journal of human nutrition and dietetics

Journal of Human Nutrition and Dietetics



ANTHROPOMETRY
Body composition and morphological assessment of
nutritional status in adults: a review of anthropometric
variables
A. M. Madden1 & S. Smith2
1
School of Life and Medical Sciences, University of Hertfordshire, Hatfield, UK
2
School of Health Sciences, Queen Margaret University, Edinburgh, UK




Keywords Abstract
anthropometry, body composition, height,
nutritional status, sagittal diameter, skinfolds, Evaluation of body composition is an important part of assessing nutritional
waist circumference, weight. status and provides prognostically useful data and an opportunity to moni-
tor the effects of nutrition-related disease progression and nutritional inter-
Correspondence vention. The aim of this narrative review is to critically evaluate body
A. Madden, School of Life and Medical Sciences,
composition methodology in adults, focusing on anthropometric variables.
University of Hertfordshire, Hatfield AL10 9AB,
The variables considered include height, weight, body mass index and alter-
UK.
Tel.: +44 (0)1707 281385 native indices, trunk measurements (waist and hip circumferences and sagit-
E-mail: tal abdominal diameter) and limb measurements (mid-upper arm and calf
circumferences) and skinfold thickness. The importance of adhering to a
How to cite this article defined measurement protocol, checking measurement error and the need
Madden A.M. & Smith S. (2016) Body to interpret measurements using appropriate population-specific cut-off val-
composition and morphological assessment of ues to identify health risks were highlighted. Selecting the optimum method
nutritional status in adults: a review of
for assessing body composition using anthropometry depends on the pur-
anthropometric variables. J Hum Nutr Diet 29,
pose (i.e. evaluating obesity or undernutrition) and requires practitioners to
7–25
doi: 10.1111/jhn.12278 have a good understanding of both practical and theoretical limitations and
to be able to interpret the results wisely.




ing on anthropometric variables and part two focussing
Introduction
on the use of bioelectrical impedance analysis, markers of
Technological advances have increased knowledge and muscle strength, functional status and imaging techniques
understanding of body composition and its influence on with particular reference to developments relevant to
health risk and clinical outcome. As a consequence of practice.
these advances, new concepts have emerged, such as sar-
copenia, dynapenia, obesity paradox and intermuscular
Height and weight
adipose tissue. For healthcare practitioners to be able to
evaluate body composition correctly, there is a need for a Height
critical understanding of the strengths, limitations and Height is used in public health and clinical nutrition to
issues for practice, of both current and emerging meth- assess risk of undernutrition and obesity (Elia, 2003), to
ods. Furthermore, as healthcare becomes more outcome- estimate basal metabolic rate (Henry, 2005) and to deter-
driven, it is important that practitioners strive to identify mine drug dose (Pai, 2012). Accurate measurement
and use valid methods that can not only evaluate baseline requires a standardised procedure and the use of appro-
nutritional status and effects of nutritional interventions, priate, calibrated measuring equipment. Surveys of nutri-
but also contribute to the development of practice. The tional status use standard measurement techniques and,
aim of this two-part review is to critically evaluate body for standing height, require shoes to be removed, with
composition methodology in adults, with part one focus- the measured person standing upright with arms loosely


ª 2014 The British Dietetic Association Ltd. 7

, 1365277x, 2016, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jhn.12278 by Edinburgh University Library, Wiley Online Library on [24/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Anthropometric assessment of nutritional status A. M. Madden and S. Smith


to the side, back straight, heels against a vertical measure Schwab et al., 2005). Corrections to height measurements
and the head in the Frankfort plane (Fig. 1). Height is in scoliosis may be made using stereophotogrammetric
measured after a deep in-breath, ensuring that the head ISIS scanning (Carr et al., 1989), although this may not
remains in the correct position (Department of Health, be practical. A method for estimating height in patients
2012). The careful following of a standardised protocol is with contractures has been proposed recently by Finch &
recommended to minimise intra-observer technical error Arumugam (2014) and may comprise a more useful
of measurement, which may be as high as 1.3 cm for approach. An inability to stand for height measurement
adult height (Ulijaszek & Kerr, 1999). has been reported in many elderly people in nursing
Height can be measured using a free-standing or porta- homes and in hospitalised patients (Berkhout et al., 1989;
ble stadiometer, or a wall-mounted measure. Compari- Elia, 2003). In practice, deciding whether a patient has
sons of equipment indicate no significant difference in scoliosis or whether they are able to stand for measure-
height measured (Voss & Bailey, 1994; Geeta et al., 2009). ment may be subjective and so practitioners are advised
However, incorrectly assembled or positioned measuring to carefully consider each patient’s circumstances and
equipment leading to inaccurate measurements have been clearly document their observations, as well as how height
reported and thus regular calibration is required (Voss was derived.
et al., 1990; Biehl et al., 2013). When height cannot be measured, an approximation
Variation in standing height during the day has been can be derived from self-reported values and observer
reported in healthy volunteers with afternoon measure- estimation or be calculated from other body measure-
ments of approximately 6 mm less than those recorded ments using prediction equations. A systematic review of
7 h earlier (Coles et al., 1994). Conversely, resting supine studies comparing self-reported and measured height
for approximately 50 min is associated with significantly found an overall tendency to overestimate height, with
greater height (>5 mm) in women than preresting values studies reporting mean differences of up to 7.5 cm (Con-
during osteoporosis screening (Coles et al., 1994). This nor Gorber et al., 2007). Loss of height with increasing
indicates the need for paying careful attention to a stan- age is associated with greater inaccuracies of self-reported
dardised procedure when accurate serial measurements height, with studies of adults aged ≥65 years reporting
are required (Stothart & McGill, 2000). Longitudinal mean overestimates of 2.3–5 cm and a worst individual
studies indicate a loss of height with increasing age in overestimate of 18.5 cm; greater differences in women
adults of approximately 1 mm per year after age 40 years, were probably associated with greater osteoporosis-related
with an increasing rate of loss with age (Dey et al., 1999; bone loss (Payette et al., 2000; Frid et al., 2013; Reidlin-
Sorkin et al., 1999). ger et al., 2014).
Factors that may impede accurate measurement of The implications of using self-reported height may
standing height range from minor confounders (e.g. hair depend on what the values are used for. For example,
braiding) to abnormal spinal curvature (e.g. idiopathic a study of 146 patients with a mean (SD) age of
scoliosis, spinal injury, muscular dystrophy and Marfan 56 (15) years and body mass index (BMI) of
syndrome), which precludes adherence to the measure- 27.9 (5.7) kg m–2 found that using self-reported height
ment protocol. The prevalence of scoliosis in otherwise and weight do not appear to influence malnutrition
healthy adults is estimated at 8–30% but, in older adults, screening outcome (Stratton et al., 2003a). A study of 15
may be approximately 68% (Carter & Haynes, 1987; men and 22 women aged ≥70 years observed no significant




Figure 1 Position of head for measuring height
using (a) Frankfort plane where lower eye socket is
(a) (b) horizontally level with upper ear canal and (b) typical
but incorrect position (Madden et al., 2012).


8 ª 2014 The British Dietetic Association Ltd.

, 1365277x, 2016, 1, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/jhn.12278 by Edinburgh University Library, Wiley Online Library on [24/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
A. M. Madden and S. Smith Anthropometric assessment of nutritional status


difference in BMI when calculated from self-reported or select methodology on the basis of practicality and an
measured height in men but significantly lower BMI calcu- equation derived in a comparable population.
lated from self-reported height in women (Reidlinger
et al., 2014). Further research in a wider population is
Weight
needed to confirm the usefulness of self-reports. A study
comparing measured height with values estimated by Body weight represents the sum of all body compartments
healthcare professionals reported that these were less accu- (i.e. fat-free mass and fat mass) but does not discriminate
rate than self-reports with only 41% of estimates within between these. Therefore, changes in body weight may
2.54 cm of measured values (Hendershot et al., 2006). represent alterations in muscle, fat, water or a combina-
Evidence to date does not support the routine use of tion of these and so, from a nutritional perspective, they
self-reported or observer estimated height. provide limited information. Despite this, body weight is
Published equations allow estimated height to be calcu- routinely measured in healthcare and used to assess
lated from a range of different body measurements, health status and future clinical risk.
including knee height (Chumlea & Guo, 1992; Han & A standardised weighing technique requires the
Lean, 1996; Ritz, 2004), arm span (Brown et al., 2000; removal of shoes, outer garments such as jackets and car-
Mohanty et al., 2001; de Lucia et al., 2002; Capderou digans, heavy jewellery, loose change and keys. Partici-
et al., 2011), demi-span (Bassey, 1986; Hirani & Aresu, pants then stand with their feet together in the centre of
2012), ulna length (Elia, 2003; Auyeung et al., 2009) and the scales with heels against the back edge with arms
hand length (Guerra et al., 2014) (Table 1). The relation- hanging loosely by their sides and head facing forward,
ship between height and other body variables is influ- not down (Department of Health, 2012). The weight
enced by several factors, including age and ethnicity recorded includes light clothing. Records from the 1960s
(Steele & Chenier, 1990; Launer & Harris, 1996; Reeves indicate this is approximately 0.9 kg, with men tending
et al., 1996; Chumlea et al., 1998; Mohanty et al., 2001; to wear slightly heavier clothes than women but, cur-
Madden et al., 2012). For example, arm span is approxi- rently, this may be lighter (Stevens et al., 2006). Provid-
mately equal to height in White adults but greater than ing that a consistent approach is taken, no allowance
height in Black Africans and Asians (Steele & Chenier, should be made for the weight of clothes worn during
1990; Reeves et al., 1996). Some published equations have weighing. Similarly, no allowance is made for diurnal var-
been derived in young and healthy populations and so iation, which may be as much as 2 kg as a result of food
their use in hospitalised patients has been questioned and fluid intake and bladder and bowel evacuation (Loh-
(Hickson & Frost, 2003). Studies evaluating the accuracy man et al., 1988).
and precision of calculated height have been undertaken Fluctuation in body weight associated with physiologi-
in different populations and with varying conclusions cal changes in fluid balance in healthy adults may lead to
(Hickson & Frost, 2003; Shahar & Pooy, 2003; Van Lier small inaccuracies but is unlikely to mask systematic
et al., 2007; Auyeung et al., 2009; Reidlinger et al., 2014). changes in body weight as a result of a loss or gain of
Overall, these indicate that equations that are derived in a muscle or fat mass. For example, changes in fluid weight
population with age and ethnicity comparable to the peo- measured across the menstrual cycle in 98% of healthy
ple in whom they will be used are most likely to yield young women were <0.75 kg or 1.2% (Watson & Robin-
accurate estimates of height. At present, it is not possible son, 1965), whereas dehydration that is sufficient to
to make a globally useful recommendation for the best invoke thirst is likely to be associated with a weight
prediction method of predicting height and a systematic change of up to 1.5% (Stevens et al., 2006). Body weight
review of comparison studies is needed. fluctuation of 1.1–3.6% over a 3-day period has been
When measuring other body dimensions to enable reported in well-hydrated patients aged ≥60 years but
height to be calculated, practicality should also be consid- variation in weight can be reduced to ≤0.4 kg if repeat
ered especially because this is often required in bed- measurement is undertaken at the same time of day
bound or frail individuals. As a result, procedures that (Vivanti et al., 2013).
require little effort from the subject and minimal undress- Pathological changes in fluid balance may be greater
ing are more useful. From this perspective, measuring and have the potential to obscure nutritionally important
ulna length and knee height may be more practical than changes in other body compartments even when fluid
arm span or demi-span when an older person is unable changes are not clinically detectable (Bellizzi et al., 2006;
to stretch out or hold their arms for measurement. In the Morgan et al., 2006). In haemodialysis, a mean (SD)
absence of clear evidence of superior validity of any single interdialytic weight change of 1.9 (1.6) kg has been
proxy height measure or equation, practitioners are observed (Chan et al., 2008) but may be higher with
advised to view all estimates of height with caution and gaining ≥4.0 kg between dialysis associated with adverse


ª 2014 The British Dietetic Association Ltd. 9