Physical activity and the risk of dementia in end-stage renal disease patients undergoing hemodialysis: a nationwide population-based study
Article information
Abstract
Background
Several studies have reported that dementia has a high prevalence in end-stage kidney disease (ESKD) patients. However, the relationship between physical activity (PA) and the risk of dementia has not been elucidated for hemodialysis patients.
Methods
A total of 11,724 patients aged ≥40 years who started hemodialysis between 2012 and 2017 were identified from the Korean National Health Insurance Service database. Individuals with PA were defined as meeting the following criteria: 1) 1 or more days per week of vigorous activity of at least 20 minutes per day or 2) 1 or more days per week of moderate-intensity activity of at least 30 minutes per day. The occurrence of dementia was monitored until the end of 2018 based on specific codes of International Classification of Diseases, 10th Revision.
Results
During the 1.9-year follow-up, 489 hemodialysis patients developed dementia. PA was associated with a lower risk of any dementia (hazard ratio, 0.686; 95% confidence interval, 0.553–0.85) even after adjusting for confounding factors. The risk of dementia was lower in hemodialysis patients with PA when the risk of Alzheimer disease was analyzed separately, but vascular dementia was not. PA was associated with a lower risk of dementia in a dose-dependent manner when stratified by the energy expenditure level. In subgroup analyses stratified by age, sex, income level, smoking, drinking, diabetes mellitus, hypertension, dyslipidemia, and cerebrovascular accident, the risk of dementia tended to be lower in hemodialysis patients with PA than in those without PA.
Conclusion
PA was associated with a lower risk of dementia development in ESKD patients undergoing hemodialysis.
Introduction
Chronic kidney disease (CKD) is a pervasive global health concern, affecting millions of individuals worldwide and imposing a substantial burden on healthcare systems [1]. Among the advanced stages of CKD, end-stage kidney disease (ESKD) is characterized by a significant decline in kidney function, necessitating renal replacement therapies such as hemodialysis. ESKD accounts for a significant proportion of the world disease burden, which not only threatens global health but also contributes to increased socioeconomic costs [1]. Hemodialysis, the most widely used renal replacement therapy, is a very expensive treatment modality [2]. The number of ESKD patients undergoing hemodialysis is increasing markedly. Patients undergoing hemodialysis are exposed to various systemic challenges, including cardiovascular comorbidities and cognitive impairment.
Cognitive impairment, particularly dementia, is a multifaceted issue that further compounds the complexity of managing ESKD. Dialysis patients have a high prevalence of cognitive impairment even at young ages [3,5]. It has been reported that patients of all ages receiving hemodialysis have decreased cognitive function compared with that of the general population [5,6]. In addition, not only is the number of elderly people undergoing hemodialysis continuously increasing, but the life expectancy of hemodialysis patients is also increasing, thus increasing the average age of dialysis patients [7].
Physical activity (PA) has emerged as a modifiable lifestyle factor with the potential to mitigate cognitive decline in various populations. PA has been reported as an effective nonpharmacological intervention to preserve global cognitive function [8], especially executive function in community‐dwelling older adults [9,10]. However, the relationship between PA and dementia in ESKD patients undergoing hemodialysis has not been comprehensively examined on a nationwide scale. Understanding this association is crucial not only for enhancing the holistic care of ESKD patients but also for devising preventive strategies that may alleviate the cognitive impact of this debilitating condition. In this context, we aimed to investigate the association of PA status with the risk of dementia in patients undergoing hemodialysis. To better elucidate the relationship between PA and dementia in hemodialysis patients, we analyzed large-scale nationally representative data from the Korean National Health Insurance Service (NHIS).
Methods
Data source and study population
Information on the Korean NHIS has been published previously [11]. We recruited participants who developed ESKD between 2012 and 2017 using the codes of International Classification of Diseases, 10th Revision (ICD-10) and rare intractable disease codes for hemodialysis, peritoneal dialysis, and kidney transplantation (V001, V003, and V005, respectively). The detailed definition of newly developed ESKD was as follows: 1) patients who have been prescribed a fee code for doing HD (O7020) or fee code of materials used for HD (O7021), for more than 90 days continuously and 2) patients who have been prescribed rare intractable diseases code for hemodialysis (V001) for more than 90 days [11]. Among the ESKD patients, we selected those who had undergone hemodialysis only from the beginning of renal replacement therapy (n = 73,531). Then, we excluded patients who did not receive at least one health examination following the development of ESKD. We also excluded individuals under 40 years of age, with missing data such as demographic variables, previous disease histories, or health examination data, and who converted to peritoneal dialysis or kidney transplantation. Additionally, we excluded individuals who had dementia previously and had developed dementia within 1 year of the start of hemodialysis. Finally, we enrolled 11,724 hemodialysis patients in this study (Fig. 1).
Flow diagram of the study population.
ESKD, end-stage kidney disease; NHIS, National Health Insurance Service.
Definitions of physical activity
Data on PA were obtained from a health examination questionnaire. A standardized self-reported questionnaire on PA using a 7-day recall method was distributed at the time of the health checkup (Supplementary Fig. 1, available online). The questionnaire was in Korean and consisted of three items similar to the assessment by Smith et al. [12], and the validity and reliability have been established [13]. PA level was classified as “active” or “inactive” based on the number of days per week that patients engaged in PA similar to previous research [14,15]. Active was defined as 1) 1 or more days per week of vigorous activity of at least 20 minutes per day or 2) 1 or more days per week of moderate-intensity activity of at least 30 minutes per day. Regular PA was defined as 1) 3 or more days per week of vigorous activity of at least 20 minutes per day or 2) 5 or more days per week of moderate-intensity activity of at least 30 minutes per day.
To estimate PA quantitatively, we calculated the metabolic equivalent of tasks (METs) of participants, expressed as MET-min/week. We calculated vigorous PA, moderate-intensity PA, and walking as 8, 5, and 3 METs, respectively, according to the 2011 updates on compendium of PA [16]. Using the minimum time consumed for each PA category, we calculated the overall amount of weekly PA (MET-min/week) as follows: (8 MET × 20 minutes × a) + (5 MET × 30 minutes × b) + (3 MET × 30 minutes × c), where a, b, and c are the frequencies of vigorous PA, moderate PA, and walking, respectively [17]. Subsequently, MET-min/week was categorized into four groups: <500, 500–999, 1,000–1,499, and ≥1,500 MET-min/week.
Definitions of variables and outcomes
Data on smoking and status of alcohol consumption were obtained from a health examination questionnaire. A standardized self-reported questionnaire was used for the following variables: alcohol consumption (none, mild: <30 g of alcohol/day, heavy: ≥30 g of alcohol/day) and smoking status (never, former, current). Body mass index (BMI) was calculated as the subject’s weight in kilograms divided by the square of the subject’s height in meters. Blood samples for the measurement of serum total cholesterol and glucose were drawn after an overnight fast. The above variables were extracted from health examination data provided by the participants to the NHIS biennially. Hospitals that performed these health examinations were certified by the NHIS and subjected to regular quality control evaluations. The level of income was divided into quartiles, and the lowest quartile was defined as low income. The presence of diabetes mellitus was defined according to the following criteria: at least one claim per year under ICD-10 codes E11–14 and at least one claim per year for the prescription of antidiabetic medication or a fasting serum glucose level ≥126 mg/dL in the health examination database. The presence of hypertension was defined according to the presence of at least one claim per year under ICD-10 codes I10–I13 and I15 and at least one claim per year for the prescription of an antihypertensive agent or systolic/diastolic blood pressure ≥140/90 mmHg in the health examination database. The presence of dyslipidemia was defined according to the presence of at least one claim per year under ICD-10 code E78 and at least one claim per year for the prescription of a lipid-lowering agent or a total cholesterol level ≥240 mg/dL. The presence of previous cerebrovascular accident (CVA) was defined according to the presence of at least one claim per year under ICD-10 codes I60–I64.
The study population was followed up from the time of baseline measurement until the date of dementia diagnosis or December 31, 2018, whichever came first. The primary endpoint was incident dementia. The development of dementia was monitored until the end of 2018 with the ICD-10 codes F00, F01, F02, F03, G30, and G31. Alzheimer disease was identified with the ICD-10 codes F00 and G30, and vascular dementia was identified with the ICD-10 code F01. A list of ICD-10 codes used to define variables and outcomes is provided in Supplementary Table 1 (available online).
Statistical analyses
Continuous variables are presented as mean ± standard deviation, and categorical variables are presented as number (%). Inter-group differences were estimated using chi-square test or analysis of variance, as appropriate. The incidence rates of dementia are presented per 1,000 person-years. Multivariable Cox proportional hazard regression analysis was used to estimate the hazard ratio (HR) and 95% confidence interval (CI) of the risk of dementia associated with PA along with adjustment for age, sex, smoking, alcohol consumption, income status, and previous history of hypertension, diabetes mellitus, dyslipidemia, and CVA. Sensitivity analyses were performed with the exclusion of subjects who underwent peritoneal dialysis or kidney transplantation during the follow-up period. Subgroup analyses were performed to assess the effect modification on the risk of dementia in patients undergoing hemodialysis according to age (<65 and ≥65 years), sex, income status, smoking, alcohol consumption, history of hypertension, diabetes mellitus, dyslipidemia, and CVA. Interaction terms were added to test for effect modification across subgroups. Statistical analyses were performed using SAS version 9.3 (SAS Institute), and a p-value of <0.05 was considered to indicate statistical significance.
Ethical approval
The requirement for ethical approval for this study was waived by the Institutional Review Board of Chonnam National University Hospital (No. CNUH-EXP-2022-267). The requirement for obtaining informed consent was also waived; hence, consent was not obtained because the participants’ records and information were anonymized and de-identified prior to analysis.
Patient and public involvement
Patients or the public were not involved in the design, conduct, reporting, or dissemination plans of this research.
Results
Baseline characteristics of study population
The baseline characteristics of the participants according to their PA status are shown in Table 1. Active patients were more likely to be younger and male. In addition, active patients were less likely to have diabetes mellitus, dyslipidemia, CVA, and low income. In comparison with physically inactive patients, active patients had a lower waist circumference; however, there was no significant difference in BMI between the groups. Active patients were likely to have lower fasting glucose, systolic blood pressure, and triglyceride levels and higher high-density lipoprotein cholesterol level.
Baseline characteristics of the study population according to physical activity status
Risk of dementia according to physical activity status in hemodialysis patients
During the 1.9 years of follow-up, a total of 489 incident cases of dementia occurred. Data on the association between PA status and the risk of dementia in hemodialysis patients are presented in Table 2. The incidence rates of any dementia, Alzheimer disease, and vascular dementia were lower in active patients than in inactive patients. In Cox proportional hazard regression analysis, active patients had a significantly lower risk of dementia (adjusted HR, 0.69; 95% CI, 0.55–0.85). The risks of Alzheimer disease were also lower in active patients (adjusted HR, 0.70; 95% CI, 0.55–0.89) (Table 2). The risk of vascular dementia tended to be lower in active patients than in inactive patients but was not statistically significant after adjustment for covariates (adjusted HR, 0.57; 95% CI, 0.30–1.09) (Table 2). For sensitivity analyses, we analyzed the association between PA status and the risk of dementia in hemodialysis patients after excluding patients who underwent peritoneal dialysis or kidney transplantation during the follow-up period (Supplementary Table 2, available online), which were also performed using more strict criteria for PA (regular PA) (Supplementary Tables 3 and 4, available online). The association between PA status and the risk of dementia was consistent in both sensitivity analyses. Kaplan-Meier analysis showed a significant difference in the risk of dementia according to PA status (Fig. 2).
Incidence rates and HRs for dementia according to PA status
Cumulative probability of dementia development in hemodialysis patients according to PA status.
(A) Any dementia. (B) Alzheimer disease. (C) Vascular dementia.
PA, physical activity.
We analyzed the association between PA status and the risk of dementia in hemodialysis patients according to PA-related energy expenditure (MET-min/week), divided into <500 MET-min/week, 500–999 MET-min/week, 1,000–1,499 MET-min/week, and ≥1,500 MET-min/week (Table 3). The risk of any dementia and Alzheimer disease tended to be lower with increasing MET-min/week, and the association was dose-dependent (Table 3).
Incidence rates and HRs for dementia according to study participants’ energy expenditure assessed by MET
Subgroup analyses of risk of dementia in hemodialysis patients according to physical activity
We further analyzed the association between PA status and the risk of dementia according to subgroups stratified by age, sex, income, smoking and drinking status, hypertension, diabetes mellitus, dyslipidemia, and CVA (Fig. 3). The risk of any dementia was lower in active patients than in inactive patients in all subgroups, except for current smokers and heavy drinkers (Fig. 3). This tendency was also observed for the risk of Alzheimer disease (Fig. 3).
HR (95% CI) for dementia according to physical activity status in subgroups.
(A) Any dementia. (B) Alzheimer disease. (C) Vascular dementia.
CI, confidence interval; CVA, cerebrovascular accident; DM, diabetes mellitus; HR, hazard ratio; NA, not available.
Discussion
In the present nationwide population-based study, we investigated the effect of PA on the incidence of dementia in hemodialysis patients. PA was significantly associated with a lower risk of any dementia and Alzheimer disease, but not with vascular dementia. When the effect of PA on dementia was analyzed according to the level of energy expenditure, PA was associated with a lower risk of dementia in a dose-dependent manner. Furthermore, PA was associated with a lower risk of dementia in patients undergoing hemodialysis in most subgroup analyses. In this study, we intentionally excluded dementia that developed within 1 year of starting hemodialysis to avoid misattributing dementia due to causes other than hemodialysis as the outcome. Nevertheless, it cannot be denied that the risk of dementia even in the first year of hemodialysis is high. Previous study using the U.S. Renal Data System reported that 4.6% of women and 3.7% of men who are older than 66 years old were diagnosed with dementia within the first year after initiating hemodialysis [18].
Cognitive impairment and executive dysfunction can cause major functional problems in dementia and affect the lives of hemodialysis patients. Existing studies have indicated that the cognitive function of hemodialysis patients is considerably poor. McAdams-DeMarco et al. [5] showed that the global cognitive function of hemodialysis patients with a mean age of 55 years was as low as that of the most cognitively vulnerable community‐dwelling older adults aged 60 years and older. Kurella et al. [19] reported that only 13% of hemodialysis patients had normal cognitive function. Some studies found that the cognitive function of patients starting hemodialysis may already be impaired and may worsen as dialysis continues [20,21]. In particular, patients with cognitive impairment are less likely to undergo peritoneal dialysis or kidney transplantation, which may be one of the reasons why hemodialysis patients have poor cognitive function [22]. Executive function is known to be the poorest among the various cognitive domains of patients with ESKD [23]. In comparison with the general population of the same ages, hemodialysis patients have been found to have a 3‐fold higher rate of executive function impairment [4]. Impairment of executive function makes it difficult for hemodialysis patients to perform daily functions such as maintaining independent living, controlling diet, taking complex medications, and managing dialysis schedules [24]. Such problems ultimately lead to increased patient mortality [25,26].
The cause of the high prevalence of dementia and cognitive impairment in dialysis patients has not been clearly identified. However, risk factors such as a high frequency of CVAs and increased vascular stiffness are presumed to contribute to the development of dementia [27,28]. In addition, white matter disease and cerebral atrophy are much more common in hemodialysis patients than in patients without kidney disease, and the prevalence of unrecognized infarction is high [29]. It has been reported that the accumulation of uremic toxin or cerebral ischemia due to recurrent hypotension in hemodialysis patients and activities such as sleeping or watching TV during hemodialysis also contribute to cognitive dysfunction [4,30,31]. When renal function is restored through kidney transplantation, cognitive function is partially recovered, which could provide an indirect glimpse into the effects of renal function decline and dialysis itself on cognitive function [32,33].
Given that dementia can increase the mortality of hemodialysis patients, doctors treating these patients should pay more attention to the management of dementia [26]. Our study suggests increasing PA as one of the ways to manage dementia and improve the prognosis of hemodialysis patients. Our study showed that PA was associated with a lower risk of dementia in hemodialysis patients, even when adjusting for age, gender, or underlying diseases such as diabetes mellitus and hypertension, which may affect PA. This association appeared stronger as PA-related energy expenditure (expressed as MET-min/week) was increased. A similar trend was observed in subgroup analyses; however, no association between PA and dementia was found among current smokers and heavy drinkers. This result is somewhat expected because smoking and heavy drinking are strong risk factors for dementia and may have a negative effect on the relationship between PA and dementia [34]. The presence of these two well-known dementia risk factors may have prevented the risk of dementia from being reduced despite the significant positive effects of PA. Several studies have reported that exercise could contribute to the preservation of cognitive function in dialysis patients [35–37]. Manfredini et al. [36] reported that a 6-month home-based personalized exercise program significantly improved the self-reported cognitive function and quality of social interaction of dialysis patients. In a secondary analysis of the same study, exercise was found to preserve cognitive function in elderly hemodialysis patients aged 65 years or older [35]. Stringuetta et al. [37] reported that intradialytic exercise with cycling increased cerebral blood flow and cognitive function. Additionally, some studies showed that erythropoietin-stimulating agents or vitamin D supplements contributed to an improvement in dementia in hemodialysis patients [38,39]. However, no study has reported that PA is associated with a lower risk of dementia in hemodialysis patients. This study is the first large-scale nationwide study to report that PA may reduce the risk of dementia in hemodialysis patients. In large-scale cohort studies targeting dialysis patients, it is difficult to collect data on cognitive function. Our study overcame these hurdles by utilizing well-collected nationwide claims data.
There are several points of particular interest in our study. First, our study showed that a higher level of PA was associated with a lower risk of dementia in hemodialysis patients. However, even a low level of PA also showed a tendency to reduce the risk of dementia. This is consistent with the results of a previous study targeting the general population, which showed that low-intensity exercise such as walking could reduce the risk of dementia [40]. Considering that hemodialysis patients often have difficulty performing high-intensity exercise due to multiple comorbidities, our results are of significant value. Second, in subgroup analyses, PA was associated with low dementia risk especially in elderly patients over 65 years of age. Age is the strongest risk factor for dementia. However, according to our study results, even in the elderly population, PA may help improve the prognosis of hemodialysis patients in terms of cognitive function. Third, the association of PA with vascular dementia appeared to be relatively weaker than that with any dementia or Alzheimer disease. In particular, subgroup analyses showed somewhat inconsistent results for each subgroup. It is presumed that the reason for the heterogeneous results is that patients receiving hemodialysis are likely to have advanced vascular problems. Nevertheless, our study is the only study that compared and analyzed Alzheimer disease and vascular dementia in hemodialysis patients.
Although this study provides valuable insights, there are some limitations. First, we identified the occurrence of dementia using ICD-10 codes. However, there is a possibility of underestimation or overestimation in selecting dementia patients using only the diagnosis indicated by the ICD-10 codes. The diagnosis of dementia could be delayed or missed by physicians especially in ESKD patients and early symptoms are vague and undetectable. It may take several years from the onset of dementia symptoms to diagnosis. Second, this study retrospectively formed a virtual cohort from a large database and tracked the occurrence of outcomes. However, there are still limitations in asserting the causality between PA and dementia. Undiagnosed dementia might have led to decreased PA and incident dementia by national data may not be consistent with the exact time of dementia onset. Third, the self-reported questionnaire used to define PA is known as a relatively accurate evaluation tool [13]; however, there may be bias in patient’s recall. Lastly, any dementia is simply a collection of all diseases grouped under the ICD-10 code. Dementia is the collection of symptoms accompanied by cognitive and memory dysfunction and has many heterogeneous etiologies. However, we wanted to emphasize that dementia, a degenerative mental health disease, is important in the care of hemodialysis patients.
In conclusion, PA was significantly associated with a lower risk of dementia in hemodialysis patients. PA was significantly associated with a lower risk of Alzheimer disease but not vascular dementia. Therefore, PA recommendation for hemodialysis patients within the range permitted by the condition of patients could improve prognosis. Physicians caring for hemodialysis patients need to actively provide not only medical treatments but also non-pharmacologic treatments, such as exercise.
Supplementary Materials
Supplementary data are available at Kidney Research and Clinical Practice online (https://doi.org/10.23876/j.krcp.24.197).
Notes
Conflicts of interest
All authors have no conflicts of interest to declare.
Funding
This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (RS-2023-00217317 and 2022R1C1C1007573) and a grant (BCRI23046) from Chonnam National University Hospital Biomedical Research Institute.
Data sharing statement
Anonymized data are publicly available from the National Health Insurance Sharing Service and can be accessed at https://nhiss.nhis.or.kr/bd/ab/bdaba000eng.do.
Authors’ contributions
Conceptualization: HSC, SHS, CSK, EHB, SKM, SWK
Data curation, Formal analysis: BK, KDH
Funding acquisition: HSC, SWK
Supervision: SWK
Writing–original draft: HSC
Writing–review & editing: HSC, SHS, CSK, EHB, SKM, SWK
All authors read and approved the final manuscript.