Annals of Tropical Paediatrics (2008) 28, 87–101

Protein metabolism in severe childhood malnutrition

FAROOK JAHOOR, ASHA BADALOO*, MARVIN REID* &

TERRENCE FORRESTER*

USDA/Agricultural Research Service, Children’s Nutrition Research Center, Department of Pediatrics,

Baylor College of Medicine, Houston, Texas, USA and *Tropical Metabolism Research Unit, Tropical
Medicine Research Institute, University of the West Indies, Mona, Kingston, Jamaica

(Accepted March 2008)

Abstract The major clinical syndromes of severe childhood malnutrition (SCM) are marasmus (non-
oedematous SCM), kwashiorkor and marasmic-kwashiorkor (oedematous SCM). Whereas treatment of marasmus
is straightforward and the associated mortality is low, kwashiorkor and marasmic-kwashiorkor are difficult to treat
and have high morbidity and mortality rates. Despite extensive research, the pathogenic factors which cause a child
to develop the oedematous instead of the non-oedematous form of SCM in response to food deprivation are still
not clear. Over the years, two attractive hypotheses have been put forward. The first proposed that a dysadaptation
in protein metabolism was involved and the second proposed that free radical damage of cellular membranes might
be involved. To address aspects of these hypotheses, in this article we have reviewed work done by our group and by
others on protein metabolism and pro-oxidant/anti-oxidant homeostasis in children with the oedematous and non-
oedematous syndromes of SCM. A significant finding is that when there is chronic food deprivation children with
non-oedematous SCM can maintain body protein breakdown at the same rate as when they are well nourished, but
children with oedematous SCM cannot. The slower protein breakdown rate of children with oedematous SCM
reduces the supply of most amino acids, resulting in decreased availability for the synthesis of plasma proteins
involved in nutrient transport and the acute phase response to infection. Another consistent finding is that children
with oedematous SCM have oxidative stress as there is evidence of oxidant-induced cellular damage and impaired
synthesis of the primary cellular anti-oxidant glutathione.

Introduction kwashiorkor and marasmic-kwashiorkor are

The major clinical syndromes of severe
childhood malnutrition (SCM) are maras-
mus (non-oedematous SCM), kwashiorkor
and marasmic-kwashiorkor (oedematous
SCM). Whereas treatment of marasmus is
straightforward and the associated mortality
is low, kwashiorkor and marasmic-kwashior-
kor are difficult to treat and have high
morbidity and mortality rates.1,2 While
wasting characterises all syndromes of SCM,
Reprint requests to: Professor Farook Jahoor,
Children’s Nutrition Research Center, Department of
Pediatrics, Baylor College of Medicine, 1100 Bates
Street, Houston, TX 77030, USA. Fax: z001 713 798
7119; email: fjahoor@bcm.tmc.edu
# 2008 The Liverpool School of Tropical Medicine
DOI: 10.1179/146532808X302107
characterised additionally by oedema, anor-
exia, dermatitis, hypopigmented skin and
hair, neurological abnormalities, and a
higher incidence of hepatic steatosis.
Although an overall deficiency of dietary
energy and protein plus clinical and/or
subclinical deficiencies of most micronu-
trients underlie all syndromes, the aetiology
of oedematous SCM is more complex and
might involve the added physiological
stresses of environmental toxins and/or
infections.3 With respect to the pathogen-
esis of oedematous SCM, despite extensive
research, the underlying mechanism(s)
that cause these additional pathophysiolo-
gical changes remain unclear. About 40
years ago, a hypothesis was proposed that
88 F. Jahoor et al.
kwashiorkor results from a dysadaptation of
protein and lipid metabolism to chronic
food deprivation,4,5 and later, in the 1980s,
observations that children with oedematous
SCM had lower blood glutathione (GSH)
and evidence of oxidant-induced cellular
damage led to another hypothesis that free
radical damage of cellular membranes
plays an important role in the pathogenesis
of the disease.6 In this review, we shall
focus on protein metabolic and pro-oxidant-
anti-oxidant differences between the oede-
matous and non-oedematous syndromes of
SCM to better understand their divergent
clinical courses and outcomes.
Protein and Amino-Acid Metabolism
In the dysadaptation hypothesis proposed
by Whitehead & Alleyne,5 they reasoned
that adaptation to food deprivation, as seen
in marasmus, involved the gradual wasting
of muscle and fat to provide energy for
survival and amino acids to protect various
metabolic processes such as synthesis of
proteins essential for homeostasis. In con-
trast, in kwashiorkor, tissue catabolism does
not occur to the same extent, perhaps
because sufficient carbohydrate is consumed
for energy maintenance. Hence, there is an
insufficient supply of amino acids and fatty
acids from muscle and adipose tissue break-
down to fill the shortage created by inade-
quate dietary intakes. As a consequence,
there is a shortage of amino acids to
synthesise the proteins, peptides and bio-
molecules that are necessary for adaptation
to chronic inadequate food intake. Although
this hypothesis was attractive at the time, it
was never tested. In our ongoing studies of
protein and amino-acid metabolism in
children with SCM, we have aimed to
determine: (i) whether there are differences
in protein breakdown rates between chil-
dren with oedematous and non-oedematous
SCM and, if so, (ii) whether these are
associated with slower production rates
and smaller pools of amino acids. Also,
(iii) are they associated with slower synthesis
rates of plasma proteins?
Protein breakdown rate
To address the first point, we used stable
isotope tracer methods to measure endo-
genous leucine flux, an index of whole-body
protein breakdown rate, in children with
oedematous and non-oedematous SCM in
the malnourished and recovered states.
Studies were conducted in the fed and
post-absorptive states within the resuscita-
tive phase of treatment at ,3 days after
admission (malnourished state) and at ,55
days post-admission after a weight-for-
length of at least 90% of expected had been
reached (recovered state). Leucine flux
was slower in the oedematous than in the
non-oedematous SCM children in the mal-
nourished state and, compared with the
corresponding values at recovery, leucine
flux was markedly slower in the oedematous
group but not in the non-oedematous group
(Table 1). Interestingly, when the children
were fully recovered, leucine flux was faster
(p,0.05) in the oedematous group than in
the non-oedematous group. The slower
leucine flux of the oedematous SCM chil-
dren corroborates the findings of Manary
et al.7 who reported that leucine flux from
protein breakdown was 55% slower in
children with kwashiorkor than in those
with marasmus. In other studies,8,9 we have
reported slower endogenous fluxes of two
other essential amino acids, phenylalanine
and methionine, indicating slower protein
breakdown rate in oedematous than in non-
oedematous SCM. At first glance these
findings seem to support the proposal of a
dysadaptation in protein metabolism in
kwashiorkor.4,5
The finding that protein breakdown rate
increased significantly in the oedematous
group when they recovered, but did not
change in the non-oedematous group, sug-
gests that, when there is chronic food
deprivation, children with non-oedematous
SCM can maintain body protein breakdown
 Protein metabolism & severe malnutrition 89

at the same rate as when they are well
nourished, but children with oedematous
SCM cannot. Is this difference, however,
really due to a ‘dysadaptation’ in protein
breakdown in children with oedematous
SCM? The observation that protein turn-
over is slower in the malnourished state than
in the recovered state in children with SCM
has been interpreted as a necessary adapta-
tion to conserve energy and protein and,
hence, prolongs survival in the face of
chronic reduced food intake.10,11 Our data
indicate that this down-regulation of protein
breakdown occurred only in the oedematous
group, indicating that they had the appro-
priate adaptive response to food deprivation.
In the non-oedematous group the rates were
not different, indicating that children with
marasmus break down their body proteins
at the same rate as when they are well
nourished. This apparent ‘lack of adapta-
tion’ of protein breakdown in response to
food deprivation by children with maras-
mus, however, seems to confer a metabolic
advantage that enables them to cope with
and survive chronic food deprivation better
than their oedematous counterparts.
A surprising finding was that children
who had recovered from non-oedematous
SCM were breaking down their body
proteins 25% more slowly than children
who previously had oedematous SCM.
Based on evidence that a slower protein
turnover rate improves the efficiency of
dietary protein utilisation, hence, N balance
in children10 and adults on a marginal
protein intake,12 it can be argued that an
inherently slower protein turnover rate, as
seen in the children who had recovered from
non-oedematous SCM, confers a metabolic
advantage which enables them to better
adapt to chronically inadequate dietary
protein intake. The converse would be true
for children with oedematous SCM.
Amino acid production rates and plasma pool
sizes
Because the supply of essential amino acids
(EAA) is derived almost exclusively from
breakdown of body proteins, the slower
protein breakdown rate of children with
oedematous SCM suggests a reduced sup-
ply of all EAAs for metabolic purposes and
for maintenance of intracellular and extra-
cellular pools. In the case of the non-
essential amino acids (NEAA), however, it
is possible for de novo synthesis to fill the gap
created by reduction in the amount released
from protein breakdown. This might not be
true for NEAAs such as tyrosine and
cysteine which are synthesised from the

TABLE 1. Leucine kinetics in children with oedematous and non-oedematous severe childhood malnutrition [mean
(SEM)].

Malnourished Recovered

Leucine kinetics, mmol/kg fat-free wt21?h21 Non-oedematous Oedematous Non-oedematous Oedematous

Fasted state n57 n516 n57 n516

Endogenous flux*{ 205 (24) 138 (9){,1 202 (13) 267 (34){
Fed state n59 n515 n59 n515
Total flux*{ 146 (8) 118 (5.8){,1 165 (10) 180 (5.5)
Endogenous flux*{ 92 (8.1) 66 (5.5){,1 109 (9.8) 125 (4.8)
Splanchnic uptake
Absolute rate 13.5 (2.5) 16.1 (1.6) 11 (1.1) 17 (2.1)1
% of enteral intake 29 (5) 35 (3) 22 (3) 34 (4)

Means were compared by repeated-measures ANOVA; * main effect of clinical state, p,0.001; { diagnosis by
clinical state interaction, p,0.003; { within clinical state, values are significantly different from similar state non-
oedematous value, p,0.05 (post hoc pair-wise comparison by Tukey’s method); 1 significantly different from
corresponding recovered-state value, p,0.05 (post hoc pair-wise comparison by Tukey’s method).
90 F. Jahoor et al.

EAAs phenylalanine and methionine. To
determine whether NEAAs were being
produced in adequate quantities in children
with oedematous versus non-oedematous
SCM, the kinetics of tyrosine, cysteine and
glycine were measured in the acutely mal-
nourished state ,3 days post-admission and
after nutritional recovery at ,60 days post-
admission.8,13,14 Glycine is an NEAA pro-
duced and utilised in large quantities
because it is a precursor of numerous
specialised proteins, peptides and other
biomolecules critical for normal health.15
As shown in Table 2, in the malnourished
state, both groups of children had slower
cysteine production and slower cysteine
release from protein breakdown compared
with the recovered state values. De novo
cysteine synthesis in the malnourished state
was actually faster compared with the rate at
recovery in the oedematous SCM group,
indicating that all children with SCM have
slower cysteine production because of
decreased contribution from protein break-
down, not from decreased de novo synthesis.
The magnitude of this reduction, however,
was much greater in the children with
oedematous SCM than in those with non-
oedematous SCM. Similarly, endogenous
tyrosine production, i.e. tyrosine derived
from protein breakdown plus de novo synth-
esis, was 35% slower in the children with
oedematous SCM than in those with non-
oedematous SCM in the malnourished
state. Because protein breakdown is ,30%
slower in children with oedematous SCM
than in non-oedematous SCM (Table 1),
this finding suggests that the slower flux of
tyrosine is entirely owing to reduced release
from protein breakdown, a deficit in tyr-
osine supply that is not made up for by
increased de novo synthesis from phenylala-
nine. On the other hand, there were no
differences in glycine flux between the
children with oedematous and non-
oedematous SCM in either the malnour-
ished or recovered state (Table 2). There
were also no differences in glycine flux
between the malnourished and the recov-
ered state in either group. These findings
suggest that there is increased contribution
of glycine from de novo synthesis to make up
for the deficit created by its reduced release
from the slower protein breakdown in

TABLE 2. Cysteine, tyrosine and glycine kinetics in children with oedematous and non-oedematous SCM in the fed
state [mean (SEM)].

Malnourished Recovered

Kinetic parameter, mmol/kg fat-free wt21?h21 Non-oedematous Oedematous Non-oedematous Oedematous

Cysteine n511 n511 n511 n511

Exogenous inflow 7.9 (0.06) 7.8 (0.02) 7.9 (0.01) 7.8 (0.02)
Total flux 37.2 (2.5){ 27.9 (2)*{ 48.4 (3.4) 44.8 (2.9)
Endogenous flux 29.3 (2.6){ 20 (2)*{ 40.6 (3.4) 37 (2.9)
De novo synthesis 8.6 (0.4) 9.4 (0.8){ 7.9 (0.4) 7.6 (0.3)
Protein-derived flux 20.7 (2.5){ 11.4 (1.7)*{ 32.7 (3.3) 29.4 (2.4)
Tyrosine n56 n56 n56 n56
Exogenous inflow 10.6 (0.1) 10.6 (0.04) 18.8 (0.7) 18.4 (1.5)
Total flux 51.4 (5.6) 37.2 (5)*{ 68 (6) 61 (12)
Endogenous flux 40.8 (5.6) 26.5 (5)*{ 49.5 (5.8) 43 (11)
Glycine n59 n510 n59 n510
Exogenous inflow 41 (0.6) 41 (1.0) 44 (0.9) 44 (0.4)
Total flux 331 (12) 290 (13) 332 (16) 293 (24)
Endogenous flux 291 (12) 249 (13) 291 (16) 249 (24)

{
* Significantly different from same-state, non-oedematous value, p,0.05; significantly different from recovered
value, p,0.05.
 Protein metabolism & severe malnutrition 91

children with oedematous SCM. These find-
ings also suggest that individual NEAAs
have unique metabolic responses to chronic
food deprivation. For example, the ability
to maintain glycine production in the face
of chronic food deprivation might be a
necessary adaptation because of the para-
mount importance of glycine as a precursor
of the synthesis of numerous specialised
proteins, peptides and other biomolecules
critical to survival. These findings also suggest
that the pool size of some NEAAs, such as
tyrosine, might be smaller in oedematous
than in non-oedematous SCM.
To this end, we have confirmed that
most plasma amino-acid concentrations
are lower in children with oedematous
SCM than in their non-oedematous coun-
terparts (Table 3). Compared with values at
recovery, only the concentrations of the
three branched chain amino acids were
significantly lower in children with non-
oedematous SCM. On the other hand, with
the exception of alanine, glycine and serine,
the concentrations of all other amino acids
were significantly lower in the children with
oedematous SCM compared with their
values at recovery. The greater depletion of
the plasma amino-acid pools in oedematous
SCM suggests a shortage in the supply of
most amino acids which will negatively
affect the synthesis rates of proteins, espe-
cially those with fast turnover rates such as
the plasma proteins.
Synthesis of plasma proteins
To this end we have measured the synthesis
rates of representative positive and negative
acute-phase proteins (APP) in children with
oedematous and non-oedematous SCM.
The positive APPs participate and assist in
different aspects of host defences. Hence,
inability to mount an adequate APP
response to injury or infection will compro-
mise host defences. This is a distinct
possibility in children with oedematous
SCM in whom we have shown that the
production of EAAs and some NEAAs are
markedly slower. The negative APPs, on the
other hand, are transporters of nutrients,
hormones, metabolites and drugs to sites of

TABLE 3. Plasma amino acids in children with oedematous and non-oedematous SCM in the fed state [mean
(SEM)].

Malnourished Recovered

Amino acid mmol/L Non-oedematous, n57 Oedematous, n57 Non-oedematous, n57 Oedematous, n57

Essentials
Leucine* 82 (9) 50 (9) 119 (13) 122 (16)
Isoleucine*{ 58 (8) 34 (3) 74 (10) 75 (7)
Valine{ 203 (18) 111 (11) 285 (23) 282 (40)
Histidine 60 (6) 80 (8) 61 (3) 66 (4)
Methionine*{ 9 (1) 5 (6) 11 (0.5) 11 (0.5)
Lysine 106 (11) 78 (9) 117 (8) 105 (7.6)
Phenylalanine*{ 41 (3) 28 (8) 45 (1.4) 42 (1.8)
Threonine 91 (15) 71 (9) 107 (15) 101 (9)
Non-essentials
Alanine 223 (20) 223 (50) 247 (38) 247 (50)
Glycine 190 (22) 261 (32) 205 (18) 249 (29)
Serine 148 (18) 150 (19) 129 (13) 138 (10)
GLX 489 (58) 379 (50) 536 (37) 526 (34)
Cysteine*{ 24 (9) 2.3 (1.4) 49 (5) 35 (8)
Tyrosine*{ 46 (5) 11 (1) 50 (3.7) 47 (3.3)

{
GLX, glutamate plus glutamine; * main effect of clinical state, p,0.005; main effect of diagnosis, p,0.04.
92 F. Jahoor et al.

FIG. 1. Plasma albumin concentration, fractional synthesis rate (FSR) and absolute intravascular synthesis rate
(ASR) in children with oedematous (n57, &) and non-oedematous (n57, %) SCM on post-admission days 2
(experiment 1), 8 (experiment 2) and 59 (experiment 3). Values are means (SEM). a, significantly different from
experiment 3 value, p,0.01 or 0.05; b, significantly different from experiment 1 value, p,0.05; c, significantly
different from the same-state value of the non-oedematous group, p,0.05.

utilisation. Hence, impaired synthesis will ,8 days post-admission (experiment 2),

further compromise metabolic/physiological
capacity. Indeed, reduction of nutrient
transport proteins correlates with a high
mortality rate in children with SCM.16 To
determine whether APP synthesis rates were
compromised in children with oedematous
SCM, we measured the synthesis rates of
the nutrient transport proteins, albumin and
transthyretin and the positive APPs, a1-
antitrypsin and haptoglobin in children with
oedematous and non-oedematous SCM. All
had evidence of one or more infections on
admission. The first measurement was made
in the infected-malnourished state ,2 days
post-admission (experiment 1), the second
when their infections were clear and the
oedematous children had lost their oedema,
and the third measurement was performed
when the children were fully recovered, ,54
days post-admission (experiment 3).17–20
Albumin and transthyretin (TTR)
In experiment 1, the plasma albumin con-
centrations of both the oedematous and
non-oedematous groups were significantly
lower than the values at recovery, but there
was no difference between the concentra-
tions of the two groups (Fig. 1). After
infections had cleared (experiment 2), the
albumin concentrations of both groups were
still significantly lower than on recovery, but
the concentration of the oedematous group
had a modest and significant (p,0.05)
 Protein metabolism & severe malnutrition 93

FIG. 2. Plasma transthyretin concentration, fractional synthesis rate (FSR) and absolute intravascular synthesis
rate (ASR) in children with oedematous (n57, &) and non-oedematous (n57, %) SCM on post-admission days 2
(experiment 1), 8 (experiment 2) and 59 (experiment 3). Values are means (SEM). a, significantly different from
experiment 3 value, p,0.01; b, significantly different from experiment 1 value, p,0.01; c, significantly different
from the same-state value of the non-oedematous group, p,0.05.

increase. On recovery, the albumin concen-
tration of the oedematous group was sig-
nificantly higher (p,0.05) than in the
non-oedematous group. There was no dif-
ference in the fractional synthesis rate (FSR)
of plasma albumin between the two groups
in any of the three experiments. In the
oedematous group, there was no difference
in the FSR of albumin in the three
measurements but in the non-oedematous
group albumin FSR was significantly faster
(p,0.05) in the malnourished/infected
state than at recovery. In the oedematous
group, the absolute synthesis rate (ASR) of
albumin in experiment 1 was significantly
slower (p,0.01) than the rate on recovery,
but, in the non-oedematous group, ASR in
experiments 1 and 2 were not different from
the ASR on recovery. When both groups of
patients had recovered, the oedematous
group of children had a significantly faster
albumin ASR than the non-oedematous
group.
As shown in Fig. 2, the transthyretin
kinetics of the two groups were almost
identical to their albumin kinetics. Plasma
TTR concentrations in both groups of
children in experiment 1 were significantly
lower (p,0.05) than the concentrations in
experiments 2 and 3. The plasma TTR
concentrations in the two groups were not
different in experiments 1 and 2, but, at
recovery, the TTR concentration of the
non-oedematous group was significantly
94 F. Jahoor et al.
lower (p,0.03) than in the oedematous
group. There was no difference between
the FSR of TTR of the two groups of
patients in any of the three experiments. In
the oedematous group, FSR was unchanged
from experiment 1 to experiment 3, but, in
the non-oedematous group, FSR in experi-
ment 1 was significantly faster (p,0.03)
than in experiments 2 and 3. Whereas in
the oedematous group the ASR of TTR
in experiment 1 was significantly slower
(p,0.01) than in experiment 3, in the non-
oedematous group, the ASR of TTR was
not different in any of the three experiments.
When both groups of patients had recov-
ered, the oedematous group had a signifi-
cantly faster ASR than the non-oedematous
group.
These findings indicate two primary
differences in the parameters of albumin
and TTR metabolism between the two
groups. Firstly, whereas repletion of the
intravascular albumin and TTR pools of the
children with oedematous SCM was asso-
ciated with parallel increases in synthesis
rates when they recovered, there was no
change in synthesis rates during repletion of
the pools in the children who previously had
non-oedematous SCM. Secondly, at recov-
ery, the children who previously had oede-
matous SCM had larger intravascular
albumin and TTR pools and synthesised
the proteins faster than children who had
recovered from non-oedematous SCM.
From these findings it can be deduced that
replenishment of the albumin and TTR
pools of the children with non-oedematous
SCM during nutritional rehabilitation was
owing to a decrease in the rate of catabolism
and/or loss from the intravascular space and
not to an increase in the synthesis rates. The
same mechanism was also responsible for
the increase in intravascular albumin con-
centration in the oedematous group in
experiment 2 as there was no difference in
albumin synthesis from experiments 1 to 2.
In the oedematous group, however,
increases in albumin and TTR synthesis
rates also contributed to repletion of the
pool as synthesis rates of both proteins
in experiment 3 were faster compared
with the rates in experiment 1. Hence, one
can reason but not prove that the albumin
and TTR pools became depleted during
inadequate food intake because of an
increase in the rate of catabolism and/or
loss in the non-oedematous group.
Additionally, in the oedematous group, a
reduction in synthesis rate might also have
been a contributing factor. Finally, in
experiment 2, when both groups of children
were still severely malnourished but their
infections had cleared, the plasma concen-
tration of TTR increased to a value that was
not different from the value on recovery.
This indicates that the presence of infection
plays a major role in mediating depletion of
the plasma TTR pool. Hence, use of this
protein concentration as an indicator of
protein nutritional status in clinical practice
is not valid.
At recovery, the slower rates of
synthesis of albumin and TTR in the non-
oedematous group than in the oedematous
group is similar to the slower whole body
protein breakdown rate observed in children
who had recovered from non-oedematous
SCM. This seems to be a general phenom-
enon that might apply to proteins other than
albumin and TTR. The slower whole-body
protein turnover and slower synthesis rates
of albumin and TTR (and possibly other
plasma proteins) suggest that some factor(s)
might be restricting turnover of whole-body
and specific plasma proteins, hence their
pool sizes, in the recovered children who
were previously marasmic. This is probably
related to a programmed effect which
confers a survival benefit as these children
are able to adapt to reduced food intake,
which results in marasmus rather than
kwashiorkor and marasmic-kwashiorkor. It
is well known that children with the latter
syndromes have a higher morbidity and
mortality rate.1,2
A surprising finding in the present series
of studies was that the children with
marasmus had faster FSR of albumin and
 Protein metabolism & severe malnutrition 95

FIG. 3. Plasma a1-antitrypsin concentration, fractional synthesis rate (FSR) and intravascular absolute synthesis
rate (ASR) in children with oedematous (n514, &) and non-oedematous (n59, %) SCM on post-admission days
2 (experiment 1), 8 (experiment 2) and 54 (experiment 3). Values are means (SE). a, significantly different from
experiment 3 value, p,0.05; b, significantly different from experiment 2 value, p,0.05; c, significantly different
from the same-state value of the non-oedematous group, p,0.05.

TTR in experiment 1 than in experiments 2 degree of depletion in kwashiorkor and

and 3. The faster FSR in the marasmic
patients might be an adaptation which
enables them to lessen the degree of albumin
and TTR depletion, compared with
kwashiorkor and marasmic-kwashiorkor
patients, in response to severe undernutri-
tion. Although in the present study the
albumin concentration in the marasmic
group of patients was not significantly
higher than in the oedematous group on
admission, other studies have consistently
shown that the decreased plasma albumin
concentrations in marasmic patients is
always less marked compared with the
marasmic-kwashiorkor.20
a1-antitrypsin and haptoglobin
In both the non-oedematous and oedema-
tous groups, the plasma concentrations of
a1-antitrypsin and haptoglobin were signifi-
cantly higher (p,0.05) in experiment 1 than
on recovery (Figs 3 & 4). However, the
values in the oedematous group were signi-
ficantly lower than in the non-oedematous
group. In experiment 2, when the signs and
symptoms of infection had cleared, the
plasma concentration of haptoglobin had
96 F. Jahoor et al.

FIG. 4. Plasma haptoglobin concentration, fractional synthesis rate (FSR) and intravascular absolute synthesis
rate (ASR) in children with oedematous (n514, &) and non-oedematous (n59, %) SCM on post-admission days
2 (experiment 1), 8 (experiment 2) and 54 (experiment 3). Values are means (SE). a, significantly different from
experiment 3 value, p,0.05; b, significantly different from experiment 2 value, p,0.05; c, significantly different
from the same-state value of the non-oedematous group, p,0.05.

decreased to a value that was not different
from that on recovery. The plasma concen-
tration of a1-antitrypsin, however, was still
significantly higher (p,0.05) than the value
on recovery.
The FSR of a1-antitrypsin did not differ
between groups in any experiment.
However, in both experiments 1 and 2, the
ASR was faster than in experiment 3
(p,0.05). Although the oedematous group
synthesised a1-antitrypsin only 69% as fast
as the non-oedematous group in experiment
1, this difference failed to reach statistical
significance. In both groups the FSR of
haptoglobin tended to be slower in experi-
ments 1 and 2 than in experiment 3. The
ASR of haptoglobin was significantly faster
in experiment 1 than in experiment 3 in the
non-oedematous group. It was also faster
than the ASR of the oedematous group in
experiment 1.
In agreement with the findings of
others (e.g. Schelp et al.21), these results
suggest that children with oedematous SCM
can mount an APP response to infection
that is similar to the response of non-
oedematous SCM children. However, the
magnitude of the response is less in the
children with oedematous SCM. For exam-
ple, a1-antitrypsin and haptoglobin concen-
trations of the non-oedematous children
when they were infected and malnourished
were ,90% and 178% higher, respectively,
than on recovery whereas those of the
 Protein metabolism & severe malnutrition
oedematous group were only 42% and
55% higher, respectively. Similarly, the
magnitude of changes in synthesis rates of
these two proteins was higher in the non-
oedematous group. Hence, although the
oedematous group mounted an APP
response similar qualitatively to that of the
non-oedematous group, the magnitude of
the response was smaller. This might con-
tribute to the more severely impaired host
defences of children with kwashiorkor and
marasmic-kwashiorkor.
The higher plasma concentrations of a1-
antitrypsin and haptoglobin observed in the
non-oedematous children when they were
both infected and malnourished were asso-
ciated with higher absolute rates of synthesis
of these proteins. Furthermore, plasma
concentrations of these proteins and their
rates of synthesis were lower after infections
had resolved. These findings suggest that
expansion of the protein pools in response to
infection and their contraction as the infec-
tion resolved were mediated by changes in
the rates of synthesis of the proteins. The
same was true for a1-antitrypsin in children
with oedematous SCM. However, the
higher plasma concentration of haptoglobin
observed in these children when they were
both infected and malnourished was not
associated with a higher absolute rate of
synthesis compared with the rate observed
after infection had resolved. This finding
suggests that expansion and contraction of
the haptoglobin pool in response to the
presence or absence of infection in children
with oedematous SCM are mediated by a
mechanism other than synthesis rate. The
most likely mechanism by which the hap-
toglobin pool of individuals with oedema-
tous SCM expands in response to infection
is a reduction in its rate of catabolism
relative to its rate of synthesis. An adaptive
mechanism whereby the severely malnour-
ished individual can increase availability of
APPs by reducing their rates of catabolism
rather than increasing their rates of synthesis
clearly has the advantage of conserving the
limited supply of amino acids and other
97
nutrients needed to synthesise APPs. The
weaker APP response of the oedematous
group was not surprising as other aspects of
host defence, relating to immune structure
and function, are more compromised in
children with oedematous SCM than in
non-oedematous SCM.22 Although the pre-
cise reasons for this weaker APP response
are not known, our data on whole-body
protein breakdown rate and production of
amino acids in children with oedematous
SCM suggest that a shortage of amino acids
is one possibility.
Pro-oxidant/Anti-oxidant Homeostasis
A consistent finding by our colleagues in
earlier studies at the Tropical Metabolism
Research Unit is that erythrocyte and whole
blood concentration of glutathione (GSH),
the primary intracellular anti-oxidant/
detoxicant, is lower in children with oede-
matous than with non-oedematous
SCM.6,23 This observation led to the
hypothesis that stresses caused by infections
and other noxious stimuli elicit the expected
increase in free radical production, but the
mechanisms responsible for removing the
increased oxidant load are compromised. As
a consequence, there is free radical damage
of cellular membranes, resulting in clinical
and pathophysiological manifestations of
the oedematous malnutrition syndromes.
Several groups have supported this hypoth-
esis by finding elevated plasma and urinary
levels of biomarkers of oxidant-induced
cellular damage.24,25 However, whether this
increased oxidant damage was consequent
to increased production of oxidant species
or to an underlying defect in anti-oxidant
capacity, including GSH synthesis, or to a
combination of both factors was debatable.
In particular, Golden & Ramdath pro-
posed that depletion of erythrocyte GSH
in children with oedematous SCM was
owing primarily to increased consumption
rather than decreased synthesis.6 This pro-
posal was based on their observation that
98 F. Jahoor et al.

FIG. 5. Erythrocyte glutathione concentration, fractional and absolute (FSR, ASR) synthesis rates in children
with non-oedematous (n57, %) and oedematous (n57, &) SCM when they are malnourished and infected
(experiment 1), when infections are cleared and oedema is lost (experiment 2) and when fully recovered
(experiment 3). Values are means (SE). a, significantly different from experiment 3 value, p,0.05; b, significantly
different from experiment 2 value, p,0.05; c, significantly different from the same-state value of the non-
oedematous group, p,0.05.

intracellular erythrocyte GSH concentra- cysteine. We therefore measured erythrocyte

tions increased when whole blood from
children with oedematous SCM was incu-
bated for 6 hours at 25uC without exogen-
ous cysteine, glycine and glutamine. In a
later study, they reported elevated erythro-
cyte GSH S-transferase activity in children
with oedematous SCM, further supporting
their notion of increased consumption of
GSH.26 On the other hand, our finding that
plasma cysteine concentration was markedly
lower in children with oedematous SCM
but not in non-oedematous SCM (Table 3)
strongly supported the notion of decreased
GSH synthesis owing to a shortage of
GSH kinetics in vivo in children with
oedematous and non-oedematous SCM
three times during hospitalisation, shortly
after admission when the children were
infected and malnourished (experiment 1),
,8 days post-admission, when they were no
longer infected (experiment 2) and when
recovered (experiment 3).27 Children with
oedematous SCM had significantly lower
erythrocyte GSH concentration and slower
absolute rates of synthesis than children
with non-oedematous SCM both shortly
after admission and ,9 days later when
their infections had resolved (Fig. 5). At
 Protein metabolism & severe malnutrition 99

these times, the oedematous group also had

significantly lower erythrocyte GSH con-
centrations and absolute rates of synthesis
than on recovery. Plasma and erythrocyte-
free cysteine concentrations of the oedema-
tous group were significantly lower in
experiments 1 and 2 than on recovery
(Table 4). In contrast, erythrocyte GSH
concentrations, rates of GSH synthesis and
plasma as well as erythrocyte-free cysteine
concentrations in the non-oedematous
group were similar at all three time points
and greater in experiments 1 and 2 than in
the oedematous group. These results con-
firmed that GSH deficiency is characteristic
of oedematous SCM and suggest that it is
owing to a reduced rate of synthesis
secondary to a shortage in cysteine supply.
This study also showed that the children
with non-oedematous SCM, despite having
similar infections and degree of severity of
malnutrition, were still able to maintain
normal concentrations and rates of synthesis
of GSH, suggesting that the GSH precursor
supply was adequate.
If inadequate cysteine supply were mainly
responsible for the slower rate of synthesis of FIG. 6. Changes in erythrocyte glutathione concen-
GSH, then early nutritional resuscitation of tration, fractional and absolute (FSR, ASR) synthesis
children with oedematous SCM should rates in children with oedematous SCM after ,9 days
of supplementation with N-acetylcysteine (n58) and
include cysteine to accelerate restoration alanine (n58). Values are means (SE). a, alanine vs
and maintenance of anti-oxidant capacity. cysteine group, p,0.01.
We therefore investigated the effect of diet-
ary cysteine supplementation on GSH concentration can be restored during the
synthesis in children with oedematous early phase of treatment if SCM patients are
SCM. Two groups of children with oede- supplemented with cysteine. In addition to
matous SCM were supplemented with this early restoration of GSH synthesis rate
either 0.5 mmol/kg-1/d-1 of N-acetylcysteine and concentration, the children had faster
or isonitrogenous alanine as control.28 After resolution of oedema (,5 days less) than the
7 days the cysteine supplement elicited a control group, suggesting early restoration
514% increase in intracellular cysteine con- of cell integrity and function.
centration which was associated with a 76% Though our findings clearly show that
faster rate of synthesis of GSH and a 30% children with oedematous SCM have
increase in erythrocyte GSH concentration. impaired ability to synthesise the anti-
By contrast, cysteine concentration in the oxidant GSH, we cannot conclude that
alanine-supplemented group increased by this impairment existed prior to the disease
only 166% and GSH concentration and and therefore caused it. More recently, a
synthesis rate increased by only 15 and study in Malawi29 failed to prevent
27%, respectively (Figs 6, 7). These results kwashiorkor by supplementing the diets
confirmed that GSH synthesis rate and of marginally nourished 1–4-year-old
100 F. Jahoor et al.

FIG. 7. Erythrocyte cysteine concentration in oedematous SCM children supplemented with N-acetylcysteine
(n58) and alanine (n58) when malnourished and infected (experiment 1), when infections are cleared and oedema
is lost (experiment 2) and when fully recovered (experiment 3). Values are means (SEM). a, vs experiment 1,
p,0.05; b, vs experiment 2, p,0.01p; c, vs alanine, p,0.03.

TABLE 4. Concentrations of the glutathione precursor amino acids in plasma and red blood cells of children with
SCM.

Experiment 1 Experiment 2 Experiment 3

Oedema No oedema Oedema No oedema Oedema No oedema

Amino acid, mmol/L n57 n57 n57 n57 n57 n57

Plasma-free
Glycine 260 (33) 190 (22) 335 (41) 276 (54) 249¡29 205¡18
GLX 379 (61){1 489 (58) 647 (83){{ 559 (79) 526¡34 536¡37
Cysteine plus 2 6 cystine 6.9 (2.8){{1 61 (19) 16 (11){,{{ 55 (11) 38¡8 56¡9
Methionine 5.3 (0.5){{1 9.1 (1.0) 6.7 (0.7){,{{ 11 (1.7) 10.5¡0.5 11.3¡0.9
Serine 150 (20) 148 (19) 192 (21){{ 180 (28) 138¡10 129¡13
RBC-free
Glycine 721 (95)1 565 (75) 1000 (103){,{{ 490 (70) 688¡118{ 490¡11
Glutamate 565 (75) 779 (74) 695 (61) 666 (130) 538¡49 602¡48
Glutamine 183 (37){{1 386 (68) 271 (26) 392 (93) 263¡32 373¡69
Cysteine* ,1.0{{ 17.5 1.7{,{{ 16.2 22.1 30.9
(,1–11.3) (3.4–44) (,1–11.0) (8.2–40) (9.8–35) (5.8–49)

All values are mean (SE), except * values which are median (range); { oedema vs no oedema, p,0.05; { experiment
1 vs 3, p,0.001; 1 experiment 1 vs 2, p,0.05; {{ experiment 2 vs 3, p,0.05; GLX, glutamate plus glutamine; RBC,
red blood cell.

children with a cocktail of anti-oxidant References

vitamins and N-acetylcysteine, suggesting
that anti-oxidant depletion might be a
consequence rather than the cause of
kwashiorkor.
Acknowledgments
We are grateful to the National Institutes of
Health and the United States Department
of Agriculture, Agriculture Research Service
for funding these studies.
1 Alleyne GAO, Hay RW, Picou DI, Stanfield JP,
Whitehead RG. Protein-Energy Malnutrition.
London: Edward Arnold, 1977.
2 Waterlow JC. In: Protein-Energy Malnutrition.
Waterlow JC, McGregor S, Tomkins AM,
eds. London: Edward Arnold, 1992; 89–
99.
3 Jackson AA. The aetiology of kwashiorkor. In: Diet
and Disease in Traditional and Developing Societies.
GA Harrison, JC Waterlow, eds. Cambridge:
Cambridge University Press, 1990; symposium 30,
76–113.
4 Gopalan C. Kwashiorkor and marasmus: evolution
and distinguishing features. In: McCance RA,
 Protein metabolism & severe malnutrition
Widdowson EM, eds. Calorie Deficiencies and Protein
Deficiencies. London: Churchill, 1968; 49–58.
5 Whitehead RG, Alleyne GA. Pathophysiological
factors of importance in protein-calorie malnutri-
tion. Br Med Bull 1972; 28:72–9.
6 Golden MH, Ramdath D. Free radicals in the
pathogenesis of kwashiorkor. Proc Nutr Soc 1987;
46:53–68.
7 Manary MJ, Broadhead RL, Yarasheski KE.
Whole-body protein kinetics in marasmus and
kwashiorkor during acute infection. Am J Clin
Nutr 1998; 67:1205–9.
8 Jahoor F, Badaloo A, Reid M, Forrester T. Sulfur
amino acid metabolism in children with severe
childhood undernutrition: cysteine kinetics. Am J
Clin Nutr 2006; 84:1400–5.
9 Jahoor F, Badaloo A, Reid M, Forrester T. Sulfur
amino acid metabolism in children with severe
childhood undernutrition: methionine kinetics. Am
J Clin Nutr 2006; 84:1393–9.
10 Golden MH, Waterlow JC, Picou D. Protein
turnover, synthesis and breakdown before and after
recovery from protein-energy malnutrition. Clin Sci
Mol Med 1977; 53:473–7.
11 Tomkins AM, Garlick PJ, Schofield WN, Waterlow
JC. The combined effects of infection and malnu-
trition on protein metabolism in children. Clin Sci
1983; 65:313–24.
12 Gibson N, Jahoor F, Ware L, Jackson AA.
Endogenous glycine and tyrosine production is
maintained in adults on a marginal protein diet.
Am J Clin Nutr 2002; 75:511–18.
13 Reid M, Forrester T, Badaloo A, Heird WC, Jahoor
F. Supplementation with aromatic amino acids
improves leucine kinetics but not aromatic amino
acid kinetics in infants with infection, severe
malnutrition, and edema. J Nutr 2004; 134:3004–
10.
14 Jahoor F, Badaloo A, Reid M, Forrester T. Glycine
production in severe childhood undernutrition. Am
J Clin Nutr 2006; 84:143–9.
15 Jackson AA. The glycine story. Eur J Clin Nutr
1991; 45:59–65.
16 Coward WA. Serum colloidal osmotic pressure in
the development of kwashiorkor and in recovery: its
relationship to albumin and globulin concentrations
and oedema. Br J Nutr 1975; 34:459–67.
17 Morlese JF, Forrester T, Badaloo A, Del Rosario
M, Frazer M, Jahoor F. Albumin kinetics in
oedematous and non-oedematous protein-energy
101
malnourished children. Am J Clin Nutr 1996;
64:952–60.
18 Morlese JF, Forrester T, Del Rosario M, Frazer M,
Jahoor F. Repletion of the plasma pool of nutrient
transport proteins occurs at different rates during
the nutritional rehabilitation of severely malnour-
ished children. J Nutr 1998; 128:214–19.
19 Morlese JF, Forrester T, Jahoor F. The acute-phase
protein response to infection in severe malnutrition.
Am J Physiol 1998; 257:e112–17.
20 Reid M, Badaloo A, Forrester T, Morlese JF, Heird
W, Jahoor F. The acute-phase protein response to
infection in oedematous and non-ooedematous
protein-energy malnutrition. Am J Clin Nutr 2002;
76:1409–15.
21 Schelp FP, Thanagul O, Supawan V, et al. Serum
proteinase inhibitors and acute-phase reactants
from protein-energy malnutrition children during
treatment. Am J Clin Nutr 1979; 32:1415–22.
22 Cunningham-Rundles S. Effects of nutritional
status on immunological function. Am J Clin Nutr
1982; 35:1202–10.
23 Jackson AA. Blood glutathione in severe malnutri-
tion in childhood. Trans R Soc Trop Med Hyg 1986;
80:911–13.
24 Becker K, Leichsenring M, Gana L, Bremer HJ,
Schirmer RH. Glutathione and association antiox-
idant systems in protein energy malnutrition: results
of a study in Nigeria. Free Radic Biol Med 1995;
18:257–63.
25 Lenhartz H, Ndasi R, Anninos A, et al. The clinical
manifestation of the kwashiorkor syndrome is
related to increased lipid peroxidation. J Pediatr
1998; 132:879–81.
26 Ramdath DD, Golden MH. Elevated glutathione
S-transferase activity in erythrocytes from malnour-
ished children. Eur J Clin Nutr 1993; 47:658–65.
27 Reid M, Badaloo A, Forrester T, et al. In vivo rates
of erythrocyte glutathione synthesis in children with
severe protein-energy malnutrition. Am J Physiol
Endocrinol Metab 2000; 278:e405–12.
28 Badaloo A, Reid M, Forrester T, Heird WC, Jahoor
F. Cysteine supplementation improves the erythro-
cyte glutathione synthesis rate in children with
severe edematous malnutrition. Am J Clin Nutr
2002; 76:646–52.
29 Ciliberto H, Ciliberto M, Briend A, Ashorn P,
Bier D, Manary M. Antioxidant supplementation
for the prevention of kwashiorkor in Malawian
children: randomised, double blind, placebo con-
trolled trial. Br Med J 2005; 330:1109–11.