Whole-body MRI measurements took place at a separate appointment not more than 4 days apart. Waist circumference was measured midway between the lowest rib and the uppermost boarder of the iliac crest in the medial axillary line and at the end of normal expiration using a non-stretchable tape circumference measuring tape seca The seca mBCA and use four pairs of electrodes eight electrodes in total that are positioned at each hand and foot.

The 8-electrode technique enables segmental impedance measurement of the arms and legs. The mBCA is designed for measurements in the standing position and consists of a platform with an integrated scale, a handrail system and a display and operation unit.

According to the manufacturer´s instructions, arms should be held straight. The mBCA is designed for measurements in the supine position and can be operated using either four adhesive electrodes on the right side of the body or eight electrodes on both sides while the subject is lying on a non-conductive surface.

The distal edge of the toe electrode was placed from an imaginary line through the middle of the metatarsophalangeal joints of the second and third toes. The proximal edge of the ankle joint electrode was attached along a line through the highest points of the outer and inner ankle bones.

Intra- and interoperator reproducibility for total SMM BIA calculated as coefficient of variation from three replicate measurements by two observers in four participants BMI A whole-body DXA scan was performed to measure appendicular LST using a Hologic Discovery A densitometer and the whole-body software In Kiel : Measurements of SMM and visceral adipose tissue VAT volumes were performed in a supine position with arms extended above the heads using a Magnetom Avanto 1.

For measurements at the trunk participants were required to hold their breath. All images were segmented manually using Slice-O-Matic, Tomovision 4. VAT was evaluated in slices from the diaphragm top of the liver or the base of the lungs, T10 to the femur heads. The software employed knowledge-based image processing to label pixels as fat and nonfat components using a threshold for adipose tissue on the basis of the gray-level histograms of the images.

Each slice was manually reviewed and voxels arising from fatty bowel content were deleted. Total SMM excluding head and neck and VAT were determined from the sum of all tissue areas cm 2 multiplied by the slice thickness. Coefficients of variation for repeated measurements of SMM and VAT were 1.

In New York : Total body skeletal muscle SMM and VAT were measured using whole-body multislice MRI as described previously. SliceOmatic Image Analysis Software version 4.

MRI-volume estimates were converted to mass using the assumed density of 1. The technical error coefficients of variation for three repeated readings of the same scan by the same analyst for skeletal muscle and VAT were 2.

Data analyses were performed with R Software version 3. Descriptive statistics are presented as means±s. Differences between independent samples e. men and women or ethnic groups were analysed using unpaired t -test. Differences between parameters of body composition assessed by BIA and DXA or MRI were tested using paired t -test.

Stepwise multiple-regression analysis was applied to the data of the Caucasian subjects to derive the best-fitting regression equations to predict total and segmental SMM MRI and LST DXA from conventional wrist-ankle and segmental impedance data obtained in a standing or supine position.

Goodness of fit was assessed by determination coefficient R 2 , proportion of the total variance in the dependent variable that is explained by the independent variables and precision by the pure error.

The pure error was calculated as the root mean square of the differences between predicted and measured data the smaller the pure error, the greater the precision of the tested equation. Optimal combination of predictor variables was selected by consideration of the correlation between predicted and observed data which should be maximized.

The prediction accuracy was compared between BIA equations derived from the supine vs standing and conventional wrist-ankle vs segmental measurements and was tested by comparison of correlation coefficients. To determine the effect of ethnicity on the accuracy of the BIA prediction equation, the equations derived from the Caucasian subjects were applied to the Caucasian, Afro-American, Asian and Hispanic subjects in an independent population from New York.

Analysis according to Bland and Altman was used to determine absolute agreement between the body composition assessed by criterion methods MRI and DXA and BIA. According to this approach, the bias is the difference between measured and predicted values of SMM or LST and the error is the standard deviation of the bias.

error i. The validity of BIA results vs MRI and DXA was compared between different ethnic groups by comparison of the bias. The pure error accuracy statistic was used for cross-validation of BIA results i.

The pure error was calculated as the root mean square of the differences between predicted and measured data the smaller the pure error, the greater the accuracy of the tested equation.

Basic characteristics of the Caucasian study population used to generate the BIA equations are shown in Table 1 stratified by sex. Men had a higher BMI and waist circumference compared with women. Accordingly, impedance raw data obtained in the supine position with segmental measurements were lower in men.

Besides these impedance data, weight, age and gender were used as independent variable in the equations for total SMM. For prediction of SMM of the arms and legs information on segmental impedance of the limbs was used.

In addition to impedance data of the trunk, equations for prediction of VAT were obtained using waist circumference, waist circumference 2 , age, height, weight and the interaction terms waist circumference × sex and waist circumference 2 × sex as predictors.

Results of the stepwise regression analyses with total and limb SMM MRI and VAT MRI as the dependent variables are provided in Table 5. In Table 2 goodness of fit for BIA equations obtained from measurements in the standing and supine position using segmental or conventional wrist-ankle measurements are described by coefficient of determination and precision is given by the pure error.

There was also no difference between the variance in SMM MRI of the arms and legs explained by BIA in the two positions. When compared with the prediction of SMM, the accuracy of VAT prediction by BIA was considerably lower and waist circumference and gender were major predictors with impedance data having only a minor contribution to the prediction of the algorithm Table 5.

Basic characteristics of the multiethnic population used to validate the BIA equations are given in Table 3 stratified by ethnic group and gender. Age ranged from 18 to 65 years and BMI from Table 4 shows mean values for SMM, LST and VAT measured by the reference methods and predicted from BIA and compares i results from BIA vs the reference method within each ethnic group and ii the validity of SSM and LST predicted from BIA equations between different ethnic groups.

When compared with Caucasians, Asians had lower SMM of the total body, arms and legs. Hispanics had a lower and Afro-Americans a higher SMM of the legs.

VAT results from MRI were higher in Hispanics and lower in Afro-Americans when compared with Caucasians. Although the mean bias for prediction of total and regional SMM and LST was low in all ethnic groups, it was significant for all LST results and most SMM compartments, with exception of arm SMM in Asians and Hispanics and total and leg SMM in Afro-Americans.

The bias for VAT prediction was significant in Asians and Hispanics only. Comparison of the bias between Caucasians and different ethnic groups revealed a significant ethnic effect on the accuracy of BIA prediction equations i for VAT in Asians and Hispanics and ii for total and leg SMM and LST of the arms and legs for Afro-Americans and iii for leg SMM in Hispanics.

The differences between limb LST DXA and SMM MRI predicted by BIA ranged between 1. The present study investigated the impact of a standing vs supine position and conventional wrist-ankle vs segmental measurements on the goodness of fit of a BIA equation for prediction of SMM and VAT.

This result is important because it refutes the common belief that BIA analyzers that incorporate foot and hand contact points for standing on four metal plates and holding a rod with the fingers and thumb of each hand e.

This may be due to the fact that body compartments and anatomic regions are all highly correlated with each other. Because predictive accuracy reveals how well a model can predict new data points, we applied our models to the cross-validation sample and found that predictive power of plate vs gel electrode protocols was similar for all outcome parameters pure error for plate electrode protocols ranged between 0.

On the other hand, we found a significantly better prediction using segmental impedance measurements instead of a conventional wrist-ankle technology. The regional distribution of extracellular water changes from supine to standing position towards an increase in extracellular water in the limbs and a corresponding decrease in the trunk.

There are however reports that segmental impedance measurements are less accurate in predicting individual limb LST DXA than whole-body impedance 26 and early studies that estimated limb SMM MRI by segmental BIA measurements found a low bias but large limits of agreement that indicate a low precision of the method.

As a main drawback, this approach is cumbersome and error prone because it requires the correct placement of electrodes at anatomical landmarks of the trunk. Modern tetrapolar impedance devices like the seca mBCA and enable segmental impedance measurement of the right and left arm, the trunk as well as the right and left leg by using four electrodes on each side of the body at the hands and feet.

The selection or switching between the individual detecting and source electrodes allow the limbs to be used as virtual electrodes that measure the opposite side of the body.

Discrepancies in the assumptions of the homogeneous bioelectrical model that lead to a higher measurement error not only occur with changes in body position but also with differences in body shape that are associated with aging decreasing limb relative to trunk diameter , obesity apple and pear shape of body fat distribution and ethnic differences in trunk relative to leg length and regional adiposity and muscularity.

Validation of the BIA equations predicting total and limb SMM and LST in a multiethnic population with a great variance in muscularity has shown that BIA results correctly identified ethnic differences in muscularity and visceral adiposity Table 4.

However, the comparison of bias between Caucasians and different ethnic groups revealed some ethnical effects on the accuracy of BIA prediction equations that confirm previous results of other authors who used a more simple non-segmental and non-phase-sensitive without Xc output impedance measurement.

Reactance contributed 1. Although statistically significant, the differences of BIA biases between Caucasians and other ethnicities are small Table 4 and support our previous findings that a BIA equation is more device specific than population specific.

As a limitation to our study, the significant bias of all BIA results for SMM and LST in Caucasians investigated in New York suggests an additional impact of discrepancies in reference methods between the labs in Kiel and New York i.

Hologic vs GE DXA scanner and Siemens vs GE MRI machines. Finally, the differences between LST DXA and SMM MRI at the arms and legs were considerable, higher in men compared with women and more pronounced in obese participants see results due to a higher contribution of connective tissue to total lean mass.

This leads to a sizable overestimation of SMM with increasing age and adiposity when predicting appendicular LST DXA instead of SMM MRI. In conclusion, a high accuracy of phase-sensitive segmental BIA was observed with no difference in goodness of fit between standing and supine positions.

Segmental measurements improved the prediction of whole-body compartments compared with conventional wrist-to-ankle measurements. BIA equations based on MRI as a reference are more accurate for prediction of SMM when compared with DXA. Bosy-Westphal A, Müller MJ. Identification of skeletal muscle mass depletion across age and BMI groups in health and disease—there is need for a unified definition.

Int J Obes Lond ; 39 : — Article CAS Google Scholar. Kim JH, Choi SH, Lim S, Kim KW, Lim JY, Cho NH et al. Assessment of appendicular skeletal muscle mass by bioimpedance in older community-dwelling Korean adults.

Arch Gerontol Geriatr ; 58 : — Kim M, Kim H. Accuracy of segmental multi-frequency bioelectrical impedance analysis for assessing whole-body and appendicular fat mass and lean soft tissue mass in frail women aged 75 years and older.

Eur J Clin Nutr ; 67 : — Kim M, Shinkai S, Murayama H, Mori S. Comparison of segmental multifrequency bioelectrical impedance analysis with dual-energy X-ray absorptiometry for the assessment of body composition in a community-dwelling older population.

Geriatr Gerontol Int ; 15 : — Article Google Scholar. Kyle UG, Genton L, Hans D, Pichard C. Validation of a bioelectrical impedance analysis equation to predict appendicular skeletal muscle mass ASMM.

Clin Nutr ; 22 : — Pietiläinen KH, Kaye S, Karmi A, Suojanen L, Rissanen A, Virtanen KA. Agreement of bioelectrical impedance with dual-energy X-ray absorptiometry and MRI to estimate changes in body fat, skeletal muscle and visceral fat during a month weight loss intervention.

Br J Nutr ; : — Rangel-Peniche DB, Raya-Giorguli G, Alemán-Mateo H. Accuracy of a predictive bioelectrical impedance analysis equation for estimating appendicular skeletal muscle mass in a non-Caucasian sample of older people. Arch Gerontol Geriatr ; 61 : 39— Tanaka NI, Hanawa S, Murakami H, Cao ZB, Tanimoto M, Sanada K et al.

Accuracy of segmental bioelectrical impedance analysis for predicting body composition in pre- and postmenopausal women. J Clin Densitom ; 18 : — Villani AM, Miller M, Cameron ID, Kurrle S, Whitehead C, Crotty M. Body composition in older community-dwelling adults with hip fracture: portable field methods validated by dual-energy X-ray absorptiometry.

Xu L, Cheng X, Wang J, Cao Q, Sato T, Wang M et al. Comparisons of body-composition prediction accuracy: a study of 2 bioelectric impedance consumer devices in healthy Chinese persons using DXA and MRI as criteria methods. J Clin Densitom ; 14 : — Yoshida D, Shimada H, Park H, Anan Y, Ito T, Harada A et al.

Development of an equation for estimating appendicular skeletal muscle mass in Japanese older adults using bioelectrical impedance analysis. Geriatr Gerontol Int ; 14 : — Schutz Y, Kyle UUG, Pichard C.

Fat-free mass index and fat mass index percentiles in Caucasians aged 18—98 y. Int J Obes ; 26 : — Addison O, Marcus RL, Lastayo PC, Ryan AS. Intermuscular fat: a review of the consequences and causes.

Int J Endocrinol ; Schautz B, Later W, Heller M, Muller MJ, Bosy-Westphal A. Total and regional relationship between lean and fat mass with increasing adiposity—impact for the diagnosis of sarcopenic obesity.

Eur J Clin Nutr ; 66 : — Wang Z, Deurenberg P, Heymsfield SB. Cellular-level body composition model. A new approach to studying fat-free mass hydration. Ann NY Acad Sci ; : — Bosy-Westphal A, Later W, Hitze B, Sato T, Kossel E, Gluer CC et al. Accuracy of bioelectrical impedance consumer devices for measurement of body composition in comparison to whole body magnetic resonance imaging and dual X-ray absorptiometry.

Obes Facts ; 1 : — Janssen I, Heymsfield SB, Baumgartner RN, Ross R. Estimation of skeletal muscle mass by bioelectrical impedance analysis. J Appl Physiol ; 89 : — Bosy-Westphal A, Schautz B, Later W, Kehayias JJ, Gallagher D, Muller MJ. What makes a BIA equation unique?

Validity of eight-electrode multifrequency BIA to estimate body composition in a healthy adult population. Eur J Clin Nutr ; 67 Suppl 1 , S14—S Gallagher D, Kelley DE, Yim JE, Spence N, Albu J, Boxt L et al.

Adipose tissue distribution is different in type 2 diabetes. Am J Clin Nutr ; 89 : — Song MY, Ruts E, Kim J, Janumala I, Heymsfield S, Gallagher D. Sarcopenia and increased adipose tissue infiltration of muscle in elderly African American women. Am J Clin Nutr ; 79 : — Snyder WS, Cook MJ, Nasset ES, Karhausen LR, Howells GP, Tipton IH.

Report of the Task Group on Reference Man. Pergamon Press: Oxford, UK, Google Scholar. Eid M, Gollwitzer M, Schmitt M. Statistik und Forschungsmethoden. Bland JM. Perspective on the Assessment of Skeletal Muscle Mass. JMA J. Previously, measurement of skeletal muscle mass was a standalone requirement for the diagnosis of sarcopenia; moreover, it was a requirement in diagnostic criteria such as European Working Group on Sarcopenia in Older People EWGSOP , European Society for Clinical Nutrition and Metabolism-Special Interest Group ESPEN-SIG , International Working Group on Sarcopenia IWGS , and Asian Working Group for Sarcopenia AWGS , along with grip strength and walking speed.

As Muraki indicates, Computed Tomography CT and Magnetic Resonance Imaging MRI are highly accurate but difficult to implement in clinical and community settings. Conversely, Dual-energy X-ray Absorptiometry DXA and Bioelectrical Impedance Analysis BIA are easier to implement but have accuracy concerns.

D3-creatine dilution and echocardiography are promising in terms of applicability and accuracy, but there is insufficient evidence regarding cutoff values; however, future development is expected. Furthermore, when estimating skeletal muscle mass with DXA and BIA, correction is made for height squared, weight, and Body Mass Index BMI , but there is no consensus on which is the most appropriate correction method, although height squared correction has been commonly used.

For example, in the case of obesity, correction for height squared tends to overestimate skeletal muscle mass; thus, correction for body weight or BMI is preferable. Although we have demonstrated the utility of skeletal muscle measurement in the longitudinal cohort and clinical studies 2 , 3 , 4 , Sarcopenia Definitions and Outcomes Consortium SDOC has determined that skeletal muscle measurement is not necessary for the diagnosis of sarcopenia 5.

Currently, the Global Leadership Initiative on Sarcopenia GLIS group has been formed to globally discuss the definition and diagnostic criteria for sarcopenia. Liang-Kung Chen, Jean Woo, and Hidenori Arai are participating in this group from Asia, and further discussions are expected.

Muraki I. Muscle mass assessment in sarcopenia: a narrative review. View Article. Otsuka R, Matsui Y, Tange C, et al. What is the best adjustment of appendicular lean mass for predicting mortality or disability among Japanese community dwellers?

BMC Geriatr.

A less expensive, non-invasive, and reliable alternative method is bioelectrical impedance analysis BIA. From this methodology, variables such as resistance and reactance can be obtained. These variables, together with other anthropometric or sociodemographic variables, can be included in BIA predictive models or equations to estimate total or appendicular skeletal muscle mass ASM considering MRI or DXA as reference methods.

Various predictive models based on BIA have been developed worldwide 8 — Some have been validated through internal 8 , 10 , 11 , 15 , 17 or external validation procedures 18 — 20 to estimate ASM in older adults. However, high values of the coefficient of determination R 2 , low standard error of the estimate SEE of the predictive model, or the results of internal or external validation in a particular group, do not guarantee the validity of the predictive models to estimate ASM in other populations with specific or different characteristics 11 , 12 , 14 , 16 , 20 — Very few studies have considered valid some BIA equations to estimate ASM within the same ethnic group 18 , or with different health conditions 19 , In general, it is recognized that BIA equations that estimate ASM or any other body composition compartment are only precise, accurate and unbiased in populations with similar characteristics to the sample or ethnic group where it was generated 24 , These findings and others support that the published BIA equations should not be applied interchangeably.

They also highlight the need for external validation in the population of interest. This validation procedure will determine whether or not the predictive models could be generalized Currently, it is noticeable that most studies aim to generate new precise models to estimate body composition components or compartments based on the assumption 8 , 10 and their own results 9 , 11 , 12 , 14 , 16 that the existing models are not valid in certain populations.

Meanwhile, in other studies, it is possible to notice the efforts to use the existing equations, and in this way, avoid the generation of new ones unjustifiably 18 , 19 , 23 , 27 , However, based on the results of these studies, some of them have resulted inaccurate or have not achieved agreement.

In this case, an effective strategy may be the analysis of the bias during the validation process. This analysis consists on evaluating the trend of bias: verifying that it remains constant regardless of the amount of ASM presented. This allows discarding the use of published predictive equations, provide the bases for the development of new age, gender, or ethnic specific BIA equations, or determine the possibility of generating CFs for existing equations.

These CFs are derived from the mean differences between the equations' estimates and the measurements of the reference method. The condition to derive one, is that the bias must be distributed homogeneously throughout the average of both methods.

If it meets this criterion, it will be possible to add or subtract the mean difference to the estimated value of the original equation. In this way, it is possible to achieve agreement. However, it is important to clarify that a simple correction factor might not eliminate the prediction error in individuals, since these data come from the values of R 2 and the standard error in the estimation of the original equation.

In the case of Mexico, there is only one study 11 where two published BIA equations were applied to estimate ASM in healthy non-institutionalized older adults. The results of this external validation study showed that Kyle's and Sergi's equations were inaccurate in the validation sample.

Likewise, the researchers generated, and internally validated a new specific BIA equation for older Mexican adults from the center of the country. In the aforementioned study, the bias of the Kyle and Sergi's equations was not explored, and it has not currently been explored whether the equation generated in older adults from the center of Mexico and other published BIA equations could be valid for older adults from the northwest of Mexico.

This, taking into account that it was previously reported that older adults from the northwest of Mexico had less ASM compared to those from the center of the country This may probably be due to differences in total and central fat. Women and men from northwestern Mexico were fatter than those from central Mexico.

A positive association has been shown between fat mass and some markers of inflammation, such as C-reactive protein CRP , and a negative association between CRP and ASM In general, there are no studies where the equations' bias has been critically analyzed, nor where correction factors have been proposed for the existing equations based on BIA to estimate ASM.

This could stop the generation of models that may never be used. The external validation and bias analysis can provide alternatives and close the gap between equation development and implementation of equations, in this case, for estimation at the group level Moreover, the influence of the DXA model used as reference method, is a factor which has not been explored in external validation studies.

Currently, the most widely used models are DXA Lunar and DXA Hologic, of which significant differences in body composition measurements have been reported between both models 31 , Considering this evidence, it is possible that the ASM estimated by an equation may not be entirely comparable or equivalent when compared to ASM measurements with a different DXA model than the one used for the generation of the equation.

This could lead to bias in external validation studies. For all of the above, the objective of this study was to assess the agreement between six equations based on BIA and dual-energy X-ray absorptiometry to estimate ASM in non-Caucasian older adults, considering the DXA model. The bias was also analyzed in order to propose correction factors.

This is a secondary analysis generated from various studies with a cross-sectional design 33 — 35 and the baseline data of one randomized clinical trial 36 carried out in the Body Composition Laboratory of the Food and Development Research Center, CIAD, A.

This analysis included a large sample of older men and women from Hermosillo, Sonora, México. The methodology has already been described previously in the mentioned studies, but a brief description is provided. Independently of the cited studies, all participating subjects were adults over 60 years of age or older, who were invited to participate through flyers, telephone calls and home visits.

The corresponding study protocol was explained to them, as well as the procedures to which they would undergo. All volunteers underwent body composition measurement by different methodologies including DXA and BIA. Likewise, various questionnaires and scales were applied to determine the health status, including functionality and cognition.

All the subjects were free of physical disability according to the Lawton and Brody scale 38 or the Barthel Index 39 , and the majority were free of cognitive impairment according to the Pfeiffer Scale 40 or the Mini Mental State Examination Also, information on demographic and socioeconomic conditions was collected.

All these procedures were conducted at CIAD, A. From the cited studies, a primary database was built. All the volunteers selected for this study, had to have a physical file, which had to contain complete information on age, sex, waist circumference, resistance and reactance variables, and DXA scans.

They had to be free of diseases, conditions or medications that could affect body composition or hydration status. Regardless of their BMI, men and women older than 60 years were included. All those subjects who did not have complete data on the variables necessary for this external validation protocol, and those who had atypical data or outliers detected by the exploratory analysis were excluded.

The identification of outlier variables was carried out through the visual identification of variables that were separated from the set of points of the scatter plot. Height was measured in the same condition, placing the subject's head according to the Frankfurt plane and using a digital stadiometer SECA stadiometer , Hamburg, Germany.

Afterwards, the BMI was calculated. Waist circumference was measured just above the superior border of the iliac crest. The measurement was made with the subject standing and using a fiberglass measuring tape Lafayette Instruments Company Inc. Body composition in some of the cited studies was assessed using DXA Lunar Radiation Corp; Madison, WI, USA or DXA Hologic Discovery WI QDR Series; Waltham, USA.

It is important to point out that it has been reported that, quantitatively, these two models do not measure the exact same amount of body composition components such as ASM, or compartments such as fat mass. Regarding the appendicular lean mass ALM , Shepherd et al.

For this study, we analyzed the ASM measurements in a subsample of 70 older adults, who had been measured with both DXA models. DXA measurements were performed in the same day, following the same protocol for the whole scan and scan editing for ASM determination.

A paired t -test was used to determine if the mean difference of the measurements between both methods was different from zero. Protocol of DXA measurement, DXA scan edition for ASM and calibration were performed according to a published study Participants were measured wearing a disposable gown and free of plastic or metal objects.

The ASM determined by DXA ASM DXA was considered as reference. For those that did not fit in the DXA scan area, half-body scans were performed, and the remaining side was duplicated as described by Rothney et al.

In the case of two subjects who wore non-removable metal accessories, the opposite half of the body to where they had the accessory, was duplicated.

For the purposes of this secondary analysis, resistance R and reactance Xc were measured by a RJL Systems single frequency bioimpedance 50 kHz , Detroit, Mich, USA, which complied with a daily calibration protocol with a resistance of ohms.

BIA measurements were according to the methodology published previously 24 , Both the DXA and BIA measurements were performed after an 8 h fast, with an empty bladder and without having consumed food or liquid prior to the measurement.

English-language articles on topics of BIA equations or predictive models to estimate ASM published between and were identified in the PubMed database.

These cut-off points were considered since R 2 is expected to be as close to 1 and SEE as close to zero. An equation with these values can assure considerable precision. Only BIA equations generated including older adults aged 60 to 90 years, non-institutionalized of any nationality or ethnic group, and with DXA as the reference method were included.

Age, sex, and body weight are both clinically and statistically associated with ASM, and together with BIA variables, the predictive model can yield more precise and accurate results of the ASM. Finally, the selected equations must have been generated with a single or multi-frequency bioimpedance model.

Also, there was no discrimination regarding the method of generation and validation of the equations, nor the nutritional status of the subjects that integrated the generation or validation sample. The data was analyzed using STATA version 16 StataCorp LP, TX, USA.

An exploratory analysis of the primary data was carried out to observe the behavior of the data and detect atypical data or outliers. The significance of the differences between men and women was determined using an independent sample t -test and the results are presented as mean ± standard deviation.

To test if the differences between the ASM measured by DXA Lunar and Hologic were different from zero, a paired t -test was used in the sample of 70 adults. Regarding the validation procedure, the agreement between methods was evaluated using the Bland and Altman procedure, which considers that the average of the two methods is the best estimator.

Objectively agreement was tested by a paired t test and by simple linear regression analysis. The paired t -test assessed if the mean differences between the estimation of each equation and the ASM measurement by DXA were statistically different from zero, and the simple linear regression analysis, which assessed the homogeneity of the dependent variable.

To visually analyze the mean of the differences and the distribution of the differences between methods, Bland and Altman 46 plots were incorporated.

Additionally, the simple regression analysis must test that the differences are randomly distributed. This would prove the homogeneity of the bias, that is, the homogeneous distribution of the differences along the spectrum of the mean of ASM between methods.

If these two conditions were met, agreement was accomplished, meaning that the BIA equation can be considered as an interchangeable method to DXA to assess ASM in this large sample of non-Caucasian older adults. This methodology to establish agreement has been described and applied in other validation studies 33 , This bias analysis supports or rejects the possibility of deriving a CF.

In order to propose one, the bias distribution must be homogeneous, and the mean of the differences must be different from zero.

If so, the equation can be corrected by subtracting or adding the mean difference to the respective equation. This CF does not change the behavior of the variables included in the equation, but it makes it possible to reduce the average of the differences bias in the estimates at group level.

This correction has been proposed in other studies 33 , 48 , and has provided the opportunity to improve the estimates according to the equations where applicable. The initial sample made up of all the subjects participating in the previously mentioned studies was of participants.

Ninety-five volunteers were excluded due to lack of BIA data. The sample consists of women Some of them reported a previous diagnosis of hypertension, controlled type 2 diabetes, and dyslipidemia, with their respective pharmacological control.

Other diseases reported were colitis, gastritis, bronchitis, rheumatoid arthritis, bronchial asthma, or controlled hypothyroidism, with stable weight according to self-report. The mean value of BMI was According to their BMI classification, 6 subjects were underweight 1.

The mean value of ASM in the whole sample was of According to the DXA model, the mean ASM measured by DXA Hologic was The general characteristics of both samples are found in Table 1.

Regarding the BIA equations to estimate the ASM, a total of 25 equations were found, of which 10 were generated in older adults. Of these, only 5 had reported an internal validation process, and 6 have been externally validated in other studies.

Only 6 equations which met the selection criteria were selected: Kim's, Kyle's, Rangel-Peniche's, Sergi's, Toselli's and Yoshida's equations. The characteristics of these equations are shown in Table 2. These equations were applied to the complete sample, and with this, the variables ASM Kim , ASM Kyle , ASM Rangel , ASM Sergi , ASM Toselli and ASM Yoshida were obtained.

Importantly, Kim's and Toselli's equations generated with DXA Lunar, were tested on subjects measured with DXA Lunar, while BIA equations generated using DXA Hologic as the reference method, were tested on those measured with that model.

This, in order to eliminate the effect or possible bias due to DXA model in this validation procedure. The results of the paired t -test between the measurements by both DXA models in the subsample of 70 subjects, showed a mean difference different from zero These differences between DXA models support the decision to validate the equations according to the DXA model taken as reference, since the measurements between both models are not interchangeable.

The mean value of ASM estimated by the Kim's and Toselli's equations in the sample of subjects measured by DXA Lunar was Regarding the Kyle, Rangel-Peniche, Sergi and Yoshida equations, the mean value of ASM was Clearly, these results indicate that 2 equations underestimated ASM DXA , while 4 overestimated it Table 3.

Figure 1. Bland and Altman plots of the equations generated using DXA Lunar. Behavior of the mean difference against the mean of the measurements between the equations of Kim et al. Solid red lines indicate the mean difference. Solid blue lines indicate limits of agreement.

Solid black lines indicate the regression line. Dotted line indicates zero. ASM, appendicular skeletal muscle mass; MD, mean of the differences.

A Kim et al. B Toselli et al. Figure 2. Bland and Altman plots of the equations generated using DXA Hologic. Behavior of the mean difference against the mean of the measurements between the equations of Kyle et al.

A Kyle et al. B Rangel-Peniche et al. C Sergi et al. D Yoshida et al. This indicates that these equations do not significantly underestimate or overestimate as ASM increases. Having a homogeneous bias allows us to suggest a correction factor, which could correct the significant differences found in the paired t tests in these three equations.

Table 4. Comparison of the mean values of the estimated ASM and the ASM DXA. This wasn't possible for Kim's, Sergi's and Yoshida's equations. In these cases, the overestimation or underestimation of these equations as the ASM increases is significant, so they cannot be corrected. Considering the finding of homogeneous bias, correction factors were proposed by considering the mean difference between DXA and both equations.

The bias of each one of the equations was subtracted or added as following:. ASM ToselliCF , corrected Toselli's equation. ASM KyleCF , corrected Kyle's equation. ASM RangelCF , corrected Rangel-Peniche's equation. WC, waist circumference in cm. Weight in kilograms. Sex: 0 for women and 1 for men.

Age in years. The mean value of ASM estimated by the corrected Toselli's equation Toselli CF in the sample of subjects measured by DXA Lunar was On the other hand, the mean value of ASM estimated by the corrected Kyle's equation Kyle CF and the corrected Rangel-Peniche's equation Rangel CF in the sample of subjects measured by DXA Hologic was When these three corrected BIA equations were compared with their respective reference method, the mean differences were less than 0.

By carrying out the same tests applied previously paired t test and simple linear regression , and considering the criteria to determine agreement, it was possible to achieve agreement between the three corrected BIA equations and the ASM DXA.

This analysis gave us three corrected equations with a bias very close to zero, which is not statistically significant, and which maintained a homogeneous bias in the estimation. Table 5. Figure 3. Bland and Altman plots and simple linear regression of the selected equations applying the correction factors.

Behavior of the mean difference against the mean of the measurements between the corrected equations and their respective reference method. Solid red line indicates the mean difference.

Solid blue line indicates the limits of agreement. Solid black line indicates the regression line. ASM, appendicular skeletal muscle mass. MD, mean of the differences. A Corrected Toselli's equation Toselli CF.

B Corrected Kyle's equation Kyle CF. C Corrected Rangel-Peniche's equation Rangel CF. The purpose of this study was to validate some published BIA equations for estimating ASM. None of these BIA equations met the criteria for agreement in this sample.

However, the analysis of bias permitted to derive CFs, which, when applied to some equations, showed agreement with DXA. A valid corrected equation for this group of older adults can be a useful tool for epidemiological studies.

To the best of our knowledge, in Mexico, low muscle mass has only been assessed at the national level using calf circumference From our perspective, estimating it with accurate and practical tools, such as BIA equations could guarantee a better estimate of skeletal muscle, particularly ASM.

All the BIA equations selected for this study have already been tested in other populations previously, where they were discarded for its inaccuracy in certain populations due to the difference in age ranges 11 , 12 , 21 , nutritional status 20 , 50 , differences in body composition and anthropometry measurements related to ethnicity 18 , health status 19 , differences in functional status 14 , or BIA device employed For example, in other external validation studies 18 , 20 , Kim's equation was found to have the highest mean difference compared to DXA Lunar ASM estimations.

In these studies, authors discuss that it is most likely due to the fact that it was developed for an Asian population, but also because the authors used a multifrequency bioimpedance device, operating at a single frequency of Hz. It is already well recognized, that low frequencies predominantly measure extracellular water.

At higher frequencies, in contrast, cell membranes are permeable to current, so both intracellular and extracellular water are measured In this way, it is understood that multifrequency devices measure body composition in a slightly different way.

In our study, the Kim equation yielded the highest mean difference of all Both equations were generated in older Asian adults and using multi-frequency BIA devices, thus, we hypothesize that these two characteristics may have been an important factor contributing to bias in this sample as well.

Sergi's equation was generated in Caucasian subjects, and it only included older adults for its generation process. Even though their generation sample has very similar characteristics to ours, the equation had a very high bias, and like the others, the mean of the differences was significant.

It is important to remember that several studies have described the differences in body composition between different ethnic groups 42 , 52 , 53 , which could also have contributed to the bias of this equation as well.

This did not allow a correction factor to be proposed for these models. Kyle's equation was developed for Swiss adults in the age range of 22 to 94 years. Many studies have tried to validate it in external validation protocols. In almost all validation studies 11 , 12 , 14 , 16 , 18 , 19 , 21 — 23 , 50 , 54 , the equation has overestimated the ASM in different conditions, which the authors consider is due to the fact that it is not specific for a particular age group.

Therefore, this equation is usually discarded for use in certain populations. In our study, this equation overestimated 1. Toselli and Rangel-Peniche equations were the ones with the mean of the differences closest to zero In the case of the Rangel-Peniche equation, this must be since it was developed in a group of individuals of the same nationality as our sample.

Despite this, this equation does not meet the established criteria for agreement in this sample of older adults from the northwest of the same country. This confirms the nature of the equations to be specific for the population where it was generated and very similar populations.

In fact, another study by Rangel-Peniche et al. After adjusting for age, body weight, height, health status, estimated energy expenditure, and some demographic variables, ASM and the appendicular muscle mass index in older adults from central Mexico were significantly higher compared to the older adults from the northwest of Mexico.

This could be one reason why Rangel-Peniche's equation was not valid for our sample. In other studies, such as the one by Yu et al. In another study 19 , it overestimated approximately 0. In the study by Coëffier et al.

Due to these values, these studies have decided to rule out the use of this equation. On the other hand, this is the first study to externally validate Toselli's equation. This model, which includes waist circumference among the predictor variables, turned out to have a very low bias in our sample In their study, the authors discuss the relationship between waist circumference and ASM.

We believe that having taken this variable into account in this model and applying it to a sample with a high mean waist circumference, could be the reason why it had the smallest mean difference. According to our results, none of the selected equations was valid for older adults from the northwest of Mexico.

However, an important finding achieved when analyzing the bias of the equations, is that we realized that the Toselli, Rangel-Peniche and Kyle equations had a homogeneous bias.

This allowed them to be further improved to yield accurate data in this sample of older Mexican adults. By deriving a correction factor for Toselli's, Kyle's and Rangel-Peniche's equations, precise, accurate, and bias-free ASM estimates were obtained.

Importantly, this was possible after the analysis of the bias in this external validation study. This turned out to be a very useful strategy to use the existing equations in the literature, and thus not contribute to the development of more equations, which would have been generated unjustifiably and that, as mentioned in the systematic review by Beaudart et al.

This study has several advantages: to our knowledge, it is the first study to propose correction factors for BIA equations to estimate ASM, derived from a validation study with a large sample that included subjects of a wide nutritional range, age range, physically independent and without uncontrolled diseases that affected body composition.

Likewise, it is the first study that considers the DXA model in the validation process. Many external validation studies have treated the DXA model indistinctly, despite the differences that are already recognized in the literature 31 , 32 , 55 — In this study, in addition to considering these differences, we tested if the measurements taken by both DXA models were different in a subsample of subjects.

Once confirmed, we chose to separate the validation according to the DXA model: the equations generated with a model, were applied only in subjects measured with that same model.

This reduces the influence of the DXA model in the validation process, which could have been an important contributing bias factor. Another advantage is that this validation confirms that single frequency bioimpedance devices are a valid tool for ASM estimation compared to DXA.

These models are cheaper and more practical compared to others, and they can be a portable alternative for epidemiological studies. A final advantage that we find are the criteria established in this article to determine agreement between methods.

When assessing other validation studies, we noticed that some of them only carry out paired t -tests between methods, some use the pure error, or the Pearson or Lin coefficient. Some others are satisfied with only determining which was the lowest mean error of the selected equations.

We also notice that most studies do not analyze the bias distribution. We opted for the criteria mentioned in the Materials and Methods section, because, by adding paired t -tests and simple linear regression to the statistical methods, we address more than what is included in the Bland and Altman plot, testing agreement not only subjectively, but also objectively.

These steps should be fundamental in validating equations. Daniel Haigis, Silas Wagner, … the BaSAlt study team. Of particular importance for the functionality of older persons are the changes in their muscle mass and strength with age.

A reduction in strength and muscle mass destabilizes the work of muscles, resulting in the deterioration in general physical fitness, an increased risk of falls and limitations in the performance of everyday activities. Frailty is a multifaceted clinical syndrome considered in the three main aspects: physical, psychological and social.

Physical frailty is characterized by a cumulative reduction in physical functionality and a susceptibility to adverse effects in physical stress conditions, such as illness or hospitalization. The frailty phenotype increases the risk of falling ill with several acute and chronic diseases, and also the risk of death.

Frail elderly persons are exposed to sarcopenia, cachexia and wasting conditions [ 4 , 5 , 6 ]. Frailty can be assessed by various methods. The clinical assessment standard recommended by the Task Group of the International Conference of Frailty and Sarcopenia Research [ 3 ] is the highly confirmed frailty phenotype FP described by Fried et al.

Body mass loss in elderly persons is the principal component taken into account by most of the frailty state identifying methods. However, there are reports that obesity may be an important determinant of frailty in old age [ 7 , 8 ]. Therefore, the mass and quality of the skeletal muscles considering their role in body functioning seem to be more precise indicators of physical frailty than weight loss.

Among the many methods of assessing skeletal muscle mass in clinical practice and screening examinations, the bioelectrical impedance analysis BIA method is quite often used. The accuracy of the muscle mass estimates by the BIA method has not been clearly confirmed, therefore, the European Working Group on Sarcopenia in Older People EWGSOP2 recommends using of raw impedance measures along with the Sergi et al.

This method is regarded as highly promising owing to the possibility of analyzing, in a wide frequency spectrum, such impedance components as: resistance, reactance and phase angle as identifiers of the condition of tissues [ 2 , 10 , 11 ].

A particular importance, although not yet completely explained, is attributed to the impedance phase angle. It is regarded as indicator of muscle cells quality, as it depends on their number and size and the integrity of cell membranes [ 10 , 12 ].

The aim of our study was to assess the correlation between appendicular skeletal muscle mass and quality and the prevalence of frailty in elderly persons. We undertook an attempt to determine whether skeletal muscle mass and quality assessed on the basis of BIA measurements may be significant indicators of the risk of frailty syndrome prevalence in elderly persons.

In the years —, thousand-and-sixteen persons men and women aged 60—87 years Participants were assessed as subjectively healthy on the basis of declarations of good health, no difficulty walking, and no limitations in daily activities.

The study protocol was approved 18 February by the Senate Research Ethics Committee of the University School of Physical Education in Wroclaw. The research was carried out consistently with the Declaration of Helsinki recommendations in the Biokinetics Laboratory of the University School of Physical Education in Wrocław, certified according to the PN-EN ISO Quality Management System Certificate No.

The project was funded Grant No. N by the Ministry of Science and Higher Education. The participants were informed about the aim and methods of the study, the procedures used and the experimental risk.

All the persons who declared their participation in the study signed a document of voluntary and informed consent. Body height Ht and mass Wt were measured with an accuracy of, respectively, 0.

The analyser measures impedance with an accuracy of 0. Resistance, reactance and phase angle values were measured at the 50 kHz operating frequency of the 0.

The measurement was performed in standing position on a platform with built-in four electrodes 2 per foot and with two two-electrode handgrips enabling additional segmental readings separately for each limb and the trunk. Every day prior to the tests, proper repeatability of impedance measurement results was checked through two successive tests carried out on two volunteers.

The analyser software uses proprietary equations for estimating fat-free mass and intra- and extracellular water contents in the body, which due to commercial sensitivity are unavailable for publication. BIA measurements were carried out in the mornings, using the procedures indicated by the analyser manufacturer [ 13 ].

The subjects were asked not to eat, not to drink and not to undertake any physical activity at least 3 h before the test and to void the bladder immediately before the measurement.

The presence of an electronic implant e. In compliance with the latest EWGSOP2 recommendations [ 2 ], appendicular skeletal muscle mass ASMM was estimated using the predictive equation published by Sergi et al.

To minimize the differences stemming from inter-subject variability and considering the strong correlation between muscle mass and body size, the ASMM value was adjusted to the square of body height [ 2 ].

Hand grip strength HGS was measured with an accuracy of 1 kg by means of a JAMAR Sammons Preston Rolyan, USA hydraulic hand dynamometer with an adjustable handle set to position 2. The recommendations of the American Society of Hand Therapists ASHT were adopted [ 14 ].

The subjects were asked to perform two maximum grip strength tests for alternately the left hand and the right hand.

Each of the tests lasted 3—5 s and the inter-measurement interval was 15—20 s long. The highest value from all the tests was recorded as the HGS value. The number of seconds needed to get up from a seated position, walk 8 feet, turn and return to the seated position was measured total distance was 16 feet.

It was recommended to cover the distance as quickly as possible. Weekly physical activity PA was evaluated using the International Physical Activity Questionnaire IPAQ [ 17 ].

The participants would answer questions concerning the frequency and duration of their low-, moderate- and intensive-level physical activities.

Applying the criteria of the frailty phenotype model proposed by Fried et al. Frail cases were excluded from the study because the state of frailty was identified only in one person at least three frailty criteria were satisfied.

The normality of the distribution of all the variables was tested using the Shapiro—Wilk test. No normal distribution was confirmed for most of the variables, but the low asymmetry of the distributions and the possibility of comparing the results with the results reported by other authors induced us to use classical statistical description measures.

Non-parametric Kruskal—Wallis test was used to evaluate the differences between the gender and frailty groups. The differences between the non-frail persons and the pre-frail persons in the gender groups were verified using the U Mann—Whitney test.

No X c or R was included since they are trigonometric functions of the phase angle and are variables in the BIA equation for ASMM. The variables which constituted the frailty criterion HGS, Walking time and PA in our study were also not taken into account.

The statistical significance of individual regression coefficients was tested using Wald Chi-square statistics. The variables which had been found to be significantly correlated with pre-frailty in the univariate analyses were taken into account in the multiple logistic regression multivariable analysis.

The descriptive characteristics of the subjects and the differences between the non-frail persons and the pre-frail persons are presented in Table 1. One patient was frail and was excluded from the analysis.

As expected, in comparison with the women, the men were characterized by greater body mass Wt and height Ht , appendicular skeletal muscle mass ASMM and hand-grip strength HGS.

The women were characterized by higher resistance R and reactance X c values, a smaller phase angle PhA and a lower walking speed than the men.

The results of the tests of the significance of the differences between the women and the men, which had been expected and repeatedly reported in the literature, were not included in the table.

Besides the expected differences in age, mass and physical functionality parameters i. strength, walking speed and physical activity , it was found that the pre-frail persons had significantly lower phase angle values than the non-frail persons Table 1.

Differences in mean skeletal muscle mass and quality indices between groups of frailty status and age. ASMM appendicular skeletal muscle mass, Ht height, HGS hand-grip strength. H 3, , p : Kruskal—Wallis rank test results. Univariate logistic regression was used to preliminarily check the correlation between the selected variables without their mutual interaction and the probability of identifying pre-frailty.

A positive correlation of the pre-frail state with age and negative correlations with gender, phase angle and with the appendicular skeletal muscle mass and quality indices were confirmed Table 2. Gender and the functional quality of the skeletal muscles most strongly affected the odds of pre-frailty.

The variables which had been found to be significantly correlated with pre-frailty in the univariate analyses were taken into account in the multiple logistic regression Table 2. Insignificantly contributing variables were successively removed from the model.

Ultimately, the following pre-frailty predictors were found to be significant: age, appendicular skeletal muscle mass index, and muscle functional quality index Table 2.

A stronger correlation with frailty syndrome prevalence was observed for the skeletal muscle mass and quality indices than for age. The results of our investigations confirm that the adverse changes accompanying physical frailty can be observed quite early. We have shown that in the early stage of pre-frailty, a significant decrease in skeletal muscle mass both absolute and corrected for body size occurs which was not so obvious as we did not observe significant differences in body mass between the pre-frail women and the non-frail women.

We have observed that muscle quality determines the probability of pre-frailty regardless of gender and age. It is difficult to identify adverse muscle mass loss in older persons, especially women, since it can be masked by fatty tissue which enlarges with age.

On the basis of their review of literature on correlations between body composition and frailty in elderly persons. Reinders et al. established that obesity and large waist size pose a high risk of frailty [ 8 ]. Unfortunately, the correlations between the frailty syndrome and the particular body mass components mainly muscle mass and muscle fat infiltration remain unclear.

Jung et al. The state of frailty was linked with a decrease in fat-free mass corrected for squared height. In the non-frail persons, the fat-free mass loss amounted to 0. The authors assessed that in the persons with frailty, the risk of significant loss of skeletal muscle mass with age was almost three times higher than in the persons without frailty.

Recently, Ishii et al. showed that physical frailty and low muscle mass significantly contributed to disability among older community-living persons [ 20 ]. Many researchers emphasize that muscle strength is the best measure of changes in muscles and it is more closely linked with physical disability and functional limitations in instrumental activities of daily living IADL than muscle mass [ 2 , 20 , 21 , 22 , 23 ].

The rate of strength loss higher than the rate of muscle mass loss is due to changes in muscle composition and strength [ 24 , 25 ]. The muscle strength corrected for the skeletal muscle mass is a measure of functional muscle quality and is not governed by individual variability [ 2 , 15 ].

The interpretation of this index suggests that the generation of the same force at a greater ASMM means poorer quality of the skeletal muscles. This means that muscle quality can be an important indicator in identifying frailty. We found that the impedance phase angle, which is regarded as a measure of muscle cellular quality [ 10 , 12 ], was significantly lower in the pre-frail persons than in the non-frail persons: by 0.

No significant importance of PhA for prediction of pre-frailty suggests that the difference in muscle cellular quality observed by us was most probably due to aging there was a 3-year difference in age between the frailty phenotype groups.

This had also been observed in our previous study where in over year-old persons, the year difference between the age groups had generated a reduction in PhA by 0. Barbosa-Silva et al. in over year-old persons, they registered PhA values lower than the ones registered in persons 10 years younger: by nearly 0.

Moreover, the changes in PhA were found to increase with age in both gender groups [ 26 , 27 ]. In their study of patients admitted to geriatric wards, diagnosed as significantly frail and with a range of comorbidities, Slee et al.

They registered much lower phase angle values 4. Moreover, in the study carried out by Slee et al. Recently, Mullie et al.

They demonstrated that a smaller PhA would increase the odds of a higher degree of frailty and decline in physical fitness and mortality. They discovered significant correlations of PhA with the Fried score, the Essential Frailty Toolset score and the Short Physical Performance Battery SPPB score [ 29 ].

Cesari et al. Williams et al. reported that in cancer patients, skeletal muscle density was more closely correlated with the prevalence of frailty and pre-frailty than muscle mass [ 31 ]. In our study, the corrected mass of skeletal muscles and the index of their functional quality were strongly correlated with the probability of pre-frailty state, with a weaker influence of age, regardless of the gender of the subjects.

These results indicate that both BIA-assessed variables can be used could be used as additional frailty identifiers. Considering the above, it would be necessary to determine reference and normalized cut-off points for the skeletal muscle mass and quality indices, which could remove the ambiguities connected with body mass decline assessment in frailty identification.

This would have important implications especially for screening, where weaker tools are used to assess frailty. The availability, low cost and quickness of measurement are significant advantages of assessing ASMM by means of the bioelectrical impedance method. The limitation of our study is the absence of frailty in the persons subjected to Fried frailty phenotype assessment.

Thus, our results are applicable only to the assessment of the probability of pre-frailty. Second, no evaluation of parameters which can have a bearing on the relationship between muscle mass, muscle function and frailty, such as protein consumption, was carried out. Also, our use of BIA, instead of the reference method, to estimate ASMM values, can be debatable, but we wanted to indicate the potential of BIA for the routine monitoring of aging, as an alternative to the often unaffordable reference methods.

Moreover, the method and the ASMM predicting equation used by us have been taken into account in the latest recommendations of the European Working Group on Sarcopenia in Older People [ 2 ]. The strong point of our study is the large number of ethnically not diverse, independently living subjects.

We think that our project broadens the knowledge of frailty in elderly persons and it can be an important basis for orienting further research on improving frailty identification methods.

Quick diagnosis of pre-frailty and frailty contributes to the more effective prevention of the frailty syndrome and ensures successful aging. This study has indicated that the easily available and inexpensive method of BIA can be used to preventively monitor changes not only in the mass of skeletal muscles, but also in their quality, which is particularly important in the case of pre-frail older persons.

The presented results of the study confirm that the skeletal muscle quantity and quality indices based on BIA estimates can facilitate the assessment of the state of frailty in routine geriatric care. World Health Organization Global strategy and action plan on ageing and health, Geneva.

Accessed 29 Mar Cruz-Jentoft AJ, Baeyens JP, Bauer JM et al Sarcopenia: revised European consensus on definition and diagnosis. Age Ageing — Article PubMed Google Scholar. Dent E, Morley JE, Cruz-Jentoft AJ et al Physical frailty: ICFSR international clinical practice guidelines for identification and management.

J Nutr Health Aging — Article CAS PubMed PubMed Central Google Scholar.

In the years —, thousand-and-sixteen persons men and women aged 60—87 years Participants were assessed as subjectively healthy on the basis of declarations of good health, no difficulty walking, and no limitations in daily activities.

The study protocol was approved 18 February by the Senate Research Ethics Committee of the University School of Physical Education in Wroclaw. The research was carried out consistently with the Declaration of Helsinki recommendations in the Biokinetics Laboratory of the University School of Physical Education in Wrocław, certified according to the PN-EN ISO Quality Management System Certificate No.

The project was funded Grant No. N by the Ministry of Science and Higher Education. The participants were informed about the aim and methods of the study, the procedures used and the experimental risk.

All the persons who declared their participation in the study signed a document of voluntary and informed consent. Body height Ht and mass Wt were measured with an accuracy of, respectively, 0.

The analyser measures impedance with an accuracy of 0. Resistance, reactance and phase angle values were measured at the 50 kHz operating frequency of the 0. The measurement was performed in standing position on a platform with built-in four electrodes 2 per foot and with two two-electrode handgrips enabling additional segmental readings separately for each limb and the trunk.

Every day prior to the tests, proper repeatability of impedance measurement results was checked through two successive tests carried out on two volunteers. The analyser software uses proprietary equations for estimating fat-free mass and intra- and extracellular water contents in the body, which due to commercial sensitivity are unavailable for publication.

BIA measurements were carried out in the mornings, using the procedures indicated by the analyser manufacturer [ 13 ].

The subjects were asked not to eat, not to drink and not to undertake any physical activity at least 3 h before the test and to void the bladder immediately before the measurement.

The presence of an electronic implant e. In compliance with the latest EWGSOP2 recommendations [ 2 ], appendicular skeletal muscle mass ASMM was estimated using the predictive equation published by Sergi et al.

To minimize the differences stemming from inter-subject variability and considering the strong correlation between muscle mass and body size, the ASMM value was adjusted to the square of body height [ 2 ]. Hand grip strength HGS was measured with an accuracy of 1 kg by means of a JAMAR Sammons Preston Rolyan, USA hydraulic hand dynamometer with an adjustable handle set to position 2.

The recommendations of the American Society of Hand Therapists ASHT were adopted [ 14 ]. The subjects were asked to perform two maximum grip strength tests for alternately the left hand and the right hand. Each of the tests lasted 3—5 s and the inter-measurement interval was 15—20 s long.

The highest value from all the tests was recorded as the HGS value. The number of seconds needed to get up from a seated position, walk 8 feet, turn and return to the seated position was measured total distance was 16 feet. It was recommended to cover the distance as quickly as possible.

Weekly physical activity PA was evaluated using the International Physical Activity Questionnaire IPAQ [ 17 ]. The participants would answer questions concerning the frequency and duration of their low-, moderate- and intensive-level physical activities.

Applying the criteria of the frailty phenotype model proposed by Fried et al. Frail cases were excluded from the study because the state of frailty was identified only in one person at least three frailty criteria were satisfied.

The normality of the distribution of all the variables was tested using the Shapiro—Wilk test. No normal distribution was confirmed for most of the variables, but the low asymmetry of the distributions and the possibility of comparing the results with the results reported by other authors induced us to use classical statistical description measures.

Non-parametric Kruskal—Wallis test was used to evaluate the differences between the gender and frailty groups. The differences between the non-frail persons and the pre-frail persons in the gender groups were verified using the U Mann—Whitney test.

No X c or R was included since they are trigonometric functions of the phase angle and are variables in the BIA equation for ASMM. The variables which constituted the frailty criterion HGS, Walking time and PA in our study were also not taken into account.

The statistical significance of individual regression coefficients was tested using Wald Chi-square statistics. The variables which had been found to be significantly correlated with pre-frailty in the univariate analyses were taken into account in the multiple logistic regression multivariable analysis.

The descriptive characteristics of the subjects and the differences between the non-frail persons and the pre-frail persons are presented in Table 1.

One patient was frail and was excluded from the analysis. As expected, in comparison with the women, the men were characterized by greater body mass Wt and height Ht , appendicular skeletal muscle mass ASMM and hand-grip strength HGS. The women were characterized by higher resistance R and reactance X c values, a smaller phase angle PhA and a lower walking speed than the men.

The results of the tests of the significance of the differences between the women and the men, which had been expected and repeatedly reported in the literature, were not included in the table.

Besides the expected differences in age, mass and physical functionality parameters i. strength, walking speed and physical activity , it was found that the pre-frail persons had significantly lower phase angle values than the non-frail persons Table 1.

Differences in mean skeletal muscle mass and quality indices between groups of frailty status and age. ASMM appendicular skeletal muscle mass, Ht height, HGS hand-grip strength. H 3, , p : Kruskal—Wallis rank test results.

Univariate logistic regression was used to preliminarily check the correlation between the selected variables without their mutual interaction and the probability of identifying pre-frailty. A positive correlation of the pre-frail state with age and negative correlations with gender, phase angle and with the appendicular skeletal muscle mass and quality indices were confirmed Table 2.

Gender and the functional quality of the skeletal muscles most strongly affected the odds of pre-frailty. The variables which had been found to be significantly correlated with pre-frailty in the univariate analyses were taken into account in the multiple logistic regression Table 2.

Insignificantly contributing variables were successively removed from the model. Ultimately, the following pre-frailty predictors were found to be significant: age, appendicular skeletal muscle mass index, and muscle functional quality index Table 2. A stronger correlation with frailty syndrome prevalence was observed for the skeletal muscle mass and quality indices than for age.

The results of our investigations confirm that the adverse changes accompanying physical frailty can be observed quite early. We have shown that in the early stage of pre-frailty, a significant decrease in skeletal muscle mass both absolute and corrected for body size occurs which was not so obvious as we did not observe significant differences in body mass between the pre-frail women and the non-frail women.

We have observed that muscle quality determines the probability of pre-frailty regardless of gender and age. It is difficult to identify adverse muscle mass loss in older persons, especially women, since it can be masked by fatty tissue which enlarges with age.

On the basis of their review of literature on correlations between body composition and frailty in elderly persons. Reinders et al. established that obesity and large waist size pose a high risk of frailty [ 8 ]. Unfortunately, the correlations between the frailty syndrome and the particular body mass components mainly muscle mass and muscle fat infiltration remain unclear.

Jung et al. The state of frailty was linked with a decrease in fat-free mass corrected for squared height. In the non-frail persons, the fat-free mass loss amounted to 0. The authors assessed that in the persons with frailty, the risk of significant loss of skeletal muscle mass with age was almost three times higher than in the persons without frailty.

Recently, Ishii et al. showed that physical frailty and low muscle mass significantly contributed to disability among older community-living persons [ 20 ].

Many researchers emphasize that muscle strength is the best measure of changes in muscles and it is more closely linked with physical disability and functional limitations in instrumental activities of daily living IADL than muscle mass [ 2 , 20 , 21 , 22 , 23 ].

The rate of strength loss higher than the rate of muscle mass loss is due to changes in muscle composition and strength [ 24 , 25 ]. The muscle strength corrected for the skeletal muscle mass is a measure of functional muscle quality and is not governed by individual variability [ 2 , 15 ].

The interpretation of this index suggests that the generation of the same force at a greater ASMM means poorer quality of the skeletal muscles. This means that muscle quality can be an important indicator in identifying frailty. We found that the impedance phase angle, which is regarded as a measure of muscle cellular quality [ 10 , 12 ], was significantly lower in the pre-frail persons than in the non-frail persons: by 0.

No significant importance of PhA for prediction of pre-frailty suggests that the difference in muscle cellular quality observed by us was most probably due to aging there was a 3-year difference in age between the frailty phenotype groups. This had also been observed in our previous study where in over year-old persons, the year difference between the age groups had generated a reduction in PhA by 0.

Barbosa-Silva et al. in over year-old persons, they registered PhA values lower than the ones registered in persons 10 years younger: by nearly 0. Moreover, the changes in PhA were found to increase with age in both gender groups [ 26 , 27 ]. In their study of patients admitted to geriatric wards, diagnosed as significantly frail and with a range of comorbidities, Slee et al.

They registered much lower phase angle values 4. Moreover, in the study carried out by Slee et al. Recently, Mullie et al. They demonstrated that a smaller PhA would increase the odds of a higher degree of frailty and decline in physical fitness and mortality.

They discovered significant correlations of PhA with the Fried score, the Essential Frailty Toolset score and the Short Physical Performance Battery SPPB score [ 29 ]. Cesari et al. Williams et al. reported that in cancer patients, skeletal muscle density was more closely correlated with the prevalence of frailty and pre-frailty than muscle mass [ 31 ].

In our study, the corrected mass of skeletal muscles and the index of their functional quality were strongly correlated with the probability of pre-frailty state, with a weaker influence of age, regardless of the gender of the subjects.

These results indicate that both BIA-assessed variables can be used could be used as additional frailty identifiers. Considering the above, it would be necessary to determine reference and normalized cut-off points for the skeletal muscle mass and quality indices, which could remove the ambiguities connected with body mass decline assessment in frailty identification.

This would have important implications especially for screening, where weaker tools are used to assess frailty. The availability, low cost and quickness of measurement are significant advantages of assessing ASMM by means of the bioelectrical impedance method.

The limitation of our study is the absence of frailty in the persons subjected to Fried frailty phenotype assessment. Thus, our results are applicable only to the assessment of the probability of pre-frailty.

Second, no evaluation of parameters which can have a bearing on the relationship between muscle mass, muscle function and frailty, such as protein consumption, was carried out.

Also, our use of BIA, instead of the reference method, to estimate ASMM values, can be debatable, but we wanted to indicate the potential of BIA for the routine monitoring of aging, as an alternative to the often unaffordable reference methods.

Moreover, the method and the ASMM predicting equation used by us have been taken into account in the latest recommendations of the European Working Group on Sarcopenia in Older People [ 2 ]. The strong point of our study is the large number of ethnically not diverse, independently living subjects.

We think that our project broadens the knowledge of frailty in elderly persons and it can be an important basis for orienting further research on improving frailty identification methods.

Quick diagnosis of pre-frailty and frailty contributes to the more effective prevention of the frailty syndrome and ensures successful aging.

This study has indicated that the easily available and inexpensive method of BIA can be used to preventively monitor changes not only in the mass of skeletal muscles, but also in their quality, which is particularly important in the case of pre-frail older persons.

The presented results of the study confirm that the skeletal muscle quantity and quality indices based on BIA estimates can facilitate the assessment of the state of frailty in routine geriatric care.

World Health Organization Global strategy and action plan on ageing and health, Geneva. Accessed 29 Mar Cruz-Jentoft AJ, Baeyens JP, Bauer JM et al Sarcopenia: revised European consensus on definition and diagnosis.

Age Ageing — Article PubMed Google Scholar. Dent E, Morley JE, Cruz-Jentoft AJ et al Physical frailty: ICFSR international clinical practice guidelines for identification and management. J Nutr Health Aging — Article CAS PubMed PubMed Central Google Scholar.

Fried LP, Tangen CM, Walston J et al Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci M—M Article CAS PubMed Google Scholar.

Morley JE, Vellas B, van Kan GA et al Frailty consensus: a call to action. J Am Med Dir Assoc — Article PubMed PubMed Central Google Scholar. Rockwood K, Mitnitski A Frailty in relation to the accumulation of deficits.

J Gerontol A Biol Sci Med Sci — Ferriolli E, Pessanha FPADS, Moreira VG et al Body composition and frailty profiles in Brazilian older people: frailty in Brazilian older people study-FIBRA-BR. Arch Gerontol Geriatr — Reinders I, Visser M, Schaap L Body weight and body composition in old age and their relationship with frailty.

Curr Opin Clin Nutr Metab Care — Sergi G, De Rui M, Veronese N et al Assessing appendicular skeletal muscle mass with bioelectrical impedance analysis in free-living Caucasian older adults. Clin Nutr — Heymsfield SB, Gonzalez MC, Lu J et al Skeletal muscle mass and quality: evolution of modern measurement concepts in the context of sarcopenia.

Proc Nutr Soc — Yamada Y, Buehring B, Krueger D et al Electrical properties assessed by bioelectrical impedance spectroscopy as biomarkers of age-related loss of skeletal muscle quantity and quality. Norman K, Stobäus N, Pirlich M et al Bioelectrical impedance phase angle and impedance vector analysis—clinical relevance and applicability of impedance parameters.

Tanita Corp Multi-frequency body composition analyser MCMA. First, it was a retrospective study. Second, edema was not assessed. Fluid retention can only be examined in ascites. Third, the equipment used for the BIA differs between the two facilities in this study.

When evaluating SMI in pathological conditions accompanied by fluid retention such as liver cirrhosis, it is important to select an appropriate BIA measuring device and also evaluate water content by site. We will investigate this and report it in an upcoming article.

Due to the characteristics of the measurement principle of BIA, overestimation of muscle mass is predicted to be affected by fluid retention. It is necessary to assess edema before determining muscle mass using the BIA method. We thank the Statista Corporation for assistance with the statistical analyses.

We also thank Robert E. Brandt, Founder, CEO, and CME, of MedEd Japan, for editing and formatting the manuscript. This study protocol was reviewed and approved by the Institutional Review Board Ethics Committees of Tokushukai Medical Group Number: TGE and Kitasato University School of Medicine Number: C This study was a retrospective observational study.

Informed consent was obtained from all individual participants included in the study by the opt-out method of our hospital Website, which was approved by the Research Ethics Committees of Tokushukai Medical Group and Kitasato University School of Medicine.

Nahoko Kikuchi, Haruki Uojima, Hisashi Hidaka, Shuichiro Iwasaki, Naohisa Wada, Kousuke Kubota, Takahide Nakazawa, Akitaka Shibuya, Makoto Kako, Teruko Sato, and Chika Kusano contributed equally to this work; Nahoko Kikuchi and Haruki Uojima collected and analyzed the data; Haruki Uojima drafted the manuscript; Hisashi Hidaka and Makoto Kako designed and supervised the study; Shuichiro Iwasaki, Naohisa Wada, Kousuke Kubota, Takahide Nakazawa, Akitaka Shibuya, Teruko Sato, and Chika Kusano offered technical or material support.

The technical appendix, statistical code, and dataset are available from the corresponding author email: kiruha kitasato-u. All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.

Sign In or Create an Account. Search Dropdown Menu. header search search input Search input auto suggest. filter your search All Content All Journals Annals of Nutrition and Metabolism.

Advanced Search. Skip Nav Destination Close navigation menu Article navigation. Volume 78, Issue 6. Materials and Methods. Statement of Ethics. Conflict of Interest Statement. Funding Sources. Author Contributions. Data Availability Statement. Article Navigation. Research Articles August 18 Evaluation of Skeletal Muscle Mass in Patients with Chronic Liver Disease Shows Different Results Based on Bioelectric Impedance Analysis and Computed Tomography Subject Area: Endocrinology , Further Areas , Nutrition and Dietetics , Public Health.

Nahoko Kikuchi ; Nahoko Kikuchi. a Department of Nutrition, Kitasato University Hospital, Sagamihara, Japan. k kitasato-u. This Site. Google Scholar. Haruki Uojima ; Haruki Uojima. b Department of Gastroenterology, Internal Medicine, Kitasato University School of Medicine, Sagamihara, Japan.

c Department of Gastroenterology, Shonan Kamakura General Hospital, Kamakura, Japan. Hisashi Hidaka Hisashi Hidaka. Shuichiro Iwasaki ; Shuichiro Iwasaki. Naohisa Wada Naohisa Wada. Kousuke Kubota ; Kousuke Kubota. Takahide Nakazawa ; Takahide Nakazawa. Akitaka Shibuya ; Akitaka Shibuya. Makoto Kako ; Makoto Kako.

Akira Take ; Akira Take. d Department of Microbiology, Kitasato University School of Medicine, Sagamihara, Japan. Yoshihiko Sakaguchi ; Yoshihiko Sakaguchi. Teruko Sato Teruko Sato.

Chika Kusano Chika Kusano. Ann Nutr Metab 78 6 : — Article history Received:. Cite Icon Cite. toolbar search Search Dropdown Menu. toolbar search search input Search input auto suggest.

View large Download slide. Table 1. Baseline clinical characteristics. View large. View Large. Table 2. Univariate and multivariate analyses of factors affecting mismatch between SMI using BIA and CT. The authors declare that they have no conflicts of interest.

There were no funding sources. Search ADS. Prospective study for an independent predictor of prognosis in liver cirrhosis based on the new sarcopenia criteria produced by the Japan Society of Hepatology.

Effect of exercise therapy combined with branched-chain amino acid supplementation on muscle strength in elderly women after total hip arthroplasty: a randomized controlled trial. Effects of enriched branched-chain amino acid supplementation on sarcopenia. Sarcopenia: European consensus on definition and diagnosis: report of the European Working Group on Sarcopenia in Older People.

Sarcopenia in Asia: consensus report of the Asian Working Group for Sarcopenia. Japan Society of Hepatology guidelines for sarcopenia in liver disease 1st edition : recommendation from the working group for creation of sarcopenia assessment criteria. The value of L3 skeletal muscle index in evaluating preoperative nutritional risk and long-term prognosis in colorectal cancer patients.

Di Vincenzo. Bioelectrical impedance analysis BIA -derived phase angle in sarcopenia: a systematic review. Portal hypertensive bleeding in cirrhosis: risk stratification, diagnosis, and management: practice guidance by the American Association for the study of liver diseases.

Bioelectrical impedance analysis for the assessment of sarcopenia in patients with cancer: a systematic review. Biological indexes considered in the derivation of the bioelectrical impedance analysis.

Phase angle from bioelectrical impedance analysis is a useful indicator of muscle quality. Phase Angle from bioelectrical impedance for the assessment of sarcopenia in cirrhosis with or without ascites.

Prognostic value of sarcopenia in patients with liver cirrhosis: a systematic review and meta-analysis. Skeletal muscle mass influences tolerability and prognosis in hepatocellular carcinoma patients treated with lenvatinib.

Sarcopenia as a predictor of prognosis in patients following hepatectomy for hepatocellular carcinoma. Evaluation and prognosis of sarcopenia using impedance analysis in patients with liver cirrhosis.

Reference values and age differences in body composition of community-dwelling older Japanese men and women: a pooled analysis of four cohort studies.

Percentage of total body fat as estimated by three automatic bioelectrical impedance analyzers. No differences in the body fat after violating core bioelectrical impedance measurement assumptions.

Published by S. Karger AG, Basel. This article is licensed under the Creative Commons Attribution 4. Usage, derivative works and distribution are permitted provided that proper credit is given to the author and the original publisher. Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication.

However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions.

Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor s.

The publisher and the editor s disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.

This work is licensed under a Creative Commons Attribution 4. View Metrics. Email alerts Online First Alert.

Latest Issue Alert. Citing articles via Google Scholar. CrossRef 1. Latest Most Read Most Cited Eicosanoids and oxylipin signature in HH patients are similar to DIOS patients but are impacted by dietary iron absorption. Consumption Pattern of Tea Is Associated with Serum Ferritin Levels of Women of Childbearing Age in Nandi County, Kenya: A Cross-Sectional Study.

Suggested Reading Limitations of Fat-Free Mass for the Assessment of Muscle Mass in Obesity Obes Facts May, Impact of Low Skeletal Muscle Mass on the Prognosis of Patients with Head and Neck Cancer Treated Nonsurgically ORL January, The Biagram Vector: A Graphical Relation between Reactance and Phase Angle Measured by Bioelectrical Analysis in Infants Ann Nutr Metab April, Relationship between the Phase Angle and Volume of Tumours in Patients with Lung Cancer Ann Nutr Metab December, Online ISSN Print ISSN Karger International S.

Karger AG P. O Box, CH Basel Switzerland Allschwilerstrasse 10, CH Basel. Facebook LinkedIn X YouTube WeChat Experience Blog.

Privacy Policy Terms of Use Imprint Cookies © S. This Feature Is Available To Subscribers Only Sign In or Create an Account.