South African Journal of Clinical Nutrition

ISSN: 1607-0658 (Print) 2221-1268 (Online) Journal homepage: https://www.tandfonline.com/loi/ojcn20

Mid-upper arm circumference (MUAC) as a feasible

tool in detecting adult malnutrition

E Van Tonder, L Mace, L Steenkamp, R Tydeman-Edwards, K Gerber & D

Friskin

To cite this article: E Van Tonder, L Mace, L Steenkamp, R Tydeman-Edwards, K

Gerber & D Friskin (2019) Mid-upper arm circumference (MUAC) as a feasible tool in
detecting adult malnutrition, South African Journal of Clinical Nutrition, 32:4, 93-98, DOI:
10.1080/16070658.2018.1484622

To link to this article: https://doi.org/10.1080/16070658.2018.1484622

© 2019 The Author(s). Co-published by NISC

Pty (Ltd) and Informa UK Limited, trading as
Taylor & Francis Group

Published online: 26 Jul 2018.

Submit your article to this journal

Article views: 3725

View related articles

View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ojcn20
South African Journal of Clinical Nutrition 2019; 32(4):93–98
https://doi.org/10.1080/16070658.2018.1484622 SAJCN
ISSN 1607-0658 EISSN 2221-1268
Open Access article distributed under the terms of the
© 2019 The Author(s)
Creative Commons License [CC BY-NC 4.0]
http://creativecommons.org/licenses/by-nc/4.0 RESEARCH

Mid-upper arm circumference (MUAC) as a feasible tool in detecting adult

malnutrition
E Van Tondera, L Maceb, L Steenkampc, R Tydeman-Edwardsd, K Gerbere and D Friskinf
a
Department of Dietetics, Nelson Mandela University, Port Elizabeth, South Africa
b
Dietetics Department, Department of Health, Tayler Bequest Hospital, Matatiele, South Africa
c
Dietetics Department, Department of Health, Settlers Hospital, Grahamstown, South Africa
d
Dietetics Department, Department of Health, Port Elizabeth Provincial Hospital, Port Elizabeth, South Africa
e
Nursing Department, Nelson Mandela University, Port Elizabeth, South Africa
f
Department of Statistics, Nelson Mandela University, Port Elizabeth, South Africa
*Corresponding author, email: [email protected]

Objectives: This study aimed to expand on the limited South African malnutrition prevalence data and investigate the feasibility
of mid-upper-arm circumference (MUAC) as a malnutrition screening tool.
Design: A cross-sectional, multi-centre, descriptive design was adopted.
Setting: The study was undertaken in three tertiary public hospitals in the same urban area within the Eastern Cape of South
Africa.
Subjects: Adult hospitalised patients volunteered to participate (n = 266).
Methods: Data were collected using interviewer-administered questionnaires; obtaining anthropometric measurements; and
consulting medical files. For maximum accuracy of various MUAC cut-off points, receiver operating characteristic curves were
generated and area under the curve determined.
Results: Both body mass index (BMI) and MUAC identified 21% of participants as underweight or malnourished, and 39% as
overweight or obese. The Malnutrition Universal Screening Tool (MUST) found 23% at increased malnutrition risk. Nurses or
doctors detected and referred only 19% of underweight patients (BMI < 18.5 kg/m2), to dietetics services. Direct
measurements of BMI and MUST were unobtainable in 38% and 43% of patients respectively, whilst MUAC was obtainable
in 100%.
A statistically significant relationship (p < 0.001) exists between MUAC, BMI and MUST to detect malnutrition or malnutrition risk.
MUAC cut-offs for undernutrition were determined at < 23 cm (BMI < 16 kg/m2) and < 24 cm (BMI < 18.5 kg/m2), respectively, for
the study’s population groups.
Conclusion: Malnutrition prevalence was high in this study, but often unidentified, with only a fifth referred to dietetic services.
MUAC is a feasible method to identify adult malnutrition and should be considered as a malnutrition screening tool and key
nutritional status indicator in South African public hospitals.

Keywords: adults, hospitalised, malnutrition, mid-upper arm circumference, MUAC, referral rates, screening

Introduction and objectives on admission or thereafter.2,6 The most frequently cited reason
The unfavourable effects of malnutrition on healthcare costs, as for this is understaffing.2 Nurses also report low confidence
well as patient quality of life, are well established.1 A recent levels in calculating parameters such as body mass index (BMI)
study conducted in the Eastern Cape, South Africa,2 shows a and percentage weight loss,2 which often form part of malnu-
high prevalence of malnutrition in three public hospitals. The trition screening. The decision to refer a perceived malnourished
Malnutrition Universal Screening Tool (MUST) found 27% of par- patient to dietetic services may therefore be purely subjective,
ticipants to be underweight, 48% at high malnutrition risk and lacking substantiated parameters such as low BMI or significant
33% overweight or obese. unplanned weight loss.

A well-established system across the continuum of care is An uncomplicated method to identify malnourished patients for
necessary to minimise malnutrition risk, through efficient nutritional intervention is needed in South African public hospi-
implementation of prevention and treatment measures, includ- tals. Mid-upper-arm circumference (MUAC) may be a feasible
ing: malnutrition screening; triaging of nutrition care; and an evi- option as it is relatively easy to measure,7 requires little equip-
dence-based nutrition intervention.1 ment and calculations, and is transportable and inexpensive.7
Its diagnostic value in determining malnutrition among children
Nursing staff usually obtain information regarding appetite and is already a globally accepted practice, including in South
feeding requirements on admission, and have the ideal opportu- Africa.8 However, global MUAC cut-offs for adult malnutrition
nity to screen patients. In low resource areas, the lack of staff, classification have not yet been established.7 Namibia, Ethiopia,
skills, equipment and time can collectively hinder or inhibit Uganda and Zambia have developed their own cut-offs to
the screening process.2–5 Currently, nutritional screening is not screen for programme eligibility. However, the optimal cut-off
mandatory in South African public hospitals, with reported low values are unconfirmed.7 In pregnancy, MUAC correlates
incidences of weighing patients or measuring of height reported strongly with BMI up to 30 weeks’ gestation, and could

South African Journal of Clinical Nutrition is co-published by NISC (Pty) Ltd, Medpharm Publications, and Informa UK Limited
(trading as the Taylor & Francis Group)
94 South African Journal of Clinical Nutrition 2019; 32(4):93–98

postnatal (n = 2), some surgical (n = 2), some medical (n = 2),

burns (n = 2), urology (n = 1), psychiatric (n = 1), critical care (n
= 1), and high-care (n = 2) wards were excluded from the
study. The critical care and burns wards were excluded due to
the difficulty in obtaining accurate anthropometric measure-
ments and the urology ward formed part of the pilot study.
Some medical wards were excluded due to time constraints.

Anthropometric measurements and nutritional

screening
A calibrated SECA scale (Seca GmbH, Hamburg, Germany), SECA
stadiometer, and a non-stretchable measuring tape were used
to obtain anthropometric measurements,11 including: weight;
height; and MUAC. BMI, MUAC and MUST scores were used to
Figure 1: Method of sample selection identify malnutrition risk.7,12,13 Body mass index (< 18.5 kg/m2)
and MUAC (< 23 cm),14,15 were used to determine malnutrition
prevalence and MUST to determine malnutrition risk prevalence.
substitute for BMI estimation.9 MUAC measurement has been Demi-span or ulna length was used to determine height for non-
included in the Maternal Care Record nationally in South ambulatory patients, using validated formulae.11,12 Body mass
Africa.9 MUAC is also recommended for patients with ascites index was then estimated, from which the current weight was
or oedema in the legs or trunk, to gauge dry weight and BMI.10 extrapolated.11,12 To improve intra-observer reliability, dietetic
interns were trained in relevant measuring techniques by a
This article therefore proposes to expand MUAC as an adult mal- registered biokineticist, after which an acceptable level of tech-
nutrition screening tool, thereby improving malnutrition detec- nical error of measurement (TEM) was obtained.11
tion rates and nutritional intervention referrals.
The use of a validated screening tool (MUST) further improved
Methods the validity of the data.

Study design and participants Statistical analysis

A cross-sectional, multi-centre investigation using an explora- Statistica® (version 13) (www.statsoft.com) was used to analyse
tory descriptive research design was conducted in three the data. Descriptive statistics were used for categorical data,
Eastern Cape tertiary public hospitals within the same urban and means, standard deviations and ranges for continuous
area. The study consisted of Phase 1, which entailed an initial data. Groups were compared by means of chi-square tests16
screening phase determining malnutrition prevalence and exist- with a p-value < 0.05 considered as statistically significant.
ing dietetic service referral rates; and Phase 2, which determined
MUAC’s feasibility as a malnutrition screening tool and appli- Receiver operating characteristic (ROC) curves were generated
cable cut-off values. and area under the curve (AUC) calculated17 to determine
global accuracy measurements for MUAC cut-off points to clas-
Participants sify BMI classes (sub-sample i only). Youden’s18 index was used
An initial study sample A was recruited to determine malnu- to obtain optimal MUAC cut-off point accuracy.
trition prevalence in the Eastern Cape, including participants
on whom either direct or indirect (i.e. immobile patients) anthro-
Ethical considerations
pometric measurements could be obtained. A sub-sample (i)
Permission was granted by the Nelson Mandela University
was then extracted to contain only those participants from
Research Ethics Committee (Human) (H15-HEA-DIET-005), the
whom direct anthropometric measurements were obtained
Eastern Cape Department of Health, as well as the CEO of
and who were between 18 and 65 years of age (Figure 1).
each participating hospital.

Study sample A The Helsinki declaration guidelines were followed, including

Two trained dietetic interns collected data via interviewer-admi- voluntary written informed consent and maintenance of
nistered questionnaires, anthropometric measurements and confidentiality.19
consulting medical files. The questionnaires included: demo-
graphic and anthropometric information; disease condition;
and dietitian referral status during the current admission. The Results
face and content validity of the interviewer-administered ques- Study sample A
tionnaire were tested in a pilot study, involving six participants Study sample A consisted of n = 266 patients with a mean age of
in one of the hospitals’ urology wards, after which no changes 47.11 ± 15.4 years with a relatively even distribution of males
were required. All patients who met the inclusion criteria (≥ 18 (52%) and females (48%). The ethnic distribution was 68% (n
years of age, not pregnant, able to provide informed consent = 181) black, 26% (n = 69) coloured, 6% (n = 15) white and <
and without psychiatric illness) in the pre-selected wards were 1% (n = 1) Indian patients.
invited to participate in the study. Consecutive purposive
sampling was used in the following number of in-patient adult The sample distribution according to disease-specific wards was
wards: surgical (n = 2), medical (n = 7), orthopaedic (n = 2), oncol- as follows: surgical 17% (n = 45), medical 43% (n = 114), ortho-
ogy (n = 2), haematology (n = 1), cardiothoracics (n = 1) and neu- paedic 15% (n = 39), oncology/haematology 12% (n = 31), cardi-
rology (n = 1). All paediatrics (n = 10), antenatal (n = 2) and othoracic 10% (n = 26) and neurology 4% (n = 11).
Mid-upper arm circumference (MUAC) as a feasible tool in detecting adult malnutrition 95

Table 1: Malnutrition prevalence and risk according to nutrition-related which is reliant on advanced clinical skills, in which non-nutri-
indicators tional health professionals are not specifically trained. A
further challenge was to obtain patients’ usual weight in order
Prevalence of malnutrition according to BMI
to calculate percentage weight loss, an important parameter
n Percentage (%) for determining a patient’s malnutrition risk.20 The latter was
Underweight (< 18.5 kg/m2) 57 21.4 due to patients’ weights not routinely being documented in
Normal weight (18.5–24.9 kg/m2) 105 39.5 their medical files, and many patients not knowing their usual
Overweight (25.0–29.9 kg/m2) 45 16.9 weight. Consequently, 43.2% of participants’ MUST scores
Obesity (30.0–39.9 kg/m2) 49 18.4
could not be calculated. Conversely, MUAC was obtainable in
100% of the sample, which made it a more feasible screening
Morbidly obese (≥ 40 kg/m²) 10 3.8
tool for non-nutritional healthcare professionals such as nurses.
Total 266 100
Prevalence of malnutrition according to mid-upper arm circumference
(MUAC) cut-offs Study sub-sample (i)
Direct anthropometric measurements were obtained in 181 par-
n Percentage (%) ticipants from study sample A, which was included in sub-sample
Normal (MUAC > 23 cm) 208 78.2 (i). Fifteen of these patients were > 65 years of age, and were
excluded. Therefore 166 participants were included in sub-
Malnutrition (MUAC ≤ 23 cm) 58 21.8
sample (i), to statistically compare MUAC values with BMI and
Total 266 100
MUST, and to determine MUAC cut-off values (see Figure 1).
Prevalence of malnutrition according to MUST score

Malnutrition risk n Percentage (%) MUAC as a feasible screening tool

The performance of MUAC as a screening tool to classify partici-
Low Risk (Score = 0) 89 33.5
pants as malnourished, was further explored (based on data
Medium Risk (Score = 1) 19 7.1
from sub-sample (i)) for feasibility and accessibility, in compari-
High Risk (Score ≥ 2) 43 16.2 son with the barriers experienced with both BMI and MUST. The
Unable to determine 115 43.2 mean age of sub-sample (i) (n = 166) was 41.72 ± 12.8 years with
Total 266 100 a relatively even distribution of males (52%) and females (48%).
BMI: body mass index; MUST: Malnutrition Universal Screening Tool. Sub-sample (i) consisted of 71% (n = 118) black, 25% (n = 42)
coloured, 3% (n = 5) white and <1% (n = 1) Indian patients. The
ability of MUAC to accurately detect patients with malnutrition
Outcomes of the nutritional status indicators (MUAC, BMI and
or malnutrition risk was statistically compared with BMI and
MUST) of study sample A are summarised in Table 1. Some
MUST respectively.
21% (n = 57) of participants were classified as underweight
(BMI < 18.5 kg/m2), whilst 39% (n = 104) were classified as either
overweight (BMI 25–29.9 kg/m2) or obese (BMI > 30 kg/m2). A statistically significant relationship (chi-square = 66.816; df = 4;
MUAC measurements showed that 22% (n = 58) of the sample p < 0.001) existed between MUAC and BMI in classifying under-
were malnourished with a measurement of ≤ 23 cm. According nourished patients, with Cramer’s V = 0.634, indicating large
to MUST, 23% (n = 158) of patients were at increased malnu- practical significance.
trition risk (MUST score ≥ 1), with 16% at high risk (MUST
score ≥ 2). Mid-upper arm circumference showed a significant strong corre-
lation with BMI (r = 0.93, p < 0.001). The relationship is illustrated
in Figure 2.
Referral to dietetics services
As shown in Table 2, only 19% of underweight patients accord-
ing to BMI, 22% of malnourished patients according to MUAC Similarly, a statistically significant relationship was found
and 21% of patients with a MUST score of ≥ 2 were referred to between MUAC and MUST (chi-square = 38.816; df = 2; p <
dietetics services. Similarly, only 14% of overweight or obese 0.001) with Cramer’s V = 0.623, also indicating large practical
patients were referred. significance.

Factors limiting malnutrition screening

Malnutrition screening was limited to mobility of participants for
direct measurements of weight and height to calculate BMI. Of
the 266 participants, 37.6% of BMIs had to be estimated,

Table 2: Referral rates to dietetics services based on nutritional status or

risk

Not referred to a
Referred to a dietitian dietitian
Factor
n Percentage (%) n Percentage (%)
2
BMI < 18.5 kg/m 11 19 46 81
BMI ≥ 24.9 kg/m2 15 14 89 86
MUAC ≤ 23 cm 13 22 45 78
MUST ≥ 2 9 21 34 79
BMI: body mass index; MUAC: mid-upper arm circumference; MUST: Malnutrition Figure 2: Relationship between body mass index (BMI) and mid-upper
Universal Screening Tool.
arm circumference (MUAC) according to gender
96 South African Journal of Clinical Nutrition 2019; 32(4):93–98

Table 5: Predicted MUAC cut-offs for the detection of malnutrition

according to gender

MUAC Sensitivity Specificity

BMI (kg/m2) (cm) AUC (%) (%)
Males:
< 16 Severely < 22.6 0.96 91 88.9
underweight
< 18.5 < 23.7 0.88 86.4 78.6
Underweight
≥ 25 > 29.0 0.97 92.3 95.9
Overweight
≥ 30 Obese > 29.9 0.97 100 91.6
Females:
< 16 Severely < 21.1 0.99 96.1 100
underweight
< 18.5 < 23.5 0.98 93.1 100
Underweight
≥ 25 > 28.0 0.94 95.7 81.8
Overweight
≥ 30 Obese > 29.4 0.96 100 82
Figure 3: ROC curve for MUAC classification of BMI < 16 kg/m2 (Sub- BMI: body mass index; MUAC: mid-upper arm circumference; AUC: area under the
sample i) ROC: Receiver operating curve, MUAC: mid-upper arm curve.
circumference

by the United Nations International Children’s Emergency

Table 3: Optimal MUAC cut-offs to detect malnutrition based on study Fund (UNICEF) to identify malnutrition in children (see Table 4).
sub-sample (i)
In addition to the proposed MUAC cut-off values for undernutri-
MUAC Sensitivity Specificity
BMI (kg/m2) (cm) AUC (%) (%) tion, which were similar to the results of Ferro-Luzzi and James,14
Chakraborty et al.20 and Sultana et al.,21 the current study pre-
< 16 Severely 22.6 0.97 91.6 91.7
underweight dicted MUAC cut-off values for the overweight population as
well.
< 18.5 23.7 0.92 89.3 82.9
Underweight
≥ 25 28.1 0.96 94.9 86.9 MUAC cut-offs according to gender and race
Overweight The MUAC cut-off points with the highest sensitivity and speci-
≥ 30 Obese 29.4 0.97 100 87.2 ficity to correspond with BMIs of < 16 kg/m2, < 18.5 kg/m2, ≥
BMI: body mass index; MUAC: mid-upper arm circumference; AUC: area under the 25 kg/m2 and ≥ 30 kg/m2, were also determined according to
curve. gender and race. The proposed cut-off values for each corre-
sponding BMI value are summarised in Tables 5 and 6. These
cut-off values were again rounded off for practical purposes
MUAC cut-offs to identify malnutrition risk and colour coded for ease of use (Table 7).
Receiver operating curves (ROC) were plotted and area under
the curve (AUC) was calculated to determine global accuracy
Table 6: Predicted MUAC cut-offs for the detection of malnutrition
measurements for using MUAC cut-off points to classify BMI according to race
classes. Figure 3 provides an example of the ROC for using
MUAC to classify a BMI of < 16 kg/m2. MUAC Sensitivity Specificity
BMI (kg/m2) (cm) AUC (%) (%)
Based on the coordinates of the curve, Youden’s index was used Black
to determine the optimal cut-off points (highest combined < 16 Severely < 23.5 0.97 85.5 100
sensitivity and specificity) to classify BMI’s of < 16 kg/m2, underweight
< 18.5 kg/m2, ≥ 25 kg/m2 and ≥ 30 kg/m2 (Table 3). < 18.5 < 23.7 0.92 90.7 85.7
Underweight
The MUAC values derived for the whole sample population were ≥ 25 > 28.1 0.95 92.9 85.5
rounded off to < 23 cm, < 24 cm, > 28 cm and > 29 cm and Overweight
colour coded for ease of classification purposes in the form of ≥ 30 Obese > 29.9 0.98 100 89.2
a MUAC tape, similar to what has previously been developed Coloured
< 16 Severely < 22.6 0.96 0.868 100
underweight
Table 4: Proposed MUAC tape for nutritional screening < 18.5 < 22.9 0.92 0.964 78.6
Underweight
2
BMI category (kg/m )
≥ 25 > 28.1 0.99 100 93.3
< 16 < 18.5 18.5–24.9 ≥ 25 > 29.9 Overweight
MUAC (cm) < 23.0 < 24.0 24.1–27.9 > 28.0 > 29.0 ≥ 30 Obese > 29.4 0.99 100 94.4
BMI: body mass index; MUAC: mid-upper arm circumference; MUST: Malnutrition BMI: body mass index; MUAC: mid-upper arm circumference; AUC: area under the
Universal Screening Tool. curve.
Mid-upper arm circumference (MUAC) as a feasible tool in detecting adult malnutrition 97

Table 7: Proposed MUAC cut-off values according to gender and race The current study confirms previous findings 7–9,14,15,20,21,29–32
that MUAC correlated well with BMI, and could be used to ident-
BMI category (kg/m2) ify patients as underweight and overweight. In addition, the
MUAC (cm) < 16 < 18.5 18.5–24.9 ≥ 25 > 29.9 current study also demonstrated that MUAC correlated well
Males <23 <24 24.1–28.9 >29 >30 with the validated malnutrition screening tool (MUST),10 albeit
Females <21 <24 24.1–27.9 >28 >29 to a lesser extent than BMI. Therefore, MUAC may be a feasible
Black <24 <24 24.1–27.9 >28 >30
nutritional screening tool to be used in both clinical and other
settings, especially where frontline professions (e.g. nurses) do
Coloured <23 <23 23.1–27.9 >28 >29
no not routinely use nutritional indicators, e.g. weight, height,
BMI: body mass index; MUAC: mid-upper arm circumference.
BMI and percentage weight loss.

The insufficient number of white and Indian patients was a limit- Inhibiting factors to malnutrition screening methods
ation of the study, which inhibited the prediction of MUAC cut- A limitation to the current study was that more than a third of
offs for these groups. Further research is needed in these groups. patients in the initial study, sample A were not ambulant, there-
fore direct measures of weight or height could not be measured.
BMI had to be estimated in such circumstances and could have
Discussion
impacted on the accuracy of the findings. Some 43% of partici-
pants in study sample A were unable to recall their usual body
Nutritional status of the sample weight. The data were also unavailable from the medical file,
The world is faced with a double burden of disease encompass- which made accurate calculation of percentage weight loss
ing both spectrums of malnutrition (undernutrition and overnu- impossible, thereby inhibiting MUST score calculations in a
trition) within the same communities.22 The current study large sample of patients. The current study’s challenges com-
illustrated that the double burden of disease also exists in the pared with those experienced by Powell-Tuck and Hennesy,29
hospitalised population that was studied. On screening, 21% where BMI was obtainable in only 44% of their study population,
were classified as underweight (BMI < 18.5 kg/m2), whilst 23% and recent percentage weight loss in only 33%, but MUAC in
were at increased malnutrition risk (score ≥ 1) by MUST. Accord- 95% of their patients.
ing to BMI, a staggering 39.1% of participants were overweight
or obese. The malnutrition prevalence rates found in this study
From the above data it can be seen that the use of screening
agree with previously reported global prevalence of hospital
tools, where more advanced anthropometric skills and calcu-
malnutrition, varying between 13 and 78%23, as well as a
lations may be required, is not currently feasible in South
recent study in the Eastern Cape (2016) in the same study
African public hospitals. Undergraduate and in-service training
population.2
of frontline workers (e.g. nurses) will have to be up-skilled first.
Additional inhibiting factors are the time constraints and avail-
Referral to dietetic services ability of nursing staff and equipment,2,3,5 thereby reinforcing
Only approximately 20% of underweight and high malnutrition the need for a relatively easy and quick malnutrition risk screen-
risk patients were referred to dietetic services, and even fewer ing tool in the acute setting, such as MUAC.10
overweight or obese patients. This is lower than an Australian
study (2011) that found 36% of malnourished patients were Since MUAC is already used as a screening tool in children under
referred to a dietitian in a tertiary teaching hospital.24 Since five years old and pregnant women in South Africa, it may be
the current lack of human resources in the public health easier for nursing staff to adopt this method of blanket nutri-
service impedes the establishment and implementation of tional screening of adult patients.9 Mid-upper arm circumfer-
blanket malnutrition screening and referral policies or pro- ence is not without its limitations, and does not account for
cedures, many patients in the three public hospitals who recent significant unplanned weight loss, an important indicator
required at least some form of nutrition intervention remained of nutritional risk.33 However, even after taking the above-men-
undetected. tioned inhibiting factors into consideration, MUAC remains feas-
ible as a stepping stone towards nutritional screening in South
By proposing the MUAC measurement as a malnutrition screen- African public hospitals.
ing method in South African public hospitals, many more
patients could be screened, detected and referred to a dietitian
Predicting MUAC cut-offs
for a more thorough assessment and nutritional intervention.
The strong correlation between BMI and MUAC allowed for
This can lead to multiple benefits for both the patient and insti-
optimal MUAC cut-offs to be derived for identifying malnour-
tution,25 but will also need a larger nutrition workforce to cope
ished patients. Depending on a facility’s resources, a MUAC cut-
with an increased patient load.
off of either < 23 cm (corresponding to a BMI of < 16 kg/m2) or
< 24 cm (corresponding to a BMI < 18.5 kg/m2) is proposed (see
Although there has been progress in the development of South Table 3). More specific cut-offs have been proposed that consider
African national policies26,27 and guidelines28 that emphasise demographic characteristics (see Table 6), again depending on
nutrition screening, more collaboration with other professional facilities’ preferences and resources.
societies and organisations is needed in terms of screening,
management and referral pathways of malnourished or nutri-
The current study’s cut-offs are similar to those proposed by
tionally at-risk patients.
Chakraborty et al. 20 (< 24 cm in males as an indicator of malnu-
trition and illness), Sultana et al. 21 (< 25 cm in males and < 24 cm
Relationship of MUAC with BMI and MUST in females to correspond to a BMI < 18.5 kg/m2), Brito et al.13
The use of the MUAC as a screening tool to assess adult nutri- (≤22.5 cm to correspond highly to a BMI of <18.5 kg/m2 and
tional status, especially in low resource areas, has steadily Ferro-Luzzi and James14 (< 23 cm for men and < 22 cm for
increased since its first application in the late 1900s.7,14,15 women based on a BMI < 18.5 kgm2)7.
98 South African Journal of Clinical Nutrition 2019; 32(4):93–98

Sultana et al. 21 (n = 1373) and Chakraborty et al. 20 (n = 205) both 11. Lee RD, Nieman DC. Nutritional Assessment. 5th ed. McGraw-Hill
used ROC and Youden’s index, similar to the current study, to Publ, Boston; 2010. ISBN: 9780071267724.
determine optimal MUAC cut-offs, whilst Ferro-Luzzi and 12. The British Association for Parenteral and Enteral Nutrition. A guide
James14 analysed and extrapolated anthropometric data from to the malnutrition universal screening tool for adults. 2001.
13. Brito NB, Suarez Llanos JP, Fuentes Ferrer M, et al. Relationship
adults with normal nutritional statuses across nine adult
between mid-upper arm circumference and body mass index in
surveys from Asia, Africa and the Pacific. inpatients. PLoS ONE. 2016;11(8):e0160480.
14. Ferro-Luzzi A, James WPT. Adult malnutrition: simple assessment
The current study was further able to determine MUAC cut-offs techniques for use in emergencies. Br J Nutr. 1996;75:3–10.
for overweight and obese patients, due to the high prevalence in 15. James WP, Mascie-Taylor, GC, Norgan, NG, et al. The value of arm cir-
the population group. Cut-offs of > 28 cm or > 29 cm (corre- cumference measurements in assessing chronic energy deficiency in
sponding to a BMI of ≥ 25 and 29.9 kg/m2 respectively) are pro- third world adults. Eur J Clin Nutr. 1994;48(12):883–94.
posed (see Table 3), again depending on facilities’ preferences 16. McHugh ML. The Chi-square test of independence. Biochem Med
and resources. Gender and race specific cut-offs have also (Zagreb). 2013;23(2):143–149.
17. Mandrekar JN. Receiver operating characteristic curve in diagnostic
been determined (see Table 6).
test assessment. J Thorac Oncol. 2010;5(9):1315–1316.
18. Youden WJ. Index for rating diagnostic tests. Cancer. 1950;3:32–35.
Conclusion doi:10.1002/1097-0142(1950)3:1<32::AID-CNCR2820030106 > 3.0.
CO;2-3
The high prevalence of malnutrition, both as underweight and
19. World Medical Association. Declaration of Helsinki ethical principles
overweight, illustrated that the double burden of disease was
for medical research involving human subjects. JAMA. 2013;310
also endemic in this study population. Patients falling in either (20):2191–2194. doi:10.1001/jama.2013.281053
of these categories were often unidentified and untreated for 20. Chakraborty R, Bose K, Koziel S. Use of mid-upper arm circumfer-
malnutrition, with only a fifth being referred to dietetic services. ence in determining undernutrition and illness in rural adult
Further research is recommended to validate the proposed Oraon men of Gumla district, Jharkhand, India. Rural Remote
MUAC cut-offs in hospitalised patients, but also to extend it to Health. 2011;11:1754.
other settings, e.g. community clinics and long-term care set- 21. Sultana T, Karim MN, Ahmed T, et al. Assessment of undernutrition of
tings. As the selection of patients was not randomised, the Bangladeshi adults using anthropometry: can body mass index be
results could not be generalised to all hospitalised patients. replaced by mid-upper-arm-circumference? PLoS ONE. 2015;10(4):
e0121456.
Research on larger samples of the white and Indian populations
22. World Health Organization. Double burden of malnutrition. 2016.
are needed, to develop MUAC cut-offs for these groups.
[Online]. http://www.who.int/nutrition/double-burden-malnutrition/
en/.
The use of MUAC should be considered as a first step to manda- 23. Kubrak C, Jensen L. Malnutrition in acute care patients: a narrative
tory nutritional screening and a key nutritional status indicator in review. Int J Nurs Stud. 2007;44(6):1036–1054.
South African public hospitals and other settings. 24. Gout BS, Barker LA, Crowe TC. Malnutrition identification, diagnosis
and dietetic referrals: are we doing a good enough job? Nutr Diet.
2009;66:206–211.
References 25. Tappenden KA, Quatrara B, Parkhurst ML, et al. critical role of nutri-
1. Mitchell H, Porter J. The cost-effectiveness of identifying and treating tion in improving quality of care: an interdisciplinary call to action
malnutrition in hospitals: a systematic review. J Hum Nutr Diet. to address adult hospital malnutrition. J Acad Nutr Diet.
2016;29:156–164. 2013;113:1219–1237.
2. Van Tonder E, Gardner L, Cressey S, et al. Adult malnutrition: preva- 26. Department of Health. Roadmap for nutrition in South Africa 2013–
lence and use of nutrition-related quality indicators in South 2017. South Africa: Department of Health, Pretoria; 2013.
African public-sector hospitals. South Afr J Clin Nut. 2018. 27. Health Professions Council of South Africa. Professional board for die-
[Online].https://doi.org/10.1080/16070658.2017.1410003 tetics and nutrition, 2017. The scope and competencies of the new
3. Evans L, Best C. Accurate assessment of patient weight. Nurs Times. dietitian nutritionist in the well-being of the South African population
2004;110(12):12–14. with associated assessment criteria for entry-level dietitian-nutrition-
4. Porter J, Raja R, Cant R, et al. Exploring issues influencing the use of ist. Draft report. South Africa.
the Malnutrition Universal Screening Tool by nurses in two Australian 28. Department of Health. National enteral nutrition practice guidelines
hospitals. J Hum Nutr Diet. 2009;22:203–209. for adults. Department of Health, Pretoria; 2016.
5. Roller RE, Eglseer D, Eisenberger A, et al. The Graz Malnutrition 29. Powell-Tuck J, Hennessy EM. A comparison of mid upper arm
Screening (GMS): a new hospital screening tool for malnutrition. circumference, body mass index and weight loss as indices of
Br. J. Nutr. 2016;115:650–657. undernutrition in acutely hospitalized patients. Clin Nutr. 2003;22
6. Prins A, Visser J. Immunonutrition: a South African perspective. S Afr J (3):307–312.
Clin Nutr. 2012;25(3):S1–S12. 30. Mazicioglu MM, Yalkin BM, Ustunbas HB, et al. Anthropometric risk
7. Tang A, Dong K, Deitchler M, et al. Use of cut-offs for mid-upper arm factors for elevated blood pressure in adolescents in turkey aged
circumference (MUAC) as an indicator or predictor of nutritional and 11–17. Pediatr. Nephrol. 2010;25(11):2327–34.
health related outcomes in adolescents and adults: a systematic 31. Martin AC, Pascoe EM, Forbes DA. Monitoring nutritional status accu-
review. Food and Nutrition Technical Assistance (FANTA). 2013. rately and reliably in adolescents with anorexia nervosa. J Paediatr
https://www.fantaproject.org/sites/default/files/resources/MUAC% Child Health. 2009;45(1–2):53–7.
20Systematic%20Review%20_Nov%2019.pdf 32. Kaushik B, et al. Relationship of Income with Anthropometric
8. Collins S, Duffield A, Myatt M. Assessment of nutrition status in emer- indicators of chronic energy deficiency among adult female
gency affected populations. RNIS Supplement. 2000. [Online]. http:// slum dwellers of Midnapore Town. J Hum Ecol. 2007;22(2):171–
unsystems.org/scn/archivs/adults/begin.htm#contents. 176.
9. Fakier A, Petro G, Fawcus S. Mid-upper arm circumference: a surro- 33. Loh K, Vrens M, Gerritsen A, et al. Unintentional weight loss is the
gate for body mass index in pregnant women? S Afr Med J. most important indicator of malnutrition among surgical cancer
2017;107(7):606–610. patients. Neth J Med. 2012;70(8):365.
10. Elia M, Russell C, Stratton R, et al. The ‘MUST’ explanatory booklet. The
British association for parenteral and enteral nutrition. 2011. Received: 4-03-2017 Accepted: 26-05-2018