Relationship between physical strength measurements and anthropometric measurements: examples from Santal labourer population of Birbhum District, West Bengal, India

Background: Muscular/ physical strength are fundamental in order to satisfy basic needs for survival and productivity, especially among labourer population in developing countries. Research suggests physical body dimension is one of the key determinants of strength; however, the relationship between overall physical body dimension and strength is yet to explored. Aims: to explore the relationship between strength measurements and anthropometric prole of Santal labourer population. Subjects and methods: Cross sectional data on strength (handgrip and back strength) and anthropometric measurements were taken from 536 healthy adult Santal labourers (Male=251, Female=285) aged between 18 to 65 years of Birbhum district, West Bengal. Statistical analysis including Pearson’s correlation and canonical correlation analysis (CCA) were used. Results: Strength negatively correlated with age and positively correlated with most of anthropometric measurements. CCA between strength and anthropometry were 0.765 (p<0.001, explained 85.53% total variance) and 0.611 (p<0.001, explained 71.41% total variance) for males and females respectively. Cross-loading value of CCA indicate fat-free-mass, forearm circumference and weight for male and fat-free mass, chest circumference and wrist breadth for female were most important predictors of strength. Conclusion: Fat-free mass, forearm circumference, weight, chest circumference, wrist breadth were the key measurements for overall strength of Santal labourer population.


Introduction
Human survival is largely dependent on the individual's ability to perform muscular work [1,2]. The advancement of technology, mechanization and automation in many sectors has greatly reduced the physical labour/ muscular work in humans. In spite of this, physical labour is still required in many work situations, especially in developing countries. In India, most of the unorganized sectors are poorly aided with modern technology and thus, physical labour is an integral part for the productivity. Therefore, the assessment of physical labour/ muscular strength and its determinants are necessary for better productivity, sustainable development and welfare of the labourers [3].
Strength is the capacity of an individual to exert force against some external object or resistance [4]. In other words, strength is the maximum force which can be exerted against an immovable object (static or isometric strength), the heaviest weight which can be lifted or lowered (dynamic strength), or the maximal torque which can be developed against a pre-set rate-limiting device (iso-kinetic strength) [5]. Handgrip strength and back strength tests are one of the simple, non-invasive methods for testing muscular/ physical strength of the individual/ population, which is also suitable for both epidemiological and clinical setup [6]. However, variation in terms of strength tests were reported between sexes [7], geographical regions, ethnicity [8], handedness of the participants [9], genetic endowment [10], sociocultural background [11], anthropometric pro le [12], age [13], physical training [14] and so on. Besides, posture (position of body and wrist) of exerting strength [15] and occupational exposure to certain hazardous substances [16] also affects strength of the individuals. Instead of that most of the scholars argued that strength tests can be used as a predictor for overall health and well-being [17], morbidity and mortality [18,19], nutritional status [20,21] and productivity of the workforce [22].
The relationship between strength measurements and anthropometric pro le were demonstrated in several studies e.g. Malina [1] reported that excess body weight, fatness and endomorphic body shape negatively effects strength measurements. However, Kritz-Silverstein and Barrett-Connor [23] pointed out that over-weight individual had greater grip strength in both dominant and non-dominant hand than under-weight individual. Fuster and colleague [24] reported that grip strength was more related with body weight compared to height. However, Chandrasekaran and colleague [25] reported that both height and weight were equally associated with grip strength of an individual. Some studies found relationship between strength measurements and upper extremity measurements [12], measurements on wrists and hands [26], fat free mass [27] and body surface area [28]. Other studies reported that back strength was signi cantly associated with height, weight, BMI, hip circumference [29] and scapular skinfold thickness [13]. Thus, studies (mentioned above) are not consistent in their ndings, primarily to recognize the important anthropometric measurements for strength. Moreover, to our knowledge none of the study explores the overall relationship between strength and anthropometric pro le.
In view of the above, present study was conducted among the Santal labourer population of Birbhum district, West Bengal. The aim was to explore the relationship between strength measurements (viz. handgrip and back strength) and selected anthropometric measurements of the adult Santal labourer population.

Materials And Methods
Present data was collected as a part of a larger bio-medical project ('Health status of the stone mine/ quarry workers of Birbhum district, West Bengal') conducted on Santal labourer groups of Suri subdivision of Birbhum district, West Bengal between December 2012 to June 2016. Cross-sectional data on strength (handgrip and back strength) and anthropometric measurements were taken from 536 healthy adult Santal labourers (i.e. wage earner in stone mines and agricultural sectors) of both sex (251 males, 285 females) aged between 18 to 65 years. The study was restricted to single ethnic group (i.e. Santal) in order to avoid possible ethnic/ genetic effect (if any) in respect of variables under study.
Santals are the third largest marginal (schedule tribe) community and distributed in most of the districts of West Bengal [30]. They were classi ed as 'Pre-Dravidian' tribe. Their language, Santali belongs to the Mundari branch of Austro-Asiatic language family [31] and now they have their own script i.e. 'Ol-Chiki'.
No statistical sampling were attempted for the selection of study participants. Individuals who persuaded to participate and voluntarily agreed with written consent were included in the present study without any Page 4/19 bias. The research was conducted after prior approval from the Ethical Committee for the Protection of Research Risks to Humans, Indian Statistical Institute.
Strength data in terms of handgrip strength (on both hands) and back strength were collected through battery operated digital handgrip dynamometer and back strength dynamometer (manufactured by Takei Scienti c Instrument Co. Ltd., Tokyo, Japan) respectively, following standard protocols for measurements.
In measuring handgrip strength, participants instructed to pull one arm of the dynamometer as close as possible with another arm xed with palm, in standing position, using one's strength of a single hand. No part of upper or lower arm or hand may push against any object or any other part of the body. Handgrip strength measured in terms of scores observed on the dial of handgrip dynamometer and the highest score considered out of three satisfactory attempts. The measurements were taken on both the hands separately.
In case back strength, participants instructed to stand in erect posture on the base of the dynamometer with straight arms and ngers extended downward as far as possible towards thigh. The chain (bar attached to the chain) was then xed with the instrument so that it becomes 1 to 2 inches below one's ngertip. Then participant were asked to bent forward slightly, and pull the bar, which is attached through chain with the base of the dynamometer. The tests done consecutively three times and the best score was recorded.
The IBP basic list of anthropometric measurements were included like height, sitting height, weight, diameters (bi-condylar of humerus, bi-condylar of femur, bi-acromial and bi-iliac), circumferences (forearm, medial calf, chest, waist and hip) skinfold thickness (biceps, triceps, medial calf, subscapular and supra-iliac) as well as measurements on hand length and wrist breadth on both hands were taken using standard techniques and instruments [32]. Further, body mass index (BMI), fat mass (in kg) and fat free mass (in kg) were calculated using following formulae- Due to the absence of written records of age in some of the individuals, the ages were estimated with reference to important local events and crosschecked with elderly individuals, which were further compared with the ages of individuals for whom age records existed.

Statistical analysis
Descriptive statistics computed for each variable under present study. Pearson's correlation analysis performed between strength measurements and anthropometric measurements to understand the bivariate relationship.
Canonical correlation analysis (CCA) performed to understand the overall relationship between strength (both grip and back strength as a whole) and Canonical correlation value is the bivariate correlation between two composite score. We consider rst canonical function because it provides highest correlation value and can explain majority of variance of the depended set. Canonical loadings and cross-loadings value provides information about the most in uential variables for each set, which in uence the other sets.
All the statistical analyses have been done using SPSS software 16.0 (SPSS Inc., Chicago, IL, USA) and SAS software 9.0 (SAS Institute Inc., Cary, NC, USA).    Table 3 shows the result of canonical correlation analysis (CCA) between two sets (i.e. Set 1 = strength measurements [RHGS, LHGS, BS] and Set 2 = all anthropometric measurements and age) of variables obtained from Santal labourer population for both sex. The CCA is restricted to deriving three functions because the dependent set contained the maximum number of three variables. In males, the rst canonical function shows 0.7657 correlations, 85.53% variance explained from the rst canonical function. Similarly, the second and third function represented 8.53% and 5.94% respectively. Only the rst correlation was statistically signi cant (p < 0.001, F-test). Therefore, only the rst function has been elaborated and noteworthy in the context of present study. In females, the rst canonical function shows 0.6107 correlations, 71.41% variance explained from the rst canonical function. Similarly, the second and third function represented 17.98% and 10.61% respectively. Only the rst correlation was statistically signi cant (p < 0.001, F-test). Therefore, only the rst function has been elaborated and noteworthy in the context of present study. Table 4 shows the loadings and cross-loadings of the variables for the rst canonical function in canonical correlation analysis for both males and females. Canonical loadings depict correlation between observed variables (dependent/ independent) with the same set of canonical variate (dependent/ independent set). On the other, the canonical cross-loadings show correlation between observed variables (dependent/ independent) with the opposite set of canonical variate (dependent/ independent set). In females, the loading of the variables for rst function reveals that most important variables for anthropometric set was fat-free mass (loading: 0.591) followed by chest circumference (loading: 0.559), weight (loading: 0.555) and wrist breadth (loading: 0.551). On the other, loading values of the variables for rst function reveals that left handgrip strength (loading: 0.947) and right handgrip strength (loading: 0.928) both were more or less equally contributing for the strength set. The cross-loadings of the variables for rst function reveals the most important variables for strength were the following anthropometric measurements-fat free mass (cross-loading: 0.361), chest circumference (cross-loading: 0.342), weight (cross-loading: 0.339) and wrist breadth (cross-loading: 0.336).

Discussion
The aim of the present study was to explore the relationship between strength measurements with anthropometric measurements of adult Santal labourer population of Birbhum district, West Bengal. The individuals of the present study were from same ethnic origin, have more or less similar socio-economic condition. The test protocols for collection of data were similar for all the individuals and the data were collected by single investigator (BM) with single set of instrument.
The result of present study depicts higher values in strength measurements in males compared to females. Similar ndings reported in most of the previous studies [33,34] and they noted that may be due to advantages in amount of muscle mass and contractile tissue in males as compared to females.
Secondly, the advantages of height and forearm length in males favoured them for greater lever arm for force generation. Moreover, the nature of daily activity is more strenuous in males that facilitate more developed and stronger muscles than females. The pattern of strength measurements were similar in boys and girls up to puberty, and there after it diversi ed from one another as a result of predominant adipose deposition in girls and increases in muscle mass in boys [35,36]. Strength measurement reaches its peak during the middle age and then gradually declines with increment of age [37].
The result of present study indicate that the mean value of handgrip and back strength of both male and female participants were much lower as compared to male agricultural labourers of Jalpaiguri district [38], female construction labourer of Jalandhar [39] and brick eld workers of Hooghly district [40,41].
The lower values in strength measurements of the present population may be associated with their poor nutritional status and poor socio-economic condition as noted elsewhere [42]. Corroborate with present postulation, Chilima and Ismail [43] reported that poor grip strength of Malawian sample signi cantly related with their poor nutritional condition (indicated by low BMI). Further, Pieterse and colleague [44] added that poor grip strength among the Rwandan refugees was associated with long and/ or short-term effect of poor living condition, which is common in most of the developing countries. In a large-scale study, Leong et al. [8] stated that people of South East Asia generally have lower strength measurements compared to other parts of the world that attributes to their poor nutritional and socio-economic condition.
The results of Pearson's correlation analysis indicate that strength measurements negatively correlated with age. Therefore, strength measurements gradually decline with advancement of age. Empirical studies on Japanese-American [45] and U.K. [46] also gure out that grip strength declines one to two percent annually. Rosenberg [47] and Chen et al. [48] explained that aging process associated with decrease in hormonal level, loss of muscle bres and muscle mass that resulting changes in muscle strength. Moreover, the steadiness in hands decreases with increase in age that also resulting decline in strength parameters [49].
On the other, handgrip (both hands) and back strength was positively correlated with most of the anthropometric measurements that corroborates with earlier studies [50,51,53]. Coherence with present nding, Roy and Pal [13] reported that the mean value of anthropometric measurements were greater in higher strength group compared to their lower counterpart, indicative of strong positive correlation among the variables. In addition, More and Vyavahare [53] suggested that the knowledge on strength and anthropometric traits might have an impact on productivity and rehabilitation, particularly among the people who earn their livelihood on physical/ muscular strength.
The results of canonical correlation analysis indicate that present model explained around 85% and 71% variability of the strength data from anthropometric data for males and females respectively. The model also depicts that the most signi cant predictors of strength were fat-free mass, forearm circumference and weight for male data set and fat-free mass, chest circumference, weight and wrist breadth for female data set. Corroborate with present nding, Roberts and colleague [54] noted that forearm circumference of the adult naval personnel were signi cantly associated with their strength. Further, studies con rmed that anthropometric measurements related to upper extremities [12,52] including hands [49] were important predictors of strength measurements. That may be explained as the muscles used to produce grip force are primarily located in the forearm regions [55]. Many studies [19,56] found strong association between strength components and fat-free mass that corroborates with present nding. Jurimaea et al. [57] and de Souza et al. [58] also reported that fat free mass had greatest in uence on strength. Studies [40,41] conducted on labourer population narrated that greater daily physical activity associated with lower body fat deposition, resulting increase in strength parameters. Coherence with present nding, the relationship between strength measurements and weight were documented in several studies [59,60]. They elucidate that strength measurements were highest in normal weight individual followed by under-weight and obese. The reason may be energy de ciency in under-weight individual, while fatty in ltration in muscle and changes in distribution of type I and II muscle bres are the reason behind lower strength among obese individual.

Conclusion
In sum, the result of present study provides a sample of healthy adult Santal labourer population on strength measurements that may be used for epidemiological/ clinical and rehabilitation purposes. It explores the overall relationship between strength and anthropometric measurements. Strength measurements share a strong relationship with anthropometric measurements and along with age, fatfree mass, forearm circumference, weight, chest circumference and wrist breadth were the most important anthropometric measurements that were related to strength measurements. However, the result of the present study cannot be claimed to be universal because of the limitations of the present study -(a) study was restricted to a particular occupational group with cross-sectional in nature, (b) single ethnic/ genetic (endogamous) group, (c) other uncontrollable factors like diurnal variation, temperature, humidity and other concomitants were not taken into account. Therefore, future studies in different populations (ethnic/genetic) and different occupational groups with larger sample size, considering all the limitations of the present study as mentioned would provide better insights into the present problem.