Body Mass Index, Performance of Activities of Daily Living and Cognition: Analysis in Colombian and Korean Older Adults

Background: With this study, we aim to describe the associations of the different categories of the Body mass index (BMI) with ADL and cognitive performance in Colombian and South Korean older adults living in the community. Methods: A cross-sectional analysis of two surveys were analyzed; Survey on Health, Well-Being, and Aging in Colombia (SABE) (n= 23448) and Korean Longitudinal Study of aging (KLoSA) (n=4501). Participants older than 50 years were selected from rural and urban areas achieving a representative sample. Here we investigate the association between BMI categories with function, using zero-inated negative binomial regression, and cognition, using a logistic regression model. Results: In Colombia, underweight was associated with an impaired score on the MMSE and worse performance in the IADL. Being overweight was associated with better scores on the MMSE and the IADL. For both outcomes education was highly inuential. In Korea, after adjustments, there were no signicant associations for any of the outcomes. Conclusions: In Colombian population, underweight, in particular, seems to be negative to preserve both cognitive and daily functioning. Additionally, being overweight but not obese was associated with less cognitive impartment and better daily functioning.


Introduction
Due to health and social advances, older adults are living longer, and the population older than 60 is increasing [1]. This is one of the greatest human achievements. However, to keep older adults living longer but also healthy and free of dependency represents several challenges to societies. Independence in activities of daily living is an important marker of health in older adults [2]. Autonomy is the ability of a person to satisfy their needs, independently and satisfactorily.
Functional deterioration in older adults is becoming more relevant as a major public health problem since functional decline is a common endpoint of multiple comorbidities and is directly associated with frailty, poor quality of life, dependency, and high costs to health systems [3].
Another important component of older adult's health is cognition. Cognitive impairment, including dementia, is a very important risk factor for dependence, poor quality of life, and death [2,4,5]. Therefore, maintaining normal cognition and performance in Activities of daily living (ADL) should be priorities in health systems [6,7].
Maintaining an adequate nutritional status has been pointed out as an important intervention to preserve well-being and health in older adults, and several studies have related nutrition with cognition and function [8,9]. Studies in younger populations have generally shown that overweight and obesity predispose to unfavorable conditions such as cardiovascular diseases, mobility issues, depression, anxiety, dementia, and poor quality of life and death [10]. However, recent studies show that this may be different in older populations, where undernutrition is the main factor associated with negative outcomes.
Particularly, evidence has shown that a "U shaped '' relationship exists between nutritional status and unfavorable outcomes, where the people in the extremes (i.e. underweight and obesity) represent the groups with a higher risk of mortality [11,12]. Similar associations have been found for other frequent geriatric conditions such as frailty [013]. However, most of these studies are performed in developed western countries, and there is a lack of evidence regarding performance in activities of daily living.
This study describes the associations of Body mass index (BMI) categories with ADL and cognitive performance in Colombian and South Korean older adults living in the community. Korea and Colombia are countries with very different diets and cultures, thus we included both populations to evaluate if the effect of the BMI could be different.

Setting and participants
This is a secondary analysis of two national studies: The Health, Well-Being, and Aging (SABE) Colombia study and the Korean Longitudinal Study of Aging (KLoSA). The studies were designed to determine the factors that characterize aging in these countries.
In both studies, face-to-face interviews were conducted, and subjects were given sets of questionnaires, concerning sociodemographic characteristics, health-related issues, lifestyle habits, and cognitive function. Participants with incomplete data or those who refused follow-up could not be included in the analysis. SABE was performed in 2015 with a representative sample of community-dwelling Colombian older adults (age ≥ 60 years).
As of 2020, this is the largest database available regarding Latin American older adults. A set of questionnaires on different topics (socio-demographic characteristics, health-related issues, access to health services, cognitive performance, functional status, and nancial resources) was applied to all the participants by interviewers at the older adult's household. A total of 3,694 older adults were surveyed for an effective national response rate of 66%. Complete methodology, processes, and objectives are available elsewhere [14].
KLoSA started in 2006, with follow-up every 2 years. A strati ed multistage probability sampling was used to obtain a representative sample. The analysis was made on wave 2016. The recruitment of participants and methods used in KLoSA has been described in detail elsewhere [15,16].

Basic Activities of Daily Living:
The basic activities of daily living (BADL) refer to self-care and mobility, and their deterioration is closely related to clinical complications, geriatric syndromes, frailty, and dependency [19].
In SABE, BADLs were assessed using the Barthel scale (0-100) [20]. The KLoSA adapted existing BADL and IADL instruments to assess the functional status of the community-dwelling adult population. The KLoSA consists of 7 BADLs items, including dressing, washing the face, bathing, eating, getting out of bed, toileting, and bladder/bowel management. All BADLs variables from KloSA were dichotomized (not need any help = 0 and any kind of help =1) and summed [21,22].

Instrumental Activities of Daily Living:
The instrumental activities of daily living(IADLs) refer to the individual's ability to carry out actions that link the person to the environment, allowing the use of community resources to supply their own needs.
In SABE, IADLs were assessed through the question: "Can you perform the following activities?: 1) Able to manage own nances, 2) Capable of making daily purchases (especially food) 3) Able to prepare food, 4) Able to manage own medications, 5) Use of public transportation or taxi, 6) Telephone use. Answer options were codi ed as binary: 0. capable of performing the tasks alone (including both those who perform the activities alone without no di culties and those who perform the activities alone with di culty); and 1. not able to do it alone (including both those who perform the activities with any kind of help and those who cannot perform the activity).
In KloSA 10 IADLs items were evaluated, including grooming, housekeeping, preparing meals, laundering, going out, using public transportation, shopping, money management, phone use, and medication management. IADLs were further codi ed as 0 if the individual was able to perform the task alone (either with or without di culty) and as 1 if he/ she was not able to perform the task by her/himself [22].
A summary score was created, ranging between 0 and 6 for SABE and 0 to 10 for KLosa; a higher score re ected a greater impairment in IADLs. For the IADL analysis, those persons with any problems in BADLs were excluded, this to assess only those that have problems in IADL and has not progressed to having limitations on BADL.

Cognitive function:
In SABE the Mini-Mental State Examination test ( MMSE ) in its validated Spanish version was used to determine the cognitive status (score ranging from 0-30) [23].
KLoSA subjects were screened using the Korean Mini-Mental State Examination (K-MMSE). The K-MMSE is a validated measure with a score ranging from 0-30 [24].

Confounding variables
We included sociodemographic factors (age, sex, and years of schooling) and chronic diseases (hypertension, diabetes, COPD, stroke, myocardial infarction, arthritis, and cancer). To further capture the burden of chronic disease, a summary score was created, summing up each disease. Evidence suggests that multi-morbidity accounts more e ciently for the impact on global health in older adults than individual entities [25].

Statistical analysis
The descriptive analyses were performed by estimating percentages for categorical variables, and means and standard deviations for quantitative variables, and groups were compared using the Pearson chi-square test and t-student test, respectively. For BADL and IADL scores, we tted a zero-in ated negative binomial regression model to evaluate the differences between nutritional levels, due to the high frequencies of zeros in the datasets (i.e. subjects without decline on BADLs or IADLs). We inverted the Barthel test score in the SABE dataset to allow the correct estimation of the model. Moreover, we tted a logistic model for the MMSE score dichotomizing its values to normality (> 24) and cognitive decline (<= 24). Based on the literature review, we adjusted the models for the following covariates: sex, level of education, number of comorbidities, and age. We considered signi cance at P < .05 to evaluate the variables in the model. R software was used to perform all statistical analyses.

Results
The nal sample from Colombia was n=23448 and Korea n= 4501. For Colombia, the means of the MMSE was 23.08 ± 7.75, in IADL 0.88 ± 1.55, and BADL 3.41 ± 10.54; for Korea the MMSE 25.49 ± 5.38, IADL 0.63 ± 2.04, BADL 0.2 ± 1.04. In both samples, there were signi cant differences concerning sex, comorbidity, and education. More detailed characteristics of the sample can be found in Table 1. For IADL and BADL education was highly associated, a graphical representation of this is shown in Figure 3.

Korea
After adjustment for confounding variables, there were no signi cant associations for cognition, IADL, or BADL, table 2.
For complete bivariate information, see table 1.

Discussion BMI and Cognition
In Colombia, we found that underweight was associated with an impaired score at the MMSE. This is in line with studies that relate malnutrition with cognitive deterioration and as a risk factor for frailty and dementia (11,14-16). On the other hand, being overweight was associated with normal scores in the MMSE. Previous studies have reported the existence of an "obesity paradox" [26]. Some studies have found that even though the risk of certain diseases increases as the BMI rises (including cardiovascular diseases, cancer, arthritis, etc.), people tend to live longer, and in older adults being a bit on the higher side appears to give an extra protective reserve [27].
In Korea, we did not nd any signi cant associations. Kim G. et al. have studied this association longitudinally, nding that those who were underweight exhibited faster deterioration in cognitive functioning, compared to those with a normal weight.
However, overweight or obese older adults presented a slower cognitive decline than those with normal weight [28]. As shown in Figure 2, the tendency of our data goes in the same direction. Even so, our study differs in many aspects. First, our study was cross-sectional and the directionality of any association therefore could not be determined. Second, their study used a longitudinal model to evaluate the relationship between cognition and age, assuming that measurements were equidistant over time. Finally, we used MMSE as a dichotomic variable, we could not t a normal distribution in the MMSE score analysis, even though we employed the usual linear transformations recommended in the literature.

BMI and IADL
In Colombia, being underweight was found to be associated with impaired performance in the IADLs. Also, overweight was associated with better scores in the IADLs. The explanation for this could be that when weight is lost, there is also loss of muscle mass [29]. Research has shown that muscle volume protects against many kinds of negative outcomes and this might be one of the reasons why being overweight (balance between fat and muscle mass) but not obesity (too much fat and low muscle mass) may protect against the unfavorable outcomes assessed in this paper [29].
It is important to highlight that people with di culties in IADL are usually at a point that can either revert or progress to having problems in BADL. Older adults with limitations in performing BADL require more assistance from a caregiver and have more health problems related to these limitations. Especially regarding IADL, BMI values show a U-shaped curve where extremes are related to the worst performance in IADL. This, when people are losing their instrumental functional capacities but still preserve the basic, might be a crucial point of intervention, where targeting nutritional status could modify functional prognosis and prevent progression to further dependency.

BMI and BADL
In Colombia, we found signi cant associations between both underweight and obesity with high impairment in BADL.
Although this is a cross-sectional analysis, these results might be explained in the rst place by the disease(s) responsible for the functional impairment and in second place by the lack of mobility which is associated with both sarcopenias, physical wasting, undernutrition, and obesity.

Education and Functional status
It worth highlighting that in Colombia, education was a very signi cant confounder. In Figure 3, we show a graphical representation of the big difference between those with the highest education vs. illiterates. The worst status for those with a higher educational (>high school) is much better than the best status for those with low education (< high school or no education).
Education in the context of a country such as Colombia does not refer only to the knowledge acquired at school; it also refers to opportunities since early life, economic situation, and living environment.

Limitations and strengths
This study has some limitations. First, this is a cross-sectional study; therefore, causality cannot be determined. Second, this study is based on self-reported measurements of functional status, allowing memory bias. Third, we performed a secondary analysis which means that the variables were not collected for the speci c purpose of this paper. Fourth, IADL and BADL were categorized differently and measured using different methods in the studied populations. BMI in KLoSA was determined by self-report, which may have led to memory bias. Thus, due to several differences between the measurement methods used in the surveys, we did not intend to compare the two cohorts directly and all analyses have been performed separately. BMI is a widely-used instrument, useful for important health-related outcomes that become more likely when a person is underweight or obese. However, as with most measures of health, BMI is not a perfect measure, especially when referring to older adults; mainly due to the changes related to aging of the musculoskeletal system. Thus, a complete nutritional and anthropometric assessment would be the best way to approach body composition and nutrition in further studies.
This study encompasses several strengths. The data were collected from a representative sample of old adults from two countries, one in Latin America ( the biggest population-based study adults to date), and the other in South Korea in Asia.
Also, this study provides relevant information for health care providers and future strategies for the prevention of disability [13,28].

Conclusion
In Colombian older adults underweight, in particular, seems to be negative concerning the preservation of both cognitive and daily functioning. Additionally, being overweight but not obese was associated with less cognitive impartment and better daily functioning.