Normative data and associated factors of hand grip strength among elderly individuals: The Yilan Study, Taiwan

Weak grip strength is associated with subsequent mortality in elderly populations. The normative data and associated factors of HGS in community-dwelling elderly Taiwanese populations require further evaluation. From February 2012 until the end of 2016, all residents of Yilan City, Taiwan aged 65 years or older were randomly selected for a population-based community health survey. A total of 2,470 older adults were enrolled in this study. The relationships between HGS and various anthropometric and sociodemographic correlates were examined. The results showed that HGS was higher in men than in women. The mean HGS exhibited a decreasing trend with advanced age in both men and women. HGS was significantly associated with height, weight, and exercise habits. The physical as well as the mental component summary measures of health-related quality of life (HRQoL) were positively associated with HGS. After HRQoL was integrated into the regression model, female sex, age, waist circumference, and diabetes mellitus were significantly negatively associated with HGS. In conclusion, HGS significantly decreased with advanced age. among community-dwelling Taiwanese elderly people, Various factors had different effects on HGS.

than that among elderly people in some Western countries 7 . Kim et al. demonstrated that among Koreans HGS peaked at 35-39 years of age and thereafter decreased in both sexes 8 . Additional studies are warranted, especially in other regions of Asia, to evaluate the normative data of HGS.
Age, sex, race/ethnicity, education, smoking status, body mass index, comorbidities, and physical activity have been associated with HGS in community-dwelling elderly people 2 . Greater height, higher weight, and smaller waist circumference have also been independently associated with higher HGS 7 . Because HGS and its associated factors have become a major public health concern, especially among elderly populations worldwide, additional studies are required to thoroughly elucidate the various factors that influence HGS and the underlying associations. Therefore, the present study was designed to investigate the following: (1) the normative data of HGS and (2) the associated factors of HGS among community-dwelling elderly individuals in Taiwan.

Methods
Study design and participants. This study was part of the Yilan study series [9][10][11][12][13] , a population-based community health survey conducted by the Community Medicine Research Center of the National Yang-Ming University and the National Yang-Ming University Hospital in Taiwan. The study protocol was evaluated and approved by the Institutional Review Board (IRB) of National Yang-Ming University Hospital (IRB Approval No.: 2011A016). All the enrolled participants provided written informed consent. All methods were performed in accordance with the relevant guidelines and regulations.
The study methods have been described elsewhere 9,11,14 . In brief, from February 2012 to the end of 2016, all residents of Yilan City, located in the northeast of Taiwan, aged 65 years or older were randomly selected and invited to participate. Those who agreed to participate in the study were enrolled, and those who did not agree to participate were excluded. Additionally, those with preexisting hand problems were excluded from HGS measurement. Well-trained project assistants visited all the enrolled participants at their homes and conducted face-to-face interviews. Basic data and health-related information, including demographic characteristics, body weight/height, waist circumference, education level, living status (full-time living alone, part-time living alone, and not living alone), lifestyle (ie, smoking/drinking habits, exercise condition within the past 1 week, and sleep status), health-related quality of life (HRQoL) within the past 1 month, self-reported medical histories of chronic diseases (diabetes mellitus, hypertension, cardiovascular disease, hyperlipidemia, cerebrovascular accident, gout, and cataract) and current psychological status, and fall episodes within the past 1 year, were collected through physical examination and well-established questionnaires 9,11,14 . The Short-Form 12 Health Survey Version 2 (SF-12v2) was used for evaluating the HRQoL; the SF-12v2 comprises 12 items that constitute the mental component summary (MCS) and physical component summary (PCS) for measuring the participant's mental and physical function status in the past 4 weeks 15,16 .
Initially, 2,584 community-dwelling elderly individuals were randomly selected and interviewed. However, 114 of them were determined to have mild to moderate disability because of a previous cerebrovascular accident. Therefore, these participants were excluded from the present study. Finally, 2,470 subjects were enrolled.

Measurement of HGS.
HGS of both hands was measured with a hydraulic hand dynamometer (Jamar, Jackson, MI, USA). Each participant completed the trial for each hand, and the final estimate of HGS was the average of all measurements 10,17 . Statistical analysis. Descriptive statistics are presented as numbers of cases, percentages, and means with standard deviation (SD). The independent t test and the chi-squared test were used to analyze the differences in continuous and categorical variables, respectively, between groups. Box plots of HGS estimates were drawn, and their trend tests were evaluated using a simple linear regression. The beta (β) coefficient and the partial R-squared of individual independent factors of HGS were calculated using stepwise multivariate linear regression to perform multivariate analyses.

Results
Sociodemographic data of participants. A total of 2,470 adults aged 65 years and older [998 (40.4%) men and 1,472 (59.6%) women] were enrolled in this study. A comparison of the basic characteristics, including age strata, chronic physical disorders, living status, and education level, of the sexes is presented in Table 1. The sex distribution, especially in the 65-69 years stratum, was significantly different. However, no significant difference in comorbidities such as heart disease (32.5% vs 31.9%) and diabetes mellitus (22.3% vs 24.7%) was noted between men and women. Additionally, men exhibited lower rates of full-time and part-time living alone and higher rates of not living alone than women (5.4% vs 9%; 2.4% vs 5.4%; 92.2% vs 85.6%, respectively). Furthermore, men also had higher education levels (junior high school or higher education level; 44.9% vs 23.0%).
normative Data of HGS. Table 2 presents the means and SD of HGS by sex and age stratum. The mean values of HGS for men were significantly higher than those for women in sum (25.4 kg vs 15.5 kg) or in subgroups by 2-and 4-year age strata. The mean HGS dropped from 33.0 kg in men aged 65-69 years to 21.1 kg in men aged 80 years or older. Moreover, HGS dropped from 19.1 to 12.6 kg in women. In the oldest stratum (75 years or older), the mean HGS values decreased to only 23.2 and 13.9 kg in men and women, respectively.
Minimum, maximum, median, and interquartile range of all HGS point estimates among various age strata for both sexes are presented in Figs. 1 and 2. Significantly decreasing trends (p < 0.001) of HGS across all age strata were noted in both men and women. The medium values of HGS in men in all age strata were higher than those in women.
www.nature.com/scientificreports www.nature.com/scientificreports/ correlates of HGS. Stepwise multivariate linear regression analysis was conducted to evaluate the correlations of HGS with various factors. The results are presented in Table 3. For the overall sample, HGS was significantly positively associated with height (β = 0.19) and weight (β = 0.18) and slightly associated with exercise. Female sex, age, waist circumference, diabetes mellitus, and heart disease were independently associated with decreased HGS (β = −7.5, −0.35, −0.12, −1.16, and −0.63, respectively). HGS was significantly higher in those living alone full time (β = 1.17). Stratified regression analysis in men and women indicated that HGS was positively associated with height (β = 0.24 and 0.16) and weight (β = 0.30 and 0.13) and negatively associated with age (β = −0.53 and −0.23), waist circumstance (β = −0.24 and −0.09) and diabetes mellitus (β = −1.33 and −0.10) in both genders, and heart disease in women (β = −0.64). The HGS was significantly stronger (β = 2.72) in men with full-time living alone status. For the overall sample, the highest R-squared percentage (36.4% in a total of 53.3%) was noted for height. In men and women, the highest R-squared percentage values were obtained for age (20.9% in a total of 35.8% and 16.1% in a total of 26.1%).
A stepwise multivariate linear regression analysis was performed again to evaluate the associated factors of HGS with the inclusion of HRQoL measured using the SF-12v2 with MCS and PCS. For the overall sample, HGS was significantly associated with height (β = 0.  www.nature.com/scientificreports www.nature.com/scientificreports/ mellitus (β = −1.04 and −0.08). After controlling for height, age, PCS, MCS, weight, waist circumstance, and diabetes mellitus, men with full-time living alone status still had significantly stronger HGS (β = 2.35). For the overall sample, height had the highest R-squared percentage (36.4% in a total of 57.8%). In men and women, the highest R-squared percentage values were related to age (20.6% in a total of 42.8% and 16.1% in a total of 33.1%).

Discussion
HGS is a crucial and easily available indicator for predicting the ability to perform daily functions, especially among elderly individuals 18 . In our study, 2,470 community-dwelling elderly persons were recruited and we conducted physical examination and face-to-face interviews at their homes 11,14,17 . Unlike the other studies in Taiwan, we used primary data collection with a large sample size 5,[19][20][21] . Therefore, elderly individuals who previously tended to stay at home because of disability and could not go outside to undergo examinations could be assessed. Thus, even in the oldest age stratum (80 years or older), 803 residents were recruited in this study.   7 . It has been proposed that global variations in the norms of HGS were present, which might be attributable to the status of development 5 . Additionally, HGS has been determined to decrease with advanced age and among women 1,4,5,22,25 . After controlling for age and sex, several other factors have also been proposed to influence HGS, such as anthropometric elements (body weight, height, limb circumference, and palm length), testing protocol, study methods, measuring device, comorbidities, and other socioeconomic factors (ethnicity, education, marital status, and employment status) 5,7,19,23,24 . In our study, we proposed that the anthropometric differences might play a crucial role in the lower norms of HGS compared with the values reported for Caucasians in other studies because people in Taiwan have smaller physical stature. Additionally, it was reasonable to estimate that elderly individuals who accepted HGS measurement at certain community centers, similar to those in Liao's study, might have a more favorable health status because of their independent ambulation. By contrast, those who received interviews at home in our study might include those individuals with poor mobility and worse health status and with resultant lower HGS. This would indicate that the real norm of HGS in Taiwan might be lower than those reported in other studies.
HGS is a diagnostic criterion for sarcopenia, and the cutoff value in Asian populations has been proposed to be <26 kg for men and <18 kg for women. In our study, the cutoff points for both sexes were higher than the mean values (25.4 kg for men and 15.5 kg for women) 26 . As evident in Table 2, it is clear that men older than 75 years and women older than 70 years had mean HGS values below the threshold of sarcopenia. A higher percentage of elderly people in our study population might explain the lower HGS values obtained in our study.
In Figs. 1 and 2 illustrate the large gap between maximum and minimum HGS in each age stratum. Furthermore, the coefficient of variation increased with advanced age. We proposed that the increased variation in HGS with advanced age might be attributable to the accumulation of those correlation factors. Lino contended that various chronic diseases might be associated with low HGS, and thus in many elderly people, an increase in the number of comorbidities might lead to the large variation of HGS 25 .
In our study, age, sex, height, weight, hypertension, and diabetes were significantly associated with HGS, which was compatible with the results reported by Lino and Ramlagan 25,27 . As presented in Table 3, for each year increase of age after 65 years, an average reduction of 0.35 kg in HGS can be expected. A decrease of 3.5 kg in HGS for each decade of life can be estimated, which is substantially higher than that observed by Lino in Brazil 25 . Particularly, men exhibited a larger decline in HGS than women for each year of increase (0.53 kg vs 0.23 kg). Therefore, prevention of muscle strength deterioration in elderly people, especially in men, should be a focus of interventions after retirement from their careers.
Although weight was positively associated with HGS, a negative association between waist circumference and HGS was also observed. Our result was consistent with the finding reported by Ong et al. in 2017 7 . This result suggested that older adults with a higher weight might have higher HGS; however, this did not hold true for obese  www.nature.com/scientificreports www.nature.com/scientificreports/ older adults. Elderly people with a higher body weight consisting of lean muscle rather than central fat had higher HGS.
Exercise may improve muscle strength in elderly people 28 . As presented in Table 3, exercise in the past 1 week was a significant factor associated with HGS but with a low beta coefficient. According to our records, most elderly individuals performed low-or low-to-moderate-intensity physical exercise such as walking, which was apparently not sufficient to improve HGS. Therefore, health-related specialists should design appropriate beneficial programs for promoting physical fitness improvement among elderly individuals.
Furthermore, also evident in Table 3, heart disease and full-time living alone status were significantly associated with HGS. However, after the introduction of HRQoL (Table 4), the association of HGS with both the factors disappeared and the total R-squared increased. HRQoL is an objective measure of the consequence of health status on quality of life. It measures the effects of various chronic illnesses, treatments, and disabilities on the physical and mental status experienced in daily life 29 . Therefore, the PCS of HRQoL had a positive association with HGS, and HRQoL might mediate the effects of heart disease and living status on HGS 24 . However, mental status has often been overlooked as one of the associated factors of HGS in elderly people. Depression has ever been observed to be negatively associated with HGS 25,30 . Pasco suggested that sarcopenia with resultant decreased strength and depression possibly shared a common pathophysiological pathway 31 . We have also previously demonstrated that depression is negatively associated with both MCS and PCS in elderly people 11,14 . Therefore, it would not be surprising for MCS to be positively associated with HGS.
Chronic diseases have been demonstrated to be associated with HGS 12,13,15,25 . In our study, both diabetes and heart disease were associated with HGS. Even with the introduction of HRQoL, diabetes exhibited a significantly negative association with HGS. Stenholm suggested that insulin resistance might have catabolic effects on muscle and resultant sarcopenia 13 . Mainous also suggested that sarcopenia and impaired metabolic states share common cellular and molecular characteristics 12 . These findings might partially explain the direct association between diabetes and HGS.
Our study analyzed the data from a population-based community health survey in Yilan City. For a more comprehensive survey, we have conducted a home-visit intervention. However, there were still some limitations in this study. First, some factors related to muscle strength, such as the nutritional status of the participants, were not included in this study. These factors may be included in future following-up studies with home-visit interventions 32 . Second, for the HGS test, most of the older adults took the measurements in a standing position, with the exception of some older people who took the procedure in a sitting position to prevent falls. Although little difference was evident between these two test positions 23 , a more suitable sitting position should be applied to reduce the small intervention error in future research. Third, though the same two well-trained assistants could minimize the investigation bias, it took longer time for them to recruit these data from over 2,000 elderly individuals. Therefore, some changes of the community at macro level, such as the health policy of the government and the socioeconomic status of the inhabitants, might happen and affect these individuals' health outcomes during these 4 years. Forth, under the restriction of Personal Information Protection Act in Taiwan, we were not allowed to have the detailed names and addresses of these elderly individuals who were registered in Yilan city. Therefore, we could not but randomly selected and invited the old inhabitants to participate this study. It was not  surprising that there were some differences when comparing the distribution of age and sex of these participants to that of the registered inhabitants in Yilan city 33 . This difference may limit the generalizability of our findings to the whole population in Taiwan. However, on the other hand, the elderly who agreed to participate our study might be the population who were prone to receive further community interventions. Therefore, the results of this study will provide us more sound data to plan and perform the following health promotion programs in the near future. Finally, some older adults did not participate in interviews for personal reasons, especially those in the age stratum of 65-69 years. Including more participants, especially in this age stratum, is necessary to obtain reliable results in the following-up studies. Despite these limitations, the results from this study provide important and valuable information for the government to arrange further steps to increase the health status among elderly people in the community.

conclusion
Among community-dwelling Taiwanese elderly individuals, a significant decreasing trend of HGS with advanced age has been noted both in men and women. The HGS was independently positively associated with weight, height, PCS, MCS, and exercise habit and negatively associated with age, female sex, waist circumference, and a history of diabetes mellitus. Additionally, because this study recruited more elderly people with disabilities, the results suggested that the normative data of HGS in elderly Taiwanese people might be lower than those reported from previous studies.

Data availability
The datasets analyzed during the current study are available from the corresponding author on reasonable request.