Bennett2007; Takata et al., Reference Takata2008; Auyeung et al., Reference Auyeung, Lee, Kwok and Woo2011; Yassuda et al., Reference Yassuda2012; Bjork et al., Reference Bjork, Johansson and Hassing2016; Veronese et al., Reference Veronese2016). Hence, it seems an appealing question whether the role of cognitive reserve may differ between physically frail compared to less frail individuals due to the aforementioned cross-domain compensation effects. Yet, to the best of our knowledge, no study so far targeted this question.
Therefore, the present study set out to investigate the relation of key markers of cognitive reserve (i.e. education, occupation, and cognitive leisure activity) to cognitive status in old age (i.e. Mini-Mental State Examination (MMSE)) and its interplay with physical frailty (i.e. grip strength) in a large sample of older adults.
Methods
Participants
The sample comprised 701 community-dwelling older adults from Fonte Boa, Apuí, and Manaus, Brazil. Participants were volunteers recruited via advertisements distributed through newspapers, local radio, churches, and senior centers. Inclusion criteria were living in one of the three geographic areas of Brazil, being at least 60 years old, and being able to walk independently to visit the laboratory. As we aimed to take the full range of cognitive status of participants into account no inclusion/exclusion criteria based on MMSE scores were applied. All participants gave informed consent and the present study included adherence to the declaration of Helsinki and had been approved by the local ethics commission.
Data collected
Cognitive status
We used the MMSE (Folstein et al., Reference Folstein, Folstein and McHugh1975) as an indicator of cognitive status. The MMSE requires multiple abilities such as orientation, memory, counting backwards, and language. We analyzed the MMSE total score (possible range 0–30 points) to take the full range of interindividual differences available into account.
Education
Participants were asked to indicate the total time in years they had spent for education (comprising primary school, secondary school, and university).
Cognitive level of job
Participants were asked to indicate the principal profession they had practiced during their working life, which we thereafter classified into two groups of cognitive demands: 1 = higher cognitive level (white collar jobs, e.g. clerical work, medical practice, and other occupations with a university degree) versus 0 = lower cognitive level (blue collar jobs, e.g. factory work, plumbing, carpentry, farming, etc.).
Cognitive leisure activity
Participants were asked to indicate the average time per day they have spent for cognitively stimulating leisure activity such as playing cards, scrabble, etc.; doing crossword puzzles, sudoku, etc.; reading.
Grip strength
We assessed maximum static grip strength as an indicator of individuals’ physical frailty using a dynamometer with adjusted grip. Participants stood in an upright position with the dynamometer in the preferred hand, slightly away from the body and arm straight. They had to squeeze firmly and gradually, building quickly up to maximum force. The test was performed twice, with an interval of about 10 sec between trials. The best result of the two attempts was the maximum grip strength score, recorded in kilograms (Oja and Tuxworth, Reference Oja and Tuxworth1995).
Statistical analyses
First, we inspected bivariate relations between MMSE scores, education, cognitive level of job, cognitive leisure activity, and grip strength (by calculating Pearson's correlation coefficients r). Second, we investigated whether the relation of education/cognitive level of job/cognitive leisure activity to MMSE scores varied as a function of grip strength. Specifically, we regressed MMSE scores on education/cognitive level of job/cognitive leisure activity as respective predictor plus grip strength as moderator plus an interaction term. In a subsequent set of simple slope analyses, we investigated moderation patterns in more detail. Specifically, to evaluate the direction of moderation effects we estimated the relation of education/cognitive level of job/cognitive leisure activity to MMSE scores at low and high values of grip strength (i.e. −1 SD and +1 SD, respectively) in the respective moderation model (Preacher et al., Reference Preacher, Curran and Bauer2006). All analyses were performed with the R environment (R Core Team, 2014). We applied a two-tailed significance level of alpha = 0.05 for all analyses.
Results
Descriptive statistics
Two hundred sixty eight of individuals assessed (38.2%) were men and 433 (61.8%) were women. Mean age was 70.4 years (SD = 6.9). Mean MMSE score was 24.41 (SD = 4.23). Mean education was 5.36 years (SD = 5.54). Forty three point four percent of individuals assessed had a job of higher cognitive level and 56.6% of lower cognitive level. Mean cognitive leisure activity was 10.9 min per day (SD = 37.3). Mean grip strength was 26.20 kg (SD = 7.77; men: M = 32.46 kg, SD = 7.23; women: M = 22.33 kg, SD = 5.12). Mean Body Mass Index (BMI) was 28.24 kg/m² (SD = 4.94).
Bivariate relations between measures
Longer education (r = 0.54, p < 0.001), higher cognitive level of job (r = 0.36, p < 0.001), and greater engaging in cognitive leisure activity (r = 0.09, p = 0.016) were significantly related to higher MMSE scores. In addition, greater grip strength (r = 0.23, p < 0.001) was significantly related to higher MMSE scores.
Moderation analyses
Regarding our second goal, there was a significant moderation of the relations of education, cognitive level of job, and cognitive leisure activity to MMSE scores by grip strength (see left panel of Table 1). Together with these significant moderation effects, subsequent simple slope analyses indicated that the relations of education, cognitive level of job, and cognitive leisure activity to MMSE scores were significantly larger in individuals with lower, compared to those with greater grip strength (see right panel of Table 1).
Moderation analyses with education/cognitive level of job/cognitive leisure activity as predictor of MMSE scores and grip strength as moderator
Note. Left panel: Moderation analyses with education/cognitive level of job/cognitive leisure activity as respective predictor of MMSE scores and grip strength as moderator. βP = standardized coefficient of education/cognitive level of job/cognitive leisure activity as respective predictor of MMSE scores. βM = standardized coefficient of grip strength as moderator. βI = standardized coefficient of the respective interaction term.
Right panel: Subsequent simple slope analyses to estimate the relation of education/cognitive level of job/cognitive leisure activity to MMSE scores at low and high values of grip strength (i.e. −1 SD and +1 SD, respectively) in the respective moderation model. βL standardized coefficient of education/cognitive level of job/cognitive leisure activity as respective predictor of MMSE scores, estimated at a low value (−1 SD) of grip strength. βH standardized coefficient of education/cognitive level of job/cognitive leisure activity as respective predictor of MMSE scores, estimated at a high value (+1 SD) of grip strength.
***p < 0.001; *p < 0.05; ns = non-significant, p > 0.05.
Additional control analyses
We additionally repeated moderation analyses simultaneously controlling for sex, age, and BMI. In these control analyses, the moderation of the relations of education, cognitive level of job, and cognitive leisure activity to MMSE scores by grip strength remained significant (all p's < 0.034). Together with these significant moderation effects, subsequent simple slope analyses simultaneously controlling for sex, age, and BMI again indicated that the relations of education, cognitive level of job, and cognitive leisure activity to MMSE scores were significantly larger in individuals with lower, compared to those with greater grip strength. In a further set of additional control analyses, we investigated whether moderation patterns differed by sex/age/BMI. These control analyses revealed that the previously observed pattern of larger relations of education, cognitive level of job, and cognitive leisure activity to MMSE scores in individuals with lower, compared to those with greater grip strength did not vary as a function of sex, age, nor BMI (all p's > 0.247).
Discussion
The present study set out to investigate the relation of key markers of cognitive reserve to cognitive status and its interplay with physical frailty in a large sample of older adults. First of all, results showed that greater grip strength was related to higher MMSE scores. This confirms the role of physical frailty being an important correlate of cognitive impairment in old age (Buchman et al., Reference Buchman, Wilson, Boyle, Bienias and Bennett2007; Takata et al., Reference Takata2008; Auyeung et al., Reference Auyeung, Lee, Kwok and Woo2011; Veronese et al., Reference Veronese2016). Furthermore, results showed that longer education, higher cognitive level of job, and greater engaging in cognitive leisure activity were related to higher MMSE scores. This further corroborates the conceptual view that cognitive stimulation throughout the life course may contribute to cognitive reserve, thereby being related to better cognitive status in old age (Gatz et al., Reference Gatz, Svedberg, Pedersen, Mortimer, Berg and Johansson2001; Hertzog et al., Reference Hertzog, Kramer, Wilson and Lindenberger2008; Adam et al., Reference Adam, Bonsang, Grotz and Perelman2013; Opdebeeck et al., Reference Opdebeeck, Martyr and Clare2016).
Following our second and main goal, coherently for all three investigated markers of cognitive reserve, we found that the relations of education, cognitive level of job, and cognitive leisure activity to MMSE scores were larger in individuals with lower, compared to those with greater grip strength. This suggests that cognitive status in old age may more strongly depend on cognitive reserve accumulated during the life course in physically frail (compared to less frail) individuals. Thus, present findings further corroborate prior evidence suggesting that cognitive reserve may have greater compensating effects in vulnerable (compared to less vulnerable) individuals (Robertson, Reference Robertson2013; Arenaza-Urquijo et al., Reference Arenaza-Urquijo, Wirth and Chetelat2015; Ihle et al., Reference Ihle2016; Reference Ihle2017) and that this may transfer to physically frail older adults as another group of particularly vulnerable individuals.
These findings may be explained by cross-domain compensation effects in vulnerable individuals. Specifically, in line with the SOC model (Baltes and Baltes, Reference Baltes and Baltes1989) and the cognitive reserve concept (Stern, Reference Stern2009; Reference Stern2012), these cross-domain compensation mechanisms may allow the compensation of a lack of health resources in terms of physical frailty (as well as chronic diseases and metabolic syndromes such as obesity; see Ihle et al., Reference Ihle2016; Reference Ihle2017) by buffering resources in terms of cognitive reserve. Given that vulnerable individuals have greater difficulty (than less vulnerable individuals) to cope with such lack of health resources (as postulated in the vulnerability framework of Spini et al., Reference Spini, Bernardi and Oris2017), the need for cross-domain compensation by cognitive reserve may be disproportionately higher in vulnerable (compared to less vulnerable) individuals, as empirically corroborated with present data.
In terms of possible clinical and policy implications, one could argue that the present results may underline the role of cognitive reserve accumulated during the life course to compensate for vulnerability effects on cognitive status in old age. However, given the limitations of present cross-sectional data we underline that future longitudinal studies will have to further investigate the present observations before making suggestions. In the context of the latter notions, we acknowledge that the present study is limited by its cross-sectional design that does not allow for causal inferences. The directionality of observed relationships cannot be established with present cross-sectional data. Thus, we cannot clearly disentangle between individuals who decline because of reducing cognitively stimulating activity and individuals who reduce cognitively stimulating activity because of cognitive decline. Yet, for the observed relations of educational attainment in early life as well as occupation in midlife to cognitive status in old age, it is much less likely that they are driven by individuals who reduced cognitively stimulating activity because of cognitive decline. More generally, analyses of the present study give only information about interindividual differences in cognitive status but do not allow drawing conclusions regarding cognitive decline (i.e. intraindividual changes over time), for which longitudinal research is needed. One may argue that grip strength depends on many factors such as sex, age, and BMI. Yet, controlling for these covariates in additional analyses suggests that sex, age, nor BMI influenced the pattern of observed results. We acknowledge that physical frailty was investigated only by means of grip strength. Therefore, the present study may stimulate future longitudinal research assessing decline in a comprehensive set of cognitive domains as well as different markers and determinants of frailty and cognitive reserve to gain further insights into the interplay of the build-up of cognitive reserve during the life course, frailty, and cognitive impairment in old age.
Conflict of interest
None.
Description of authors’ roles
A. Ihle formulated the research question, analyzed the data, and wrote the manuscript. É. R. Gouveia, B. R. Gouveia, D. L. Freitas, and J. Jurema conceptualized the study, supervised the data collection, and participated in writing. A. P. Odim helped with data collection and assisted with writing. M. Kliegel contributed to formulate the research question and was involved in writing the manuscript.
Acknowledgments
The authors are grateful to Maria A. Tinôco, Floramara T. Machado, and Bárbara R. Muniz for technical assistance in the data collection and management. We are especially grateful to the older people for their participation and interest. A. Ihle and M. Kliegel belong to the Swiss National Center of Competences in Research LIVES – Overcoming vulnerability: life course perspectives which is financed by the Swiss National Science Foundation (Grant number: 51NF40-160590). The authors are grateful to the Swiss National Science Foundation for its financial assistance. We are also grateful for the support from City Hall of Apuí - Secretary of Health, Amazonas.
