Journal of Human Nutrition and Dietetics

CLINICAL NUTRITION
The estimation of body height from ulna length in healthy
adults from different ethnic groups
A. M. Madden, T. Tsikoura & D. J. Stott
School of Health and Emergency Professions, University of Hertfordshire, Hatfield, UK

Keywords Abstract
ethnic group, height, nutritional assessment,
prediction, ulna length. Background: Assessments of nutritional status frequently incorporate a measure
of height to evaluate a person’s relative thinness or fatness. Because height is
Correspondence often difficult to quantify, it may be predicted from alternative anthropometric
Dr A. Madden, School of Health and measurements, including ulna length. Little information is available about the
Emergency Professions, University of
accuracy of these predictions in an ethnically diverse population. The present
Hertfordshire, Hatfield AL10 9AB, UK.
Tel.: +44 (0)1707 281385
study aimed to evaluate published equations for predicting height from ulna
Fax: +44 (0)1707 284977 length in adults from different ethnic groups.
E-mail: a.madden@herts.ac.uk Methods: Ulna length and standing height were measured in a gender-stratified
sample of 60 Asian, 69 Black and 65 White healthy volunteers, aged 21–
65 years. Height was predicted from ulna length using the Malnutrition Uni-
How to cite this article
Madden A.M., Tsikoura T. & Stott D.J. (2012)
versal Screening Tool (MUST) equations and compared against the measured
The estimation of body height from ulna values. Linear regression analysis was used to develop equations to estimate
length in healthy adults from different ethnic height from ulna length and to explore the relationship between height and
groups. J Hum Nutr Diet. 25, 121–128 ulna length in subgroups.
doi:10.1111/j.1365-277X.2011.01217.x Results: The mean (SD) age for Asian, Black and White in men was 31.7
(11.0), 32.0 (10.3) and 38.6 (12.5) years and in women was 26.2 (5.4), 32.6
(8.9) and 35.7 (11.7); the mean (SD) height in men was 170.9 (5.2), 178.1
(7.3) and 176.3 (7.7) cm and in women was 157.7 (4.7), 164.0 (5.9) and 163.7
(6.2) cm. Ulna length and measured height were significantly correlated among
all subgroups, except Asian women (r = 0.11, P = 0.57). The mean (SD) differ-
ence between predicted and measured height showed significant overestimates
for Asian and Black men [4.0 (4.8) and 6.7 (5.3) cm] and Asian and Black
women [6.4 (4.9) and 4.4 (4.9) cm] but not for White men and women.
Conclusions: The MUST equations for predicting height from ulna length in
healthy adults should be used with some caution among ethnically diverse pop-
ulations, particularly in Asian women.

try, the evaluation of body dimensions, is useful because

Introduction
it is an objective measure that allows an approximate
Nutritional screening plays an important role in the iden- evaluation of body stores to be made; for example, via
tification and prevention of malnutrition in patients in body mass index (BMI = weight/height2). Although BMI
hospitals and the community (Elia, 2003). A large num- is commonly used in nutritional screening, including the
ber of screening tools have been proposed to facilitate this Malnutrition Universal Screening Tool (MUST) (Elia,
and most of them utilise a combination of variables, 2003), it requires a measure of height that might be diffi-
including clinical, dietary, biochemical and anthropomet- cult to obtain and/or of questionable accuracy (Kirk
ric measurements (Green & Watson, 2005). Anthropome- et al., 2003; Cook et al., 2005). As a result, alternative

ª 2011 The Authors

Journal of Human Nutrition and Dietetics ª 2011 The British Dietetic Association Ltd. 121
 Estimating height from ulna length A. M. Madden et al.

measurements for estimating height are required. Pub- completed, the records from subjects between the ages of
lished studies have explored the prediction of height from 21 and 65 years were categorised into one of the follow-
a range of methods (Hickson & Frost, 2003), including ing groups based on the 2001 UK Census (Office for
knee height (Han & Lean, 1996; Ritz, 2004), arm span National Statistics, 2006) and included in the subsequent
(Brown et al., 2000; Mohanty et al., 2001; de Lucia et al., analysis:
2002; Zverev, 2003), demi-span (Bassey, 1986) and ulna l Asian (Bangladeshi, Indian, Pakistani)

length (Cheng et al., 1998; Elia, 2003; Gauld et al., 2004; l Black (Black African, Black Caribbean)

Agnohotri et al., 2009). The latter has been examined in l White (English, Irish, Scottish, Welsh)

children (Cheng et al., 1998; Gauld et al., 2004), young

adults aged 18–28 years (Agnohotri et al., 2009) and in
Measurements
adults <65 and ‡65 years (Elia, 2003), where the resulting
MUST prediction equations have been recommended for Before collecting data from the study subjects, height and
predicting height in the absence of actual measurements. ulna length were measured five times in 12 separate sub-
It is well known that anthropometric measurements, jects to investigate the reproducibility of the method. On
including height, can be affected by a variety of factors, the basis of the results, data were subsequently collected
including racial and ethnic differences (Steele & Chenier, by measuring height once and ulna length three times
1990; World Health Organization, 1995; Launer & Harris, using a standardised protocol. All measurements were
1996; Reeves et al., 1996; Cheng et al., 1998; Chumlea made between 10.30 and 16.30 h by a single observer
et al., 1998; Mohanty et al., 2001). These differences relate (TT) using the same equipment.
not only to absolute anthropometric measurements, but l Height – Standing height was measured using a Leices-

also to the relationship between variables; for example, ter Height Measure portable stadiometer (Chasmors
arm span is approximately equal to height in White Limited, London, UK). The stadiometer was placed on
adults but greater than height in Black Africans and a firm, level surface with stabilisers positioned against a
Asians (Steele & Chenier, 1990; Reeves et al., 1996). The wall to ensure rigidity. The subject removed shoes,
equations describing the relationship between ulna length socks and hats, if worn. If they had a hairstyle that
and height have been published, although details concern- would affect the measurements, they were asked to
ing how they were derived have not been reported, so adjust it so that an accurate result was obtained. The
that the ethnicity of the reference population is unknown. subject stood on the platform of the stadiometer facing
An exploration of the applicability of the equations in forward with shoulders relaxed, arms hanging freely by
adults from different ethnic groups is required if predic- the sides, legs straight and close together with the
tion equations are to be useful in screening patients from upper back, buttocks and heels in contact with the
diverse backgrounds. The present study aimed to evaluate upright section of the stadiometer. The subject’s head
the MUST equations for predicting height from ulna was positioned in the Frankfort horizontal plane
length in adults from different ethnic groups. (Fig. 1) by the interviewer and the head plate lowered
until it was just brought into contact with the top of
the head (Ruston et al., 2004). The measurement was
Materials and methods
performed once and recorded to an accuracy of 0.1 cm.
Subjects l Ulna length – The forearm was measured using an

The subjects were recruited opportunistically from staff anthropometric tape (Butterfly, Shanghai, China). If
and students at London Metropolitan University by invit-
ing those who were waiting in or passing through the
main reception area to participate. Initial inclusion crite-
ria were an ability to speak and communicate in English
and an ability to stand for height and ulna measurement.
Subjects with obvious physical disabilities that might
influence their height and/or bone length (e.g. those in a
wheel-chair, stooping or walking with a limp) were

9 8
excluded. The study details were explained and data were
(a) (b)
subsequently collected from those providing their verbal
consent. Subjects were asked to provide their date of
birth, gender and ethnic background and this was Figure 1 Position of head for measuring height using: (a) the Frank-
recorded on an anonymous form before their height and fort plane where lower eye socket is horizontally level with upper ear
ulna length were measured. After all data collection was canal and (b) a typical but incorrect position.

ª 2011 The Authors

122 Journal of Human Nutrition and Dietetics ª 2011 The British Dietetic Association Ltd.
 A. M. Madden et al. Estimating height from ulna length

the subject was wearing wristbands, tight jewellery,

Results
bracelets or watch that could make the reading inaccu-
rate, they were asked to remove them or change their The measurement of both height and ulna length was
position. Subjects were asked to bend their left arm reproducible in men and women with mean coefficient of
and place it across the chest with the fingers pointing variation of 0.12 and 0.47, respectively.
to the opposite shoulder (Fig. 2). Measurements were Two hundred and eighty subjects agreed to participate
taken between the point of the elbow (olecranon pro- and were measured. Data from 86 of them were subse-
cess) and the midpoint of the prominent bone of the quently excluded from the analysis either because they
wrist (styloid process) (Elia, 2003). The procedure was were aged under 21 years, and therefore may not have
repeated three times, readings were recorded to the reached adult height (Noppa et al., 1980), or because they
nearest 0.5 cm and a mean value was calculated. described their ethnicity as mixed or other than Asian,
Black or White. Analyses were subsequently undertaken
on the remaining 194 subjects.
Calculation and analysis
Summary statistics for age, measured height and ulna
Predicted height was calculated from the mean ulna length for the six subgroups are presented in Table 1.
length using the MUST equations for adults aged There is strong evidence of difference between the
<65 years (Elia, 2003): means of the ethnic groups for ulna length and height
l Men: Predicted height (cm) = 79.2 + [3.60 · ulna among both men and women (one-way anova:
length (cm)] P < 0.001 for all four comparisons). Differences in
l Women: Predicted height (cm) = 95.6 + [2.77 · ulna mean age were highly significant among women (one-
length (cm)] way anova: P < 0.001) and also significantly different
The difference in age and measurements between sub- among men (one-way anova: P 0.029). The significance
groups was examined using one-way analysis of variance (P £ 0.05) of any paired differences (post hoc Bonfer-
(anova). Pearson’s correlation coefficient and simple lin- roni correction) are given in Table 1. Asian and Black
ear regression were used to examine the relationship groups differ in height and ulna length among both
between ulna length and measured height in each sub- men and women; the same is generally true of Asian
group. The difference between measured and predicted and White groups, with the exception of ulna length
height was examined within each subgroup using paired among women. Black and White groups differ among
t-tests. men and women with respect to ulna length. Overall,
The study protocol was approved by the Research Eth- the Asian subgroups contained the youngest, least tall
ics Committee of London Metropolitan University (23 and shortest ulna lengths of the three ethnic groups on
August 2007). the basis of mean values.
In Table 2, the relationship between measured height
and ulna length and the difference between mean mea-
sured height and predicted height calculated using the
MUST equations has been summarised. There was a
moderately strong and significant correlation (r > 0.6)
between ulna length and measured height among Black
and White subgroups for men and women. The relation-
ship was somewhat weaker among Asian men (r = 0.43)
and, among Asian women, was very weak (r = 0.11) and
was not significant.
In Fig. 3, the distribution of ulna length against height
has been plotted separately for each subgroup by gender
and ethnicity. The dotted regression line represents the
‘predicted’ relationship according to the relevant MUST
equation. The least-squares best-fit regression for each sub-
group is shown by the solid line. Any extension of these
regression lines beyond the limits of the data distribution
must be viewed with caution. There are several features to
note. The majority of the values for Asian and Black sub-
groups fall below the predicted regression. The overall
Figure 2 The position of the forearm for measuring ulna length. trend for data from the White subgroups is more concor-

ª 2011 The Authors

Journal of Human Nutrition and Dietetics ª 2011 The British Dietetic Association Ltd. 123
 Estimating height from ulna length A. M. Madden et al.

Table 1 Age and measured height and ulna length in 194 subjects by sex and ethnicity
Age (years) Measured height (cm) Ulna length (cm)

Group n Mean (SD) Range Mean (SD) Mean (SD)

Men
Asian 30 31.7 (11.0) 21–55 170.9 (5.2) 26.6 (1.0)
Black 34 32.0 (10.3) 21–58 178.1 (7.3) 29.3 (1.5)
White 30 38.6 (12.5) 21–62 176.3 (7.7) 27.5 (1.2)
Significance AB < 0.001; AW = 0.009 AB < 0.001; AW = 0.021; BW < 0.001
Women
Asian 30 26.2 (5.4) 21–45 157.7 (4.7) 24.7 (0.7)
Black 35 32.6 (8.9) 21–57 164.0 (5.9) 26.3 (1.8)
White 35 35.7 (11.7) 21–61 163.7 (6.2) 24.7 (1.4)
Significance AB = 0.018; AW < 0.001 AB < 0.001; AW < 0.001 AB < 0.001; BW < 0.001

Significance: analysis of variance (Bonferroni post hoc test).

AB, Asian–Black; AW, Asian–White; BW, Black–White.

Table 2 Correlation between measured height and ulna length and difference between measured and predicted height in 194 subjects by sex and ethnicity
Correlation
between measured Difference in height (cm)
height and ulna Regression coefficients Predicted height (cm) predicted – measured

Group r P value Constant a Slope (95% CI) b Mean (SD) Mean (SD) 95% LA

Men
Asian 0.43 0.017 109.4 2.31 (0.46, 4.17) 175.0 (3.5) 4.06 (4.87) )5.5, 13.6
Black 0.68 <0.001 77.9 3.42 (2.10, 4.73) 184.8 (5.2) 6.70 (5.33) )3.8, 17.2
White 0.62 <0.001 66.6 4.00 (2.06, 5.92) 178.2 (4.4) 1.86 (6.07) )10.0, 13.8
Women
Asian 0.11 0.570 140.8 0.68 ()1.75, 3.11) 164.0 (2.1) 6.37 (4.90) )3.2, 16.0
Black 0.61 <0.001 111.5 2.00 (1.07, 2.93) 168.4 (4.9) 4.44 (4.87) )5.1, 14.0
White 0.63 <0.001 96.3 2.72 (1.54, 3.91) 164.1 (4.0) 0.42 (4.79) )9.0, 9.8

r, correlation coefficient; Cl, confidence interval; LA, limits of agreement.

dant with the MUST prediction. Among the Black men, predicted height and measured height has been further
there is evidence of a linear distribution of measured height explored. Because the data comprise two independent
that follows the same slope as the data from the MUST variables and not two indirect measurements of the same
equation, although typically 5 cm lower across the range of variable, conventional ‘limits of agreement’ analysis are
ulna lengths. The data from the Asian subgroups, especially not directly applicable. However, the difference for each
among the women, exhibit a more clustered pattern. These participant (predicted height minus measured height) can
visual patterns are supported by the regression coefficients be interpreted as ‘residuals’ using the MUST equation as
and the difference between the measured and predicted the baseline for the calculation of these residuals. When
height in the six subgroups (Table 2). plotted against ulna length, it is evident that the magni-
The predicted mean height for Asian and Black sub- tude of these differences is not unduly affected by varia-
groups typically exceeded the measured values by approx- tions in ulna length (Fig. 4). The mean difference and the
imately 4–6 cm. The mean differences between measured 95% ‘limits of agreement’ (SD of difference · 1.96) have
and predicted values for White men and women were been added to indicate that these differences are of con-
much smaller. The slope coefficients for the subgroups siderable magnitude and do not appear to be propor-
were relatively consistent with those used in the MUST tional to ulna length in the subgroups. The limits of
equations (3.60 for men and 2.77 for women) with the agreement are stated in Table 2.
exception of Asian women, where any upward trend in
increasing height with ulna length is modest and not
Discussion
significant.
To explore further limitations of using ulna length to The present study aimed to evaluate the MUST equations
predict height, the pattern of the difference between for predicting height from ulna length in healthy adults

ª 2011 The Authors

124 Journal of Human Nutrition and Dietetics ª 2011 The British Dietetic Association Ltd.
 A. M. Madden et al. Estimating height from ulna length

200 Asian men 180 Asian women

190
170

180

160

170

150
160

24 25 26 27 28 29 30 31 32 33 20 21 22 23 24 25 26 27 28 29 30

200 Black men 180 Black women

Measured height (cm)

190
170

180

160
170

150
160

26 27 28 29 30 31 32 33 20 21 22 23 24 25 26 27 28 29 30

200 White men 180 White women

190
170

180

160
170

150
160

Dotted line MUST regression: height = 79.2 + 3.60 (ulna length) Dotted line MUST regression: height = 95.6 + 2.77 (ulna length)

24 25 26 27 28 29 30 31 32 33 20 21 22 23 24 25 26 27 28 29 30
Ulna length (cm)

Figure 3 Distribution of measured height against ulna length in 194 participants by sex and ethnicity. The dotted line represents the regression
for the relevant Malnutrition Universal Screening Tool (MUST) equation and the solid line represents the regression for the values from the
participants.

from different ethnic groups and the results obtained If a value for height cannot be obtained by measure-
show differences in accuracy, particularly in non-White ment or calculation using a prediction equation, a self-
participants. This raises some concerns about the use of reported or estimated height might be used. Systematic
these equations in some groups of the population and, review of studies examining the accuracy of self-reported
therefore, whether there is a need to develop separate, height has identified an overall tendency to overestimate
ethnic-specific equations that would provide more accu- true height, probably reflecting societal norms to value
rate predicted values. To address this, there is a need to tallness (Gorber et al., 2007). The studies included exam-
consider alternative methods of predicting height and the ined diverse populations using different methodologies
clinical implications of using an inaccurate predicted and the degree of disparity between reported and mea-
height value in nutritional screening. sured values varied considerably. In nutritional screening,

ª 2011 The Authors

Journal of Human Nutrition and Dietetics ª 2011 The British Dietetic Association Ltd. 125
 Estimating height from ulna length A. M. Madden et al.

15 Asian men Asian women +1.96 SD

+ 1.96 SD
15

10

10
5 mean
mean
5
0

–1.96 SD 0
–5
–1.96 SD

–10 –5
Difference: predicted height – measured height (cm)

24 25 26 27 28 29 30 23 24 24 24 25 26 26

20 Black men Black women

+1.96 SD
15 +1.96 SD
15

10 10
mean
5
5 mean

–1.96 SD 0
–5

–5 –1.96 SD
–10

26 28 30 32 20 22 24 26 28 30

15 White men + 1.96 SD 15 White women

10 +1.96 SD
10

5
mean 5

0
mean
0
–5

–1.96 SD –5
–10
–1.96 SD
–15 –10

25 26 27 28 29 30 22 23 24 25 26 27 28
Ulna length (cm)

Figure 4 The difference between predicted and measured height plotted against ulna length in 194 participants by sex and ethnicity. The dotted
lines represent the mean difference and 95% ‘limits of agreement’ (i.e. ± 1.96 SD).

if a patient is unable or too ill to self-report their height, Stratton et al. (2003) found that using self-reported
this may be estimated by a healthcare professional. How- height and weight in MUST screening of patients with a
ever, in a study of 110 patients attending an emergency mean (SD) age of 56 (15) years and BMI of 27.9
department, Hendershot et al. (2006) reported that only (5.7) kg m)2 was unlikely to alter the malnutrition risk
41% of healthcare professionals were able to estimate category, although similar investigations are needed in
patients’ height to within 2.54 cm of measured values older patients and those with lower BMI values.
compared to 53% of patients themselves. This compares A number of alternative methods for predicting height
favourably with the results from the present study, where have been proposed. Regression equations derived from
height predicted using the MUST equations was within knee height have been most commonly explored in differ-
2.54 cm of individual measured values in only 30%, 15% ent ethnic groups, including non-Hispanic White, non-
and 30% of Asian, Black and White men and 17%, 37% Hispanic Black and Mexican-American adults aged
and 49% of Asian, Black and White women, respectively. ‡60 years (Chumlea et al., 1998), Korean adults (Hwang

ª 2011 The Authors

126 Journal of Human Nutrition and Dietetics ª 2011 The British Dietetic Association Ltd.
 A. M. Madden et al. Estimating height from ulna length

et al., 2009) and Caucasian adults aged 30–55 years (Ce- positioning of the leg may be difficult (Cook et al.,
reda et al., 2010), all of which have been cross validated. 2005), and it may be more acceptable for the person
Chumlea et al. (1998) did not present data for the differ- being measured to bare their fore arm rather than lower
ences between height measured and predicted using their leg and foot. An ulna length can be measured using a
equations in a total validation population of approxi- standard anthropometric tape, which is cheaper and
mately 2375 adults but stated that the root mean square more available than an anthropometric calliper that has
error values varied between 3.45 for Mexican-American been used for knee height in some studies (Chumlea
women and 4.18 for non-Hispanic white women. Hwang et al., 1998; Cereda et al., 2010).
et al. (2009) reported mean differences between measured The participants studied in the present investigation,
and predicted height of <1 cm in all validation subgroups which has a small sample size compared to other anthro-
(total number 1022), with intraclass correlations between pometric validation studies, were recruited from a popu-
0.88 (0.83–0.91) for post-menopausal women and 0.92 lation that can be considered as predominantly healthy
(0.92–0.94) for men. The differences between height mea- because they were attending a university to study or
sured and predicted by Cereda et al. (2010) were also very work. Clearly, these are neither a representative sample of
small (<1 cm) in a validation population sample of 120 the whole population, nor of the people who are most
with root mean square error value of 3.2 and 95% confi- likely to be undergo nutritional screening, and this is a
dence intervals of )6.1 and 6.5 cm (combined data for significant limitation. As a result, the numerical data from
men and women). This range is greater (i.e. less precise) the present study should not be extrapolated and used to
than in any of the groups in the present study, although ‘correct’ predicted values of height obtained from the
the difference is smaller (i.e. more accurate) than MUST equations in a hospitalised population. However,
observed in Asian or Black men or women (Table 2). it is reasonable to conclude that the absence of a signifi-
The clinical implications of the results from the present cant relationship between ulna length and height observed
study can be explored by extrapolating ‘worse case scenar- in the Asian women who participated might be found in
ios’ in hypothetical screening. We found that ulna length other groups of Asian women. As such, further studies
was in the range 23.3–26.0 cm among Asian women and are required to develop more accurate equations using
that it was weakly associated with measured height. If we ulna length in a non-White population and to explore
apply this range of ulna lengths to an Asian woman alternatives to ulna length in Asian women.
weighing 50 kg and with a measured height of 157.7 cm In conclusion, the findings obtained in the present
(i.e. study mean value from Table 1), her calculated BMI study indicate that the MUST equations for predicting
would be in the range 17.8–19.5 kg m)2 instead of the height from ulna length in adults provide useful estimates
true value of 20.1 kg m)2 (i.e. she would be identified as of height in White healthy volunteers but overestimate
under nourished when she may not be). However, if an height in Asian and Black healthy volunteers. The absence
Asian woman with the same measured height and weigh- of a significant relationship between measured height and
ing 75 kg was screened using a value for height predicted ulna length raises particular concerns about the use of
from the study data, her calculated BMI would be in the prediction equations in Asian women.
range 26.7–29.3 kg m)2 instead of the true value of
30.2 kg m)2 (i.e. she would not be categorised as obese
Acknowledgments
when she actually is). These two opposite hypothetical sit-
uations use standard World Health Organization BMI The authors are grateful to the healthy volunteers who
cut-off points, which may not be appropriate in an Asian participated in the study.
population (World Health Organization, 2004), but illus-
trate the potential consequences that may arise when Conflict of interests, source of funding and
inappropriate height prediction equations are used and, authorship
in this case, how Asian women may be over-diagnosed as The authors declare that there are no conflicts of interest.
under nourished, whereas some who are overweight or No external funding was received to undertake this study.
obese may not be identified. AMM conceived the original idea, contributed to study
No prediction formula derived from regression equa- design and drafted the manuscript. TT contributed to
tions will be able to provide estimates that are 100% study design, undertook all the data collection and con-
accurate and the practicality of the method for collecting tributed to the analysis and manuscript. DJS undertook
data must be weighed against potential inaccuracies. The the data analysis and contributed to the manuscript. All
prediction of height from ulna length offers some authors critically reviewed the manuscript and approved
advantages over the use of knee height measurement, the final version submitted for publication.
particularly in bed-bound patients where correct

ª 2011 The Authors

Journal of Human Nutrition and Dietetics ª 2011 The British Dietetic Association Ltd. 127
 Estimating height from ulna length A. M. Madden et al.

Hickson, M. & Frost, G. (2003) A comparison of three meth-

References
ods for estimating height in the acutely ill elderly popula-
Agnohotri, A.K., Kachhwaha, S., Jowaheer, V. & Singh, A.P. tion. J. Hum. Nutr. Diet. 16, 13–20.
(2009) Estimating stature from percutaneous length of tibia Hwang, I.C., Kim, K.K., Kang, H.C. & Kang, D.R. (2009)
and ulna in Indo-Mauritian population. Forensic Sci. Int. Validity of stature-predicted equations using knee height for
187, 109.e1–109.e2. elderly and mobility impaired persons in Korea. Epidemiol.
Bassey, E.J. (1986) Demi-span as a measure of skeletal size. Health 31, e2009004.
Ann. Hum. Biol. 13, 499–502. Kirk, S.F.L., Hawke, T., Sandford, S., Wilks, Z. & Lawrenson,
Brown, J.K., Whittemore, K.T. & Knapp, T.R. (2000) Is arm S. (2003) Are the measures used to calculate BMI accurate
span an accurate measure of height in young and middle- and valid for use in older people? J. Hum. Nutr. Diet. 16,
age adults? Clin. Nurs. Res. 9, 84–94. 366–367.
Cereda, E., Bertoli, S. & Battezzati, A. (2010) Height prediction Launer, L.J. & Harris, T. (1996) Weight, height and body mass
formula for middle-aged (30–55 y) Caucasians. Nutrition 26, index distributions in geographically and ethnically diverse
1075–1081. samples of older persons. Age Ageing 25, 300–306.
Cheng, J.C., Leung, S.S., Chiu, B.S., Tse, P.W., Lee, C.W., Mohanty, S.P., Babu, S.S. & Nair, N.S. (2001) The use of arm
Chan, A.K., Xia, G., Leung, A.K. & Xu, Y.Y. (1998) Can we span as a predictor of height: a study of South Indian
predict body height form segmental bone length measure- women. J. Orthop. Surg. 9, 19–23.
ments? A study of 3,647 children J. Pediatr. Orthop. 18, Noppa, H., Andersson, M., Bengtsson, C., Bruce, A. & Isaks-
387–393. son, B. (1980) Longitudinal studies of anthropometric data
Chumlea, W.C., Guo, S.S., Wholihan, K., Cockram, D., Kucz- and body composition: the population study of women in
marski, R.J. & Johnson, C.L. (1998) Stature prediction equa- Goteborg, Sweden. Am. J. Clin. Nutr. 33, 155–162.
tions for elderly non-Hispanic white, non-Hispanic black, Office for National Statistics. (2006) Census 2001. A Guide to
and Mexican-American persons developed from NHANES Comparing 1991 and 2001 Census Ethnic Group Data.
III data. J. Am. Diet. Assoc. 98, 137–142. Available at: http://www.statistics.gov.uk/cci/article.asp?
Cook, Z., Kirk, S., Lawrenson, S. & Sandford, S. (2005) Use of id=1471 (accessed on 27 January 2011).
BMI in the assessment of malnutrition in older subjects: Reeves, S.L., Varakamin, C. & Henry, C.J.K. (1996) The rela-
reflecting on practice. Proc. Nutr. Soc. 64, 313–317. tionship between arm-span measurement and height with
de Lucia, E., Lemma, F., Tesfaye, F., Demisse, T. & Ismail, S. special reference to gender and ethnicity. Eur. J. Clin. Nutr.
(2002) The use of armspan measurement to assess the nutri- 50, 398–400.
tional status of adults in four Ethiopian ethnic groups. Eur. Ritz, P. (2004) Validity of measuring knee-height as an esti-
J. Clin. Nutr. 56, 91–95. mate of height in diseased French elderly persons. J. Nutr.
Elia, M. (2003) The ‘MUST’ Report. Nutritional Screening of Health Ageing 8, 386–388.
Adults: A Multidisciplinary Responsibility. Development and Ruston, D., Hoare, J., Henderson, L., Gregory, J., Bates, C.J.,
Use of the ‘Malnutrition Universal Screening Tool’ (‘MUST’) Prentice, A., Birch, M., Swan, G. & Farron, M. (2004)
for Adults. Redditch: Malnutrition Advisory Group (MAG), National Diet and Nutrition Survey: Adults Aged 19 To
a Standing Committee of the British Association of Paren- 64 Years. Vol. 4. Nutritional Status (Anthropometry and
teral and Enteral Nutrition (BAPEN). Blood Analytes), Blood Pressure and Physical Activity. Lon-
Gauld, L.M., Kappers, J., Carlin, J.B. & Robertson, C.F. (2004) don: The Stationary Office.
Height prediction from ulna length. Dev. Med. Child Neurol. Steele, M.F. & Chenier, T.C. (1990) Arm-span, height, and age
46, 475–480. in black and white women. Ann. Hum. Biol. 17, 533–541.
Gorber, S.C., Tremblay, M., Moher, D. & Gorber, B. (2007) A Stratton, R.J., Dixon, R., Longmore, D., Stroud, M. & Elia, M.
comparison of direct vs self-report measures for assessing (2003) Effect of recalled weight and height on malnutrition
height, weight and body mass index: a systematic review. risk. Clin. Nutr. 22(Suppl. 1), S9–S10.
Obes. Rev. 8, 307–326. World Health Organization. (1995) Physical Status: The Use
Green, S.M. & Watson, R. (2005) Nutritional screening/assess- and Interpretation of Anthropometry. Geneva: World Health
ment tools for use by nurses: a comprehensive literature Organization, Report of a WHO Expert Committee, Techni-
review. J. Adv. Nurs. 50, 69–83. cal Report Series, vol. 854, pp. 1–452.
Han, T.S. & Lean, M.E. (1996) Lower leg length as an index of World Health Organization. (2004) Appropriate body-mass
stature in adults. Int. J. Obes. Relat. Metab. Disord. 20, 21– index for Asian populations and its implications for policy
27. and intervention strategies. Lancet 363, 157–163.
Hendershot, K.M., Robinson, L., Roland, J., Vaziri, K., Rizzo, Zverev, Y.P. (2003) Relationship between arm span and stature
A.G. & Fakhry, S.M. (2006) Estimating height, weight and in Malawian adults. Ann. Hum. Biol. 30, 739–743.
body mass index: Implications for research and patient
safety. J. Am. Coll. Surg. 203, 887–893.

ª 2011 The Authors

128 Journal of Human Nutrition and Dietetics ª 2011 The British Dietetic Association Ltd.