INTRODUCTION
Sarcopenia is a disease characterized by a progressive and generalized disorder of skeletal muscle that is highly prevalent among older adults. It is associated with an increased risk of falls, mobility impairments, and loss of independence, contributing to poorer quality of life and a greater need for long-term care.1-3)
The diagnosis of sarcopenia is based on reduced muscle strength and mass. However, the methods used to determine low muscle mass tend to be costly and are not typically employed in clinical practice, particularly in primary healthcare settings.1) Therefore, there is a need to identify accessible and feasible methods that can be applied on a large scale to detect this condition.
Anthropometry is a low-cost, noninvasive, and easy-to-apply approach that could serve as a more accessible alternative for detecting low muscle mass and assessing the risk of sarcopenia. However, few studies have evaluated the predictive capacity of different anthropometric measures for screening this condition.4)
Arm circumference (AC) is a simple, noninvasive anthropometric indicator frequently included in geriatric health assessment scales to evaluate nutritional status. It reflects both muscle mass and subcutaneous fat. AC may be particularly useful because it is less affected by fluid retention, whereas edema is common in the lower extremities and can influence calf circumference.5) A reduction in AC has been associated with an increased risk of all-cause mortality among older adults.6)
Other markers of functional performance in older adults, such as the Timed Up and Go (TUG) test, the Short Physical Performance Battery (SPPB), and gait speed (GS), are also potential indicators of low muscle strength and, consequently, sarcopenia.1,7,8)
In this context, the aim of this study was to evaluate the diagnostic accuracy of AC compared with other anthropometric and functional markers as a predictor of sarcopenia in community-dwelling older adults.
MATERIALS AND METHODS
Study Design
This study is based on an observational, cross-sectional, analytical, and descriptive household survey of a population sample. It involves older adults and is titled “Alagoas First Diagnosis on Health, Nutrition, and Quality of Life in Older Adults,” with data collection conducted between April 2022 and January 2024.
Population and Sample
The total sample consisted of individuals aged 60 years or older, distributed across 11 municipalities in the state of Alagoas. A non-probabilistic convenience sampling plan was adopted, based on the main study population, which comprised individuals who met the eligibility criteria. All participants from the main study who had complete data for the variables of interest in this study—AC, calf circumference (CC), sit-to-stand (STS), handgrip strength (HGS), and fat-free mass index (FFMI) assessed by bioelectrical impedance analysis (BIA)—were included. Individuals lacking any of these key variables in the original research database were excluded.
Data Collection
Data collection was conducted through home visits to older adults by trained and qualified researchers. Visits were made to each participant identified in the community who agreed to take part in the study. A pre-established questionnaire was used to gather sociodemographic, economic, housing conditions, current and past health status, lifestyle information, followed by assessments of body composition and physical performance tests for sarcopenia diagnosis.
Prior to the start of the main study, all research staff were trained and qualified by an experienced researcher to minimize discrepancies in the data collected. Initially, all anthropometric measurements and functional tests were performed and repeated by the study researchers, who were organized into teams consisting of three researchers and one supervisor (an experienced researcher).
To standardize anthropometric measurement procedures and evaluation techniques, all researchers received both theoretical and practical training. After the techniques were standardized, all researchers repeated the evaluations with volunteers, and the values obtained were compared with those from the assessments conducted by the experienced researcher. Following individual evaluations, if a difference equal to or greater than 3% was identified between the supervisor’s and the researchers’ measurements, the team members underwent additional training to standardize anatomical landmarks and measurement procedures.
Specifically for mid-upper arm circumference measurement, the volunteer was instructed to keep the non-dominant arm relaxed along the side of the body. Using a measuring tape, the circumference was measured at the midpoint between the acromion and the olecranon, and the value was then recorded with precision.
Study Variables
Sociodemographic and economic variables
Demographic data collected included sex, age, marital status, education level, monthly household income, and family arrangements (whether family members were related or unrelated and residing in the same household). Clinical variables related to lifestyle (smoking, alcohol consumption, and physical activity level), the presence of chronic diseases, and polypharmacy were also recorded.
Anthropometrics
The CC measurement was taken using a flexible, non-elastic measuring tape at the most voluminous point of the calf, keeping it level and firm without compressing the skin. The cutoff point for reduced muscle mass was considered to be ≤34 cm for male and ≤33 cm for female.9)
The AC measurement was performed using a flexible, nonelastic measuring tape placed on the participant’s relaxed non-dominant arm at the midpoint between the olecranon and the acromion.10)
Assessment of FFMI
Fat-free mass was assessed using a tetrapolar BIA body composition analyzer (Model BC601G; Tanita Corporation, Tokyo, Japan). The FFMI was calculated by dividing the whole-body fat-free mass (FFM) by the square of the height (FFM/height²). The definition of low FFMI was based on the cutoff points established by Kawakami et al.,11) being <18 kg/m² for male and <15 kg/m² for female.
Strength assessment
Low muscle strength was assessed using the HGS test and the STS. To assess the HGS, three alternating measurements were performed on each hand (dominant and non-dominant), each lasting 5 seconds, and among the six measurements, the highest value recorded was considered. The cutoff points proposed by the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) were used, which establishes <27 kg for male and <16 kg for female for low muscle strength.1) The STS was performed with the aid of a chair, and the individual was instructed to stand up and sit down five times as quickly as possible without physical assistance, with the shortest time taken being considered. A time >15 seconds was classified as low strength.1)
Physical performance assessment
Physical performance was assessed using GS and the TUG test. GS is a test in which the time taken for the participant to walk 4.6 m at their usual walking pace is recorded.12) Three trials were performed, and the average of these was used as the final value. A speed ≤0.8 m/s was considered an indicator of severe sarcopenia.
The TUG test involves the older adult standing up from a chair, walking three meters, turning around, and sitting down again in the same chair. Completing the test in ≥20 seconds was considered an indicator of sarcopenia severity and a high risk of falls.1)
Diagnosis of Sarcopenia
Sarcopenia was defined based on the assessment of muscle strength, evaluated by HGS, and muscle mass, measured by the indirect assessment of FFMI, as suggested by Kawakami et al.,11) for sarcopenia screening in population-based studies. Its severity was identified based on low physical performance in the GS or TUG tests.1)
Statistical Analyses
Statistical analyses were performed using R software (version 4.4.1; https://www.r-project.org).
Absolute and relative frequencies were calculated for categorical variables, accompanied by 95% confidence intervals (CI). For proportions, Wilson's method was used. In cases of very low frequencies, Clopper-Pearson exact method was applied.
Multivariable logistic regression models were fitted to assess the associations between the predictor variables (AC, CC, STS, TUG, and GS) and the presence of sarcopenia. The regression coefficients were interpreted as odds ratios (ORs), accompanied by their 95% CI. The analysis was conducted using the glm() function in R, and model diagnostics were validated with the car package.13)
To evaluate the discriminative ability of the variables, receiver operating characteristic (ROC) curves were generated, and the area under the curve (AUC) was calculated as a measure of performance. Youden's index was used to identify the optimal cutoff point for AC, maximizing sensitivity and specificity. For the other variables, the previously defined cutoff points were considered. These analyses were conducted using the pROC package.14)
Differences in AUCs between explanatory variables for sarcopenia were evaluated only for the variables that remained significant in the multivariable logistic regression, using DeLong's test.15) This test was implemented with the roc.test() function.
RESULTS
Initially, 1,152 individuals comprised the main study. However, 444 were excluded due to physical limitations or missing data on physical and functional assessments, resulting in a final sample of 708 older adults (Fig. 1). A comparative analysis was conducted between the group of older adults excluded from the original database and those who remained in the study. The excluded group did not differ statistically in terms of sex, age, education, income, body mass index (BMI), HGS, or FFMI (p>0.05 in all analyses), ensuring the representativeness of the study sample from the original cohort.
The majority were female (n=458; 64.7%), aged between 60 and 94 years (mean 70±7.74 years), had less than 4 years of education (n=506; 72.9%), and had a family income greater than one minimum wage (n=322; 55.4%). A large proportion lived in mixed or extended family arrangements (n=505; 79.7%), had chronic diseases (n=622; 89.9%), and were sedentary (n=566; 82.5%).
The prevalence of sarcopenia was 14.7% (n=104), and 4.9% (n=30) presented with the severe form of the condition (Table 1).
In the logistic regression, AC emerged as an independent predictor of sarcopenia (OR=0.83; 95% CI 0.75–0.92; p=0.001), indicating that for each additional centimeter of AC, the odds of sarcopenia decrease by 17%. CC demonstrated an even stronger effect (OR=0.66; 95% CI 0.58–0.76; p<0.001), reducing the risk by 34% per centimeter. Among the functional indicators, only TUG remained significant (OR=1.21; 95% CI 1.08–1.34; p=0.001), increasing the odds of sarcopenia by 21% for each additional second. STS (OR=1.43; 95% CI 0.71–2.99; p=0.327) and GS (OR=2.27; 95% CI 0.48–10.61; p=0.301) did not show statistical significance (Table 2).
The ROC curve (Fig. 2) revealed cutoff points for AC by sex: 28.0 cm for male (sensitivity 84.3%, specificity 75.9%) and 27.0 cm for female (sensitivity 79.2%, specificity 76.8%). These results indicate excellent discriminative ability of AC in identifying sarcopenia, with an AUC of 0.85 for both sexes. CC also showed high accuracy: AUC = 0.83 in meale (sensitivity 88.0%, specificity 63.3%) and AUC=0.89 in female (sensitivity 94.2%, specificity 66.1%).
In contrast, TUG presented high specificity (97.7%) and low sensitivity (3.4%) using a general cutoff point. STS demonstrated a sensitivity of 83.5% and specificity of 31.8%, while GS showed a sensitivity of 35.3% and specificity of 75.7% (Table 3).
Table 4 presents the comparison between AUCs. AC and CC did not differ significantly (z=0.24, p=0.81), confirming similar performance. However, both significantly outperformed TUG (AC vs. TUG, z=5.10, p<0.01; CC vs. TUG, z=4.80, p<0.01).
Table 5 presents the confusion matrix of the predictors, by sex. AC showed slightly higher sensitivity and positive predictive value (PPV) among females (sensitivity 90.2% and PPV 43%). In male, AC demonstrated slight superiority in specificity (specificity 70.1%). The CC presented greater sensitivity and specificity among male (94.2% and 66.2%, respectively), however, the PPV was higher among female (37.6%). TUG showed the poorest performance indicators in both sexes.
DISCUSSION
This study stems from the intention to contribute to the screening and diagnosis of sarcopenia, proposing AC as a complementary or alternative indicator for other assessment methods. By investigating the accuracy of AC, we suggest that it can serve as an additional screening tool for this condition when more sophisticated methods are not feasible, allowing for early interventions that may mitigate associated complications and improve quality of life in this population. The prevalence of sarcopenia in this study was expected for older adults living in the community. Other studies have reported a prevalence of sarcopenia between 11.98% and 17.4% in the older adult population.16-19)
Our findings indicate that AC is a robust indicator of sarcopenia, comparable to CC, with both demonstrating greater discriminative ability than the other variables studied. This is particularly relevant for older adult populations, in whom objective and reliable assessments are essential for early diagnosis and the implementation of appropriate interventions. In general, older adults with sarcopenia exhibit reduced AC.20) Few studies have suggested the use of AC alone or in combination with other indicators in predicting sarcopenia in older adults.21,22) In this research, the assessment of AC as a proxy for the diagnosis of sarcopenia showed high accuracy in older adults living in the community. In China, this indicator demonstrated good screening ability in community-dwelling older adults of both sexes (AUC 0.74–0.90), with cutoff points lower than those identified in our study (25.9 cm in male and 26.5 cm in female). Furthermore, AC was able to distinguish differences in muscle mass, fat mass, and physical function in male.23)
The performance of parameters for the diagnosis of sarcopenia in French individuals undergoing hemodialysis showed moderate accuracy for AC in both sexes (AUC 0.63–0.87), with cutoff points of 26.8 cm in female (sensitivity 69%, specificity 65%) and 27.5 cm in male (sensitivity 61%, specificity 80%).24) Similar to our study, AC proved to be an indicator of sarcopenia as important as CC. Other studies21,25) have demonstrated that AC serves as a useful marker of low muscle mass in sarcopenia screening among hospitalized older adults, showing high accuracy (AUC 0.89–0.91). These studies revealed that the inclusion of AC significantly increased the diagnostic accuracy of screening instruments, reinforcing its use as a practical and accessible complement in the assessment of this condition.
It is well recognized that CC is a sensitive indicator for identifying low muscle mass in clinical practice.26) However, CC may become unfeasible in conditions such as hydroelectrolytic disorders, the presence of edema, and lower-limb amputations. AC, on the other hand, can be as useful as CC in assessing muscle mass and is less affected by these conditions.27)
The results indicate that AC (<28 cm for male and <27 cm for female) and CC are the most promising variables for predicting sarcopenia. Both indicators showed high discriminative ability and cutoff points that balance sensitivity and specificity. These findings reinforce the relevance of AC and CC as potential clinical markers for identifying sarcopenia, contributing to a more accurate and objective, data-driven diagnosis. To the best of our knowledge, this is the first study to identify the relationship between AC and sarcopenia to compare AC with other anthropometric markers in older adults on the American continent. Furthermore, the use of a robust sample, as employed in this study, enhances the relevance of the research by providing more comprehensive findings with greater scientific reliability.
From a public health perspective, it is important to highlight that AC is an easily performed measurement that can be applied by primary care professionals, including community health workers during home visits. This practical applicability expands its potential as a population-based screening tool for sarcopenia in resource-limited settings, such as many regions of Brazil.
Our findings also align with national studies. Pereira et al.18) and Sousa et al.19) reported results consistent with ours regarding the relationship between AC and sarcopenia. The study by Barbosa-Silva et al.,28) which identified a high prevalence, used different diagnostic criteria, which may explain methodological divergences and distinct percentages. This contrast reinforces the importance of studies with standardized criteria and specific cutoff points for the Brazilian context.
In the present study, the diagnosis of sarcopenia was not associated with GS or TUG test. This finding may be explained by the fact that these measures predominantly reflect overall physical performance, which is influenced not only by muscle strength and mass but also by factors such as balance, cardiorespiratory function, sensorimotor capacity, and cognition. Thus, in community-based populations composed mainly of older adults who remain independent and have good functional reserve, early alterations in muscle mass and strength may not be sufficient to cause a detectable decline in GS or TUG performance, as many of these individuals still maintain relatively preserved functional capacity through compensatory mechanisms such as gait adaptations and the use of balance strategies. Therefore, an initial loss of muscle mass and strength may not produce a detectable decline in GS or TUG until sarcopenia reaches a more advanced stage.1,25,29)
Moreover, these tests are subject to a ceiling effect, reducing their discriminative capacity in samples with high functional performance.25,30) They also tend to be more strongly associated only with more severe cases of sarcopenia, in which the reduction in muscle strength and mass is intense enough to impair physical performance and limit mobility, unlike the profile of the older adults included in the present study. This pattern is consistent with recent literature, which recognizes that physical performance decline occurs mainly in the more advanced stages of the disease, after substantial muscle loss has already taken place.1,29)
It is worth noting that this study has some limitations that should be considered when interpreting the results. Among them, the use of a convenience sampling method from a larger population-based study may limit the external validity of the findings, as our results may not reflect the conditions of other populations, particularly those in more developed regions. However, the remaining sample in this study exceeds by more than three times the minimum number expected for studies having sarcopenia as the main outcome with an estimated prevalence of 17.4%, a confidence level of 95%, and a sampling error of 5% (minimum sample size=221).
A second limitation lies in the method used for estimating muscle mass, which relied on the fat-free mass index obtained through BIA. On the other hand, in population-based studies with robust samples, muscle composition analysis by dual-energy X-ray absorptiometry (DXA) becomes unfeasible, both due to the high cost and the logistics of transporting the equipment or the individuals for the assessment. Furthermore, to minimize this limitation, we used the cutoff points for FFMI established by Kawakami et al.,11) who reported a strong correlation (r=0.95) between FFMI and muscle mass assessed by DXA, with the following AUC for screening low muscle mass defined by DXA (0.95 for male and 0.91 for female).
Finally, the assessment of raw arm circumference, measured exclusively on the non-dominant limb, could be interpreted as a limitation of this study. However, the choice of the non-dominant side in this research was intentional and followed current international anthropometric protocols to minimize bias resulting from dominance and muscle hypertrophy, since the dominant arm may present greater mass and volume due to repeated use in occupational or daily activities, particularly among older people. This tends to overestimate muscle mass and compromise comparability among individuals of the same sex and age.10,26,29-31) Furthermore, it is necessary to standardize measurements and assessments in research on sarcopenia diagnosis, facilitating the reproducibility and comparability of results identified by different researchers. Thus, the evaluation of the non-dominant arm represents an appropriate methodological approach to reduce this potential bias and ensure greater validity of muscle mass estimates among older people.
In conclusion, the data from this study indicate that AC, based on the established cutoff points (28.0 cm for male and 27.0 cm for female), is a relevant indicator of sarcopenia risk associated with reduced muscle mass in community-dwelling older adults. This makes it an important tool for healthcare professionals in detecting the risk of sarcopenia, given the similar accuracy of AC and CC. Furthermore, when compared with other components of sarcopenia (TUG, STS, and GS), AC showed greater sensitivity and specificity for detecting sarcopenia.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
