Evaluation of the Heath-Carter Somatotype Revisited: New Bioimpedance Equations for Children and Adolescents

Abstract

Based on the results of the cross-sectional anthropological study of 2364 Russian children and adolescents aged 7-17 years, we suggest simple prediction formulae for automated bioimpedance-based evaluation of endomorphy and mesomorphy components of the Heath-Carter somatotype: ENDOBIA
= 0.5282×FMi + 0.2580×BMI - 0.04822×BM - 1.881 (r2=0.81, SEE=0.65); MESOBIA
= 0.3651×FFMi + 0.42765×BMI - 0.09323×BM - 4.803 (r2=0.81, SEE=0.54), where BMI, FMi and FFMi are, respectively, the body mass, fat mass and fatfree mass indices (kg/m2), and BM is the body mass (kg). In addition, in order to avoid using indirect bioimpedance body composition estimates, alternative formulae are constructed based only on directly measured rather than estimated bioimpedance data: ENDOBIA
= -3224.7/R + 0.63867×BMI - 0.04162×BM - 2.195 (r2=0.81, SEE=0.65); MESOBIA = 2195.4/R + 0.52966×BMI - 0.09740×BM - 4.5522 (r2=0.81, SEE=0.54), where R is the whole-body electrical resistance (Ohm) at a frequency of 50 kHz. These formulae can be used for the specified age range regardless of sex and, due to relatively high proportion of the explained variance, are suitable for individual typology.

Full text

Abstract— Based on the results of the cross-sectional an-
thropological study of 2364 Russian children and adolescents
aged 7-17 years, we suggest simple prediction formulae for
automated bioimpedance-based evaluation of endomorphy and
mesomorphy components of the Heath-Carter somatotype:
ENDOBIA = 0.5282×FMi + 0.2580×BMI – 0.04822×BM – 1.881
(r2=0.81, SEE=0.65); MESOBIA = 0.3651×FFM i + 0.42765×BMI
– 0.09323×BM – 4.803 (r2=0.81, SEE=0.54), where BMI, FMi
and FFMi are, respectively, the body mass, fat mass and fat-
free mass indices (kg/m2), and BM is the body mass (kg). In
addition, in order to avoid using indirect bioimpedance body
composition estimates, alternative formulae are constructed
based only on directly measured rather than estimated
bioimpedance data: ENDOBIA = –3224.7/R + 0.63867×BMI –
0.04162×BM – 2.195 (r2=0.81, SEE=0.65); MESOBIA =
2195.4/R + 0.52966×BMI – 0.09740×BM – 4.5522 (r2=0.81,
SEE=0.54), where R is the whole-body electrical resistance
(Ohm) at a frequency of 50 kHz. These formulae can be used
for the specified age range regardless of sex and, due to rela-
tively high proportion of the explained variance, are suitable
for individual typology.
Keywords— Somatotype, Heath-Carter typology, bioelectri-
cal impedance analysis, the whole-body electrical resistance,
fat mass index, fat-free mass index, prediction formulae.
I. INTRODUCTION
The terms somatotyping and constitution study are gen-
erally used for the designation of one of the methods for the
analysis and classification of body physique [1-5]. The
Heath-Carter anthropometric somatotype [6] that was sug-
gested as the development of the classical Sheldon’s
photoscopic scheme of the assessment of body physique [1],
is one of the commonly used methods and still of important
significance for anthropology and sports science [7-9].
The Heath-Carter somatotype represent an ordered set of
three numbers: endomorphy (which is regarded as a relative
body fatness), mesomorphy (a measure of musculoskeletal
development), and ectomorphy (relative linearity of phy-
sique). Software for the Heath-Carter anthropometric soma-
totype calculation and management is available [6,10,11].
With this, the assessment of the Heath-Carter somatotype is
not always possible because a significant number of anthro-
pometric measurements is needed which require considera-
ble expertise.
Classical studies revealed significant relationships of the
Heath-Carter endomorphy component with percent body fat
both in adults and children [12,13], and of the mesomorphy
component with lean body mass in adults [12], whereas in
children the mesomorphy showed little association with
lean body mass alone or in combination with height and
weight [13]. In their study of 260 adolescent boys aged 16
to 18 years, T. Nawarycz and L. Ostrowska-Nawarycz sug-
gested an approach for the computerized analysis of the first
and the second components of the Heath-Carter somatotype
using bioimpedance analysis [14], now the most promising
simple and easy to use method of body composition as-
sessment [15]. Their regression equation for the
endomorphy component was based on the bioimpedance
percentage body fat (%BF), whereas the mesomorphy com-
ponent was determined using body height, widths of
humerus and femur epiphyses, circumferences of the upper
arm and the calf, and the BIA %BF instead of skinfold data
[14]. So, the formula for the second component of the soma-
totype included a number of parameters not routinely meas-
ured within the standard procedure of bioimpedance meas-
urements.
Our aim was to re-analyse the relationships between the
Heath-Carter somatotype and body composition and to
develop prediction formulae for automated bioimpedance-
based evaluation of the somatotype in children and adoles-
cents suitable for use in a wide range of age in both sexes.
II. SUBJECTS AND METHODS
Anthropometry was performed in 2364 apparently
healthy children and adolescents of the Russian ethnicity
(1450 boys and 914 girls) aged 7-17 years using standard
measurement protocol adopted at the Institute and Museum
of Anthropology of the Lomonosov Moscow State Univer-
sity as described in [16]. The data were collected cross-
sectionally in 2005-2013 at schools of Moscow (n=1456),
Arkhangelsk (n=357), and Arkhangelsk region (n=551).
Evaluation of the Heath-Carter Somatotype Revisited: New Bioimpedance
Equations for Children and Adolescents
A.V. Anisimova1, E.Z. Godina1, D.V. Nikolaev2,4, and S.G. Rudnev3,4
1 Institute and Museum of Anthropology, Moscow State University, Moscow, Russia
2 Scientific Research Centre ‘Medas’, Moscow, Russia
3 Institute of Numerical Mathematics, Russian Academy of Sciences, Moscow, Russia
4 Central Research Institute for Health Organization and Informatics, Moscow, Russia
© Springer Science+Business Media Singapore 2016
F. Simini and B.F. Pedro (eds.), II Latin American Conference on Bioimpedance,
IFMBE Proceedings 54,
80
DOI: 10.1007/978-981-287-928-8_21

Standing height (Ht) was accessed by the GPM (Martin
type) anthropometer. Body mass (BM) was measured on a
digital scale to the nearest 100 g. Body mass index (BMI)
was calculated as the BM relative to Ht squared (kg/m2).
Calf, triceps, subscapular and supraspinale skinfold thick-
nesses were measured on the right side of the body using
the GPM (Harpenden type) skinfold caliper to the nearest
0.1 mm. Femur and humerus biepicondylar breadths, as
well as arm and calf girths were measured using appropriate
instrumentation to the nearest 0.5 mm. Endomorphy (Endo),
mesomorphy (Meso) and ectomorphy (Ecto) components of
the Heath-Carter anthropometric somatotype were deter-
mined based on the above mentioned quantities using con-
ventional formulae as described in [6].
The whole-body impedance was measured in a supine
position on the right side of the body according to a conven-
tional tetrapolar measurement scheme by the bioimpedance
analyzer ABC-01 ‘Medas’ (SRC Medas, Moscow, Russia)
at a frequency of 50 kHz using disposable Ag/AgCl Schiller
bioadhesive electrodes.
Fat-free mass (FFM) was accessed using Houtkooper
equation [17]: FFM = 0.61 × (Ht2/R) + 0.25 × BM + 1.31,
where Ht is the standing height (cm), R is the whole-body
electrical resistance (Ohm), and BM is the body mass (kg).
Fat mass (FM) was obtained as the difference between BM
and FFM. Similarly to BMI, fat-free mass index (FFMi) and
fat mass index (FMi) were calculated as the ratio of FFM
(kg) and FM (kg), respectively, to height squared (m2).
All statistical analyses were performed using Minitab 17
and MS Excel 2007 software packages.
III. RESULTS
Basic anthropometric characteristics of the study group
are shown in Table 1, with (*) showing a statistically signif-
icant differences (p<0.05) between boys and girls for a
given age.
Table 1 Height, weight and BMI of the study group according to age and
sex, mean (SD)
Age
Body height, cm
Body mass, kg
BMI, kg/m2
Boys
Girls
Boys
Girls
Boys
Girls
7
124.3 (6.7)
124.9 (6.8)
25.8 (4.9)
25.4 (5.2)
16.6 (2.0)
16.2 (2.3)
8
129.0 (6.3)
127.9 (5.8)
28.2 (5.3)
27.1 (5.0)
16.8 (2.0)
16.5 (2.2)
9
134.9 (6.0)
133.9 (5.8)
31.9 (6.3)
30.9 (6.1)
17.4 (2.7)
17.1 (2.6)
10
139.6 (5.6)
138.4 (7.1)
34.9 (6.4)
33.0 (7.1)
17.8 (2.6)
17.0 (2.3)
11
145.5 (8.0)
146.2 (7.9)
40.3 (10.0)
39.7 (10.1)
18.8 (3.4)
18.3 (3.2)
12
151.6 (7.1)
153.4 (8.2)
44.5 (8.8)
44.2 (11.1)
19.2 (3.0)
18.6 (3.4)
13
158.3 (8.8)
157.4 (7.6)
50.4 (11.4)
49.5 (11.5)
19.9 (3.1)
19.8 (3.6)
14
165.2 (9.5)*
161.6 (6.8)
56.4 (11.3)*
53.2 (10.3)
20.6 (3.0)
20.3 (3.2)
15
171.0 (8.4)*
162.3 (6.2)
61.8 (13.3)*
54.9 (8.5)
21.0 (3.6)
20.8 (2.9)
16
173.7 (7.2)*
164.6 (6.1)
65.3 (12.6)*
56.2 (7.5)
21.6 (3.3)
20.8 (2.7)
17
175.2 (6.5)*
162.4 (6.8)
66.4 (9.8)*
55.7 (8.0)
21.6 (2.7)
21.1 (2.5)
86420
200
150
100
50
0
Endo
Freque ncy
1086420
200
150
100
50
0
Meso
Freque ncy
86420
160
120
80
40
0
Ecto
Freque ncy
Fig. 1 The histograms of endo-, meso-, and ectomorphy components
distributions of the Heath-Carter somatotype in the study group (n=2364)
Our data showed unimodal distributions of the Endo
and Meso components of the somatotype in the study group
having a pronounced positive skewness and kurtosis, re-
spectively (see Fig. 1). The distribution of the Ecto compo-
nent was also, largely, unimodal with a small additional
peak at the value of 0.1 reflecting cumulative number of
children on the left tail of the distribution, i.e. with zero or
negative calculated values of the ectomorphy. The median
somatotype of our study group was 2.5-4.5-3.2 that can be
described as ectomorphic mesomorph according to Carter
and Heath typology [6].
Fig. 2 The Heath-Carter somatocharts of the study group according to age
and sex. Black circles show the median somatotypes for certain age
(years); white star indicates the overall median somatotype
Evaluation of the Heath-Carter Somatotype Revisited: New Bioimpedance Equations for Children and Adolescents 81
IFMBE Proceedings Vol. 54

Table 2 The Heath-Carter somatotype of the study group according to age
and sex
Age,
years
Boys
Girls
n
Endo
Meso
Ecto
n
Endo
Meso
Ecto
7
50
2.2
5.0
2.5
47
2.3
4.5
2.7
8
86
2.2
5.0
2.7
94
2.6
4.4
2.9
9
79
2.2
4.9
2.9
82
2.9
4.6
2.9
10
90
2.3
4.7
3.1
43
2.8
4.4
3.4
11
103
2.3
5.1
3.1
66
2.9
4.2
3.3
12
118
2.5
5.0
3.0
97
2.7
3.8
3.7
13
152
2.4
4.9
3.2
100
2.9
4.0
3.3
14
191
2.1
4.8
3.3
98
3.4
3.9
3.0
15
221
2.1
4.5
3.5
110
3.6
3.9
3.0
16
217
2.0
4.9
3.3
103
3.6
3.9
2.9
17
143
2.0
4.5
3.4
74
3.6
4.1
2.7
In boys, our cross-sectional data showed the age trend
from balanced mesomorph to ectomorphic mesomorph
category (see Fig. 2 and Table 2), with the overall median
ectomorphic mesomorph somatotype 2.2-4.8-3.2. The stud-
ied group of girls showed a more complex pattern of
change, from balanced mesomorph to ectomorphic meso-
morph and, then, through central phenotype, to endomor-
phic mesomorph category at the age of 17 thus reflecting
adiposity traits in the somatic growth and sexual maturation.
The overall somatotype of our girls was 3.1-4.2-3.1, or
balanced mesomorph.
Table 3 Pearson’s correlations between the Heath-Carter somatotype
components and the bioimpedance body composition parameters in boys
and girls (upper right and lower left parts of the table, respectively)
Endo
Meso
Ecto
BM
BMI
FM
FMi
%FM
FFM
FFMi
Endo
x
0.69
-0.78
0.34
0.70
0.74
0.87
0.80
0.14
0.35
Meso
0.65
x
-0.89
0.31
0.72
0.53
0.64
0.49
0.19
0.57
Ecto
-0.80
-0.87
x
-0.24
-0.70
-0.57
-0.73
-0.64
-0.09
-0.47
BM
0.57
0.24
-0.41
x
0.85
0.77
0.53
0.30
0.96
0.86
BMI
0.80
0.63
-0.78
0.88
x
0.88
0.81
0.58
0.72
0.86
FM
0.73
0.38
-0.55
0.92
0.91
x
0.93
0.81
0.56
0.56
FMi
0.85
0.56
-0.72
0.79
0.92
0.95
x
0.94
0.28
0.39
%FM
0.79
0.40
-0.61
0.66
0.77
0.88
0.94
x
0.04
0.09
FFM
0.42
0.12
-0.26
0.97
0.78
0.79
0.62
0.48
x
0.87
FFMi
0.62
0.60
-0.69
0.81
0.90
0.70
0.67
0.43
0.81
x
The correlations of the Heath-Carter somatotype compo-
nents Endo, Meso and Ecto in boys and girls with the indi-
ces of fat- and fat-free mass (FMi, FFMi) were higher as
compared to absolute FM and FFM values or the %FM
(Table 3). In this regard, we proposed the following simple
prediction formulae for the bioimpedance evaluation of the
Endo and Meso components of the somatotype:
ENDOBIA = 0.5282×FMi + 0.2580×BMI – 0.04822×BM –
1.881 (r2=0.81, SEE=0.65) (1)
MESOBIA = 0.3651×FFMi + 0.42765×BMI – 0.09323×BM
– 4.803 (r2=0.81, SEE=0.54) (2)
All the components of the regression formulae (1) and
(2) were essential (see Tables 4 and 5) with the regression
lines for the residuals not significantly different from zero
(Fig. 3).
Table 4 Contribution and order of entry of predictor variables to the
regression model (1) for the endomorphy component of the Heath-Carter
somatotype
Predictor variables
r2
SEE
p
FMi
0.76
0.75
<0.001
BM
0.78
0.72
<0.001
BMI
0.81
0.65
<0.001
r2 is the proportion of explained variance; SEE is the standard error of the
model; p is the significance of contribution of the
respective parameter
to
the stepwise multiple regression model
Table 5 Contribution and order of entry of predictor variables to the
regression model (2) for the mesomorphy component of the Heath-Carter
somatotype
Predictor variables
r2
SEE
p
BMI
0.47
0.89
<0.001
BM
0.71
0.66
<0.001
FFMi
0.81
0.54
<0.001
r2 is the proportion of explained variance; SEE is the standard error of the
model;
p is the significance of contribution of the respective
parameter to
the step
wise multiple regression model
10
9876543210
4
3
2
1
0
-1
-2
-3
-4
ENDO bia
Residual
10
98765432
3
2
1
0
-1
-2
-3
MESO bia
Residual
Fig. 3 The residuals and the respective regression lines for endomorphy
and mesomorphy estimates of the Heath-Carter somatotype
One can note, due to mutual dependence of the FFM on
the impedance index Ht2/R, that the FFMi, as a ratio of FFM
to Ht2, should strongly correlate with the inverse value of
the electrical resistance R. With this idea, in order to avoid
using population-specific body composition equations, we
constructed the alternative formulae for the evaluation of
the Heath-Carter somatotype relying solely on measure-
ments of height, weight, and the electric resistance:
ENDOBIA = –3224.7/R + 0.63867×BMI – 0.04162×BM –
2.195 (r2=0.81, SEE=0.65) (3)
MESOBIA = 2195.4/R + 0.52966×BMI – 0.09740×BM –
4.5522 (r2=0.81, SEE=0.54) (4)
82 A.V. Anisimova et al.
IFMBE Proceedings Vol. 54

These formulae are similar to Eqs. (1)-(2) in structure,
have the same accuracy of the response variables approxi-
mation, and take an advantage of using only directly meas-
ured rather than estimated bioimpedance data. The relative-
ly high values of the proportion of explained variance r2
enable the use of these formulae for individual typology.
Given that the ectomorphy, i.e., the third component of the
somatotype, is calculated directly on patient’s height and
weight [6], we, thus, obtain an opportunity for automated
bioimpedance-based evaluation of the overall Heath-Carter
somatotype in children and adolescents. The respective
algorithm was embedded in the current version of the ABC-
01 ‘Medas’ bioimpedance meter software.
IV. CONCLUSIONS
The assessment of body composition and somatotyping
represent two different, but correlated, ways of describing
human physique and structure. The Heath-Carter anthropo-
metric somatotype [6] is one of commonly used methods of
somatotyping and still of important significance for anthro-
pology and sports science [7-9]. However, in practice, the
assessment of the Heath-Carter somatotype is not always
available because of the need for a significant number of
anthropometric measurements that must be performed by a
qualified specialist. In our work, based on the results of
anthropological study of a large group of ethnically Russian
children and adolescents, we suggested simple prediction
formulae for automated bioimpedance-based evaluation of
the Heath-Carter somatotype that are suitable for individual
typology. We could recommend preferential use of the
equations based on directly measured electrical resistance
rather than estimated values of fat mass index or fat-free
mass index. In contrast to the results obtained earlier by the
other authors [14], the formulae are suitable for use both in
boys and girls in a relatively wide age range, from 7 to 17
years, and rely solely on data collected within the traditional
bioimpedance measurements procedure.
ACKNOWLEDGMENT
The study was supported by the RFBR grants no. 13-06-
00702 and 15-06-06901 (for AVA and EZG), and by the
RSF grant no. 14-15-01085 (for DVN and SGR).
CONFLICT OF INTEREST
The authors declare that they have no conflict of interest.
REFERENCES
1. Sheldon W H, Stevens S S, Tucker W B (1940) The varieties of
human physique. Harper Bros., New York
2. Stefko V G, Ostrovsky A D (1929) The schemes of clinical di-
agnosis of constitutional types. Biomedgiz, Moscow (in Rus-
sian)
3. Tanner J M (1964) The physique of the Olympic athlete. George
Allen and Unwin, London
4. Heath B H, Carter J E L (1967) A modified somatotype method.
Am J Phys Anthropol 27:57–74
5. Carter J E L, Heath B H (1990) Somatotyping – development
and applications. Cambridge Univ. Press, Cambridge
6. Carter J E L (2002) The Heath-Carter anthropometric somato-
type: instruction manual at http://www.somatotype.org/Heath-
CarterManual.pdf
7. Kandel M, Baeyens J P, Clarys P (2014) Somatotype, training
and performance in Ironman athletes. Eur J Sport Sci 14:301–
308
8. Rajkumar R V (2015) Endomorphy dominance among non-
athlete population in all the ranges of body mass index. Int J
Physiother Res 3:1068–1074
9. Yang L-T, Wang N, Li Z-X et al. (2015) Study on the adult phy-
sique with the Heath-Carter anthropometric somatotype in the
Han of Xi’an, China. Anat Sci Int DOI 10.1007/s12565-015-
0283-0
10. Somatotype at http://goulding.ws/somatotype
11. Silva C A D, Mendes D D S, Oliveira E et al (2014) BodyShifter
– software to determine and optimize an individual’s somato-
type. Procedia Technol 16:1456–1461
12. Slaughter M H, Lohman T G (1976) Relationship of body com-
position to somatotype. Am J Phys Anthropol 44:237–244
13. Slaughter M H, Lohman T G (1977) Relationship of body com-
position to somatotype in boys aged 7 to 12 years. Res Quart
48:750–758
14. Nawarycz T, Ostrowska-Nawarycz L (2001) Evaluation of the
first and second components of somatotype using bioelectric
impedance analysis, Proc. of XI Int. conf. on electrical
bioimpedance. Oslo, Norway, 2001, pp 349–352
15. Kyle U G, Earthman C P, Pichard C, Coss-Bu J A (2015) Body
composition during growth in children: limitations and perspec-
tives of bioelectrical impedance analysis. Eur J Clin Nutr DOI
10.1038/ejcn.2015.86
16. Godina E (2011) Secular trends in some Russian populations.
Anth Anz 68:367–377
17. Houtkooper L B, Going S B, Lohman T G et al. (1992) Bioelec-
trical impedance estimation of fat-free body mass in children
and youth: a cross-validation study. J Appl Physiol 72:366–373
Address of the corresponding author:
Author: Sergey Rudnev
Institute: Institute of Numerical Mathematics
Street: Gubkin str., 8
City: Moscow
Country: Russia
Email: sergey.rudnev@gmail.com
Evaluation of the Heath-Carter Somatotype Revisited: New Bioimpedance Equations for Children and Adolescents 83
IFMBE Proceedings Vol. 54