The Impact of High Carbohydrate Intake on Physical Frailty in Older Korean Adults: a Cohort-based Cross-sectional Study


 Background: The relationship between macronutrients and frailty is unclear. Previous studies have confirmed the relationships between energy and protein intake and physical frailty, while few studies have examined the role of carbohydrate or fat intake in the prevalence of frailty. The aim of this study is to investigate the relationship of energy and macronutrients with physical frailty in the Korean elderly population who had a high proportion of energy intake from carbohydrates.Methods: This study included 954 adults aged 70 to 84 years who have completed the assessment of frailty and 24-h recall upon enrolment in the Korean Frailty and Aging Cohort Study and have no extreme intake under 400 kcal (n = 2). The relationship between energy or macronutrients and frailty was evaluated using multivariate logistic regression models and multivariate nutrient density models.Results: In the subjects with low energy intake (odds ratio [OR] = 2.94, 95% confidence interval [CI] = 1.34–6.45) and total subjects (OR = 2.01, 95% CI = 1.03–3.93), consuming carbohydrates above the acceptable macronutrient distribution range (65% of energy) was related to a higher risk of frailty. Substituting the energy from fat with carbohydrates was related to a higher risk of frailty (1%, OR = 1.05, 95% CI = 1.00–1.09; 5%, OR = 1.26, 95% CI = 1.02–1.56; 10%, OR = 1.59, 95% CI = 1.03–2.43).Conclusions: This study showed that the proportion of energy intake from carbohydrates and fats may be an important nutritional intervention factor for reducing the risk of frailty.

elderly population consumed more than the acceptable macronutrient distribution range (AMDR) in carbohydrate energy consumption [32][33][34]. Therefore, understanding the high carbohydrate diet affecting frailty in the elderly is crucial.
To our knowledge, only a few studies have investigated the relationship between high carbohydrate diet and frailty. Furthermore, the elderly is at risk of low energy and food intake due to anorexia of aging, chronic diseases, decreased physical activity, and decreased masticatory function [35,36], but these characteristics of the elderly are not considered, and simply high intake levels of quantity tend to be recommended to prevent frailty [25][26][27]. However, additional information is necessary to con rm the intake level of macronutrients related to the prevalence of frailty in subjects with low energy intake.
Therefore, the aim of this study is to identify energy and macronutrient intake level that may be related to the frailty status among the older adults who tended to have a high energy intake from carbohydrates in Korea. Table 1 Fried frailty index and criteria for adding 1 point.

Components
Criteria for adding 1 point Unintended weight loss [40] Unintended weight loss of 4.5 kg or more in the last year Weakness [40] Grip strength lower than 26 kg for men and lower than 18 kg for women Exhaustion [42] "I felt that everything I did was an effort" or "I could not get going" was yes for three or more days in a week Slow walking speed [40] Walking speed below 1 m/s after walking 4 m at a normal rhythm Low physical activity [43] Metabolic equivalent of task in minutes per week (MET-min/week) was below 494.65 kcal for men and below 283.50 kcal for women

Dietary intake assessment
The interviewers were trained with the standardized protocol for the 24-h recall method. Interviews were conducted based on home visiting, and visual aids, which were developed by the Korea Disease Control and Prevention Agency, were used to estimate the quantity of the food consumed during a day. Nutrient intake was calculated using the 24-h recall dietary assessment system of the National Institute of Health and the Korea Disease Control and Prevention Agency [44].
Total energy and macronutrient intake Total energy was calculated as the sum of carbohydrate (4 kcal * g/d), protein (4 kcal * g/d), and fat (9 kcal * g/d) kcal. Carbohydrates, protein, and fat were expressed as the percentage of energy to evaluate the adequacy of macronutrient intake using the AMDR (carbohydrate, 55-65%; protein, 7-20%; fat, 15-30%) [32]. The participants were categorized by the AMDR of each macronutrient. Adjusted multinomial logistic regression analysis was conducted to compare the energy and macronutrient intake of frail and prefrail older adults with those of the robust older adults. We tried to con rm the relationship between the prevalence of frail status and su cient energy intake, but because of the small number of subjects with su cient energy intake, irregularities appeared in the Hessian matrix, and the validity was uncertain, so those with su cient energy intake were excluded from the multinomial logistic regression analysis. Odds ratios (ORs) and their corresponding 95% con dence intervals (CIs) were derived, after adjusting for age, sex, education attainment, house income, living situation (alone or not), hypertension, triglyceride, number of prescription drugs, chewing status, alcohol, and PA in Model 1. Model 2 was adjusted for Model 1 covariates and total energy.
Multivariate nutrient density models were conducted to examine the relationship between macronutrients and frailty status in total subjects as well as subjects who consumed the unmet intake with the standard energy intake criteria. Detailed information on the multivariate nutrient density models is available elsewhere [45]. Brie y, in logistic regression analysis, total energy intake and macronutrient of interest were included as independent variables, and frail status was included as a dependent variable. With the total energy intake adjusted, the increase in the energy of the macronutrient of interest can be interpreted as a decrease in the iso-energy of other nutrients not included in the model. For example, to examine the substitution of fat with carbohydrates, the model includes carbohydrates, protein, total energy intake, and covariates but excludes fat. The OR for carbohydrates represents substituting fat with the same energy from carbohydrates. All statistical analyses were conducted using the IBM SPSS Statistics 25.0 Program (IBM SPSS INC, Armonk) and were tested at a P-value below 0.05.

Results
Of the 1,002 participants recruited to KFACS, 954 (95%) had complete data for all the variables of interest. Table 2 shows the relationship between the aspects of general characteristics and frailty status. The participants had an average age of 76.3 years (70-84). The proportion of women was 51.7%, of which 10.9% were frail and 50.2% prefrail. After adjusting for age and sex, frail older adults were less educated, had low income, and presented with a high prevalence of physician-diagnosed hypertension. Frail older adults also had lower HDL cholesterol and marginally higher triglyceride levels. The proportion of polypharmacy and chewing discomfort according to frail status were signi cantly different, showing a higher proportion in frail older adults. The level of PA was lower in frail older adults than in prefrail and robust older adults. Table 2 The relationship between the aspects of general characteristics and frailty status 1)  Table 3 shows the macronutrient intake of the participants according to frail status and level of energy intake. In all participants, those who were frail had, on average, lower energy intake and lower carbohydrate, protein, and fat intake than those who were prefrail and robust. Regarding the percentage of energy from macronutrients, frail participants had higher energy intake from carbohydrates and lower energy intake from protein and fat than those who were prefrail or robust. The prevalence of frailty was signi cantly high in participants with high energy intake from carbohydrates (>65%) and low energy intake from fat (<15%). A similar relationship was observed in subjects who consumed less than the EER, but not in subjects who consumed energy adequately. Table 3 The macronutrient intake of the participants according to frail status and level of energy intake 1)   Table 4 shows the adjusted multinomial logistic regression models comparing the nutrient intakes for frail and prefrail subjects with those for the robust subjects comprising the reference group. In the entire study population, energy and macronutrients were related to the risk of frailty, but after adjusting for energy intake in Model 2, most of these relationships were no longer signi cant. In the participants with low energy intake, relationships of macronutrients with physical frailty were still signi cant, even after adjusting for energy. In Model 2, the risk of frailty per 10 g increase in carbohydrates (OR =1.14, 95% CI = 1.02-1.27) and per 100 kcal increase from carbohydrates (OR = 1.38, 95% CI = 1.05-1.80) increased signi cantly. The risk of frailty decreased with the increase in the same amount of fat (OR = 0.67, 95% CI = 0.50-0.91) and energy from fat (OR = 0.64, 95% CI = 0.46-0.90). In the same model, the increase in protein intake and energy intake from protein did not show a signi cant relationship with physical frailty, but the recommended protein intake in consideration of sex and age showed a positive relationship with the prevalence of prefrailty. Notably, in both groups of subjects with low energy intake (OR = 2.94, 95% CI = 1.34-6.45) and total subjects (OR = 2.01, 95% CI = 1.03-3.93), consuming carbohydrates above the AMDR (65% of energy) was related to a higher risk of frailty, even after adjusting for energy. Table 4 The relationship between the nutrient intakes and frail status 1), 2)  carbohydrates or fat with protein showed no signi cant relationship with risk of frailty among the low energy intake groups. Table 5 The relationship between the replacement of energy from macronutrients with other macronutrients and frail status 1), 2), 3) * p value < 0.005 1) Reference category: Robust. 2) Odds ratio and 95% con dence interval were obtained by the multinomial logistic regression analysis.
3) Models were adjusted for macronutrients, age, sex, education attainment, house income, living alone or with partner, hypertension, triglyceride, number of prescription drugs, chewing status, alcohol, physical activity, total energy.

Discussion
In this study comprising community-dwelling older-aged people, our ndings show that the risk of frailty was positively related to the high level of energy intake from carbohydrates (>65% energy). Particularly, for the group with low energy intake, we found that a high energy intake from carbohydrates was related to an increased prevalence of frailty, whereas high energy intake from fat was related to a decreased prevalence in multivariate nutrient density models.
The Fried frailty phenotype includes the components for assessing physical abilities, which are related to muscle mass and strength [7]. Muscle protein is regulated and metabolized depending on the intake of protein and amino acids [46,47]. Studies on the relationship between protein intake and the prevalence of frailty have been conducted considering protein intake as the main nutritional factor, but inconsistent results have been obtained [12][13][14][15][16][17][18][19][20][21][22][23][24]. In the present study, high protein intake was related to a lower prevalence of frailty, but this was not signi cant after adjusting for total energy. This nding is consistent with the results from the study on the relationship between protein intake and grip strength in Korean elderly women, in which total protein intake had a positive relationship with grip strength, but no signi cant results were found after adjusting for energy [33]. A lower protein effect after energy adjustment may partially indicate the relationship between energy intake and frailty [18, 21,48]. Previous studies have suggested that low energy intake is positively related to the prevalence of frailty [20,21].
Several studies have shown a 5-17.9% prevalence of anorexia in the elderly [49,50]. Anorexia has been known to occur owing to various causes, such as reduction of PA, decrease in mastication function, increased levels of cholecystokinin and leptin hormones, decrease in digestive function, and drug intake, which leads to decreased nutritional intake [35,36]. Therefore, it may be important to determine the intake levels associated with the prevalence of frailty in older adults with low energy intake. In this study, the relationship between the level of macronutrient intake and the prevalence of frailty was different depending on whether the energy intake criteria were met or not. In the group with insu cient energy intake, high energy intake from carbohydrates showed a positive relationship with the prevalence of frailty, which was signi cant even after adjusting for total energy.
In the present study, high energy intake from carbohydrates was positively associated with the prevalence of frailty. Previous studies have reported the carbohydrate intake effect on chronic disease and mortality [30,31,51,52]. The Atherosclerosis Risk in Communities study reported a U-shaped association between energy intake from carbohydrates and mortality as well as a signi cantly lower risk of death, particularly with a carbohydrate energy intake of 50-55% [51]. A meta-analysis of 432,179 people from different countries (e.g., the US, Sweden, Japan, and Greece) has also shown a U-shaped association between energy intake from carbohydrates and mortality (energy from carbohydrates of <40%, OR = 1.20, 95% CI = 1.09-1.32; energy from carbohydrates >70%, OR = 1.23, 95% CI = 1.11-1.36) [51]. The Prospective Urban Rural Epidemiology (PURE) study found that high carbohydrate energy intake was associated with an increased risk of death, but no signi cant association with hypertension prevalence was found [52]. High carbohydrate intake was associated with low LDL cholesterol but was also associated with low HDL cholesterol, high triglyceride levels, and an increase in the apo B/apo A1 ratio, which is known to be a strong predictor of myocardial infarction [53]. High carbohydrate intake was also associated with elevated in ammatory responses in the skeletal muscle [54], increased insulin resistance [55], mitochondrial damage in the muscle cells [56], and chronic low levels of in ammation that contribute to sarcopenia [57].
Frailty has been reported to be associated with higher oxidative stress and lower antioxidant markers in the body [21,58,59]. Low energy intake might re ect low overall nutrient intake [34]. In this study, the high prevalence of frailty in elderly subjects with low energy intake and high carbohydrate intake may be due . These data suggest that the quality of diet might be related to physical frailty.
We found that frailty was decreased when carbohydrates were exchanged for fat in the low energy intake group. A previous meta-analysis showed the negative association between the substitution of carbohydrate with protein or fat and mortality [51]. The risk of mortality was different according to food source; substituting the carbohydrate with plant-based protein or fat was negatively associated with mortality (hazard ratio [HR] = 0.82, 95% CI = 0·78-0·87) and replacing the carbohydrates with animal-based protein or fat were positively associated with mortality (HR = 1·18, 95% CI = 1·08-1·29) [51]. These results suggest that it may be appropriate to recommend the intake of plant-based proteins or fats instead of carbohydrates. In the present study, the proportion of energy from fat in subjects with low energy intake was 13% for energy and below the lowest AMDR for fat. The PURE study represented that the highest quintile of total fat (median = 35.3%) was associated with less risk of total mortality when compared with the lowest quintile (median = 10.6%) and the highest intake of saturated fat, monounsaturated fat, and polyunsaturated fat was related to less risk of total mortality [52]. However, in a cross-sectional study of 4,724 people aged 50 years and older who participated in the US National Health and Nutrition Examination Survey, consuming more than 20% of energy from saturated fatty acids was associated with higher morbidity and mortality [62]. Therefore, the type of fat as well as the energy intake from fat within the AMDR should be considered in preventing frailty.
However, our study has some limitations. First, this is a cross-sectional study, so causality between dietary intake and frailty may not be veri ed. Second, a single 1-day 24-h recall may not be representative of the usual intake. However, the within-person variances of the energy and macronutrients have shown to be relatively low among the elderly [63,64]. Future studies will be needed to clarify the relationship between macronutrient intake and frailty by using the repeated 24-h recall survey and prospective longitudinal data. Third, this study population included ambulatory older adults who lived near the center. Our results could not be fully representative of the older adults in Korea. Fourth, in this study, we were unable to gather information on the subtypes of dietary fat associated with physical frailty because no nutrient database for each subtype of fat was available. Since the health effects of intake levels according to dietary fat subtypes are different, studies suggesting the appropriate types of fats are needed. Fifth, in this study, the statistical test for subjects who consumed the recommended energy level was limited because of the small sample size. Therefore, the analysis of the relationship between energy and macronutrients and the prevalence of frailty in these subjects could not be performed. Nevertheless, our study has shown that the intake levels of carbohydrates and fats may be important factors in preventing the risk of frailty in older adults with low energy intake.

Conclusions
In this study, we found that the inappropriate proportion of energy from carbohydrates and fat was related to a higher prevalence of frailty in the elderly with low energy intake. It was also indicated that the proportion of energy intake from carbohydrates and fats may be an important nutritional intervention factor in reducing the risk of frailty, as well as protein as previously emphasized.

Declarations
Ethics approval and consent to participate This study was performed in accordance with the Declaration of Helsinki, and it was approved by the Institutional Review Board of Dankook University, Hanyang University and Kyung Hee University. Written informed consent was obtained from all participants.