Hand-grip strength among older adults in Singapore: a comparison with international norms and associative factors

Study approval was obtained from the relevant ethics committees in Singapore -National Healthcare Group Domain Specific Review Board and the SingHealth Centralised Institutional Review Board. Older adults aged 60 years and above were randomly selected from a national database which consisted of administrative data (i.e. name, ethnicity, gender, and residential address) of all citizens and permanent residents in Singapore. These selected participants were notified of the study by mail and then approached at home by interviewing staff. Older adults who resided in day care centres, nursing homes, and institutions were also included. Individuals who were not living in Singapore and who could not be contacted due to invalid addresses were excluded from the study. Written informed consent was obtained from all the participants. In the case where participants were unable to provide consent, consent was taken from their legally acceptable representative or next-of-kin. For each older adult, an informant of each participant- “someone who knew the participant best”, was also invited to take part in the survey. The informant could be either caregivers, co-residents or someone who had close contact with the older person but was not involved in a caregiver role. Informants provided information on participants only when the participant could not provide the relevant information, such as questions on “participant background information” and “physical health condition” [23]. Disproportionate stratified sampling design was used where residents in the older age group and those from Malay and Indian ethnic group were over-sampled to ensure that sufficient sample size would be achieved to improve the reliability of estimates for the subgroups analysis. A detailed description of the methodology can be found in an earlier paper [23].

All measurements and data on sociodemographic information were collected by trained interviewers. Hand dominance was determined by asking which hand they use for writing or which hand they would predominantly use when performing a task. HGS was measured in kilograms (kg) by taking the average of the two dominant handgrip attempts using a Jamar Plus + Digital Hand Dynamometer (Pennsylvania, United States). The device measures the isometric muscle contractions as the participant grip against the resistance of the stationary grip handle. For each of the HGS assessments, subjects were instructed by trained interviewers to sit as per the American Society of Hand Therapist’s (ASHT) recommendation for HGS; with their shoulder adducted and neutrally rotated, elbow flexed at 90^o with the forearm in neutral position for HGS measurement [24].

Sociodemographic information on age, gender, ethnicity, marital status, education, employment status were collected from participants and verified with informants. Anthropometric measurements were carried out with a brief fully structured physical and neurological assessment, the NEUROEX [25], which included height, weight, HGS, upper arm circumference, and waist circumference. Upper arm circumference was measured in centimetres using a measuring tape around the thickest part of the upper arm of the dominant hand [26]. Participants were instructed to have their shoulders relaxed and both arms hanging loosely at the sides as flexing or tightening the arm muscle would results in an inaccurate measurement. Waist circumference measured in centimetres was measured at the narrowest part of the body between the chest and hips for females, and measured at the level of the umbilicus for males [27].

Data analysis was carried out using Statistical Analysis Software (SAS) system version (9.3). To ensure that the survey findings were representative of the Singapore population, all estimates were analysed using survey weights to adjust for complex survey data. We have computed survey weights that incorporate sampling weight, non-response weight and post-stratification weight which was used to weigh the sample back to the population to adjust for oversampling, non-response and post-stratify by age and ethnicity distributions between the survey sample and the Singapore elderly population. This approach has been recommended when analysing complex survey data [28]. Those diagnosed with dementia using 10/66 diagnostic criteria [29] were excluded from the analysis since dementia is associated with poor HGS [20, 30].

Mean and standard deviation were calculated for continuous variables, and frequencies and percentages were calculated for categorical variables. HGS were calculated and presented separately by age, sex and ethnicity. We used 6 age groups categories to present the data: 60–64 years, 65–69 years, 70–74 years, 75–79 years, 80–84 years and 85+ years. Multiple linear regression analyses were used to explore the sociodemographic and anthropometric correlates of HGS. Standard errors (SE) of means, regression coefficients and other statistics were estimated using the Taylor series’ linearization method to adjust for the weighting. Multivariate significance was evaluated using Wald X² tests based on design corrected coefficient variance-covariance matrices. Statistical significance was evaluated at the 0.05 level using 2-sided tests.

The descriptive data are listed in Table 1. A total of 2565 older residents participated in the study, giving a response rate of 66%. Of the 2565 participants who completed the study, 2043 participants aged 60 years and over were included in our study sample. Those with missing HGS data (n = 171, mostly because of health reasons) and dementia (n = 399) were excluded. Further, the results for those who were left-handed were excluded due to small sample size (n = 84). The mean age was 68.8 years, ranging from 60 to 105 years. 82.9% were of Chinese descent, 9.3% were Malays descent, 6.2% were of Indian descent, and 1.7% belonged to other ethnic groups. Majority of the sample were women (54.3%), married/ cohabiting (66.8%), had completed primary education (25.1%), and were employed (37.5%).

Table 2 shows the means and standard deviation of HGS by age group, sex and ethnicity. The mean HGS for the males and females participants in the youngest age group (60-64years) was 31.1 kg and 18.2 kg respectively while it dropped to 18.5 kg for males and 12.4 kg for females participants in the oldest age group (85+ years). The mean HGS showed a decreasing trend with increasing age among all three ethnic groups in both sexes.

Table 3 shows the sociodemographic correlates of HGS. HGS was significantly greater in the youngest age group (60-64) than the other older age groups, 70-74 (β = −3.29, p < 0.001), 75-79 (β = −5.28, p < 0.001), 80-84 (β = −5.94, p < 0.001), and 85+ (β = −9.15, p < 0.001) and in males (β =10.76, p < 0.001) than females. Those of Malay (β = −2.01, p < 0.001) and Indian (β = −1.64, p < 0.001) ethnicity had significantly lower HGS values than Chinese after adjusting for other sociodemographic correlates.

Table 4 shows the association between HGS with height, weight, upper arm circumference, and waist circumference after adjusting for sociodemographic and other anthropometric correlates. Multiple linear regression were conducted and for the overall sample, HGS was found to be significantly associated with height (β = 0.12, p = 0.001), weight (β = 0.09, p = 0.019), and inversely associated with waist circumference (β = −0.08, p = 0.018).

Stratified regression analysis showed that for males, only height (β = 0.17, p = 0.010) and upper arm circumference (β = 0.33, p = 0.004) were significantly positively associated with HGS. For females, the findings were similar to the overall sample where HGS was significantly associated with height (β = 0.10, p = 0.016), weight (β = 0.13, p = 0.009), and waist circumference (β = −0.09, p = 0.020).

The summary details of the eight included studies for HGS comparison are listed in Table 5. HGS data was compared with Western (i.e. US, UK) and other Asian countries (i.e. Japan, Taiwan, Hong Kong, Malaysia and Singapore). For both sexes, HGS among Singapore older adults were relatively low compared to Western and other Asian countries except for Taiwan (Figs. 1 and 2). US and UK had the highest mean HGS followed by Japan, Malaysia, Hong Kong, Singapore, and Taiwan. Only right-handed HGS data were extracted from other studies for comparison.

Overall, Singapore older adults’ mean HGS was lower compared to other countries. The HGS for both genders were generally lower compared to older adults in UK [8], US [7], Japan [9], Hong Kong [10] and Malaysia [13] but was comparable to a previous study conducted in Singapore [6] and Taiwan [13]. Several possible reasons could explain the difference in normative HGS data between countries and these include differences in body composition such as mean height, weight, body sizes, palm size and ways of measuring grip strength e.g. sitting or standing positions and the brand of dynamometer [11].

The norms of HGS may differ between populations within Europe and East Asia. The Survey of Health, Ageing and Retirement in Europe (SHARE) study conducted in 11 European countries found lower HGS in the southern countries (Spain, Italy, and Greece) compared to northern and continental countries (Sweden, Denmark, Netherlands, Germany, Austria, Switzerland, and France) [46]. In a study by Lin et al. [47] which assessed the anthropometric characteristics of adults from East Asian countries (i.e. China, Taiwan, Japan, and South Korea), significant morphological difference were reported among these peoples in the same region [47]. Clearly these differences in anthropometric measurements within regions are likely to be explained by a range of factors such as nutrition and genetic factors which may also account for the differences seen in HGS among countries [48].

Varying methods of measuring grip strength could also explain the difference in grip strength across and within countries. The previous study mentioned- SHARE study - which included data from 11 European countries, used the hand-grip dynamometer, Smedley, while another systematic review using data from 12 British general population studies different dynamometers (Smedley and Jamar) in the seated and standing position (refer to Table 5) were used for the data collection. A systematic review by Roberts et al. [49] found a wide variability in the choice of grip strength measuring equipments and protocols across clinical and epidemiological studies. Furthermore, evidence pointed that variation in approach can affect the values recorded and summary measures of grip strength varied widely including maximum or mean value, from one, two or three attempts, with either hand or the dominant hand alone [49].

Within Singapore, the present study reported similar normative HGS data as Malhotra et al. [6], yet the minimal differences in data reported could be attributed to possible reasons such as individual differences, cohort effect (national survey conducted in 2009 vs. present study: 2013), and instrument used to measure HGS (Smedley spring-type dynamometer vs. present study: Jamar digital-type dynamometer), and position (standing position vs. present study: sitting position). All these findings highlight the importance of having a standardized method of assessing HGS to facilitate comparison between studies and enable consistent measurement of grip strength [49, 50].

The findings of this study should be interpreted in the light of the following limitations. Firstly, the generalizability of the study is limited. The present paper only includes participants with HGS measurements, not diagnosed with 10/66 dementia, and right-handed individuals. Future studies could explore the anthropometric correlates of HGS for both right and left hands. Secondly, there is also the possibility that the participants, who are older, may not have understood the instructions during HGS measurement and that could affect the validity of the measurements taken [20, 51]. To reduce the likelihood of such occurrence, all interviewers received standardized instructions and training from senior researchers to ensure proper use of the equipment and demonstrations of using the Jamar dynamometer were conducted for all subjects. All participants were also briefed in the language chosen by participants according to their familiarity and comfort.

Despite these limitations stated above, results from this study hold important implications on the healthcare of the older adults in Singapore. In terms of external comparisons, Singapore older adults’ have generally lower HGS compared to other countries. The comparison allows clinicians to gauge the performance of Singapore older adults’ HGS performance and offer a better standard for treatment and interventions, and researchers could use the normative data as baseline to study the trend for comparison with future studies. For internal comparisons, the normative grip strength data allow individuals to interpret what is typical in their country. It served as a reference point for comparison to someone of the same age, gender, and ethnicity to determine if their personal HGS is higher or lower than what is typical in their country. However, care must be taken for comparison. Since height, weight, upper arm circumference, and waist circumference were found to be associated with HGS, thus the result would be more useful as a gauge than a strict benchmark [11].

Our study was the first in Singapore that examines the anthropometric correlates of HGS in Singapore older adults and makes comparison with the available HGS data of Western and other Asian countries. Our results suggest that ethnicity and gender differences exist for the anthropometric correlates of HGS (males: height and upper arm circumference; females: height, weight, and waist circumference) in Singapore, which might also explain some of the differences between Singapore older adults against other countries.