Background
Diabetes has become a major public health concern globally. Approximately 422 million people worldwide are living with diabetes, and the prevalence of it increases with age [
1]. A recent study indicated that approximately 11% of the population in China has diabetes [
2], ranking one of the top three countries in the world [
3]. As a leading cause of mortality, diabetes is a strong risk factor for a series of complications such as cardiovascular disease, kidney disease and neuropathy, contributing to significant burden at the individual and social level. Moreover, growing evidence have suggested the associations between diabetes-related diseases and cognitive disorders [
4,
5].
Cognitive disorders are a category of mental health disorders that primarily affect cognitive abilities such as learning, memory and perception [
6], including delirium and dementia. As the most serious stages in the development of cognitive dysfunction [
7], dementia would considerably affect the span and quality of life among adults [
8]. The consequences of diabetes and dementia present substantial individual, community, and societal impact. China has the largest population of patients with dementia in the world, imposing a heavy burden on the public and health care systems [
9]. Globally, various epidemiological studies and mechanism research have discovered the correlation between diabetes and cognitive dysfunction, demonstrating that diabetic patients have lower cognitive function than healthy individuals while having a higher risk of cognitive disorders and declines diabetic patients’ quality of life and leads to severe behavior disorder [
10‐
17]. Thus, effective interventions to prevent or reduce cognitive impairment is crucial to better care of diabetic patients. Previous research reported that the presence of diabetes was associated with an elevated risk for vascular brain damage and neurodegenerative changes [
10], and may accelerate the progression from mild stage of cognitive disorder to dementia [
3].
Physical activity (PA) is defined as bodily movement produced by skeletal muscles that results in energy expenditure [
18]. A growing body of research suggested that greater levels of physical activity could positively influence cognitive function across the lifespan, such as simple reaction time, response accuracy and working memory, and reduce the risk of cognitive decline in adult population [
19‐
21]. A recent meta-analysis of five randomized control trials and cohort studies involving 2581 patients with diabetes showed that physical activity was beneficial to improving cognition in patients with diabetes in studies of follow-up time less than 1 year (
https://doi.org/10.1002/dmrr.3443). However, the long-term effect with follow-up time over 1 year needs to be explored in future studies (
https://doi.org/10.1002/dmrr.3443). By contrast, recent RCTs have shown no benefit for exercise in cognitively healthy older adults [
22,
23]. This contradictory evidence might be due to the fact that the parameters of PA, measurement of cognition and quality of study design were various [
19]. Besides, most observational data came from western countries, and similar research among diabetic individuals in China remains scarce. Since the differences in social and cultural backgrounds, lifestyle and environmental factors may result in different implications on the diabetic-related cognitive impairment, it is necessary to explore preventive strategies among people at high risk to reduce the current and future burden of dementia in different countries.
Therefore, the primary aim of this study was to examine the effect of physical activity on cognitive function among middle-aged and older diabetic adults in China, using a 2-wave longitudinal national representative data. Besides, through the subgroup analysis, we identified whether the impact differs among various age groups. This study may be useful in helping government of China and other developing countries to improve the mental health status and quality of life among middle-aged and older diabetic populations by having a clear guidance of proceeding fitness programs.
Discussion
Using a 2 year-wave nationally representative dataset of a 2 years follow-up, we examined the effect of physical activity on different cognitive subdomains among diabetic participants above 45 years old in China. Additionally, we explored this relationship in different age groups and controlled for the effect of age, gender, living area, marital status, education, expenditure, baseline cognitive function, depression and chronic conditions. Results displayed that diabetic participants who had higher level of physical activity were associated with less decline in episodic memory function in 2 years. The associations were also significant among diabetic participants aged 45 ~ 65 years old. Our findings added to the evidence of long-term effect of physical activity on improving cognition function in patients with diabetes in a longitudinal study of two-years follow-up.
Our study found that in all the 862 diabetic participants, physical activity scores were consistently associated with episodic memory function in crude and fully adjusted analysis. Additionally, greater level of physical activity was also significantly associated with less decline in episodic memory in a 2-year period. Previous evidence on this relationship was limited and controversial. The Lifestyle Interventions and Independence for Elders (LIFE) trial enrolling adults aged 70–89 years who were sedentary reported better global cognitive function and delayed memory among diabetic individuals after physical activity intervention [
32], while two cohort studies indicated little overall associations between physical activity and higher mental function measured by Mini-Mental State Exam (MMSE) among diabetic patients [
33,
34]. Although recent reviews also reported the positive effect of physical activity on cognition among diabetic population, none of them have statistically quantified the findings into a numerical estimate of effect, and the conclusions were controversial [
35‐
37]. The inconsistency might be due to the difference in characteristics of participants, types and duration of physical activity, and measures of cognition.
Moreover, after exploring the associations between physical activity and different domain of cognitive function, we observed evidence of benefit for episodic memory rather than executive function. In CHARLS questionnaires, episodic memory mostly focused on memorizing ability, while executive function involved other mental abilities, such as reading and calculating. Previous studies showed that physical activity was linked with reduced brain atrophy, which might support the decline in executive control process and memory function among older adults [
38‐
40], and Baker, et al. also found that physical activity was associated with improvement in executive function, but not memory in 28 newly-diagnosed adults (aged 57 ~ 83 years old) [
41]. One cross-sectional study among middle-aged and old-aged participants in China reported that untreated diabetes or elevated HbA1c concentrations had a larger effect on participants’ episodic memory function rather than executive function [
10], which might partly explain our results that the effect of intervention on physical activity was mostly on episodic memory function. Differences among the health conditions of subjects, study period, included physical exercise and measurement of cognition may account for the inconsistent findings.
To assess the different effect of age in the associations, we classified the full sample into two groups (45 ~ 64, ≥65), and conducted multiple regression models separately while controlling the same covariates as in full sample analysis. Consequently, the association between physical activity and episodic memory could only be seen among middle-aged adults (aged 45 ~ 65 years) and the associations was stronger than the full sample. We observed no significant relationship among older adults aged ≥65 years. Previous research reported inconclusive findings on the effect of physical activity on cognitive function in older adults: a prospective study of rural elderly people aged over 65 years found that all levels of exercise participation could prevent the decline of cognition measured by MMSE over a 2-year interval [
42], and the meta-analyses also reported that physical exercise interventions (eg. resistance training and taichi) were effective at improving the cognitive function of older adults. However, another research in people older than 50 demonstrated little benefit of exercise (eg. resistance training and taichi) on cognitive function [
43]. Since the above studies were mostly conducted in healthy older adults with interventions restricted to only one type of exercise, the applicability of these findings to older adults with metabolic illness is less well defined. By contrast, we measured physical activity based on the metabolic equivalent (MET) multipliers of all types of exercise, considering the duration and frequency of different exercise at the same time. Therefore, our conclusions might be more robust compared with previous studies.
The possible mechanism linking physical activity and cognition among diabetic participants might primarily through biomarkers such as elevated neurotrophin levels, improved vascularization, better blood glucose other signaling pathways that could benefit brain function [
44,
45]. Among diabetic patients, it is found that energy expenditure brought by physical activity might decrease insulin resistance (IR),which contributed to cognitive impairment through a vascular mechanism [
46]. Physical activity also reduced the risk of other vascular complications regardless of the intensity and type of exercise [
33]. Since hyperinsulinemia is neurotoxic, it is possible that improved insulin sensitivity after physical activity would favor neurogenesis, whereby elevated cognitive function [
37,
47]. Further studies are needed to explore the mechanism and pathway of physical activity on effect of reduce cognitive impairment.
Our studies have several strengths. Firstly, this might be the first longitudinal study to examine the associations between physical activity and different domains of cognitive function among middle-aged and old-aged diabetic individuals in China, which provided extra evidence for the existing literature. Secondly, we conducted the study in a nationally representative cohort of community-dwelling Chinese adults, so our findings are generalizable to middle-aged and older adults in China. Finally, stringent quality control and quality assurance measures were implemented in every stage of the CHARLS study, so the quality of current study can be guaranteed. Nevertheless, some limitations should also be acknowledged. First, we could not examine the potential mediating effects or pathway from physical activity to cognitive decline among diabetic patients in the present study, because that the treatment and biological assays (eg. FPG, HbA1c) were not included in CHARLS second wave and biomarkers (such as IR, neurotrophin levels, vascularization and signaling molecules) were not tested in CHARLS survey. Second, PA related questions were not asked to all responses because of missing data on PA in CHARLS, selection bias might exist. Moreover, since the questions about physical activity in CHARLS questionnaire only included vague amount of time a person spent on different types of physical activities in a usual week, the exact amount of METs of physical activity could not be calculated. Therefore, the participants could not be categorized as previous studies [
48] and the further dose-response effect between physical activity and cognitive function could not be examined in this study. Third, since our study was conducted among Chinese population, the conclusions might not be extrapolated to other countries. Evidence from prospective cohort studies and RCT in other countries is needed to further exemplify these conclusions.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.