Rev Bras Fisiol exerc 2021;19(5):369-76
doi: 10.33233/rbfex.v19i5.4106
ORIGINAL
ARTICLE
Validation
of a bioelectrical impedance scale for the estimation of body fat in
adolescents
Validação de uma
balança de impedância bioelétrica para a estimativa de gordura corporal em
adolescentes
Andreia Pelegrini1,
André de Araújo Pinto1, Hector Cris Colares de Angelo1,
Gaia Salvador Claumann1, Diego Augusto Santos Silva1,
Mateus Augusto Bim1
1Universidade do Estado
de Santa Catarina, Florianópolis, SC, Brazil
Received
on: May 7, 2020; Accepted on: September 15, 2020.
Corresponding author: Andreia Pelegrini,
Grupo de Estudos e Pesquisa em Cineantropometria – GEPECIN, Centro de Ciências
da Saúde e do Esporte (CEFID), Universidade do Estado de Santa Catarina
(UDESC), Rua Pascoal Simone, 358 Coqueiros 88080-350 Florianópolis SC,
andreia.pelegrini@udesc.br
Andreia Pelegrini: andreia.pelegrini@udesc.br
André de Araújo Pinto:
andrefsaude@hotmail.com
Hector Cris Colares de Angelo: hector.edf@hotmail.com
Gaia Salvador Claumann: gaiasclaumann@hotmail.com
Diego Augusto Santos
Silva: diegoaugustoss@yahoo.com.br
Mateus Augusto Bim:
mateus.bim@unoesc.edu.br
Abstract
Introduction: Several
different instruments available on the market have been used for the estimation
of body fat. However, many of these instruments have not been compared with
reference criteria to verify their true accuracy. Aim: To verify
the validity of a bioelectrical impedance scale (OMRON-514C) for the estimation
of body fat. Methods: Forty-four overweight adolescents (25 females) participated in
this study, with an average age of 12.3 ± 1.1 years. All were submitted to body
fat evaluations by air displacement plethysmography and bioelectrical
impedance. Results: Higher values of relative and absolute body fat were estimated by
bioelectrical impedance compared to plethysmography (p < 0.05). There was no correlation
between the relative body fat measurements between the two methods (r = 0.185;
p = 0.228). The absolute measurements of body fat were correlated (r = 0.497, p
= 0.001). Both in the measurements of relative (p = 0.034) and absolute body fat
(p = 0.021), the bioelectrical impedance overestimated the measured values. Conclusion: In adolescents with characteristics like the present study, the estimate
of body fat by the bioelectrical impedance (OMRON-514C) should be used with
caution.
Keywords: plethysmography; bioelectrical impedance; adolescents; overweight.
Resumo
Introdução: A estimativa da gordura
corporal tem sido realizada por diversos equipamentos disponíveis no mercado.
Entretanto, muitos desses equipamentos não foram comparados com critérios de
referência que verificassem a sua real acurácia. Objectivo: Verificar a validade de uma balança de impedância bioelétrica (OMRON-514C) para
a estimativa de gordura corporal. Métodos: Quarenta e quatro adolescentes com excesso de
peso (25 do sexo feminino) participaram deste estudo, com idade média de 12,3 ±
1,1 anos. Todos foram submetidos à avaliação da gordura corporal por pletismografia de deslocamento de ar e impedância
bioelétrica. Resultados: Valores mais elevados de gordura corporal relativa e absoluta
foram estimados por impedância bioelétrica em relação à pletismografia
(p < 0,05). Não houve correlação entre as medidas de gordura corporal
relativa entre os dois métodos (r = 0,185; p = 0,228). As medidas absolutas de
gordura corporal foram correlacionadas (r = 0,497, p = 0,001). Tanto nas
medidas de gordura corporal relativa (p = 0,034) quanto absoluta (p = 0,021),
nas quais, para ambas as medidas, a impedância bioelétrica superestimou os
valores medidos. Conclusão: Em adolescentes com características semelhantes ao
presente estudo, a estimativa de gordura corporal pela impedância bioelétrica
(OMRON-514C) deve ser usada com cautela.
Palavras-chave: pletismografia,
impedância bioelétrica; adolescents; excesso de peso.
Increasing interest has been observed in studies that investigate body
composition at different stages of life [1]. This is mainly due to the epidemic
proportions of obesity [2], which has been considered a public health problem
[3] and which is related to other chronic non-communicable diseases (NCDs) [4].
Such facts are due to a contemporary lifestyle, composed mainly by the
excessive intake of ultra-processed foods and sedentary behaviors [5].
In Brazilian adults, more than 50% are overweight and 18.9% are obese.
It should be noted that in a decade (2006-2016), overweight increased by 26.3%
and obesity by 60% [6]. When analyzing cases of overweight and obesity in
adolescence, the National School Health Survey (PeNSE)
revealed that 23.7% and 7.8% of Brazilian students, between 13 and 17 years
old, were overweight and obese, respectively [7]. These data are of concern,
especially due to the tendency of these prevalences
to remain during adulthood [8]. Furthermore, if current trends continue,
estimates point to a higher number of children and adolescents with obesity
than with moderate and severe malnutrition by 2022 [9].
Given the worrying prevalence of overweight in the pediatric population,
the assessment of body composition is fundamental, mainly, based on valid and
reliable techniques that estimate body fat. Traditionally, body fat measurement
techniques can be classified as direct, indirect and doubly indirect [10].
Among the indirect techniques, Dual Energy X-Ray Absorptiometry (DXA) and
air-displacement plethysmography (ADP) stand out [11]. These methods are
accurate, but have a high cost, in addition to requiring trained personnel to
carry out the measures, limiting their use in epidemiological studies [12].
Thus, it is interesting to have simpler, faster and less costly methods,
which make it possible to carry out studies with such quality observed by
doubly indirect methods. In this case, bioelectrical impedance (BIA),
considered one of these techniques, becomes an important tool for assessing and
monitoring body fat for healthcare professionals. Therefore, this technique
should be used from the validation with reference criteria [13].
Even with the practicality of BIA to other means of more accurate
assessment of body fat, the validation results are controversial, especially
because different BIA devices are being marketed. A previous study, conducted
in children, when comparing the values of body fat estimated by BIA (TANITA
SC-240) and ADP, revealed an underestimation of the values estimated by BIA to
those of ADP [14]. In French adolescents, BIA (Tanita MC-780) overestimated
body fat values [15]. On the other hand, in American children and adolescents,
it was noted that BIA (TANITA SC-240) presented acceptable precision for
estimating body fat when compared to DXA [16].
No studies were found in the literature that evaluated the validity of
the BIA scale, model OMRON-514C, with ADP and for this to be possible, the
accuracy of BIA must be determined. Therefore, the present study aimed to
verify the validity of a commercial BIA scale (OMRON-514C) for the estimation
of body fat in Brazilian adolescents, using ADP as a reference criterion.
Study
participants
This is a cross-sectional observational study, linked to a macro project
entitled “Bone mineral density in adolescents: what is the relationship with
body fat, physical activity, and sedentary behavior?”, carried out in a public
elementary school in São José, SC, Brazil, selected intentionally. The
research was approved by the Human Research Ethics Committee (opinion No.
1,468,045/2016).
Adolescents with ages between 10.0 and 14.9 years, who were overweight,
regularly enrolled in elementary school, who accepted to participate
voluntarily in the research with the signature of the term of consent and that
the parents or guardians signed the Term of Free and Informed Consent (ICF),
were considered eligible for this study.
The following exclusion criteria have been used: 1) inability to stand
and/or move; 2) full speech and/or hearing disability; 3) refer to diseases
that alter body composition, such as malignant neoplasms, chromosomal changes,
paralysis, renal or liver failure, hyper or hypothyroidism; 4) pregnancy; and
5) use of diuretic medications.
The sampling procedure took place by saturation, and all adolescents
within the pre-established age group were invited to participate in the
research. From an initial sample of 1002 subjects, 433 were excluded for being
underweight or obese, 440 refused to participate in the study and 11 were
excluded due to the exclusion criteria, totaling 118 adolescents. Of these, 44
were overweighed and have been included in the present study. A posteriori
sample power calculation was performed considering the comparison of the mean
values and standard deviation of body fat assessed by BIA (17.1 ± 6.6) and ADP
(14.2 ± 5.4) and the correlation (r = 0.497). An effect size of 0.47 was found
with a significance level of 0.05, with a beta value of 0.87.
Variables
The measurement of body mass and height, following standardized
procedures [17], of all students was performed by a trained evaluator. Body
mass (kg) was measured using a portable digital scale from Tanita (BF683W,
Arlington Heights USA). Height (m) was measured using a stadiometer (Altura Exata, Minas Gerais, Brazil). With these data, the Body
Mass Index (BMI) was calculated and classified according to Age-and
Sex-specific cutoff points [18].
The assessment of body fat measurements was carried out at the
Anthropometry Laboratory of the Federal University of Santa Catarina (UFSC) on
days previously scheduled with the school and those responsible for the
students, always during the morning shift. Transportation for driving school
students to UFSC was provided by the State University of Santa Catarina
(UDESC). Before the collection date, the guardians of the students were asked
to follow some instructions, which were necessary for the evaluations of the
students that would be carried out on the following day: a) be fasting for at
least ten hours before the exam; b) using appropriate clothing, boys in swim
trunks and girls in tops and shorts that are not loose on the body; c) not
carrying metal objects attached to the body, such as earrings, necklaces, and
piercings; d) not performing physical exercise eight hours before the exams; e)
not drinking alcohol forty-eight hours before, and f) not being in the
menstrual period.
Body fat was estimated using the air displacement plethysmography (ADP)
method (Bodpod, Life Measurements, Concorde, CA,
USA), previously calibrated, which is considered a reference for the estimation
of body fat [19], and its validity for this measure in adolescents has already
been evidenced [20].
For the BIA method, the body control scale was used (OMRON-514C, São
Paulo, Brazil), which has a system of eight electrodes (two in each hand and
foot) and performs resistance and reactance measurements in each of the segments
using electric current at the intensity of 50 kHz (kilohertz) and 500 µA
(microamperes) throughout the body (from arms to feet). For the evaluation of
the students, the manufacturer's instructions were used, thus, they should
remain in the orthostatic position, with their feet touching the electrodes of
the base of the scale, with their knees and back straight looking towards the
horizon, keeping the arms horizontal and the hands in contact with the
electrodes of the support display unit, with the shoulders flexed forming an
angle of 90º to the body and elbows extended.
Statistical
analysis
Statistical analyses were conducted using the IBM SPSS Statistics 20
software, with a significance level of 0.05. Data normality was examined using
the Shapiro Wilk test. To compare BIA's estimated body fat values with the
reference method (ADP), the paired t-test (systematic bias) was used. Pearson's
correlation was used to verify possible relationships between the methods and
the agreement between the measurements was analyzed using the Bland and Altman
plots [21].
Descriptive data of the study participants are shown in Table I. 44
adolescents participated in the study, of which 25 were female, with a mean age
of 12.3 ± 1.1) years. Higher (p < 0.05) values for a relative (28.6% and
24.6%) and absolute (17.1 kg and 14.2 kg) body fat estimated by the BIA in
relation to ADP, respectively, were found.
Table
I - Anthropometric characteristics and body fat of
overweight adolescents. São José/SC (2016)
SD
= standard deviation; BM = body mass; kg: kilograms; BIA = bioelectrical
impedance; ADP = air-displacement plethysmography; BMI = body mass index;
*p< 0.05 (comparison of average values between BIA and ADP)
There was no correlation between the measures of relative body fat
obtained between the two methods (r = 0.185, p = 0.228), however, the measures
of absolute body fat were correlated between the methods, showing a weak to
moderate correlation (r = 0.497, p = 0.001) (Figure 1). Regarding the male
gender, no correlation between the methods was observed, while in the female
gender there was a moderate to strong correlation between the relative (r =
0.565, p = 0.003) and absolute (r = 0.753, p = 0.001) measures.
%
fat = relative fat; kg = absolute fat
Figure
1 - Correlation between methods for body composition
in overweight adolescents. São José/SC (2016).
Figure 2 shows the analysis of agreement between the relative and
absolute body fat measurements obtained by the two methods. A systematic bias
was observed for both relative (p = 0.034) and absolute (p = 0.021) body fat
measurements, in which, for both measures, the BIA overestimated the measured
values, in addition to presenting broad limits of agreement.
Figure
2 - Difference between body fat (relative and
absolute) obtained between bioelectrical impedance and air-displacement
plethysmography over their averages
The main result of this study indicated that the values of relative and
absolute body fat estimated by BIA OMRON-514c were overestimated in comparison
to those of ADP (reference criterion). The existence of numerous brands and
models of BIA has been made comparisons between studies difficult, and even
those that compared the values obtained by BIA with those of DXA present
different results [15,16]. Despite this, Barreira et
al. [16], when investigating African American children, found acceptable
accuracy in estimating body fat using BIA Tanita SC-240 and DXA. In a study with
Swedish children [14], no significant differences were identified when
analyzing BIA Tanita SC-240 and ADP. On the other hand, in a study with French
adolescents, the body fat values were overestimated by BIA Tanita MC-780
compared to DXA [15]. However, when analyzing two BIA scales (Biodynamics Model
450 and InBody 230) compared to DXA, Faria et al. [22] identified that only BIA InBody230
presented results close to those of DXA, although both models overestimated the
values of body fat compared to the reference criterion (DXA).
Researchers verified the validation of methods to estimate body fat in
children and adolescents, such as DXA, underwater weighing, skinfold thickness
measurement, isotope dilution, dilution of deuterium, and BIA using the 4C
model (four compartments) as a reference method [23]. The results demonstrated
that BIA generated the least satisfactory results of body fat compared to the
other methods investigated [23]. Such results can be explained by the nature of
the method that evaluates body water by the resistance to the passage of
alternating electric current [24]. Thus, as the amount of body fat increases,
there is resistance to electrical conductivity, as the fat mass presents less
hydration than lean tissues [24].
It should also be noted that the accuracy of BIA measurements is
influenced by factors such as body hydration level, body shape and even
ethnicity [25]. For that reason, studies of validation of these instruments
would require larger samples. Other studies have also reinforced that the
different BIA models available on the market estimate body composition using
equations from the manufacturers, with no validity for specific ethnic groups
[26,27]. Thus, it is necessary to consider that the BIA model chosen to assess
body fat presupposes the knowledge of this information so as not to incur
incorrect estimates. For this, a careful analysis of these available data is
important in search of more accurate results, to support its use both in
epidemiological and clinical practice. However, it should be noted that even
generating inaccurate estimates, their results can be valuable, especially in
assessments with adolescents who have a high volume of body fat due to
difficulties with the use of anthropometry.
The strengths of this study are the use of a homogeneous sample
regarding BMI (overweight), as well as the use of a model considered a
reference criterion for estimating body fat. Besides, even though studies are
comparing the performance of different BIA models with other reference methods
for estimating body fat in adolescents, relatively little work has been done to
date, which makes these findings important. Among the main limitations of the
study, we highlight the lack of knowledge of the BIA ONROM-514c equation for
estimating absolute and relative body fat, as well as the small sample size and
the heterogeneity of the group to age (10 to 14 years).
The results of this study suggest that body fat estimates are influenced
using different assessment methods (ONROM-514c body control scale and ADP).
Thus, for overweight adolescents who have similar characteristics to those of
the present study, the use of the OMRON-514c scale should be used with caution.
In parallel with the use of BIA, the use of another method of estimating body
fat in adolescents is recommended to obtain a more accurate assessment of body
fat.
Acknowledgements
Thanks to CNPq (number 445657/2014-6) and
FAPESC (number 2017TR646) for the financial support, and Capes for the
scholarships granted.
Conflict
of interest
Authors state no conflict of interest.
Authors’
contribution
AAP, HCCA, GSC and MAB analyzed the data and drafted the paper,
including writing. AP and DASS assisted in analyses, writing and contributed to
the conception and protocol of the study. All authors aided in preparing the
manuscript.