Abstract
[Purpose]
This study aimed to identify the body composition profiles of older adults with hypertension and examine their associations with physical fitness and cardiometabolic characteristics.
[Methods]
This cross-sectional study included 491 community-dwelling older adults with hypertension. Latent profile analysis was conducted using eight body composition indicators. Associations between age, physical fitness components, and profile membership were examined using multinomial logistic regression with a three-step approach. Systolic blood pressure, diastolic blood pressure, and fasting plasma glucose levels were compared across profiles using one-way analysis of variance.
[Results]
Four profiles were identified: low muscle mass, balanced, high adiposity, and high muscle/low fat ratio. Handgrip strength, balance, and flexibility were significantly associated with profile membership, whereas age and aerobic endurance were not. The high adiposity profile had significantly higher systolic blood pressure than the low muscle and high muscle/low fat profiles, and higher diastolic blood pressure than the low muscle profile. The high muscle/low fat profile was associated with significantly lower fasting plasma glucose levels than the low muscle and high adiposity profiles.
[Conclusion]
Distinct body composition profiles in older adults with hypertension were associated with differences in physical fitness and cardiometabolic characteristics. These findings support the value of a person-centered approach that integrates body composition and physical fitness.
Keywords: hypertension, older adults, body composition, latent profile analysis, physical fitness, blood pressure, fasting plasma glucose
INTRODUCTION
The global population is rapidly aging, and the increasing prevalence of hypertension has emerged as a major public health concern. From 1990 to 2019, the number of adults aged 30-79 years with hypertension doubled worldwide, from approximately 634 million to 1.28 billion [1]. Hypertension is a major risk factor for cardiovascular disease and premature mortality, and its health burden is expected to increase further as population aging accelerates [2]. A meta-analysis of older adults in China reported a prevalence of hypertension of 47% [3]. Similarly, in Korea, the prevalence of hypertension among adults aged ≥ 65 years is > 60%, highlighting hypertension management as a key priority for promoting health in older adults [4].
A more serious concern regarding hypertension is that its adverse effects are not limited to elevated blood pressure (BP); particularly in older adults, hypertension is associated with a complex burden of cardiovascular and metabolic diseases. Approximately 80% of adults aged ≥ 80 years have hypertension, which is recognized as a major risk factor for cardiovascular disease, stroke, kidney disease, and cognitive decline [5]. When hypertension and diabetes coexist, the risk of cardiovascular disease has been shown to exceed the sum of the risks associated with each condition individually [6], with approximately 28-31% of the cardiovascular disease risk attributable to their interaction [7]. These findings highlight the importance of considering metabolic characteristics when assessing the cardiovascular and metabolic risks in older adults with hypertension.
Aging is accompanied by a decline in skeletal muscle mass and an increase in body fat with physical changes characterized by visceral fat accumulation and intramuscular fat infiltration [8]. Chronic inflammation in the adipose tissue and insulin resistance in skeletal muscles form a mutually reinforcing cycle, which has been proposed as a key mechanism linking age-related changes in body composition to cardiovascular and metabolic abnormalities [9]. These changes increase the risk of sarcopenic obesity, characterized by the coexistence of sarcopenia and obesity, which is associated with a greater risk of cardiovascular diseases, metabolic disorders, and mortality than sarcopenia or obesity alone [9,10].
These differences in body composition and physical fitness are also associated with BP levels and risk of developing hypertension. A study of 17,110 Korean older adults aged ≥ 65 years showed significant differences across BP levels in body composition as well as in physical fitness components, including muscular strength, flexibility, cardiorespiratory endurance, muscular endurance, and balance [11]. Similarly, a study on older European adults reported that greater handgrip strength was associated with a lower risk of developing hypertension [12]. However, traditional variable-centered approaches primarily examine average associations between individual body composition indicators and outcome variables, limiting their ability to capture complex patterns in which multiple body composition characteristics coexist within individuals [13].
To overcome these limitations, person-centered approaches, which classify individuals into subgroups based on combinations of individual characteristics, have recently gained increasing attention. Latent profile analysis (LPA) is a data-driven statistical technique that classifies individuals into distinct subgroups based on patterns across multiple continuous indicators [13]. Sharafi et al. (2025) applied LPA to body composition indicators measured using bioelectrical impedance analysis (BIA) and identified five latent profiles. They reported that a profile characterized by a high percentage of body fat was associated with significantly higher risks of diabetes, cardiovascular disease, and hypertension by approximately fivefold, twofold, and eightfold, respectively. However, because this study included adults aged 35-70 years, the findings may not fully reflect the characteristics of the older population. Moreover, the study did not specifically focus on older adults with hypertension or directly compare BP and fasting plasma glucose (FPG) levels across body composition profiles within this population.
The association between sarcopenia and hypertension has been confirmed in a meta-analysis [14], and older adults with sarcopenic obesity have been reported to be at a greater risk of cardiovascular disease [15]. However, previous studies have primarily focused on dichotomous classifications based on the presence or absence of sarcopenia and obesity, and the relationships among multidimensional patterns of body composition, BP, and FPG levels remain insufficiently understood.
Therefore, this study aimed to apply LPA to body composition indicators in older adults with hypertension to identify distinct subgroups, examine the associations between age and physical fitness components and latent profile membership, and determine whether BP and FPG levels differ across latent profiles. Based on these objectives, the following hypothesis is proposed: First, the body composition characteristics of older adults with hypertension were classified into multiple distinct latent profiles. Second, age and physical fitness components including muscular strength, balance, flexibility, and cardiorespiratory endurance were significantly associated with latent profile membership. Third, BP and FPG levels differed significantly across the latent profiles.
METHODS
Study Design and Participants
This cross-sectional observational study identified latent profiles based on body composition indicators among community-dwelling older adults with hypertension, examined differences in BP and FPG levels across latent profiles, and investigated the associations between physical fitness components and latent profile membership. Participants were recruited from six of the 24 districts in Seoul: Jungnang-gu, Geumcheon-gu, Yeongdeungpo-gu, Dobong-gu, Eunpyeong-gu, and Seodaemun-gu. Official letters requesting cooperation were sent to public health and senior welfare centers in these districts, and participants were subsequently recruited with the assistance of staff at affiliated senior centers and welfare facilities. Data were collected between May 2026 and July 2026.
The sample size required for LPA varies depending on factors such as the degree of separation between profiles, number of indicators included, and relative sizes of the latent subgroups; therefore, no single sample size criterion can be uniformly applied [16]. Previous simulation studies evaluated the classification performance of mixture models using sample sizes of 200, 500, and 1,000 [17], and the median sample size of studies applying LPA was reported to be 494 [18]. Considering these findings, data were collected from 526 participants to ensure the stability and interpretability of the latent profiles. Of these, eight participants with missing body composition data and 27 who did not participate in the physical fitness assessments were excluded, resulting in a final analytic sample of 491 participants (56 men and 435 women).
The inclusion criteria were as follows: (1) age 65 years or older; (2) diagnosis of hypertension by a physician; (3) ability to fully understand the explanation of the study purpose and procedures and voluntarily provide written informed consent; and (4) ability to independently perform activities of daily living and communicate without difficulty.
The exclusion criteria were as follows: (1) a cardiovascular event, such as myocardial infarction or stroke, within the previous six months; (2) a musculoskeletal disorder that could interfere with the assessments; and (3) a history of major neurocognitive or neurological disorders, such as dementia, Parkinson’s disease, or cognitive impairment severe enough to interfere with questionnaire completion or participation in the assessments.
This study was approved by the Institutional Review Board of the University of Seoul before data collection (IRB No. 2026-04-014-001). All participants were provided with detailed information regarding the study purpose and procedures, potential benefits and risks of participation, protection of personal information, and right to withdraw from the study. Only individuals who voluntarily provided written informed consent were enrolled.
Measurements
All participants fasted for at least 12 h before the assessments and were instructed to refrain from vigorous physical activity the day before testing. All measurements were conducted between 8:00 a.m. and 10:00 a.m. Participants wore light clothing and removed all jewelry and metal accessories before the assessments.
2.1. Body composition
Height and body weight were measured using a digital stadiometer (BSM 370; InBody, Seoul, Korea), and body composition was assessed using a BIA system (InBody 270; InBody, Seoul, Korea). Muscle mass, body fat percentage, estimated bone mass, estimated basal metabolic rate (EBMR), and waist-to-hip ratio (WHR) were obtained from the analyzer. Estimated bone mass refers to the value derived from the device using BIA. Body mass index (BMI) was calculated as body weight in kilograms divided by height in meters squared (kg/m²).
2.2. BP and FPG
BP was measured using an automated BP monitor (Jawon Medical, Daejeon, Korea) in accordance with the BP measurement guidelines of the European Society of Cardiology and the European Society of Hypertension [19]. The participants rested in a comfortable seated position for at least 5 min before BP was measured in the brachial artery. Two measurements were obtained at 3-min intervals, and the mean of the two measurements was used for the analysis. If the difference between the two measurements was ≥ 10 mmHg, a third measurement was obtained, and the mean of the two closest values was used for the final analysis.
Approximately 10 mL of venous blood was collected from the antecubital vein into heparinized tubes by a clinical laboratory technologist. Blood samples were centrifuged at 3,000 rpm for 10 min using a Combi-514R centrifuge (Hanil, Korea). The separated plasma was stored at ≤ −70°C until further analysis. FPG concentrations were measured in heparinized plasma using the hexokinase method (Glucose/HK Kit; Roche Diagnostics).
2.3. Physical fitness
2.3.1. Handgrip strength
Handgrip strength was assessed using a handgrip strength test [20] with a handgrip dynamometer (TKK 5401, Takei Scientific Instruments Co., Japan). Participants stood upright with their feet approximately shoulder-width apart. After the handle width was appropriately adjusted, the participants held the dynamometer with their arms naturally extended downward, ensuring that the device did not touch the body. Two trials were performed alternately with the left and right hands, and the grip strength was recorded to the nearest 0.1 kg. The highest value obtained for either hand was used for the analysis.
2.3.2 Static balance (One-leg standing test with eyes closed)
Static balance was assessed using the one-leg standing test with the eyes closed [21]. Participants were instructed to close their eyes, extend both arms to the sides, lift one foot off the floor, and maintain balance for as long as possible. Timing began when the participant lifted the foot following the “start” signal and ended when the raised foot touched the floor. Two trials were performed alternately on the right and left legs, and the duration was recorded in seconds. The longest duration was used for the analysis.
2.3.3 Upper-body flexibility (Back scratch test)
Upper body flexibility was assessed using the backscratch test [22]. While standing comfortably, the participants reached one hand over the shoulder and down the back, while reaching for the other hand behind the back with the elbow flexed and the back of the hand facing the body. Participants attempted to bring the middle fingers of both hands as close as possible. The distance between the middle fingers was measured to the nearest 0.1 cm using a ruler. Negative values were recorded when the fingers were not touched, zero when the fingers touched, and positive values when the fingers overlapped. The higher value was used for the analysis.
2.3.4 Aerobic endurance (2-min step test)
Aerobic endurance was assessed using a 2-min step test [22]. While the participants stood upright, the midpoint between the patella and iliac crest was marked on a wall with masking tape as the target knee height. Participants were instructed to march in place for 2 min, raising each knee to a marked height. At the “start” signal, the number of times one designated knee reached the target height was counted. The test was performed once after a brief practice trial to ensure the correct technique and measurement accuracy.
Statistical Analysis
Statistical analyses were performed using SPSS version 27.0 and Mplus version 7.0. The specific analytical procedure was as follows.
First, frequency and descriptive analyses were conducted to examine the general characteristics of the participants and distribution of the primary variables. Correlation analyses were also performed to explore the relationships between the study variables and to assess potential multicollinearity.
Second, LPA was conducted using Mplus version 7.0 to identify distinct body composition profiles among older adults with hypertension. To determine the optimal number of profiles, the models were estimated sequentially, beginning with a one-profile model and progressively increasing the number of profiles. Model fit was evaluated using the Akaike Information Criterion (AIC [23];), Bayesian Information Criterion (BIC [24];), and sample-size-adjusted Bayesian Information Criterion (SABIC [25];), with lower values indicating better model fit. The classification accuracy was assessed using entropy, with values closer to 1 indicating greater classification accuracy. Based on the criterion proposed by Clark (2010), an entropy value of ≥ 0.80 was considered indicative of good classification accuracy [26].
Model comparisons were further conducted using the Lo-Mendell-Rubin adjusted likelihood ratio test (LMR-LRT) and the parametric bootstrapped likelihood ratio test (BLRT). Specifically, a model with k profiles was compared with a model with k−1 profiles. A significant p value indicated that the k-profile model provided a better fit and was therefore retained, whereas a nonsignificant p value supported retention of the k−1-profile model.
After identifying the latent profiles, the three-step approach proposed by Asparouhov and Muthén (2014) was used to examine the associations between participant characteristics (age), physical fitness components (handgrip strength, balance, flexibility, and aerobic endurance), and latent profile membership [27]. In Step 1, the latent profile model is estimated using the raw scores of the eight body composition indicators. In Step 2, each participant was assigned to the most likely latent profile based on posterior class membership probabilities, and the classification errors were estimated. In Step 3, a multinomial logistic regression analysis was conducted, while accounting for classification errors to examine the associations of age, handgrip strength, balance, flexibility, and aerobic endurance with latent profile membership.
To visually compare body composition characteristics across latent profiles on a common scale, Z-scores were calculated using the mean and standard deviation (SD) of the total sample and plotted for each profile. Finally, one-way analysis of variance (ANOVA) was performed to examine differences in systolic blood pressure (SBP), diastolic blood pressure (DBP), and FPG across the latent profile groups. When the overall ANOVA result was significant, Scheffé’s post-hoc test was conducted for pairwise comparisons between profiles. The level of statistical significance was set at p < 0.05. These comparisons were conducted without adjusting for sex or other clinical covariates.
RESULTS
Descriptive Statistics and Correlation Analysis
Descriptive statistics for the indicators included in the LPA and associated factors are presented in Table 1. The latent profile indicators included height, body weight, BMI, body fat percentage, muscle mass, estimated bone mass, EBMR, and WHR. The associated factors included age, handgrip strength, balance, flexibility, and aerobic endurance.
Table 1.
Descriptive statistics of the study variables
| Variable | M | SD | Min | Max | |
|---|---|---|---|---|---|
| Profile indicators | Height (cm) | 155.24 | 6.05 | 140.00 | 173.00 |
| Weight (kg) | 59.18 | 7.31 | 41.20 | 84.70 | |
| BMI (kg/m2) | 24.61 | 2.92 | 17.83 | 37.90 | |
| PBF (%) | 34.58 | 6.41 | 15.20 | 48.60 | |
| Muscle mass (kg) | 19.21 | 4.19 | 13.10 | 39.30 | |
| EBM (kg) | 2.19 | 0.30 | 0.80 | 3.10 | |
| EBMR (kcal/day) | 1139.06 | 127.65 | 878.00 | 1651.00 | |
| WHR | 0.90 | 0.04 | 0.77 | 1.05 | |
| Predictors of profile membership | Age (years) | 73.10 | 4.65 | 66.00 | 90.00 |
| Handgrip strength (kg) | 21.84 | 4.70 | 10.30 | 39.50 | |
| Static balance (sec) | 4.39 | 2.74 | 1.00 | 18.00 | |
| Upper body flexibility (cm) | -7.18 | 7.50 | -30.00 | 12.00 | |
| Aerobic endurance (reps) | 104.44 | 14.97 | 45.00 | 135.00 | |
Note: BMI = body mass index; PBF = percentage body fat; WHR = waist-to-hip ratio; EBM = estimated bone mass; EBMR = estimated basal metabolic rate.
Correlation analysis was conducted to examine the relationships among the variables included in the analysis, and the results are presented in Table 2. Significant correlations were observed among most variables, except for a few associations, such as that between age and aerobic endurance. In particular, strong correlations were observed between estimated bone mass and height (r = 0.806, p < 0.001), estimated bone mass and muscle mass (r = 0.836, p < 0.001), EBMR and muscle mass (r = 0.895, p < 0.001), and EBMR and estimated bone mass (r = 0.932, p < 0.001). These correlations reflect the shared physiological basis of body size and lean tissue, which underlie muscle mass, estimated bone mass, and EBMR. Although these indicators were interrelated, they were retained because they represented theoretically distinguishable characteristics, including body size, muscularity, skeletal characteristics, and metabolic requirements. Height was included because although it remains relatively stable in adulthood, it reflects the overall body frame and skeletal size, which are relevant to the interpretation of absolute muscle mass, estimated bone mass, and EBMR. Accordingly, all indicators were included in the LPA to capture multidimensional within-person patterns of body size and composition.
Table 2.
Correlations among the study variables
| Height | Weight | BMI | PBF | MM | EBM | EBMR | WHR | Age | HGS | SB | UBF | AB | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Height | 1 | ||||||||||||
| Weight | 0.402*** | 1 | |||||||||||
| BMI | -0.248*** | 0.745*** | 1 | ||||||||||
| PBF | -0.572*** | 0.346*** | 0.744*** | 1 | |||||||||
| MM | 0.752*** | 0.658*** | 0.151** | -0.324*** | 1 | ||||||||
| EBM | 0.806*** | 0.767*** | 0.256*** | -0.191*** | 0.836*** | 1 | |||||||
| EBMR | 0.753*** | 0.832*** | 0.334*** | -0.187*** | 0.895*** | 0.932*** | 1 | ||||||
| WHR | 0.041 | 0.436*** | 0.390*** | 0.240*** | 0.225*** | 0.253*** | 0.308*** | 1 | |||||
| Age | 0.054 | -0.008 | -0.030 | -0.066 | 0.052 | 0.044 | 0.039 | -0.042 | 1 | ||||
| HGS | 0.529*** | 0.221*** | -0.143*** | -0.485*** | 0.537*** | 0.433*** | 0.475*** | 0.047 | -0.053 | 1 | |||
| SB | 0.190*** | -0.163*** | -0.289*** | -0.334*** | 0.068 | 0.080 | 0.016 | -0.156** | -0.047 | 0.158*** | 1 | ||
| UBF | -0.012 | -0.185*** | -0.194*** | -0.035 | -0.189*** | -0.142** | -0.183*** | -0.243*** | 0.012 | 0.129** | 0.178*** | 1 | |
| AE | -0.068 | 0.020 | 0.063 | 0.043 | -0.046 | -0.031 | -0.017 | -0.082 | -0.181*** | 0.080 | -0.077 | 0.066 | 1 |
Note: BMI = body mass index; PBF = percent body fat; MM = muscle mass; EBM = estimated bone mass; EBMR = estimated basal metabolic rate; WHR = waist-tohip ratio; HGS = handgrip strength; SB = static balance; UBF = upper body flexibility; AE = aerobic endurance.
p < 0.01,
p < 0.001.
LPA Based on Health-Related Characteristics
To determine the optimal number of latent profiles based on the health-related characteristics of participants, models with progressively increasing numbers of profiles were compared by examining changes in information criteria, model comparison tests, and entropy values. Table 3 presents the results of the study.
Table 3.
Model fit indices and classification quality for latent profile solutions
| Model selection criteria | Number of profiles |
||||||
|---|---|---|---|---|---|---|---|
| 2 | 3 | 4 | 5 | 6 | 7 | ||
| Information criteria | AIC | 18398.136 | 17447.016 | 17147.340 | 16619.913 | 16250.013 | 15857.029 |
| BIC | 18503.047 | 17589.695 | 17327.787 | 16838.128 | 16505.996 | 16150.780 | |
| SABIC | 18423.697 | 17481.779 | 17191.305 | 16673.081 | 16312.383 | 15928.601 | |
| Model comparison tests (p-value) | LMR | 0.174 | 0.112 | 0.020 | 0.055 | 0.034 | 0.005 |
| BLRT | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | |
| Classification quality | Entropy | 0.927 | 0.911 | 0.890 | 0.915 | 0.934 | 0.948 |
| Profile size, n (%) | Profile 1 | 395(80.4) | 176(35.8) | 195(39.7) | 49(10.0) | 125(25.5) | 20(4.1) |
| Profile 2 | 96(19.6) | 255(51.9) | 112(22.8) | 101(20.6) | 118(24.0) | 21(4.3) | |
| Profile 3 | 60(12.3) | 128(26.1) | 126(25.7) | 104(21.2) | 106(21.6) | ||
| Profile 4 | 56(11.4) | 87(17.72) | 20(4.1) | 95(19.3) | |||
| Profile 5 | 128(26.1) | 87(17.7) | 86(17.5) | ||||
| 37(7.5) | 35(7.1) | ||||||
| 128(26.1) | |||||||
Note: AIC = akaike information criterion; BIC = bayesian information criterion; SABIC = sample-size adjusted BIC; LMR = Lo-Mendell-Rubin adjusted likelihood-ration test; BLRT = parametric bootstrapped likelihood ratio test.
The results showed that the AIC, BIC, and SABIC, which were used as information criteria, generally decreased as the number of profiles increased (Figure 1). Lower values of these information criteria indicated a relatively better model fit. However, because the information criteria often continue to decrease as model complexity increases, these indices are not used as the sole criteria for determining the optimal number of profiles.
Figure 1. Information criteria according to the number of latent profiles.
AIC = akaike information criterion; BIC = bayesian information criterion; SABIC = sample-size adjusted BIC.
For the model comparison tests, a significant p-value for the LMR and BLRT indicates that the model with a greater number of latent profiles provides a better fit than the model with fewer profiles. In the present study, the LMR test was significant for the four-, six-, and seven-profile solutions but not for the five-profile solution. This finding indicated that the four-profile solution provided a significant improvement over the three-profile solution, whereas the five-profile solution did not provide a significant improvement over the four-profile solution. In contrast, BLRT was significant for all models and therefore did not provide a clear basis for selecting a specific model.
The classification quality was also evaluated using entropy, with values closer to 1 indicating greater classification accuracy. Overall, all models exhibited high entropy values, indicating good classification accuracy. Because entropy assesses classification precision instead of determines the optimal number of profiles, it was considered together with other model selection criteria.
The profile sizes of the four profile solutions were 195 (39.7%), 112 (22.8%), 128 (26.1%), and 56 (11.4%), indicating that all four profiles contained an adequate number of participants. In contrast, the six- and seven-profile solutions included small profiles consisting of only 20 (4.1%) and 20-21 participants (4.1%-4.3%), respectively. Although the LMR test was significant for the six- and seven-profile solutions, the addition of these small profiles increased the model complexity without providing sufficiently distinct or substantively meaningful body composition patterns. After comprehensively considering the information criteria (AIC, BIC, and SABIC), model comparison tests (LMR and BLRT), entropy, profile size, parsimony, and substantive interpretability, a four-profile solution was selected as the final model.
After selecting the final model, the classification quality was further evaluated using the posterior class membership probabilities, as shown in Table 4. The average posterior probabilities for the most likely profile membership were 0.954 for Profile 1, 0.909 for 2, 0.933 for 3, and 0.993 for 4. Generally, posterior probabilities of ≥ 0.70 are considered to indicate relatively accurate classification [28]. Therefore, the posterior probabilities ranging 0.909 to 0.993 in the present study indicate that the latent profiles were classified with a high degree of accuracy.
Table 4.
Average posterior probabilities for latent profile membership
| Most likely profile | Profile 1 | Profile 2 | Profile 3 | Profile 4 |
|---|---|---|---|---|
| Profile 1 | 0.954 | 0.030 | 0.016 | 0.000 |
| Profile 2 | 0.027 | 0.909 | 0.064 | 0.000 |
| Profile 3 | 0.009 | 0.057 | 0.933 | 0.000 |
| Profile 4 | 0.000 | 0.005 | 0.003 | 0.993 |
Characteristics and Description of the Latent Profiles
The results of the comparisons across the four identified latent profile groups are presented in Table 5. Significant differences were observed across the four profiles for all eight indicators (p < 0.001), the patterns of the latent profiles are illustrated in Figure 2.
Table 5.
Sex distribution and differences in indicators across the identified latent profiles
| Profile 1 (low muscle) |
Profile 2 (balance) |
Profile 3 (high-adiposity) |
Profile 4 (high muscle/low fat) |
F | |||||
|---|---|---|---|---|---|---|---|---|---|
| M | SD | M | SD | M | SD | M | SD | ||
| Sex, n(%) | |||||||||
| Male | 1(0.2) | 6(1.2) | 1(0.2) | 48(9.8) | |||||
| Female | 194(39.5) | 106(21.6) | 127(25.9) | 8(1.6) | |||||
| n (%) | 195 (39.7) | 112 (22.8) | 128 (26.1) | 56 (11.4) | 491 (100) | ||||
| Body composition indicators | |||||||||
| Height (cm) | 150.78 | 3.71 | 159.10 | 2.95 | 154.05 | 3.62 | 165.73 | 3.62 | 320.352*** |
| Weight (kg) | 52.91 | 4.60 | 58.86 | 3.27 | 65.48 | 3.93 | 67.17 | 6.85 | 269.455*** |
| BMI (kg/m2) | 23.38 | 2.51 | 23.33 | 1.49 | 27.73 | 2.10 | 24.27 | 2.78 | 114.513*** |
| PBF (%) | 34.18 | 4.19 | 33.22 | 4.26 | 40.75 | 2.59 | 24.56 | 7.51 | 185.701*** |
| MM (kg) | 16.09 | 1.48 | 19.90 | 2.00 | 19.66 | 1.74 | 27.59 | 5.15 | 354.144*** |
| EBM (kg) | 1.92 | 0.18 | 2.30 | 0.12 | 2.28 | 0.15 | 2.69 | 0.20 | 377.223*** |
| EBMR (kcal/day) | 1028.30 | 55.48 | 1153.13 | 42.61 | 1181.54 | 56.70 | 1399.48 | 77.90 | 678.878*** |
| WHR | 0.88 | 0.04 | 0.88 | 0.04 | 0.92 | 0.037 | 0.92 | 0.93 | 42.115*** |
Note: Values are presented as means and standard deviations for continuous variables and as n (%) for sex. Percentages for sex were calculated within each profile.
Sex was not included as an indicator in the latent profile analysis and is presented descriptively. BMI = body mass index; PBF = percentage body fat; WHR = waist to hip ratio; MM = muscle mass; EBM = estimated bone mass; EBMR = estimated basal metabolic rate.
p < 0.001.
Figure 2. Standardized body composition scores across the four latent profiles.
Values represent the standardized mean scores (z-scores) of body composition indicators in each latent profile. The horizontal line at zero represents the mean of the total sample. Positive and negative values indicate scores above and below the total sample mean, respectively. Profile 1, low muscle type; Profile 2, balanced type; Profile 3, high adiposity type; Profile 4, high muscle/low fat type. BMI = body mass index; PBF = percent body fat; WHR = waist to hip ratio; MM = muscle mass; EBM = estimated bone mass; EBMR = estimated basal metabolic rate.
Regarding the body composition patterns of each profile, Profile 1 was characterized by relatively low muscle mass and average body fat percentage. Profile 2 showed average levels of both muscle mass and body fat percentage with a higher BMI and EBMR than Profile 1. Profile 3 exhibited the highest body fat percentage and a relatively low muscle mass, resulting in the highest BMI among the four profiles. Profile 4 was characterized by the highest muscle mass, low body fat percentage, and highest EBMR. However, this profile was predominantly male (48 of 56 participants, 85.7%).
Based on these characteristics, Profile 1 was labeled as the low muscle type, Profile 2 was the balanced type, Profile 3 was the high adiposity type, and Profile 4 was the high muscle/low fat type. Sex distribution varied markedly across profiles. Men accounted for 0.5%, 5.4%, 0.8%, and 85.7% of Profiles 1, 2, 3, and 4, respectively. Accordingly, Profile 4 was predominantly male.
Factors Associated with Latent Profile Membership
Multinomial logistic regression analysis was conducted to identify the factors associated with membership in the four groups classified through LPA, and the detailed results are presented in Table 6. Age and aerobic endurance were not significantly associated with latent profile membership (p > 0.05).
Table 6.
Associations between physical fitness factors and latent profile membership
| Predictors | Profile 2 vs. Profile 1 |
Profile 3 vs Profile 1 |
Profile 4 vs Profile 1 |
Profile 3 vs Profile 2 |
Profile 4 vs Profile 2 |
Profile 4 vs Profile 3 |
||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| β (OR) | SE | β (OR) | SE | β (OR) | SE | β (OR) | SE | β (OR) | SE | β (OR) | SE | |
| Age | 0.035 (1.036) | 0.027 | 0.000 (1.000) | 0.026 | 0.061 (1.063) | 0.041 | -0.036 (0.965) | 0.030 | 0.025 (1.025) | 0.042 | 0.061 (1.063) | 0.042 |
| Handgrip strength | 0.150*** (1.162) | 0.035 | 0.084* (1.088) | 0.034 | 0.467*** (1.595) | 0.052 | -0.066 (0.936) | 0.037 | 0.316*** (1.372) | 0.048 | 0.383*** (1.467) | 0.051 |
| Static balance | 0.091* (1.095) | 0.045 | -0.127* (0.881) | 0.060 | 0.001 (1.001) | 0.077 | -0.218*** (0.804) | 0.062 | -0.090 (0.914) | 0.073 | 0.128 (1.137) | 0.085 |
| Upper body flexibility | 0.006 (1.006) | 0.017 | -0.036* (0.965) | 0.017 | -0.164*** (0.849) | 0.036 | -0.042* (0.959) | 0.019 | -0.170*** (0.844) | 0.036 | -0.128*** (0.880) | 0.036 |
| Aerobic endurance | -0.002 (0.998) | 0.009 | -0.012 (0.988) | 0.008 | -0.020 (0.980) | 0.014 | -0.011 (0.989) | 0.009 | -0.018 (0.982) | 0.014 | -0.007 (0.993) | 0.014 |
Note: In each comparison, the profile presented after “vs.” was used as the reference group.
p < 0.05,
p < 0.001.
For handgrip strength, each one-unit increase was associated with a 16.2% increase in the odds of belonging to Profile 2 compared with Profile 1 (OR = 1.162, p < 0.001), an 8.8% increase in the odds of belonging to Profile 3 compared with Profile 1 (OR = 1.088, p < 0.05), and a 59.5% increase in the odds of belonging to Profile 4 compared with Profile 1 (OR = 1.595, p < 0.001). In addition, each one-unit increase in handgrip strength was associated with a 37.2% increase in the odds of belonging to Profile 4 compared with Profile 2 (OR = 1.372, p < 0.001) and a 46.7% increase in the odds of belonging to Profile 4 compared with Profile 3 (OR = 1.467, p < 0.001).
For balance, each one-unit increase was associated with a 9.5% increase in the odds of belonging to Profile 2 compared with Profile 1 (OR = 1.095, p < 0.05). In contrast, each one-unit increase in balance was associated with an 11.9% decrease in the odds of belonging to Profile 3 compared with Profile 1 (OR = 0.881, p < 0.05) and a 19.6% decrease in the odds of belonging to Profile 3 compared with Profile 2 (OR = 0.804, p < 0.001).
For flexibility, each one-unit increase was associated with a 3.5% decrease in the odds of belonging to Profile 3 compared with Profile 1 (OR = 0.965, p < 0.05) and a 15.1% decrease in the odds of belonging to Profile 4 compared with Profile 1 (OR = 0.849, p < 0.001). Furthermore, each one-unit increase in flexibility was associated with a 4.1% decrease in the odds of belonging to Profile 3 compared to Profile 2 (OR = 0.959, p < 0.05), a 15.6% decrease in the odds of belonging to Profile 4 compared to Profile 2 (OR = 0.844, p < 0.001), and a 12.0% decrease in the odds of belonging to Profile 4 compared with Profile 3 (OR = 0.880, p < 0.001).
Differences in BP and FPG Across Latent Profiles
One-way ANOVA was conducted to examine differenc-es in SBP and DBP as well as FPG levels among the four groups identified through LPA. The results are presented in Table 7.
Table 7.
Differences in blood pressure and fasting plasma glucose across the latent profiles
| Dependent variable | Profile | Mean (SD) | F | Post-hoc |
|---|---|---|---|---|
| SBP (mmHg) | Profile 1 | 132.53(13.106) | 7.363*** | Profile 1 < Profile 3*** |
| Profile 2 | 135.14(14.976) | |||
| Profile 3 | 138.68(12.997) | Profile 4 < Profile 3** | ||
| Profile 4 | 130.28(13.000) | |||
| Total | 134.47(13.777) | |||
| DBP (mmHg) | Profile 1 | 79.33(7.796) | 3.517* | Profile 1 > Profile 3* |
| Profile 2 | 79.03(7.695) | |||
| Profile 3 | 81.97(9.274) | |||
| Profile 4 | 79.67(7.135) | |||
| Total | 79.99(8.179) | |||
| FPG (mg/dL) | Profile 1 | 122.08(21.940) | 3.657* | Profile 4 < Profile 1* |
| Profile 2 | 120.35(15.192) | |||
| Profile 3 | 123.44(16.918) | Profile 4 < Profile 3* | ||
| Profile 4 | 114.00(14.996) | |||
| Total | 121.12(18.708) |
Note: Values are presented as mean (SD). P1 = Profile 1 (Low Muscle); P2 = Profile 2 (Balanced); P3 = Profile 3 (high adiposity); P4 = Profile 4 (high muscle/low fat); SBP = systolic blood pressure; DBP = diastolic blood pressure; FPG = fasting plasma glucose; η² = eta squared. Pairwise comparisons were performed using Scheffé’s post hoc test. Exact p values are presented for all pairwise comparisons.
The results showed statistically significant differences in SBP, DBP, and fasting blood glucose levels among the four groups (p < 0.05). For SBP, the mean values were 132.53 ± 13.106 mmHg for Profile 1, 135.14 ± 14.976 mmHg for Profile 2, 138.68 ± 12.997 mmHg for Profile 3, and 130.28 ± 13.000 mmHg for Profile 4. For DBP, the mean values were 79.33 ± 7.796 mmHg for Profile 1, 79.03 ± 7.695 mmHg for Profile 2, 81.97 ± 9.274 mmHg for Profile 3, and 79.67 ± 7.135 mmHg for Profile 4. For FPG, the mean values were 122.08 ± 21.940 mg/dL for Profile 1, 120.35 ± 15.192 mg/dL for Profile 2, 123.44 ± 16.918 mg/dL for Profile 3, and 114.00 ± 14.996 mg/dL for Profile 4.
The post-hoc analysis showed that for SBP, Profile 3 had significantly higher values than Profiles 1 and 4, whereas no significant differences were observed among the other profiles. Profile 3 had significantly higher DBP values than Profile 1, with no significant differences between the other profiles. For FPG, Profile 4 had significantly lower values than Profiles 1 and 3, whereas no significant differences were observed among the other profiles.
DISCUSSION
This study applied LPA to body composition indicators in community-dwelling older adults with hypertension to identify heterogeneous body composition profiles, examine the associations of age and physical fitness components with profile membership, and determine differences in BP and FPG levels across profiles. Four distinct profiles were identified: low muscle mass, balanced, high adiposity, and high muscle/low fat ratio. Handgrip strength, balance, and flexibility were significant factors associated with profile membership, and significant differences in BP and FPG levels were observed across profiles. These findings reveal that, even among older adults with hypertension, physical and metabolic characteristics may differ according to body composition patterns and physical fitness levels.
The four profiles identified reveal that age-related changes in body composition do not occur uniformly in older adults. Aging involves a combination of muscle loss and fat accumulation, and these changes may manifest differently depending on factors such as physical activity, nutritional status, and underlying health conditions [29,30]. A recent study applying LPA to BIA-derived body composition indicators identified five distinct profiles and showed that profiles characterized by higher body fat levels were more strongly associated with diabetes, cardiovascular disease, and hypertension [13]. In this study, even within a population of older adults with hypertension, individuals with relatively low muscle mass but average body fat percentage (low muscle profile) could be distinguished from those characterized by both high body fat percentage and low muscle mass (high adiposity profile).
The high adiposity profile was characterized by a high body fat percentage and BMI, relatively low muscle mass and physical fitness levels, and higher systolic and diastolic BP than the other profiles. Obesity-related increases in BP involve several mechanisms, including activation of the sympathetic nervous system and renin-angiotensin-aldosterone system, increased renal sodium reabsorption, and adipose tissue-derived inflammatory responses [31]. Older adults with both obesity and reduced muscle mass or strength are at greater risk of physical functional limitations [30]. In contrast, the high muscle/low fat profile was characterized by high muscle mass, EBMR, low body fat percentage, relatively high handgrip strength and balance, and the lowest FPG level. Because of the important role of skeletal muscles in insulin-stimulated glucose uptake and utilization, the combination of greater muscle mass and lower body fat may contribute to a more favorable glycemic profile [32].
The low muscle profile was characterized by low muscle mass, handgrip strength, and balance despite an average body fat percentage, whereas FPG levels were comparable to those observed in the high adiposity profile. Considering that the skeletal muscle is the major tissue responsible for glucose uptake and utilization, lower muscle mass and function may be associated with FPG levels, even in the absence of marked fat accumulation [32,33]. In contrast, although the balanced profile showed muscle mass and body fat percentages close to the overall sample means, the systolic BP was higher than that of the high-muscle/low-fat profile. These findings reveal that average body composition does not necessarily correspond to the most favorable cardiovascular and metabolic characteristics. Therefore, in addition to evaluating individual indicators such as BMI or body fat percentage, it may be important to consider how muscle mass and body fat coexist in an individual.
Handgrip strength is a key physical fitness factor associated with healthy body composition profiles. The European Working Group on Sarcopenia in Older People 2 (EWGSOP2) identifies low muscle strength as a key characteristic of sarcopenia and recommended handgrip strength as a representative measure of muscle strength [34]. Handgrip strength is a simple indicator of overall muscle strength in older adults, and low handgrip strength is associated with adverse health outcomes, including functional decline and mortality [35,36]. Handgrip strength was significantly associated with latent profile membership in the present study, indicating that profiles derived solely from body composition indicators are distinguishable in terms of muscle function.
Balance and flexibility were associated with profile membership in the selected pairwise comparisons. The high adiposity profile was characterized by lower handgrip strength and poorer balance in the relevant comparisons. However, balance did not significantly distinguish the high-muscle/low-fat profile from other profiles in pairwise regression analyses. Therefore, this profile cannot be interpreted as an independently superior balance.
The direction of the flexibility results also differed from that of handgrip strength. In several pairwise comparisons, greater flexibility was associated with lower odds of belonging to a high-muscle/low-fat profile. Because this profile was predominantly composed of men, this finding may partly reflect sex-related differences in flexibility and body composition and should not be interpreted as indicating that lower flexibility is beneficial.
Flexibility is an important component of physical fitness that reflects joint range of motion and has been reported to progressively decline with age [37]. Reduced flexibility may lead to limitations in the movements required for the activities of daily living and overall physical function. Accordingly, the American College of Sports Medicine (ACSM) recommends incorporating flexibility exercises, along with aerobic and resistance exercises, into exercise prescriptions for older adults [38]. Collectively, these findings indicate that different dimensions of physical fitness are associated with profile membership in different ways. A comprehensive exercise approach addressing handgrip strength, balance, and flexibility may be relevant for older adults with hypertension; however, profile-specific exercise strategies should be confirmed in longitudinal and interventional studies.
In contrast, age and aerobic endurance were not significantly associated with profile membership. One possible explanation for this is that most participants were concentrated within an older age range, which may have limited the discriminatory ability of chronological age. Considerable individual variability in muscle mass, body fat, and functional fitness exists even among individuals of similar ages; therefore, chronological age alone may not adequately explain the physical heterogeneity among older adults with hypertension [39]. The lack of a significant association with aerobic endurance may also have been influenced by the participant characteristics or sample composition. In addition, the 2-min step test may have had limited discriminatory ability because of a potential ceiling effect, which could have reduced the variability in aerobic endurance scores and attenuated its association with profile membership. Therefore, the null association should be interpreted with caution and further studies using more sensitive measures of aerobic endurance are required to verify this finding.
The significance of applying LPA in the present study lies in its ability to move beyond examining average associations between individual body composition indicators and health outcomes and instead identify heterogeneous patterns in which multiple body composition characteristics coexist within individuals. In particular, the identification of high adiposity and low muscle profiles as distinct subgroups indicated that unfavorable health characteristics among older adults with hypertension do not necessarily occur in conjunction with excessive fat accumulation. Assessments focusing primarily on BMI or body fat percentage may fail to adequately identify older adults who are not markedly obese but have low muscle mass and poor muscle function. Therefore, muscle mass, physical fitness, and body fat should be considered when assessing the physical and metabolic characteristics of older adults with hypertension.
Collectively, management priorities for older adults with hypertension may need to differ according to their body composition profiles, instead of applying a uniform exercise management strategy to all individuals. For those with high adiposity profiles, strategies addressing both body fat reduction and improvement in muscle function may be warranted, whereas for those with low muscle profiles, maintaining and improving muscle mass and function may be prioritized over weight loss. For individuals with a high-muscle/low-fat profile, maintaining a favorable body composition and physical fitness levels may be particularly important. However, because the present study did not examine the effects of exercise intervention, the effectiveness of these profile-specific exercise strategies and the potential for transition between profiles should be investigated in future longitudinal and interventional studies.
This study had certain limitations. First, the sex distribution was highly unbalanced; women constituted most of the overall sample, whereas 85.7% of the participants in Profile 4 were men. This imbalance limits the generalizability of the findings and indicates that the body composition and physical fitness characteristics of Profile 4 may partly reflect sex-related differences. Therefore, Profile 4 should not be interpreted as a favorable sex-independent phenotype. Second, due to the cross-sectional study design, causal relationships between body composition, physical fitness characteristics, and metabolic indicators could not be established. Third, although information on sex was available, it was not accounted for in the comparisons of BP and FPG levels across latent profiles. In addition, information on antihypertensive medication use, treatment duration, diabetes status, and glucose-lowering medication use was unavailable, and therefore could not be accounted for in the analyses. Accordingly, the observed differences in BP and fasting levels should be interpreted as unadjusted between-profile differences. Therefore, future studies should include sufficient older men with hypertension to examine whether the identified profile structure differs by sex and should account for relevant clinical and medication-related factors. Longitudinal follow-up studies are warranted to investigate changes in profile membership over time and their relationships with subsequent cardiovascular outcomes.
In conclusion, this study identified distinct body composition profiles among community-dwelling older adults with hypertension, and revealed that handgrip strength, balance, and flexibility were significantly associated with profile membership. BP and FPG levels also differed across profiles, highlighting the heterogeneity in physical and metabolic characteristics within this population. These findings reveal that a person-centered approach integrating body composition and functional fitness may provide a more comprehensive assessment of older adults with hypertension than a reliance on single indicators such as BMI alone.
Footnotes
FUNDING
This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF- 2023S1A5B5A17089122) by Jae-Hoon Lee.
The authors declare no conflict of interest.
REFERENCES
- 1.NCD Risk Factor Collaboration (NCD-RisC) Worldwide trends in hypertension prevalence and progress in treatment and control from 1990 to 2019: a pooled analysis of 1201 population-representative studies with 104 million participants. Lancet. 2021;398:957–80. doi: 10.1016/S0140-6736(21)01330-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kario K, Okura A, Hoshide S, Mogi M. The WHO global report 2023 on hypertension warning the emerging hypertension burden in globe and its treatment strategy. Hypertens Res. 2024;47:1099–102. doi: 10.1038/s41440-024-01622-w. [DOI] [PubMed] [Google Scholar]
- 3.Wang Z, Wang S, Lin H, Wang C, Gao D. Prevalence of hypertension and related risk factors in older Chinese population: a meta-analysis. Front Public Health. 2024;12:1320295. doi: 10.3389/fpubh.2024.1320295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kim HC, Lee H, Lee HH, Ahn SV, Lee JM, Cheon DY, Jhee JH, Yoon M, Shin MH, Heo J, Kim E, Lee SW, Lee J, Oh YW, Jeon J, Cho M, Son D, Ahn NY NY, Korean Society of Hypertension (KSH) - Hypertension Epidemiology Research Working Group Korea hypertension fact sheet 2024: nationwide population-based analysis with a focus on young adults. Clin Hypertens. 2025;31:e11. doi: 10.5646/ch.2025.31.e11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Supiano MA, Ascher SB, Rich MW. Hypertension in the oldest old. JACC Adv. 2025;4:102306. doi: 10.1016/j.jacadv.2025.102306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tu Q, Hyun K, Lin S, Hafiz N, Manandi D, Li E, Wang X, Wu H, Redfern J. Impacts of hypertension and diabetes on the incidence of cardiovascular diseases and all-cause mortality: findings from the China Health and Retirement Longitudinal Study cohort. J Hypertens. 2025;43:623–30. doi: 10.1097/HJH.0000000000003946. [DOI] [PubMed] [Google Scholar]
- 7.Su Q, Huang WZ, Huang Y, Mo LR, Lin JH, Qin Z. Synergistic effects of hypertension and diabetes on cardiovascular risk in elderly: comparative longitudinal analysis of CHARLS and HRS. Diabetes Obes Metab. 2026;28:3317–34. doi: 10.1111/dom.70528. [DOI] [PubMed] [Google Scholar]
- 8.Palmer AK, Jensen MD. Metabolic changes in aging humans: current evidence and therapeutic strategies. J Clin Invest. 2022;132:e158451. doi: 10.1172/JCI158451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Liu C, Wong PY, Chung YL, Chow SKH, Cheung WH, Law SW, Chan JCN, Wong RMY. Deciphering the “obesity paradox” in the elderly: a systematic review and meta-analysis of sarcopenic obesity. Obes Rev. 2023;24:e13534. doi: 10.1111/obr.13534. [DOI] [PubMed] [Google Scholar]
- 10.Roh E, Choi KM. Health consequences of sarcopenic obesity: a narrative review. Front Endocrinol (Lausanne) 2020;11:332. doi: 10.3389/fendo.2020.00332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kim A, Kim E, Baek S, Shin J, Choi S. Analysis of health-related, skill-related physical strength, and physique by blood pressure levels of Korean seniors. PLoS One. 2022;17:e0279264. doi: 10.1371/journal.pone.0279264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Polo-López A, Calatayud J, Núñez-Cortés R, Andersen LL, Moya-Ramón M, López-Bueno R. Dose-response association between handgrip strength and hypertension: a longitudinal study of 76,503 European older adults. Curr Probl Cardiol. 2023;48:101813. doi: 10.1016/j.cpcardiol.2023.101813. [DOI] [PubMed] [Google Scholar]
- 13.Sharafi M, Razavi FS, Baeradeh N, Bagheri P, Rousta A, Afrashteh S, Shiravani M, Haghjoo E. Latent profiles of body composition indices and their associations with cardiovascular diseases, hypertension, and diabetes: results from the Fasa Adult Cohort study (FACS) J Health Popul Nutr. 2025;44:228. doi: 10.1186/s41043-025-00980-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Quan Y, Wang C, Wang L, Li G. Geriatric sarcopenia is associated with hypertension: a systematic review and meta-analysis. J Clin Hypertens (Greenwich) 2023;25:808–16. doi: 10.1111/jch.14714. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Evans K, Abdelhafiz D, Abdelhafiz AH. Sarcopenic obesity as a determinant of cardiovascular disease risk in older people: a systematic review. Postgrad Med. 2021;133:831–42. doi: 10.1080/00325481.2021.1942934. [DOI] [PubMed] [Google Scholar]
- 16.Tein JY, Coxe S, Cham H. Statistical power to detect the correct number of classes in latent profile analysis. Struct Equ Modeling. 2013;20:640–57. doi: 10.1080/10705511.2013.824781. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Nylund KL, Asparouhov T, Muthén BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Modeling. 2007;14:535–69. [Google Scholar]
- 18.Spurk D, Hirschi A, Wang M, Valero D, Kauffeld S. Latent profile analysis: a review and “how to” guide of its application within vocational behavior research. J Vocat Behav. 2020;120:103445. [Google Scholar]
- 19.Williams B, Mancia G, Spiering W, Agabiti Rosei E, Azizi M, Burnier M, Clement DL, Coca A, de Simone G, Dominiczak A, Kahan T, Mahfoud F, Redon J, Ruilope L, Zanchetti A, Kerins M, Kjeldsen SE, Kreutz R, Laurent S, Lip GYH, McManus R, Narkiewicz K, Ruschitzka F, Schmieder RE, Shlyakhto E, Tsioufis C, Aboyans V, Desormais I. 2018 ESC/ESH guidelines for the management of arterial hypertension: the Task Force for the management of arterial hypertension of the European Society of Cardiology and the European Society of Hypertension. J Hypertens. 2018;36:1953–2041. doi: 10.1097/HJH.0000000000001940. [DOI] [PubMed] [Google Scholar]
- 20.Rantanen T, Guralnik JM, Foley D, Masaki K, Leveille S, Curb JD, White L. Midlife hand grip strength as a predictor of old age disability. JAMA. 1999;281:558–60. doi: 10.1001/jama.281.6.558. [DOI] [PubMed] [Google Scholar]
- 21.Springer BA, Marin R, Cyhan T, Roberts H, Gill NW. Normative values for the unipedal stance test with eyes open and closed. J Geriatr Phys Ther. 2007;30:8–15. doi: 10.1519/00139143-200704000-00003. [DOI] [PubMed] [Google Scholar]
- 22.Rikli RE, Jones CJ. Development and validation of a functional fitness test for community-residing older adults. J Aging Phys Act. 1999;7:129–61. [Google Scholar]
- 23.Akaike H. A new look at the statistical model identification. IEEE Trans Automat Control. 1974;19:716–23. [Google Scholar]
- 24.Schwarz G. Estimating the dimension of a model. Ann Stat. 1978;6:461–4. [Google Scholar]
- 25.Sclove SL. Application of model-selection criteria to some problems in multivariate analysis. Psychometrika. 1987;52:333–43. [Google Scholar]
- 26.Clark SL. Los Angeles (CA): University of California, Los Angeles; 2010. Mixture modeling with behavioral data [dissertation] [Google Scholar]
- 27.Asparouhov T, Muthén B. Auxiliary variables in mixture modeling: three-step approaches using Mplus. Struct Equ Modeling. 2014;21:329–41. [Google Scholar]
- 28.Nagin DS, Tremblay RE. What has been learned from groupbased trajectory modeling? Examples from physical aggression and other problem behaviors. Ann Am Acad Pol Soc Sci. 2005;602:82–117. [Google Scholar]
- 29.Baumgartner RN, Wayne SJ, Waters DL, Janssen I, Gallagher D, Morley JE. Sarcopenic obesity predicts instrumental activities of daily living disability in the elderly. Obes Res. 2004;12:1995–2004. doi: 10.1038/oby.2004.250. [DOI] [PubMed] [Google Scholar]
- 30.Stenholm S, Harris TB, Rantanen T, Visser M, Kritchevsky SB, Ferrucci L. Sarcopenic obesity: definition, cause and consequences. Curr Opin Clin Nutr Metab Care. 2008;11:693–700. doi: 10.1097/MCO.0b013e328312c37d. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hall JE, do Carmo JM, da Silva AA, Wang Z, Hall ME. Obesity-induced hypertension: interaction of neurohumoral and renal mechanisms. Circ Res. 2015;116:991–1006. doi: 10.1161/CIRCRESAHA.116.305697. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Goodpaster BH, Sparks LM. Metabolic flexibility in health and disease. Cell Metab. 2017;25:1027–36. doi: 10.1016/j.cmet.2017.04.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Hong SH, Choi KM. Sarcopenic obesity, insulin resistance, and their implications in cardiovascular and metabolic consequences. Int J Mol Sci. 2020;21:494. doi: 10.3390/ijms21020494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Cruz-Jentoft AJ, Bahat G, Bauer J, Boirie Y, Bruyère O, Cederholm T, Cooper C, Landi F, Rolland Y, Sayer AA, Schneider SM, Sieber CC, Topinkova E, Vandervo u de M, Visser M, Zamboni M; Writing Group for the European Working Group on Sarcopenia in Older People 2 (EWGSOP2), and the Extended Group for EWGSOP2. Sarcopenia: revised European consensus on definition and diagnosis. Age Ageing. 2019;48:16–31. doi: 10.1093/ageing/afy169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Rijk JM, Roos PR, Deckx L, van den Akker M, Buntinx F. Prognostic value of handgrip strength in people aged 60 years and older: a systematic review and meta-analysis. Geriatr Gerontol Int. 2016;16:5–20. doi: 10.1111/ggi.12508. [DOI] [PubMed] [Google Scholar]
- 36.McGrath RP, Kraemer WJ, Al Snih S, Peterson MD. Handgrip strength and health in aging adults. Sports Med. 2018;48:1993–2000. doi: 10.1007/s40279-018-0952-y. [DOI] [PubMed] [Google Scholar]
- 37.Stathokostas L, Little RM, Vandervoort AA, Paterson DH. Flexibility training and functional ability in older adults: a systematic review. J Aging Res. 2012;2012:306818. doi: 10.1155/2012/306818. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.American College of Sports Medicine. Chodzko-Zajko WJ, Proctor DN, Fiatarone Singh MA, Minson CT, Nigg CR, Salem GJ, Skinner JS. American College of Sports Medicine position stand. Exercise and physical activity for older adults. Med Sci Sports Exerc. 2009;41:1510–30. doi: 10.1249/MSS.0b013e3181a0c95c. [DOI] [PubMed] [Google Scholar]
- 39.Toyoshima K, Seino S, Tamura Y, Ishikawa J, Chiba Y, Ishizaki T, Fujiwara Y, Shinkai S, Kitamura A, Araki A. Difference between “physical fitness age” based on physical function and chronological age is associated with obesity, hyperglycemia, depressive symptoms, and low serum albumin. J Nutr Health Aging. 2022;26:501–9. doi: 10.1007/s12603-022-1786-8. [DOI] [PMC free article] [PubMed] [Google Scholar]