Corresponding author: Saejong Park Tel +82-2-970-9562 Fax +82-2-970-9502 E-mail INTRODUCTION
Metabolic syndrome (MetS) is a cluster of abdominal obesity, elevated blood pressure, dyslipidemia, and hyperglycemia that is associated with higher risks of type 2 diabetes and cardiovascular disease [1]. Because MetS develops through the accumulation of interrelated lifestyle, adiposity, and fitness-related risk factors, identifying simple physical-function indicators that can be used before overt cardiometabolic disease becomes clinically apparent is important. In community and exercise-screening settings, such indicators should be inexpensive, easy to administer, and interpretable across age and sex groups.
Handgrip strength is a practical marker of muscular strength and has been associated with mobility limitation, morbidity, and mortality [2-4]. Its value is particularly relevant for population-based health screening because measurement requires minimal time and equipment. However, the interpretation of handgrip strength is complicated by the fact that larger individuals often have higher force production even when they also carry greater cardiometabolic risk. Therefore, strength indices that account for body size may be more informative than raw force values when the outcome of interest is MetS.
Absolute handgrip strength (AHS) is strongly influenced by body size. In cardiometabolic research, this can be problematic because larger body size may be associated with both higher AHS and greater metabolic risk. Relative handgrip strength (RHS), commonly operationalized as AHS normalized to body mass index (BMI), partly addresses this issue and has shown stronger associations with cardiometabolic biomarkers than AHS in U.S., Taiwanese, and Chinese adults [5-7]. Although recent normalization studies indicate that height- or surface-area-based scaling may reduce body-size effects more completely than BMI in some settings [8-10], BMI-based RHS remains a widely used and practical index for large-scale epidemiological analyses and exercise-health screening because BMI is routinely measured and easily understood.
The need for Korean-specific evidence is also important because age, sex, body size, and population context influence grip-strength levels and their metabolic interpretation. Korean population-based studies already support an inverse association between RHS and MetS. Analyses of Korea National Health and Nutrition Examination Survey data reported lower odds of MetS at higher levels of RHS in Korean adults and Korean older adults [11,12]. Other Korean adult studies likewise found that lower relative handgrip strength was associated with a higher prevalence of MetS [13,14], and a recent national study proposed age- and sex-specific cutoff values for screening purposes [15].
Despite this foundation, several gaps remain. Many prior Korean studies focused on associations across quartiles or tertiles of RHS, which is useful etiologically but less directly applicable to fitness counselling or field-based screening. In practice, health professionals often need interpretable thresholds that identify individuals who may warrant further metabolic assessment or targeted exercise intervention. Age- and sex-specific RHS thresholds may therefore provide a more usable framework than a single pooled value, especially because both muscular strength and MetS prevalence change substantially across adulthood.
Cardiorespiratory fitness (CRF) is a major determinant of metabolic health and could confound strength-MetS associations. Prospective and cross-sectional studies have shown that higher CRF is associated with lower incident or prevalent MetS, including Korean studies using measured or estimated CRF [16-19]. Another reason to examine RHS thresholds in the Korea Institute of Sport Science Fitness Standards (KISS FitS) project is that the data were obtained through standardized anthropometric, blood, and physical fitness assessments rather than a general health survey alone [20]. This setting is close to the context in which exercise professionals may use handgrip strength screening.
Therefore, the present study aimed to derive age- and sex-specific exploratory RHS thresholds for identifying prevalent MetS in Korean adults from the KISS FitS study and to examine whether lower RHS categories were associated with higher odds of prevalent MetS after adjustment for selected covariates, including estimated CRF. We also compared the discriminatory performance of RHS and AHS.
METHODS
1. Study design and participants
This study was a cross-sectional analysis of adults who participated in the 2014-2015 KISS FitS project, which was conducted to develop criterion-referenced health-related fitness standards for the National Fitness Award program [20]. Adults aged 19 years or older were recruited from the Seoul metropolitan area and four Korean provinces (Gangwon, Jeolla, Gyeongnam, and Gyeongbuk). The parent project used Neyman allocation and a stratified-cluster sampling framework [20]. Exclusion criteria included unstable cardiovascular disease, acute systemic infection, and musculoskeletal disorders that precluded fitness testing. Among 3,496 adults initially enrolled, 189 participants (57 men and 132 women) were excluded because of missing fitness tests, failure to meet the exercise-test completion criteria, or missing blood samples. The final analytic sample included 3,307 participants (1,427 men and 1,880 women). Medication use, alcohol intake, smoking status, family history, and physical activity were recorded by trained research staff using standardized questionnaires.
The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of the Korea Institute of Sport Science (KISS-201504-EFS-002; approval date: May 4, 2015). Written informed consent was obtained from all participants.
2. Clinical and laboratory assessments
Weight and body composition were measured with a bioelectrical impedance analyzer (InBody 720; Biospace, Seoul, Korea), and height was measured with a stadiometer to the nearest 0.1 cm. BMI was calculated as weight divided by height squared (kg/m2). The validity of the multifrequency bioelectrical impedance approach has been reported previously [21]. Waist circumference was measured at the level of the umbilicus. Resting systolic and diastolic blood pressure were measured twice in the seated position after at least 5 minutes of rest, and participants were asked to avoid caffeine and smoking for at least 30 minutes before measurement (HEM-7080IC; Omron, Kyoto, Japan). Fasting blood samples were collected after an overnight fast of at least 12 hours. Total cholesterol, triglycerides, high-density lipoprotein cholesterol, glucose, and insulin were analyzed by standardized enzymatic methods at the Seegene Medical Foundation (Seoul, Korea).
All participants completed a standardized medical history and lifestyle questionnaire administered by trained research staff. The questionnaire recorded self-reported health-history and cardiovascular risk-factor items, including family history of cardiovascular disease and histories of obesity, hypertension, and hyperlipidemia, as binary variables (0 =no, 1=yes). Additional binary items assessed histories of stroke, myocardial infarction, and angina, as well as cardiovascular signs or symptoms; women also reported menopausal status. Physical activity was available in the analytic dataset as a binary sedentary-lifestyle/physical-inactivity item (0 =no, 1=yes), rather than as frequency, duration, or weekly exercise volume. Smoking was not categorized as never, former, or current in the analytic dataset; instead, a dichotomous smoking-related risk factor was defined as present when any of the three smoking-related binary items was positive and absent when all three were negative. Alcohol intake was coded dichotomously using the drinking-status item (1=reported alcohol intake, 2 =no alcohol intake), and additional ordinal items on drinking frequency and amount were not used as covariates in the present analyses.
3. Physical fitness assessments
Handgrip strength was measured with a digital dynamometer (Takei Scientific Instruments Co., Ltd., Niigata, Japan). Participants performed two trials with each hand, and the highest value obtained from either hand was recorded as AHS (kg). RHS was calculated as AHS divided by BMI (kg/BMI). CRF was assessed using a symptom-limited maximal treadmill test based on the Bruce protocol for adults and the modified Bruce protocol for older adults (Takei Scientific Instruments Co., Ltd., Niigata, Japan) [22]. Tests were terminated for volitional exhaustion or other standard indications, and peak oxygen uptake (VO2peak) was estimated from total treadmill time using the equations applied in the KISS FitS project [20].
4. Classification of metabolic syndrome
MetS was defined according to the modified National Cholesterol Education Program Adult Treatment Panel III criteria [1]. Participants were classified as having MetS when they met at least three of the following criteria: (1) waist circumference 90 cm or greater in men or 85 cm or greater in women, based on Korean-specific cutoffs [23]; (2) systolic/diastolic blood pressure 130/85 mmHg or greater or use of antihypertensive medication; (3) triglycerides 150 mg/dL or greater or treatment for dyslipidemia; (4) high-density lipoprotein cholesterol lower than 40 mg/dL in men or lower than 50 mg/dL in women; and (5) fasting glucose 100 mg/dL or greater or use of diabetes medication.
5. Statistical analysis
Continuous variables are presented as mean±standard deviation, and questionnaire-derived categorical variables are presented as absolute frequencies and percentages. Participants’ characteristics were compared by sex and age group (50 years or younger vs. older than 50 years) using independent t-tests for continuous variables and chi-square tests for categorical variables. Receiver operating characteristic (ROC) curve analysis was used to compare the discriminatory performance of AHS and RHS for prevalent MetS and its individual components. Area under the curve (AUC) values with 95% confidence intervals (CIs) were calculated separately for men and women. Additional sex- and age-specific ROC analyses were then used to derive exploratory RHS thresholds for MetS in participants aged 50 years or younger and older than 50 years.
Because the purpose of this analysis was practical risk stratification rather than establishment of a single clinical decision limit, two thresholds were selected from each receiver operating characteristic curve: a sensitivity-oriented threshold and a specificity-oriented threshold [24]. These thresholds were used to construct three RHS categories within each sex and age stratum: reference (RHS above the sensitivity-oriented threshold), intermediate RHS (RHS between the sensitivity- and specificity-oriented thresholds), and low RHS (RHS at or below the specificity-oriented threshold). Finally, logistic regression was used to estimate odds ratios and 95% CIs for prevalent MetS across RHS categories. Model 1 adjusted for age, cigarette smoking, alcohol intake, family history, systolic blood pressure, and physical activity. Model 2 additionally adjusted for estimated CRF. Because MetS includes a blood pressure component, the adjusted models that include systolic blood pressure are presented as exploratory rather than etiologic models. All analyses were performed using SPSS version 21.0 (IBM Corp., Armonk, NY, USA) and MedCalc version 17.2 (MedCalc Software Ltd., Ostend, Belgium), and p<.05 was considered statistically significant.
RESULTS
The descriptive characteristics of the participants are presented in Table 1. Most anthropometric, metabolic, and fitness variables differed significantly between the younger and older age groups in both sexes. In men, BMI was lower in participants older than 50 years than in those aged 50 years or younger, whereas in women BMI was higher in the older group. AHS and RHS, as well as estimated VO2peak, were lower in older adults of both sexes. Several questionnaire-derived categorical variables also differed by age group. In men, smoking-related risk-factor positivity and alcohol intake were less frequent in the older group, whereas hypertension and hyperlipidemia histories were more frequent. In women, older participants had higher prevalences of family history of cardiovascular disease, obesity history, hypertension history, and hyperlipidemia history, but a lower prevalence of alcohol intake. Physical inactivity did not differ significantly by age group in either sex.
The prevalence of MetS was 18.15% in men and 24.36% in women. As shown in Table 2, RHS discriminated prevalent MetS better than AHS in both men (AUC, 0.612 compared with 0.533) and women (0.717 compared with 0.588). In women, RHS outperformed AHS across all individual MetS components. In men, RHS showed better discrimination for waist circumference, blood pressure, high-density lipoprotein cholesterol, and glucose, whereas it did not outperform AHS for triglycerides/dyslipidemia.
Age- and sex-specific exploratory RHS thresholds are shown in Table 3. Thresholds were consistently higher in younger adults than in older adults. The underlying receiver operating characteristic discrimination for prevalent MetS was modest in men (area under the curve, 0.610 for age 50 years or younger and 0.598 for age older than 50 years) and moderate in women (0.682 and 0.664, respectively). The sensitivity-oriented and specificity-oriented thresholds were 1.85 and 1.57 kg/BMI in men aged 50 years or younger, 1.64 and 1.36 kg/BMI in men older than 50 years, 1.21 and 1.01 kg/BMI in women aged 50 years or younger, and 1.02 and 0.84 kg/BMI in women older than 50 years.
The associations between age- and sex-specific RHS categories and prevalent MetS are presented in Table 4. In unadjusted analyses, both intermediate and low RHS categories were associated with higher odds of prevalent MetS in men and women. The graded association remained in the adjusted exploratory models. In Model 2, compared with the reference group, men in the intermediate and low RHS groups had odds ratios of 1.916 (95% CI, 1.161–3.163) and 2.209 (95% CI, 1.251–3.900), respectively. The corresponding odds ratios in women were 1.786 (95% CI, 1.089–2.930) and 2.077 (95% CI, 1.176–3.668).
DISCUSSION
In this cross-sectional analysis of a standardized fitness-testing cohort, RHS discriminated prevalent MetS better than AHS in both sexes, with stronger discrimination in women than in men. We also found that exploratory RHS thresholds decreased with age and that lower RHS categories were associated with higher odds of prevalent MetS in exploratory adjusted models. Rather than establishing definitive reference values, our findings provide cohort-specific screening bands derived from the KISS FitS study.
Our results are broadly consistent with previous Korean studies. Korea National Health and Nutrition Examination Survey analyses have shown inverse associations between RHS and MetS in Korean adults [12,14], Korean older adults [11], and, most recently, a national sample that also proposed age- and sex-specific cutoffs [15]. Smaller Korean adult studies reported similar discrepancies between AHS and RHS [13], and a recent longitudinal Korean analysis found that RHS predicted incident MetS better than AHS over 8 years [25]. The present study does not replace that literature; instead, it extends it by showing similar patterns in a cohort assembled for standardized fitness assessment and by examining threshold-based categories in models that also considered estimated CRF.
The better performance of RHS than AHS is biologically plausible. Previous studies from U.S., Taiwanese, Chinese, and Korean populations have shown that RHS rather than AHS is more consistently linked to a favorable cardiometabolic profile or lower MetS prevalence [5-7,13,14]. At the same time, BMI-normalized RHS should be viewed as a pragmatic rather than definitive scaling strategy. Recent allometric and normalization studies suggest that height-based or surface-area-based scaling may remove body-size effects more completely than BMI in some settings [8-10]. Because BMI is in the denominator of RHS and MetS includes an adiposity-related component, part of the observed discrimination may reflect body size and adiposity rather than muscular strength alone. Accordingly, the thresholds reported here should be interpreted as pragmatic screening indicators rather than pure markers of muscle function.
The stronger discrimination observed in women is compatible with known sex differences in body composition and fat distribution [26]. The age-related decline in thresholds was also expected and aligns with normative data showing lower grip strength with advancing age across adulthood [27-29]. The dual-threshold approach may be useful for communication in community or fitness settings. A sensitivity-oriented threshold can flag individuals who may warrant further metabolic assessment, whereas a specificity-oriented threshold identifies a smaller group with a higher likelihood of prevalent MetS. However, these bands are exploratory and should not be treated as clinical diagnostic cutpoints.
Our area under the curve values deserve cautious interpretation. In men, discrimination was modest, and in women it was moderate. This pattern is compatible with the broader literature, including a recent systematic review and meta-analysis showing an inverse association between lower handgrip strength and higher MetS risk [30], earlier Korean work linking muscle strength to MetS in older adults [31], and studies that derived strength cut-points for MetS in Korean adults or in male cohorts [15,32]. It also remains plausible that CRF partly explains or modifies some of the observed strength-MetS relationship, because CRF is itself strongly related to metabolic health [16-19,33]. Thus, grip-based measures should be viewed as low-burden screening aids rather than standalone diagnostic tools.
This study has several strengths, including a relatively large sample size, standardized anthropometric and fitness assessments, sex- and age-specific analyses, and consideration of estimated CRF. Several limitations should also be acknowledged. First, the cross-sectional design precludes causal inference and does not support claims about prevention or incident disease. Second, the participants were drawn from a fitness-standardization project rather than a nationally representative survey, which may limit generalizability. Third, the age split at 50 years was pragmatic and may not represent the optimal biological threshold for change in the strength-MetS relationship. Fourth, the adjusted models are exploratory because MetS includes a blood pressure component and the outcome was modeled with adjustment for systolic blood pressure. Fifth, CRF was estimated rather than directly measured, and some covariates were based on self-report. Sixth, BMI-based RHS was chosen for comparability with prior Korean MetS studies, but alternative normalization strategies may deserve direct comparison in future work [8-10]. Finally, the thresholds were derived and tested within the same cohort and therefore require external validation before broader application.
An additional limitation is the elapsed time since data collection. The KISS FitS measurements were obtained in 2014-2015, more than a decade before the current 2026 public-health context. National fitness levels, physical activity patterns, and cardiometabolic profiles in Korean adults may have changed during this interval because of population aging, secular changes in obesity and health behaviors, and especially COVID-19-related disruptions to exercise and daily mobility. Recent nationally representative Korean studies reported reductions in sufficient aerobic physical activity during the pandemic and changes in health behaviors including inadequate physical activity [34,35]. Consequently, the present thresholds should be interpreted as historical, pre-pandemic cohort-derived screening values, and their direct application to contemporary Korean adults should be made cautiously until they are validated using more recent post-pandemic datasets.
CONCLUSION
RHS outperformed AHS for identifying prevalent MetS in this standardized fitness-testing cohort and showed modest discrimination in men and moderate discrimination in women. The age- and sex-specific thresholds reported here provide a practical framework for exploratory exercise-health screening in community and fitness settings, although confirmation in independent cohorts would strengthen their broader applicability.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Print 
