Normative values of hand grip strength in a large unselected Chinese population: Evidence from the China National Health Survey

Abstract Background Hand grip strength (HGS) is a powerful indicator of sarcopenia and other adverse health outcomes. Normative values for HGS for general Chinese people with a broad age spectrum are lacking. This study aims to establish normative values of HGS and explore the correlations between HGS and body composition among unselected people aged 8–80 in China. Methods From 2012 to 2017, 39 655 participants aged 8–80 years in the China National Health Survey were included. Absolute HGS was measured using a Jamar dynamometer. The relative HGS was normalized by body mass index. Body composition indexes included body mass index, body fat percentage, muscle mass, fat mass index (FMI) and muscle mass index (MMI). Sex‐specific smoothed centile tables for the P1, P5, P25, P50, P75, P95 and P99 centiles of HGS and body composition were generated using lambda‐mu‐sigma method. The correlations between muscle strength and body composition were estimated by partial Spearman correlation analysis. Results The median values (25th and 75th percentile) of HGS in boys and girls (8–19 years old) were 22 (14, 34) kg and 18 (12, 22) kg, respectively; in men and women aged 20–80 were 39 (33, 44) kg and 24 (20, 27) kg, respectively. Values of upper and lower HGS across ages had three periods: an increase to a peak in the 20 s in men (with the 5th and 95th values of 30 and 55 kg, respectively) and 30 s in women (with the 5th and 95th values of 18 and 34 kg, respectively), preservation through midlife (20s–40 s), and then a decline after their 50 s. The lowest HGS values in both sexes were in the 70‐ to 80‐year‐old group, with the 5th and 95th percentile values of 16 and 40 kg in men, and 10 and 25 kg in women. There were substantial sex differences in body composition in the life course (all P values <0.001). In ageing, the decrease of muscle strength was faster than that of muscle mass in both sexes. The correlations between muscle mass and HGS were most robust than other correlations, especially in women (0.68 vs. 0.50), children and adolescents. Conclusions Our study established the age‐ and sex‐specific percentile reference values for hand grip strength in an unselected Chinese population across a broad age‐spectrum. The rich data can facilitate the practical appraisal of muscle strength and promote early prediction of sarcopenia and other impairments associated with neuromuscular disorders.


Introduction
Defined as a progressive and generalized skeletal muscle disorder that involves the accelerated loss of muscle mass and function, sarcopenia has raised tremendous concern world widely because of its high prevalence. 1,2 In the function-centred model for older people healthcare, muscle strength is included in the construct of intrinsic capacity that could merit lifelong monitoring. 3 Published research demonstrated that hand grip strength (HGS) is acceptable to be used as an indicator for an individual's overall muscle strength 4 and is currently recommended to measure muscle function in clinical practice. 5 Furthermore, as an inexpensive risk stratifying test, HGS measurement may be best suited to resource-challenged settings. Hence, the normative value of HGS can be used as a valuable reference to monitor muscle loss and identify early sarcopenia. HGS measured by handheld dynamometry is a simple but powerful predictor of future morbidity and mortality and can also be considered as an excellent noninvasive means to predict lifetime health status. 6,7 Body composition mainly encompasses both fat mass (FM) and muscle mass (MM) and is observed to vary with age. 8 As previous studies have found that there were significant associations between body composition and muscle strength, [9][10][11] it is essential to explore their co-change pattern and age-and sex-specific correlations to conduct a more comprehensive health assessment.
Several normative values for HGS have been established in people from Western countries. 4,[12][13][14][15] However, there are variations in HGS values among countries, which could be attributed to significant differences in lifetime exposures, body size and composition. 16 The PURE study that assessed grip strength in adults aged 35-70 who reside in 21 countries suggested that individual HGS measurements should be interpreted using region-specific reference ranges. 17 Although some studies have explored the HGS values among Asian people, 11,18,19 they focused on elders or in a selected population, for example, in healthcare industry workers, lacking data with a broad age range in a general unselected population.
Currently, the normative values of HGS for a general Chinese population with a broad age spectrum are still unclear. The China National Health Survey (CNHS) provides an optimal study sample to investigate such a study topic. From 2012 to 2017, CNHS established a representative general population sample of Chinese people, and one of the purposes of CNHS is to investigate reference intervals for physiological constants. 20 Therefore, using large-scale population-based data from CNHS, we aimed to learn the normative values of muscle strength and their correlations with body composition among Chinese people aged 8-80.

Data resource and study population
Data in this study were from China National Health Survey. The protocol of CNHS has been published previously. 20 Briefly, using a multistage stratified cluster sampling method, 11 provinces from mainland China were selected to conduct the health survey. Individuals from the selected communities and villages were all invited to participate. The inclusion criteria were adults aged 20-80 and those having lived in the local area for at least 1 year. The exclusion criteria were people with severe mental or physical disorders, pregnant women, or person on active military duty, or foreigners. In four out of the selected 11 provinces, children and adolescents aged 8-19 years were additionally recruited. The study has been carried out in accordance with the Declaration of Helsinki. Ethical approval was obtained from the Bioethical Committee of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (No. 029-2013). Written informed consent was obtained from the parent/legal guardian of participants younger than 16 and participants above 16.

Hand grip strength measurement
Hand grip strength was measured by a qualified staff after training, using Jamar Hydraulic Hand Evaluation Kit (JAMAR, UK). After completing a practice test, each participant was asked to squeeze the dynamometer twice as hard as possible for 3 seconds, with at least 30 s rests between measurements, using the dominant arm, in a standing position with the arms extended straight down to the side. Participants were excluded if they reported hand or wrist surgery in the preceding 3 months or could not hold the dynamometer with the testing hand. Absolute hand grip strength was calculated as the largest reading and expressed in kilograms. Relative hand grip strength was calculated as absolute grip strength divided by body mass index (BMI).

Anthropometry and body composition measurement
Height was measured by a fixed stadiometer. Body composition (body fat percentage, total body water, fat mass, fat free mass and muscle mass) was measured by a body composition analyser (TANITA BC-420, Japan), according to the manufacturers' recommendations. 21 Participants were asked to step on the electrodes with bare feet and light clothing. Individuals with a pacemaker or other internal medical devices were excluded. The 50 kHz impedance was used for the calculation of body composition indexes. The records were accurate to one decimal place. BMI was calculated as weight in kilograms divided by the square of height in meters (kg/m 2 ). Fat mass index (FMI) was calculated as body fat mass in kilograms divided by the square of height in meters (kg/m 2 ), and MMI was muscle mass (MM) in kilograms divided by the square of height in meters (kg/m 2 ).

Measurement of other covariates
A standardized questionnaire survey was conducted by faceto-face interview. Demographic information including sex, birthday, residential areas (living in urban or rural), and educational level were obtained. In addition, personal disease histories such as cardiovascular disease, cerebrovascular diseases, respiratory diseases, musculoskeletal disorders, fracture history, neurological disorders, and cancer information were collected. Before the survey, all interviewers and technicians underwent a training program to guarantee their capability to use specific tools and methods.

Statistical analyses
After excluding missing values (missed in HGS and body composition, n = 616), people diagnosed with diseases including cardiovascular diseases, respiratory diseases, cerebrovascular disease, musculoskeletal disorders, neurological disorders, had fracture history or cancer (n = 12 079), people who did not live in the local areas (n = 55), the final analytic sample came up to 39 508 subjects. After adjusting for potential confounders, Spearman partial correlations were performed to examine the correlations between HGS and body composition. Dixon-Reed method was used to identify the outliers of body composition and HGS. 22 The LMS (lambda-mu-sigma) method was used to construct growth reference charts and is an extension of regression analyses that includes three indexes 23 : The median (mu), which represents the corresponding change when the explanatory variable changes. The coefficient of variation (sigma), which models the spread of values around the mean and adjusts for nonuniform dispersion. The skewness (lambda), which models the departure of the variables from normality using a Box-Cox transformation. Using cubic natural smoothing spline functions, the LMS models smooth the percentile curves of HGS and body composition indexes. 24 We performed the LMS method using the GAMLSS package in R (version 4.0). As there were substantial sex differences in both HGS and body composition, the analyses were all performed by sex separately. In addition, we did further age-stratified analyses to investigate the possible age-modified relationships between HGS and body composition. The descriptive and correlation analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA). A P-value <0.05 (two-tailed) was considered as statistically significant.

The sex-specific values of body composition
The sex and age-specific BMI, FMI and MMI values were summarized in Table 1, and their growth curves were shown in Figure 1. There were substantial sex differences in the body composition indexes in the life course. In general, men/boys had higher BMI, MMI levels but lower FMI than their female counterparts. In both boys and girls, BMI increased with age. The average value of BMI in children and adolescents aged 8-19 years increased from 14.6 kg/m 2 in the 8-9 years group to 19.0 kg/m 2 in the 18-19 years group. FMI and MMI both showed an increasing trend with age during childhood. Men achieved their peak values of BMI, FMI, and MMI at 40-49 years, then declined slightly with age. However, in women, BMI, FMI and MMI kept increasing with age until their 60-70 years.

The sex-specific absolute and relative values of hand grip strength
A total of 39 508 participants completed the HGS test. The sex-specific absolute and relative values of HGS are shown in Table 2.  (Figure 2A,B). Compared the LMS curves in Figure 1E, Figure 1F with Figure 2A,B, we supposed that, in the ageing process, the decrease of muscle strength was more significant than that of muscle mass in both sexes. Figure 2C,D illustrated the sex-and age-specified trajectories of relative HGS from 8 to 80 years old. The relative values of HGS revealed slightly different trajectories compared with their absolute values. As BMI normalized relative HGS, the different shapes of relative HGS and absolute HGS curves, to some extent, can be attributed to the coinstantaneous  change of grip strength and BMI with ageing. For example, the peak value of absolute HGS in men was at the age of 20-39 years ( Figure 2A and Table 2), but the peak value for relative HGS was in the earlier age, 16-19 years group ( Figure 2C and Table 2), which suggested a higher level of BMI in the 20-39 years group than in the 16-19 years group.

The correlations between body composition and hand grip strength
The correlations between body composition indexes (BMI, body fat percentage, FMI, muscle mass, MMI) and HGS are shown in Figure 3. The correlations between HGS and body composition indexes were statistically significant (all P values <0.001). Furthermore, compared with other body composition indexes, muscle mass was found to have the strongest correlations with HGS in both men (correlation coefficient: 0.422, P < 0.001) and women (correlation coefficient: 0.504, P < 0.001). In addition, the age-stratified analysis revealed that, the correlations between HGS and body composition indexes were more substantial among younger people (before 16 years old).

Discussion
To our best knowledge, this is the first study to explore the age-and sex-specific percentile reference values of hand grip strength across a broad age spectrum in a large unselected Chinese population. The normative values suggested in this study could be a valuable source of information on health assessment of muscle strength and body composition and a reference for comparison with other researches from different populations. Our findings revealed sex disparities in the growth trajectory of HGS and body composition with ageing. In general, HGS declined with ageing, especially in men. Fat mass increased with ageing in women, but decreased in men. Muscle mass declined slowly with ageing in men but kept relatively steady from 30-40 years in women. Muscle mass and strength varied among countries and ethnicities. Data from some studies showed that the population living in western countries had higher levels of muscle strength than people from Asian countries. 17,25 Variations in muscle strength between people from different regions may be attributed in part to genetic background, anthropometric factors, dietary patterns, socio-economic status, and so on. 17 Within the same country, variations on HGS have also  19 Although they revealed strong age-dependent HGS values among the study population, comparing the results with ours may not be appropriate, given their selected study population and different device used to measure HGS. We measured both absolute and relative muscle strength and mass in the present study. Because body weight is closely associated with muscle mass and strength 10 and is related to metabolic disturbances, using the relative value of HGS to assess muscle health is more appropriate from the perspectives of public health and clinical practice. 4 Muscle strength tends to decline with ageing, which is explicable by loss of muscle mass or increased fat mass, 26 therefore, the normalized values could offset the interference caused by the coinstantaneous change of body composition. Furthermore, as there are sex-difference in the pattern of normative values of HGS, it is reasonable to develop sex-specific instrumental tools for the practical interpretation of muscle fitness, to help identify probable sarcopenia in clinical practice. The pattern of changes in muscle strength, measured by HGS, was similar to other research findings among other populations, that HGS peak in young adulthood and, after a plateau, start decreasing gradually during ageing. 2,14,27 In line with previous studies, our study revealed a higher HGS value in men than in women. 4,11,19,[28][29][30] Data from some studies showed that muscle mass and strength seem to decline with age. However, our data revealed that the decline in muscle strength was more significant than in muscle mass, similar to findings from the Italians and Koreans. 26,31 Demonstrated by other studies, during ageing, the loss of mobility and the onset of physical disability was correlated with loss of muscle mass and increased fat mass. 32 However, the role of muscle strength has also been examined by its independent associations with physical function, mobility, and mortality in cohort studies. 33 Therefore, early identification of reduced muscle function from both muscle mass and strength perspectives is vital to prevent further health damage.
Muscle health and functional capacity are influenced by the dynamic interaction of various factors, including age, sex, physical activity, nutritional status, genetic backgrounds, and the like. 11,32,34 As we developed sex-and age-specific muscle and body composition normative values, the diverse data can provide a precise tool for assessing the growth and development of skeletal muscle in healthy people from children to elders. They can also provide baseline data for the practical evaluation of interventions and early predictions of sarcopenia, or other functional and metabolic disorders. It is worth noting that cut-offs for muscle strength and mass are not currently applicable, so an algorithm developed using the charts generated by this study is necessary for the healthcare practice among the unselected population in the communities.
Although magnetic resonance imaging (MRI) and computed tomography (CT) are gold standards for assessing muscle mass, these tools are not commonly used in population studies due to the high costs of equipment, lack of portability, and the requirement for well-trained personnel. 35 On the contrary, BIA is a noninvasive, easily-mastered, quick-to-use, safe, inexpensive, and practical method to assess body composition in clinical practice and in population-based research. 36 Several studies have investigated the validity of BIA in Asian populations. For example, Chen et al. created BIA regression equations in Chinese and Southeast Asian populations, revealing excellent lean body mass validity when validated against dual-energy X-ray absorptiometry (DXA). 37 Nevertheless, differences in body proportion, fat-free body Figure 3 The adjusted correlations between hand grip strength and body composition indexes. Covariates adjusted in the overall analyses included age, urban-rural areas, and study sites; in the age-stratified analyses, the adjusted covariates were urban-rural areas and study sites; in the urbanrural stratified analyses, the adjusted covariates were age and study sites. The colour of each cell indicated the level of the correlation coefficient. Blue represented men, orange represented women. The number in each cell represented the correlation coefficient (on the left) and the P value (on the right), respectively. BMI, body mass index, kg/m 2 ; BFP, body fat percentage, %; FMI, fat mass index, kg/m 2 ; MM, muscle mass, kg; MMI, muscle mass index, kg/m 2 .
density, and hydration may impact the validity of body composition measurements in diverse populations. 37 Therefore, our study's BIA measurements and reference values may not be appropriate to extrapolate to other populations, or to studies using different body composition devices. Among the multiple body composition indexes, MM and MMI were found to have stronger correlations with HGS, which is reasonable because muscle mass remains the primary factor of muscle strength. However, other studies showed that muscle strength was only moderately correlated with muscle thickness and muscle cross-sectional areas. 38 Our study also revealed that the relationships between muscle indexes and HGS were moderate in children and adolescents and even weaker in adults. The relationships between fat mass, muscle mass, and muscle strength are complicated. Low muscle mass and strength share some underlying pathophysiological pathways with high-fat mass. 39 It is notable that, as revealed by our data, the loss of muscle mass and strength was along with increased fat mass during ageing, thus raising the concern of early identification and intervention of sarcopenic obesity in the elders.
Our study has several strengths. First, the large sample size of more than 30 000 representative general population with a wide age range provided rich data on muscle fitness and body composition assessment, which can be used to assess the growth and variations of muscle health across the life course (age 8-80 years). The established normative values will help assess the proportion of individuals with low muscular strength levels and identify target populations to prevent sarcopenia and compromised intrinsic capacity. Secondly, we measured multiple indexes that reflected body composition, and in this respect, our data provided a unique opportunity to assess the co-change of muscle strength and body composition in age and gender stratum. The limitations of our study should also be acknowledged. First, the sample size for children and adolescents was relatively small, thus may lead to unstable estimation in the LMS procedure. Second, as the validity of BIA measurements varies among populations, our findings on the correlations between body composition and absolute/relative HGS may not apply to other populations. Third, as current data from CNHS only covered part areas in China, further research should be conducted that covers more areas in China to make the dataset more representative of the country. Nevertheless, during the sampling procedure, we considered geographic and socio-economic characteristics in the target population, and enrolled subjects according to the local age and sex distribution to achieve better representativeness.
In conclusion, this study established age-and sex-specific normative values for hand grip strength for Chinese people aged 8-80 years. The findings of this study offered a unique opportunity to assess upper and lower muscle strength in a large and unselected population across life-course. This study's rich and diverse data can facilitate the practical appraisal of muscle fitness and promote early prediction of sarcopenia or other impairments associated with neuromuscular disorders.