DEVELOPMENT AND VALIDATION OF EQUATION FOR PREDICTING BODY COMPOSITION IN ELDERLY WOMEN

DESARROLLO Y VALIDACIÓN DE ECUACIONES PARA LA PREDICCIÓN DE LA COMPOSICIÓN CORPORAL EN ANCIANAS

ABSTRACT

Introduction:

Body composition assessment (BCA) using anthropometric measurements (AM) is used to monitor the nutritional and health status of the elderly. As predictor variables, MAs must be valid, practical, and quick, as they favor adherence and avoid possible resistance and embarrassment on the part of those being assessed and being minimally invasive.

Objective:

To develop and validate equations using accessible and minimally invasive anthropometric measurements for BCA in elderly women.

Methods:

100 women (68.1±6.15 years) were randomly assigned to two groups: validation (n=40; 68.1±6.15 years); and estimation (n=60; 68.4±6.70 years). DXA was selected as the criterion measure, and MAs (body mass, height, skinfolds, circumferences) were selected as predictor variables. Means were compared using the paired Student's t-test; correlations were verified using Pearson's r-test; equations using Multiple Linear Regression. The level of agreement between the groups’ results was checked using the Bland-Altman technique.

Results:

Two equations developed and tested (E3 and E4) met the validation criteria as they showed adequate correlation coefficients (E3: r=0.73; E4: r=0.70), low constant errors (E3: EC= −0.56; E4: EC=-0.90), total error (E3: ET=3.22; E4: ET=3.06) lower than the Standard Error of Estimate (E3: EPE=3.24; E4: EPE=3.21), indicating no statistically significant difference between the two BCA techniques observed (p>0.05). The Bland-Altman technique showed good agreement between the results of the two techniques.

Conclusion:

Two were validated: E3 (%Gdxa= −41.556 + 4.041(BMI) + 0.165(DcCox) − 0.440(CircCox) + 0.269(CircQuad) − 0.053(BMI)²); and E4 (%GdxaE4= 15.329 + 1.044(BMI) −1.055(CircAbra) + 0.282(CircQuad) + 0.164(DcCox) − 0.262(CircCox)). Notably, the small number of measurements were located in areas of the body that are easily accessible and have little body exposure, which minimizes possible embarrassment and favors adherence by the elderly. Level of Evidence IV; Correlational study to build a predictive equation.

Keywords:
Body Fat Distribution; Aging; X-Rays; Anthropometry

RESUMEN

Introducción:

La evaluación de la composición corporal (ECC) mediante medidas antropométricas (MA) se utiliza para monitorizar el estado nutricional y de salud de las personas mayores. Las MA, como variables predictoras, deben ser válidas, prácticas y rápidas, ya que favorecen la adherencia y evitan posibles resistencias y vergüenzas por parte de los evaluados, además de ser mínimamente invasivas.

Objetivo:

Desarrollar y validar ecuaciones utilizando medidas antropométricas de fácil acceso y mínimamente invasivas para la ECC en mujeres de edad avanzada.

Métodos:

100 mujeres (68,1±6,15 años) asignadas aleatoriamente a dos grupos: validación (n=40; 68,1±6,15 años); estimación (n=60; 68,4±6,70 años). Se seleccionó la DXA como medida criterio, y las MA (masa corporal, estatura, pliegues cutáneos, circunferencias) como variables predictoras. Las medias se compararon mediante la prueba t de Student emparejada; las correlaciones se comprobaron mediante la prueba r de Pearson; las ecuaciones mediante Regresión Lineal Múltiple. El nivel de concordancia entre los resultados de los grupos se comprobó mediante la técnica de Bland-Altman.

Resultados:

Dos ecuaciones desarrolladas y probadas (E3 y E4) cumplieron los criterios de validación, ya que mostraron coeficientes de correlación adecuados (E3: r=0,73; E4: r=0,70), errores constantes bajos (E3: EC= −0,56; E4: EC=-0,90), error total (E3: ET=3,22; E4: ET=3,06) inferior al error estándar de estimación (E3: EPE=3,24; E4: EPE=3,21), lo que indica que no se observaron diferencias estadísticamente significativas entre las dos técnicas de ECC (p>0,05). La técnica de Bland-Altman mostró una buena concordancia entre los resultados de las dos técnicas.

Conclusión:

Se validaron dos: E3 (%Gdxa= −41,556 + 4,041(IMC) + 0,165(DcCox) − 0,440(CircCox) + 0,269(CircQuad) − 0,053(IMC)²); y E4 (%GdxaE4= 15,329 + 1,044(IMC) −1,055(CircAbra) + 0,282(CircQuad) + 0,164(DcCox) − 0,262(CircCox)). Es de destacar que el reducido número de mediciones se localizaron en zonas del cuerpo de fácil acceso y con poca exposición corporal, lo que minimiza posibles situaciones embarazosas y favorece la adherencia por parte de los ancianos. Nivel de Evidencia IV; Estudio correlacional para construir una ecuación predictiva.

Descriptores:
Distribución de la Grasa Corporal; Envejecimiento; Rayos X; Antropometría

RESUMO

Introdução:

A avaliação da composição corporal (ACC) com auxílio das medidas antropométricas (MA) é utilizado no acompanhamento do estado nutricional e de saúde das populações idosas. As MAs como variáveis preditoras, devem ser válidas, práticas e rápidas, pois favorecem a adesão e evitam possíveis resistências e constrangimento das avaliadas, além de minimamente invasivas.

Objetivo:

Desenvolver e validar equações utilizando medidas antropométricas de fácil acesso e minimamente invasivas, para ACC de idosas.

Métodos:

100 mulheres (68,1±6,15 anos) distribuídas aleatoriamente por dois grupos: validação (n=40; 68,1±6,15 anos); estimativa (n=60; 68,4±6,70 anos). Como medida critério selecionou-se a DXA, como variáveis preditoras as MAs (massa corporal, estatura, dobras cutâneas, circunferências). As médias foram comparadas com o teste t de Student pareado; as correlações verificadas com o teste r Pearson; as equações com a Regressão Linear Múltipla. O nível de concordância entre os resultados dos grupos foi verificado com a técnica Bland-Altman.

Resultados:

Duas equações desenvolvidas e testadas (E3 e E4) atenderam aos critérios de validação, pois, apresentaram coeficientes de correlação adequados (E3: r=0,73; E4: r=0,70), erros constantes baixos (E3: EC= −0,56; E4: EC=-0,90), erro total (E3: ET=3,22; E4: ET=3,06) menores que o Erro Padrão de Estimativa (E3: EPE=3,24; E4: EPE=3,21), indicando não existir diferença estatisticamente significativa entre as duas técnicas de ACC observadas (p>0,05). A técnica Bland-Altman demonstrou boa concordância entre os resultados das duas técnicas.

Conclusão:

Duas equações foram validadas: E3 (%Gdxa= −41,556 + 4,041(IMC) + 0,165(DcCox) – 0,440(CircCox) + 0,269(CircQuad) – 0,053(IMC)²); e E4 (%GdxaE4= 15,329 + 1,044(IMC) −1,055(CircAbra) + 0,282(CircQuad) + 0,164(DcCox) – 0,262(CircCox)). Destaca-se o fato do reduzido número de medidas situarem-se em áreas corporais de fácil acesso e de pouca exposição corporal, que minimizam possíveis constrangimentos e favorecem a adesão das idosas. Nível de Evidência IV; Estudo correlacional para construção de equação preditiva.

Descritores:
Distribuição da Gordura Corporal; Envelhecimento; Raios-X; Antropometria

INTRODUCTION

Although the timing, speed, and severity of the changes in body composition (BC) during the aging process may differ between people, they will occur in all of them, whether healthy or athletes.¹

Changes in BC throughout life are strictly related to health status. The increase in fat mass and the decrease in muscle mass and strength, a phenomenon known as sarcopenia, are typical of aging and are associated with an increase in age-related pathologies, a decline in functionality, an increase in the risk of falls, hospitalizations, morbidities, and mortality. For these reasons, the study of BC is so important because it allows changes to be monitored, preventing pathologies and age-related limitations.² Because it allows changes in fat mass (FM) and lean mass (LM) to be monitored, the assessment of BC in the elderly is an important resource for monitoring nutritional status, controlling various metabolic diseases, and maintaining health and quality of life.¹,³ The above applies to both sexes; however, since elderly women have less absolute muscle mass when compared to men of the same age, for females, BC assessment can bring additional advantages in the sense of predicting future functional limitations.

Several laboratory methods are used as the ‘gold standard’ and as a yardstick for other tests to assess BC (Dual Energy X-ray Absorptiometry, Computed Tomography, and Hydrostatic Weighing, among others). However, although they have greater validity and accuracy, in addition to only allowing individual approaches, these techniques are expensive, require high levels of cooperation from those being assessed, and require specialized technicians and laboratory environments.²,⁴–⁶

Due to its validity and characteristics, Dual Energy X-ray Absorptiometry (DXA) has been increasingly used as a criterion measure for assessing BC.⁷,⁸ However, due to the characteristics described above, this test is expensive for most people and makes it difficult to use in population studies.⁹ Given the above, the search for and need for valid, simpler, and less expensive tests has become essential for clinical practice and studies with large populations.⁴

In this direction, anthropometric measurements (AM) such as body mass, height, skinfolds, and circumferences have been increasingly used to construct equations to predict BC in different populations.¹⁰–¹² Despite the possibility of using such equations to assess BC, it is important to consider that their generalized use is neither automatic nor advisable. In other words, the predictive equations must respect the characteristics of the population studied, considering, among other things, the different age groups, ethnicities, cultural habits, genders, and levels of physical fitness. The widespread use of predictive equations can make their validity unfeasible.¹³

When using predictive equations to assess BC in elderly Brazilian women, the number of validated equations available is still insufficient. Most validation studies of this type carried out in Brazil used people from the South,¹⁴–¹⁶ and Southeast¹⁷,¹⁸ regions of the country as their samples. Therefore, during the cross-validation process, it is not uncommon to see significant differences between the results of the CCA in different regions.

Given the above, the main objective of this study was to test, develop, and validate anthropometric equations capable of assessing the BC of elderly women living in the municipality of Maceió/AL. Secondly, the least invasive anthropometric measurements were evaluated and selected for the equations.

MATERIALS AND METHODS

The study sample consisted of 100 elderly women (68.1±6.15 years old) selected from leisure physical activity programs offered by 22 institutions for the elderly (ITI) in the municipality of Maceió. All the participants signed an informed consent form (ICF) by the recommendations of Resolution 466/12 of the Brazilian National Health Council for research involving human beings. All project stages were approved by the Human Research Ethics Committee of the Universidade Federal de Alagoas (protocol No. 23065.020769/2009-31).

The elderly women who made up the sample were randomly assigned to two different groups: the validation group - GV (n=40; mean/DP= 68.1±6.15 years) and the estimation group - GE (n=60; mean/DP= 68.4±6.70 years).

Data collection

Anthropometric measurements were taken at various times in the 22 institutions investigated. A LANGE adipometer (Beta Technology, Santa Cruz, CA, USA) was used to measure skinfolds (SC): thigh, leg, abdomen, triceps, biceps, mid-axillary, and supra-iliac; for circumference measurements (CM): brachial, forearm, abdominal, waist, hip, thigh and leg circumferences, a steel tape measure five millimeters wide and two meters long (Sanny - American Medical do Brasil, Ltda, São Bernardo do Campo/SP, Brazil) was used; height was measured with Seca brand portable stadiometers (Baystate Scale & Systems, USA) and body mass with the aid of Tanita brand portable digital scales (Mod: UM 080W Scale Plus Body Fat Monitor with Body Water - American Medical do Brasil, Ltda, São Bernardo do Campo/SP, Brazil).

Measurements of WC and HC were taken as recommended by Pollock and Jackson.¹⁹ BMI (ratio of body mass to height) was obtained in kg/m².

A Lunar Prodigy Advanced series DPX-YZB/2099 scanner (Madison, WI) was used to measure the DXA, and the BC variables were calculated using the software provided by the manufacturer (version 3.0). The equipment was calibrated daily by a specialized technician from an imaging clinic in Maceió/AL, following the manufacturer's recommendations.

Statistical Analysis

The sample was characterized using descriptive statistics, using the arithmetic mean (X) as a measure of central tendency and the standard deviation (SD) as the main measure of dispersion. The normality of data distribution was assessed using the Kolmogorov-Smirnov test, and the homogeneity of variances using the Levene test.

The equations were developed using the stepwise-multiple regression technique, and the paired student's t-test was used to compare the means of % fat assessed by DXA and the equations developed.

To carry out the cross-validation, in addition to Pearson's r correlation coefficient and the paired student's t-test, the total error (TE), the constant error (CE), and the standard error of estimate (SEM) were also calculated. All statistical calculations were carried out on the two groups separately.

The Bland-Altman test was used to analyze the agreement between the body fat values estimated by the equations developed and those measured by DXA. Limits of agreement were defined with a mean standard deviation of ±1.96 between the equation and the standard test and a 95% confidence interval.

All statistical tests were done using the Statistical Package for the Social Science, version SPSS® 20.0 (Chicago, IL, USA).

Ethics Statement

The Universidade Federal de Alagoas (UFAL) ethics committee approved the research under protocol number 020769/2009-31.

RESULTS

Table 1 shows the anthropometric characteristics of the total sample (AT) and the estimation (GE) and validation (GV) groups.

Table 1
Descriptive characteristics of the sample used in the study.

Variable	Total sample (AT)		Estimation group (EG)		Validation Group (GV)
	n = 100		n = 60		n = 40
	X ± Dp	mv – MV	X ±Dp	mv – MV	X ± Dp	mv – MV
Age	68.1±6.15	60.0-89.7	68.4±6.70	60.0-89.7	67.6±5.26	60.0-79.5
MC (Kg)	63.7±9.27	45.5-90.0	64.3±9.67	46.0-90.0	62.9±8.69	45.5-88.7
Est (cm)	150.2±6.4	1.34-1.66	150.7±6.7	1.34-1.66	149.6±5.9	1.35-1.61
BMI Kg/m2	28.2±3.85	19.1-38.5	28.3±3.91	21.9-38.5	28.1±3.81	19.1-36.6
%G DXA	42.2±5.42	27.7-53.2	42.4±6.06	27.7-53.2	41.9±4.37	34.1-50.4
MGDXA(Kg)	26.1±6.37	14.0-42.9	26.5±6.80	14.0-42.9	25.5±5.69	14.9-42.9
MMDXA(Kg)	35.1±4.17	27.1-48.5	35.3±4.43	27.1-48.5	34.8±3.78	28.4-44.2

• X = arithmetic mean; n: sample; MV: highest value; mv: lowest value; SD: standard deviation; MC: body mass; Est: height; BMI: Body mass index; %g DXA: fat percentage measured by DXA; MGDXA: fat mass measured by DXA; MMDXA: lean mass measured by DXA. (Maceió, 2011).

Table 1 shows the characteristics of the subjects in the validation (GV) and estimation (GE) groups.

To develop the regression equations, in addition to body mass (BM), height (HGS), and BMI (kg/m²), circumference measurements were used: brachial (CircBra), forearm (CircAbra), thigh (CircCox), leg (CircPer), abdomen (CircAbd), waist (CircCint) and hip (CircQuad). Some skinfold measurements were also used: triceps (DcTri), biceps (DcBi), subscapular (DcSE), mid-axillary (DcAx), abdominal (DcAbd), supra-iliac (DcSI), thigh (DcCox) and leg (DcPer).

In addition, as proposed by Prothro and Rosenbloom,²⁰ in an attempt to verify the best correlations, all the measures were evaluated in their simple form and squared (X²) and cubed (X³).

Age and height were excluded from the equation models because they had the lowest correlations with DXA. The measurements of WC, BMI, circumference, skinfolds, and some combinations of skinfolds and circumference showed significant correlations and were used to develop the regression equation models.

In developing the equations, eight linear regression equation models were obtained. All models were subjected to validation analysis using recommended criteria.²⁰, ²¹

Of the equations developed, four were discarded (E1, E6, E7, and E8) as they showed statistically significant differences (p<0.05) for the fat percentage (%F) averages when compared to the criterion test (DXA). They also had a multiple correlation coefficient (Pearson's r) of less than 0.70. The remaining four equations, E2, E3, E4, and E5, were submitted to the cross-validation stage in the validation group (Tables 2 and 3). As the correlation coefficients were lower than recommended (r=0.64) and the total error was greater than the standard estimation error, equations E2 and E5 were discarded. Therefore, at the end of the validation process, only two equations, E3 and E4, met all the validation criteria and were recommended for use.

Table 2
Best equation models developed in this study and validated internally in the validation group (GV).

Equations	R	T	R	R2	EPE
	p-value	p-value
E2	.693**	2.127	.693	0.48	3.30
	.001	.040
E3	.732**	0.952	.732	0.53	3.24
	.001	.347
E4	.703**	1.645	.703	0.49	3.21
	.001	.108
E5	.624**	.608	.624	0.38	3.21
	.001	.547

• Pearson's R (p< 0.05); paired t-test (p> 0.05); R: correlation coefficient; R²: coefficient of determination; EPE: Standard error of estimate. (Maceió, 2011).

Table 3
Comparison of the Total and Constant Error with the Standard Error of Estimate.

Equations	Constant Error	Total Error	Standard Error estimate
E2	- 1.2012	3.72	3.30
E3	−0.5672	3.22	3.24
E4	−0.9078	3.06	3.21
E5	−0.3951	4.07	3.21

• Total error =√∑(y1-y2)² / n; Constant error= ((%Gm) - (%Ge)); Standard error of estimate = s√ 1 - ^r2; Y1= % fat measured by DXA; Y2= % fat estimated by the equations developed; n = Sample %Gm = Average fat percentage measured by the equations; %Ge = Average fat percentage estimated by DXA; S = standard deviation; r²= coefficient of determination. (Maceió, 2011).

Figures 1 and 2 show the dispersions between the differences in mean %G measured by DXA and those estimated by equations E3 and E4 obtained from the dispersion diagram.

Figure 1
Scatterplot of Estimated error for Equation E3.

Figure 2
Scatterplot of Estimated error for Equation E4.

After all these validation stages, two anthropometric equations were validated for use because they were simple to apply and had high predictive power: E3: %Gdxa= −41.556 + 4.041 (BMI) + 0.165 (DcCox) − 0.440 (CircCox) + 0.269 (CircQuad) − 0.053 (BMI)²; and E4: %GdxaE4= 15.329 + 1.044 (BMI) −1.055 (CircAbra) + 0.282 (CircQuad) + 0.164 (DcCox) − 0.262 (CircCox).

DISCUSSION

This study aimed to develop and validate new anthropometric equation models for assessing BC, specifically for elderly Brazilian women.

Eight equation models were obtained after analyzing the correlations between the MAs that could serve as independent variables for predicting BC. However, after evaluating the proposed validation criteria, two equations, E3 and E4, were considered valid.

E3 and E4 had moderate coefficients of determination (R²), explaining 53.39% and 50.49% of the BC predicted by DXA, respectively. Rech et al,¹⁴ using circumference and skinfold measurements, validated two equations with higher R² values (0.77 and 0.81) than those observed in this study; however, in addition to using participants from the southern region of the country, the proposed equations had higher EPEs (3.51 and 3.81) than those observed in the equations in this study. Heyward and Stolarczyk²² note that when evaluating the predictive capacity of a regression equation, the size of the EPE is more important than the correlation and determination coefficients, as these can be highly affected by the size and variability of the sample and do not represent measures of accuracy between two means.

Among the independent variables used in the two equations validated in this study, BMI was the variable that correlated best with DXA (r=0.75). The same was observed in other studies¹¹,²³,²⁴ which also found a strong correlation between BMI and CSA.

One resource used to try to increase the correlation coefficient of some variables in equation,²⁰ is to raise the results of some measurements to some power. BMI² was, therefore, also included in E3. It should be noted that the inclusion of this measure increased the percentage of variance explained, consequently increasing the accuracy in predicting BC.

Salamat et al²⁵ suggest that to increase the predictive power of BMI, combining it with other anthropometric measures is important. This study found that BMI correlated well with hip (r=0.763) and thigh (r=0.735) circumferences and also with thigh skinfold (r=0.693). Therefore, it was found that when using these measures in the same model, the correlation coefficient increased while the EPE decreased, increasing the predictive power of the equation.

The main advantages of using equations that use anthropometric measurements and BMI as independent variables lie in that these measurements are easy to measure, do not require sophisticated equipment and are not invasive.

Another important point is that several studies have validated equations for predicting BC using skinfold measurements as independent variables.¹⁵,²⁴,²⁶ Corroborating these studies about the use of skinfolds, DcCox was the measurement that individually correlated best with DXA (r=0.42).

The use of anthropometric equations is a possible substitute for laboratory tests. However, for this purpose, it is important that the equation uses easy-to-access and measure measures and has a small number of variables to facilitate calculations, especially in studies with elderly populations. Concerning the number of independent variables used in this study, it should be noted that neither of the two validated equations used more than five measures.

In contrast to the results presented here, several studies⁹,¹⁴,²⁴,²⁷,²⁸ have validated regression equations to predict BC in the elderly using an equal or greater number of independent variables. However, the variables used in the models proposed in this study (skinfolds, circumferences, weight, and height) are easy to measure and also have low measurement errors (intra or inter-assessor).

CONCLUSION

The study has successfully validated two-equation models that use anthropometric measurements to predict body composition in elderly women. It has been established that the proposed measurements are taken in easily accessible areas of the body, requiring only a minimum number of measurements. This minimally invasive approach is particularly beneficial for the elderly population.

• Associate Editor responsible for the review process: André Pedrinelli

REFERENCES

• ¹
  Buckinx F, Landi F, Cesari M, Fieding RA, Visser M, Engelke K, et al. Pitfalls in the measurement of muscle mass: A need for a reference standard. J Cachexia Sarcopenia Muscle. 2018;9(2):269-78.
• ²
  Ponti F, Santoro A, Mercatelli D, Gasperini C, Conte M, Martucci M, et al. Aging and Imaging Assessment of Body Composition: From Fat to Facts. Front Endocrinol. 2020;10:861.
• ³
  Janssen I, Heymsfield SB, Wang ZM, Ross R. Skeletal muscle mass and distribution in 468 men and women aged 18-88 yr. J Appl Physiol. 2000;89(1):81-8.
• ⁴
  Comini CC, Melo JVM, Ferreira FR, Lima-Costa MF. Capacidade funcional de idosos: análise das questões de mobilidade, atividades básicas e instrumentais da vida diária via Teoria de Resposta ao Item. Cad Saúde Pública. 2015;31(5):931-45.
• ⁵
  De Oliveira MH, Cheliga FSJ, Cheliga FR, Cruz CJ, Pereira DS, Melo DS, et al. Composição Corporal em Mulheres: Comparação entre Métodos Avaliativos. Nutr Clín Diet Hosp. 2019;39(2):165-70.
• ⁶
  Antonio J, Kenyon M, Ellerbroek A, Carson C, Burgess V, Tyler-Palmer D, et al. Comparison of Dual-Energy X-ray Absorptiometry (DXA) Versus a Multi-Frequency Bioelectrical Impedance (InBody 770) Device for Body Composition Assessment after a 4-Week Hypoenergetic Diet. J Funct Morphol Kinesiol. 2019;4(2):23.
• ⁷
  Melo GF de, Américo DR, Glaner MF. Fat-free mass in overweight and obese older women: analysis of concurrent validity of bioelectrical impedance equations. Rev Bras Cineantropom Desempenho Hum. 2015;17(5):583-91.
• ⁸
  Borga M, West J, Bell JD, Harvey NC, Romu T, Heymsfield SB, et al. Advanced body composition assessment: from body mass index to body composition profiling. J Investig Med. 2018;66(5):1-9.
• ⁹
  Rech CR, Silva AT, Lunardi CC, Bohrer T, Petroski EL. Comparação da absortometria radiológica de dupla energia, antropometria e impedância bioelétrica na avaliação da composição corporal em mulheres. Revista Digital - Buenos Aires. 2005;10(91).
• ¹⁰
  Ortiz-Hernández L, Vega López AV, Ramos-Ibáñez N, Cázares Lara LJ, Medina Gómez RJ, Pérez-Salgado D. Equations based on anthropometry to predict body fat measured by absorptiometry in school children and adolescents. J Pediatr (Rio J). 2017;93(4):365-73.
• ¹¹
  Damasceno VO, Barros TAR, Gomes WS, Santos JVP, Ferreira DKS, Campos EZ, et al. Development and validation of an equation to estimate body fat in elderly women. Rev Bras Cineantropom Desempenho Hum. 2020;22:e73625.
• ¹²
  Leal LLA, Barbosa GSL, Ferreira RLU, Avelino EB, Bezerra AN, Vale SHDL, et al. Cross-validation of prediction equations for estimating body composition in ballet dancers. PLoS One. 2019;14(7):e0219045.
• ¹³
  Aristizabal JC, Estrada-Restrepo A, García AG. Development and validation of anthropometric equations to estimate body composition in adult women. Colomb Med (Cali). 2018;49(2):154-9.
• ¹⁴
  Rech CR, Santos DL, Silva JCN. Desenvolvimento e validação de equações antropométricas para predição da gordura corporal em mulheres entre 50 e 75 anos de idade. Rev Bras Cineantropom Desempenho Hum. 2006,8(1):5-13.
• ¹⁵
  Martins GQ, Matheus SC, Santos DL, Both DR, Farinha JB, Martins MS. Comparação de equações antropométricas para estimativa da gordura corporal em indivíduos com excesso de peso. Nutr Clín Diet Hosp. 2015;35(3):27-33.
• ¹⁶
  Ripka WL. Modelos matemáticos para estimativa da gordura corporal de adolescentes utilizando dobras cutâneas, a partir da absorciometria de raios-X de dupla energia [tese]. Curitiba: Universidade Tecnológica Federal do Paraná; 2017.
• ¹⁷
  Contino MBS. Validação de equações antropométricas para estimativa da composição corporal em adultos e idosos residentes no município de Niterói, Rio de Janeiro [dissertação]. Rio de Janeiro: Universidade Federal Fluminense, Faculdade de Medicina; 2018.
• ¹⁸
  Santi Maria T. Composição corporal e densidade mineral óssea de jovens futebolistas: propostas de equaçõespreditivas [tese]. Campinas: Universidade Estadual de Campinas, Faculdade de Educação Física; 2018.
• ¹⁹
  Pollock ML, Jackson AS. Research progress in validation of clinical methods of assessing body composition. Med Sci Sports Exerc. 1984;16(6):606-15.
• ²⁰
  Prothro JW, Rosenbloom CA. Physical measurements in an elderly black population: knee height as the dominant indicator of stature. J Gerontology. 1993;48(1):15-8.
• ²¹
  Lohman TG. Advances in Body Composition Assessment. Current Issues in Exercise Science. Champaign, Illinois: Human Kinetics Publishers; 1992.
• ²²
  Heyward VH, Stolarczyk LM. Applied body composition assessment. Champaign, Illinois: Human Kinetics Books; 1996.
• ²³
  Lee D, Keum N, Hu F, Orav E, Rimm E, Sun Q, et al. Development and validating anthropometric prediction equations for lean body mass, fat mass, and percent fat in adults using the National Health and Nutrition Examination Survey (NHANES) 1999–2006. Br J Nutr. 2017;118(10):858-66.
• ²⁴
  Hastuti J, Kagawa M, Byrne NM, Hills AP. Anthropometry to assess body fat in Indonesian adults. Asia Pac J Clin Nutr. 2018;27(3):592-8.
• ²⁵
  Salamat MR, Shanei A, Salamat AH, Khoshhali M, Asgari M. Anthropometric predictive equations for estimating body composition. Adv Biomed Res. 2015;4:34.
• ²⁶
  Henry CJ, Ponnalagu SD/O, Bi X, Tan SY. New equations to predict body fat in Asian-Chinese adults using age, height, skinfold thickness, and waist circumference. J Acad Nutr Diet. 2018;118(7):1263-9.
• ²⁷
  Kanellakis S, Skoufas E, Khudokonenko V, Apostolidou E, Gerakiti L, et al. Development and validation of two equations based on anthropometry, estimating body fat for the Greek adult population. Obesity (Silver Spring). 2017;25(2):408-16.
• ²⁸
  Stevens J, Ou F-S, Cai J, Heymsfield SB, Truesdale KP. Prediction of percent body fat measurements in Americans 8 years and older. Int J Obes (Lond). 2016;40(4):587-94.

Publication Dates

• Publication in this collection
  11 Dec 2023
• Date of issue
  2024

History

• Received
  20 Apr 2023
• Accepted
  02 Aug 2023