Socioeconomic differences in handgrip strength and its association with measures of intrinsic capacity among older adults in six middle-income countries

Handgrip strength, a measure of muscular strength is a powerful predictor of declines in intrinsic capacity, functional abilities, the onset of morbidity and mortality among older adults. This study documents socioeconomic (SES) differences in handgrip strength among older adults aged 50 years and over in six middle-income countries and investigates the association of handgrip strength with measures of intrinsic capacity—a composite of all the physical and mental capacities of an individual. Secondary data analysis of cross-sectional population-based data from six countries from the WHO’s Study on global AGEing and adult health (SAGE) Wave 1 were conducted. Three-level linear hierarchical models examine the association of demographic, socioeconomic status and multimorbidity variables with handgrip strength. Regression-based Relative Index of Inequality (RII) examines socioeconomic inequalities in handgrip strength; and multilevel linear and logistic hierarchical regression models document the association between handgrip strength and five domains of intrinsic capacity: locomotion, psychological, cognitive capacity, vitality and sensory. Wealth quintiles are positively associated with handgrip strength among men across all countries except South Africa while the differences by education were notable for China and India. Work and nutritional status are positively associated with handgrip strength. Our findings provide new evidence of robust association between handgrip strength and other measures of intrinsic capacity and confirms that handgrip strength is a single most important measure of capacity among older persons.

www.nature.com/scientificreports/ countries except Ghana, more than half of the women including in the analysis were not engaged in any form of formal or informal work. The prevalence of underweight was highest in India. Pooled data indicates, although one-third of participants reported at least one chronic disease-, around three-fourths rated their self-reported health status as good (except in Russia), providing evidence that healthy ageing does not require being disease free.
Cross-national differences in handgrip strength by sex. The mean age-adjusted handgrip strength ranged from 22.8 kg in Mexico to 40.9 kg in South Africa (Table 1). Compared to all other countries, on an average older men and women in India and Mexico had lower handgrip strength (Table 1). Both older men and women in South Africa had higher handgrip strength than older people in other countries. The range of cross-national differences in handgrip strength was substantially greater for men (28.9 kg in Mexico and India and 44.7 kg in South Africa), than for women (18.8 kg in Mexico and 37.6 kg in South Africa). The difference between male-female handgrip strength was higher in Russia and China than other countries. When further disaggregated by 5-year age groups and by sex groups, these differences remain; however, the age-associated decline is stronger for men and women in Russia (Supplementary Figure 1).

Socioeconomic differences in handgrip strength.
The Relative Index of Inequality (RII) scores for years of education and wealth quintile stratified by sex are shown in Table 2. In the age-adjusted model, wealth quintile was significantly and positively associated with handgrip strength among men in all six countries. Educational differences in handgrip strength were significant among men in China and India. Educational attainment and wealth quintile showed significant inequalities in handgrip strength among women in India, China, Russia, and South Africa. In the model fully adjusted for sociodemographic correlates, wealth quintile-based inequalities in handgrip strength were significant among men in all countries except South Africa; whereas

Association between handgrip strength and measures of intrinsic capacity. Handgrip strength
showed significant positive association with measures of cognitive capacity among both men and women in all six countries (Table 3). Handgrip strength was inversely associated with depression in Ghana, India, and Russia. In China, Ghana, Russia and South Africa, higher handgrip strength among men was negatively associated with perceived stress, a second measure of the psychological domain. A similar association was found among women in China, India, Mexico, and Russia. Higher handgrip strength was associated with significantly greater gait speed, a measure of locomotor capacity, in China, Ghana, and Russia. The association between handgrip strength and sensory function measured by visual impairment was significant and negative in China, Ghana, India, Russia, and South Africa, even when many adults used a corrective aid. In China, Ghana, India, Russia, and South Africa, the association between handgrip strength and lung function, an alternative measure of vitality, was significant and positive.

Discussion
Addressing our first objective, findings of this study confirm substantial cross-national differences in mean handgrip strength for men and women. Among the six countries, older adults aged 50 and over in India and Mexico had much lower handgrip strength compared to their counterparts in the other four countries of South  www.nature.com/scientificreports/ Africa, Russia, Ghana, and China. Such cross-national variations in handgrip strength have been reported previously across the globe suggesting lower handgrip strength in low, middle, and high income countries 16,20 . A study based on older adults from India and the United States found older adults in India had lower handgrip strength 21 . Studies have shown cross-national differences in handgrip strength may be shaped by differences in stature and body size 19 . A few studies have investigated differences by race and ethnicity 34 however differences most likely reflect within-country heterogeneity in individual socioeconomic factors, nutrition, diet and health behaviour and environmental characteristics reflecting the social determinants of health 33,35 . That being noted, identifying pathways that lead to lower hand grip strength and policy options to increase equity, is highly relevant. For example, in relation to nutritional status, studies have shown a substantial proportion of older adults in India were underweight (38%), anaemic and experiencing food insecurity (17%) 36,37 . In addition, the prevalence of Vitamin D deficiency was reported to be higher in South Asian countries such as India 38 . This is consistent with other studies that documented a significant relationship between Vitamin D deficiency with measures of intrinsic capacity including lower handgrip strength and gait speed 39 , and outcomes specific to older adults, such as sarcopenia 40 . Moreover, anaemia in older populations particularly women, is common in India and in other LMICs 37 ; studies also document its significant association with sarcopenia and handgrip strength 41 .
Unsurprisingly, increasing age showed a consistent inverse association with handgrip strength in each of the six countries. In addition, the variations in handgrip strength at each age are also documented across the six countries (Supplementary Figure 1). Men have on an average, higher handgrip strength than women 3,23,26 . Differences between men and women in handgrip strength was notable across all countries but this difference was smaller in Ghana and South Africa in comparison to other countries. These differences by gender include biological factors related to stature, as well as social and economic determinants that men and women experience differentially. However, reflecting cross-sectional data, the age-associated decline was faster among men consistent with a previous study analysing European populations 42 .
Across the six countries, underweight older adults had lower handgrip strength, similar to a previous study in Indonesia 31 . The findings of our study are important for countries like India and Ghana where a larger proportion of older adults are underweight and have lower handgrip strength. Underweight is shown to be a significant predictor of health outcomes such as anaemia, osteoporosis, reduced cognitive function, depression, and common illnesses in both developing and high income countries 36 . In addition, supporting results of a previous study 31 older adults who were not engaged in work had lower handgrip strength across the six countries and those who reported poor self-rated health had lower handgrip strength in India, China, and Russia. Longitudinal data are needed to disentangle the determinants and consequences of lower hand grip strength.
Our study reports significant wealth-based inequalities in handgrip strength, particularly among men. The results are consistent with previous studies mainly from high-income countries which document a strong relationship between wealth and handgrip strength compared to other factors such as education and work status 24,29 . The stronger effect of wealth on handgrip strength in old age also confirms wealth as a more relevant measure of socio-economic status than education, particularly in low resource settings 24 . Although the accumulation of wealth takes place across the life course, the benefits of greater wealth are stronger in the later stage of life 42 . For example, higher socioeconomic status has been associated with several advantages such as better access to diet and nutrition 33 . Higher wealth promotes better health outcomes including handgrip strength through intake of diverse nutrition-rich foods. We found a stronger effect of wealth in young old ages of 50-60 among men, but this effect narrowed with older ages over 60, supporting literature that suggests the role of mortality selection and age-as-leveller hypothesis that socioeconomic differences in health weakens with age 43,44 .
Implications of these findings for public policy require further discussion in each context. Nevertheless, there are indications on what can improve the lives of older adults, including policies and interventions promoting improved nutrition that reach older populations in low-resource settings. Existing interventions, such as increasing protein intake, can improve handgrip strength 27 . Further, measures to reduce the inequity of opportunity for appropriate nutrition, for example by targeting individuals with lower socio-economic status, are also necessary to optimize healthy ageing 34 .
Addressing our second objective, our study documents in six middle-income countries, a robust and comprehensive association of handgrip strength with measures of intrinsic capacity. These correspond to each of the five important domains identified by WHO: locomotion, psychological, vitality and cognitive capacity supporting the findings of previous fragmented studies [12][13][14]17,45,46 . For example, a study conducted among older adults aged 60 and above in Colombia showed significant association between handgrip strength and measures of intrinsic capacity such as vitality, sensory, cognition, and psychological capacities 13 . Our findings of inverse association of handgrip strength with depressive disorders, sensory impairments, and positive association with measures of cognition and gait speed are also consistent [12][13][14]17,45,46 .
Handgrip strength is a significant marker of intrinsic capacity, its interrelated domains, and is essential for daily functioning. Weaker handgrip strength is a measure of sarcopenia which reflects poor intrinsic capacity and contributes to lower functional ability of older adults 47,48 . In addition, a growing body of literature suggests a significant association between handgrip strength and cardiometabolic disease risk 49 which further increases the rate of cognitive decline and depending upon the environment, can negatively impact functional ability 50 . Focusing on cognitive capacity, findings from this cross-sectional study show a consistent and positive association between handgrip strength and cognition in six middle income countries. This extends the generalizability of results from studies mainly from high-income countries that reported a significant association between handgrip strength and cognitive capacity. Longitudinal analysis in LMICs, however is needed to confirm whether those with higher handgrip strength experiencing slower cognitive decline 7,9 and improved psychological health, for example by lowering the risk of depression 13 and disability 10  www.nature.com/scientificreports/ and cognitive capacity suggests the role of fluid intelligence (e.g. comprehension, reasoning and problem solving) 51 and nutritional status 9 and underlines what is good for the body is also good for the mind 52,53 . Lastly, the healthy ageing looks at the whole person in their unique environment. Describing and improving functional ability, intrinsic capacity, and environments, the three components of healthy ageing, represents a paradigm shift in thinking about older people and ageing. Information on intrinsic capacity, measured through five domains, provides an important basis to describe comprehensively the capacities of older adults irrespective of disease status. This comprehensive assessment is highly relevant for person centred interventions. Along with enabling environments, these capacities can interactively help to improve the functional ability of the older populations in low-and middle-income country settings. Our results suggest on one hand, handgrip strength is an important measure of overall intrinsic capacity, and on the other, approaches to improve handgrip strength such as through interventions that improve nutritional status that reach all older adults who would benefit, are important for longevity, increasing equity, and promoting healthy ageing 46 .

Strengths and limitations
The main strength of this study is twofold. It is the first study that documents the pattern of socioeconomic differences in handgrip strength in six middle-income countries and provides cross-national comparative results by age and sex. These results not only provided new insights about the significance of handgrip strength as a marker of overall intrinsic capacity, but also increased our understanding of determinants of handgrip strength, including what is generalizable across low-, middle-and high-income countries and those that are particularly important for populations with lower socioeconomic status.
The main limitation reflects that data is cross-sectional and results do not distinguish between determinants and impacts, which limits causal interpretation. Second, while we investigated the association of handgrip strength with socio-economic status and different measures of each domain of intrinsic capacity, we did not account for different levels of socio-economic status across countries. Recent studies based on longitudinal data from high-income countries have documented possible bi-directional association of the relative effect of handgrip strength on changes in wealth and income, was overall greater than the corresponding effect of income and wealth on health changes 54 . Other studies suggested a positive association between handgrip strength and work participation and economic activity 55 which contributes to economic wellbeing in old age. Future studies using longitudinal datasets are needed to understand the direction of the association between socioeconomic status and handgrip strength, and how this may differ by gender or other markers of social position. Lastly, similar to other studies based on survey data, findings of our study are subject to possible self-reported bias in reporting of health outcomes. For instance, the prevalence of chronic diseases (arthritis, stroke, hypertension, angina pectoris, diabetes mellitus, asthma, chronic lung disease), and edentulism were based on self-reported data. These limitations notwithstanding, this study sheds important new insights.

Conclusion
Our study provides new insights of significant wealth-based inequalities in handgrip strength among older adults across six middle-income countries with substantial cross-national differences. In addition, the study provides new evidence of the robust association of handgrip strength with measures of intrinsic capacity across five domains of locomotion, cognitive capacity, psychological, vitality and sensory confirming the multi-dimensional potential of using handgrip strength as a single indicator of intrinsic capacity and a stronger confirmation ageing. Our findings extend the importance of handgrip strength to monitor the progress in healthy ageing in middleincome countries. We conclude by noting that the WHO Guidelines on Integrated Care for Older People (ICOPE) advocate improving physical and mental capacity a comprehensive approach tailored to the specific needs and goals of each older adults-including multimodal exercise, nutritional interventions, and cognitive stimulation, supported by appropriate health and social care systems and service providers 56,57 .

Methods
Source. This study uses cross-sectional, population-based survey data from six countries: China, Ghana, India, Mexico, the Russian Federation, and South Africa from WHO's Study on global AGEing and adult health (SAGE) Wave 1 (conducted during 2007-2010). A multistage cluster sampling strategy was adopted in all countries except Mexico. SAGE included representative samples of persons aged 18-50 years and over in each country with a smaller representative sample of adults aged 18-49 years in each country for comparison. This study included on older adults aged 50 and above in 6 countries (n = 33,878). Household-level and person-level analysis weights were calculated for each country, which include sample selection and a post-stratification adjustment. Detailed information can be accessed from Kowal et al. 58 .
SAGE measures are comparable with other studies from low-, middle-and high-income countries such as the US Health and Retirement Study (HRS) and the family of similar studies such as the Survey of Health, Ageing and Retirement in Europe (SHARE), the English Longitudinal Study of Ageing (ELSA) and the China Health and Retirement Survey (CHARLS). Face to face interviews were conducted to obtain data on sociodemographic characteristics, work history, lifestyles, health risk factors, self-reported and symptomatic assessment of chronic conditions, subjective health, quality of life, cognitive functioning and other domains of IC, disability, and healthcare utilization. In addition, performance or assessed measures of health and anthropometric measures such as height, weight, handgrip strength, lung function, hypertension, waist and hip circumference, timed walk, and vision test, were collected. A detailed description and documentation of data are described elsewhere 58  www.nature.com/scientificreports/ dropped to the side. Respondents were asked to keep their upper arm against their body and bend their elbow to 90° with palm facing in (like shaking hands). Subsequently, respondents were asked to squeeze the dynamometer as hard as possible for a few seconds. Overall, two measurements were taken for each hand. In the analysis, we considered the best of the four measurements 59 . Since handgrip strength is the main outcome measure in this study, we excluded missing cases in the analysis (n = 2750), with a final sample of 31,128 for analysis.
Measures of intrinsic capacity. Locomotion. Gait speed. In SAGE survey, 4 m gait speed was assessed as a measure of locomotion capacity. Participants were asked to complete the 4-m distance (one attempt) in a normal pace and were permitted to use any mobility aids and the time (in seconds) taken to complete 4 m taken in the analysis 60 . For older adults who used a mobility aid (cane or walker, for example), this was instead a measure of functional ability and is not distinguished in this analysis from IC.
Cognition. Cognitive ability. In the analysis, we generated a standardised cognitive index with four itemsverbal fluency, verbal recall, digit span forward and digit span backward combining these variables covering three domains of cognition using principal components analysis and finally converted index score ranges from 0 to 100, higher scores represent higher cognitive functioning. Detailed description about the construction of cognition index is provided in the supplementary file.
Psychological. Depression. Depression was assessed through a set of symptomatic questions based on the World Mental Health Survey version of the Composite International Diagnostic Interview 61 . Diagnosis of major depressive episode was derived from an algorithm that accounted for reporting symptoms of depression during the past 12 months. The detailed symptomatic questions and algorithm used are provided in the supplementary file Table S5. Perceived stress (control) was assessed on a five-point scale based on the following question "How often have you felt that you were unable to control the important things in your life?" options were (1) Never, (2) Almost never, (3) Sometimes, (4) Fairly often and (5) Very. Perceived stress (coping) was assessed with the following question "How often have you found that you could not cope with all the things that you had to do?" options were (1) Never, (2) Almost never, (3) Sometimes, (4) Fairly often and (5) Very often. In the analysis, we generated a composite perceived stress score index variable based on two questions using factor analysis with polychoric correlations. The scores ranged from 0 to 100, higher scores indicating higher perceived stress 62 .
Vitality. Lung function. Forced Vital Capacity (FVC) measured in litres were used in the study. Lower scores indicated weaker lung capacity. In SAGE, three measurements were taken, and in the analysis, we considered the best one.
Sensory. Vision impairment. Visual acuity was measured for both near and distance vision with best possible corrections in each eye using a tumbling "E" logMAR chart. Measured near and distance visual acuity was classified into normal vision (0.32-1.6 decimal) and low vision (0.01-0.25 decimal) 63 . In this study, a respondent was categorised with low vision if they had either low near or distance vision in one or both eyes. For older adults who used a correct aid (spectacles or contact lens, for example), this was instead a measure of functional ability and is not distinguished in this analysis from IC.
Demographic and socioeconomic factors. The sociodemographic covariates included in the study are age (years), sex (male and female), place of residence (rural and urban), marital status (currently married and others), education (no schooling-category 1), 1-4 years (category-2), 5-9 years (category-3), and 10 + years (category-4), with the exception of Russia where education is categorised as 0-9 years (category-1), 10-12 years (category-2), 13-15 years (category-3) and 16 + years (category-4). Work status is categorised into currently working and not working/never worked. Wealth quintiles represent household economic status assessed using an index of household assets ranging from possessions, amenities and construction type. Principal Component Analysis was used to generate a wealth index and categorised into five categories (quintiles) ranging from poorest to richest within each country. List of wealth variables included in the index is provided in the Supplementary file Table 7.

Self-rated health (SRH).
In the SAGE survey, respondents were asked 'In general, how would you rate your health today? The response categories were: 'very good' , 'good' , 'moderate' , 'bad' and 'very bad' . We combined, 'bad' and 'very bad' health categories to represent poor self-rated health.

Chronic diseases (multimorbidity).
Multi-morbidity is defined as the presence of one or more chronic health condition at the time of data collection. In this analysis, we have included eight chronic health conditions: arthritis, stroke, hypertension, angina pectoris, diabetes mellitus, asthma, chronic lung disease and edentulism. Detailed information about the assessment of chronic diseases is described in the supplementary file Table S2. Statistical analyses. First, we assessed the association of socioeconomic status, multi-morbidity and handgrip strength using three-level linear hierarchical models, with state/province at the highest level, Primary Sam- www.nature.com/scientificreports/ 64. WHO Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet 363(9403), 157-163 (2004).