Published by Oxford University Press on behalf of the International Epidemiological Association
ß The Author 2012; all rights reserved. Advance Access publication 17 January 2012
International Journal of Epidemiology 2012;41:531–539
doi:10.1093/ije/dyr190
Comparative effects of vivax malaria, fever and
diarrhoea on child growth
Gwenyth Lee,1 Pablo Yori,1,2 Maribel Paredes Olortegui,2 William Pan,3,4 Laura Caulfield,1
Robert H Gilman,1 John W Sanders,5 Hermann Silva Delgado6,7 and Margaret Kosek1,2*
1
Department of International Health, Johns Hopkins School of Public Health, Baltimore, MD, USA, 2Biomedical Research Unit,
AB PRISMA, Ramirez Hurtado 622, Iquitos, Peru, 3Duke Global Health Institute, Duke University, Durham, NC, USA,
4
Nicholas School of the Environment, Duke University, Durham, NC, USA, 5Naval Medical Research Center Detachment,
Lima, Peru, 6Cuidados Intensivos Neonatales, Servicio de Neonatologı́a Hospital Apoyo Iquitos and 7Facultad de Medicina,
Universidad Nacional de la Amazonia Peruana
*Corresponding author. Department of International Health, Johns Hopkins School of Public Health and Hygiene,
615 N. Wolfe St w5515, Baltimore, MD 21205, USA. E-mail: [email protected]
Accepted
27 October 2011
Background The adverse impact of Plasmodium vivax on child health beyond
acute febrile illness is poorly studied. The effect of vivax malaria
on child growth was evaluated and compared with diarrhoeal disease and non-specific fever.
Methods
Using data from a 43-month longitudinal cohort of children 0–72
months of age (n ¼ 442) in the Peruvian Amazon, ponderal and
linear growth velocities over 2-, 4- and 6-month periods were examined using longitudinal models and related to the incidence of disease during the same period.
Results
An episode of vivax malaria led to 138.6 g (95% confidence interval
(CI) 81.9–195.4), 108.6 g (62.8–153.2) and 61 g (20.9–101.1) less
weight gain over 2-, 4- and 6-month intervals, respectively. These
deficits were larger than both diarrhoea (21.9, 17.2 and 13.8 g less
weight gain, respectively) and fever (39.0, 30.3 and 25.6 g less
weight gain, respectively). An incident episode of vivax also led to
0.070 cm (0.004–0.137) and 0.083 cm (0.015–0.151) less linear
growth over 4 and 6 months, respectively, which were also larger
than deficits from diarrhoea (0.029 and 0.028 cm, respectively) and
fever (not associated with linear growth deficits).
Despite the larger effect of P. vivax incident episodes on growth of a
particular child, diarrhoeal disease had a larger cumulative impact
on growth deficits as diarrhoeal incidence rates in this community
are 410-fold higher than vivax malaria.
Conclusions Disease control measures for vivax malaria and diarrhoeal disease
have the potential to improve the growth of children in endemic
areas.
Keywords
Vivax malaria, diarrhoea, fever, human growth, paediatric
531
532
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Introduction
Undernutrition is a major underlying cause of mortality in the low- and middle-income countries, responsible for 21% of deaths in children aged under 5
years principally through the modulation of the outcomes of major infectious diseases. It has been consistently shown that undernutrition is associated with
poorer cognitive development,2 decreased adult stature and work capacity3 and poorer maternal outcomes,1 leading to a lifelong, multi-generational
health burden. Height not gained during critical first
years of life is not completely recovered and therefore
linear deficits in early childhood can be seen as markers of human potential lost.4 The interactions of infection and undernutrition are complex and
bidirectional. However, the implementation of disease
control strategies for infectious diseases linked with
early growth deficits merits prioritization.
Diarrhoea is associated with decreases in ponderal
velocity (lowered weight gain) and linear velocity
(lowered length/height gain) and attained weight
and height in children.5,6 Other principal illnesses of
childhood such as respiratory infections have been
studied to a more limited extent.7 A few studies
have examined the impact of fever on growth, with
inconsistent findings.8–11
Considering the global burden of malaria, the number
of well-designed studies quantifying its effects on
growth is limited.12 Several cross-sectional studies
have reported a relationship between undernutrition
and falciparum malaria,13,14 and between current
nutritional status and recent prior incidence of
Plasmodium falciparum or Plasmodium vivax.15,16 Early
longitudinal studies in the Gambia and Uganda10,17
and a vitamin A intervention trial have found that
falciparum malaria among infants was associated with
poorer weight gain.18 A variety of malaria prevention
programmes, utilizing insecticide-treated bednets,
prophylactic treatment or indoor spraying, have also
been found to have a positive effect on growth.19,20
Loreto and the area surrounding Iquitos in the
Peruvian Amazon have high rates of childhood diarrhoea21 and febrile disease22 and are hypoendemic for
malaria. In cross-sectional thick smear surveys of
asymptomatic individuals, the rates of parasitaemia
(for either P. vivax or P. falciparum) are only 2.5%23
and asymptomatic parasitaemia is even less common
in children.24 This relative paucity of asymptomatic
parasitaemia makes the region an ideal site for the
determination of the effect of vivax malaria on early
childhood growth, as well as the cumulative impact of
multiple childhood infections on growth.
Methods
Site and surveillance
Data were collected from a prospective, communitybased study of 442 children 0–72 months of age in the
community of Santa Clara, located 15 km southeast of
Iquitos, in the Loreto province of Peru, between
October 2002 through April 2006. The study was established to explore the association between common
aetiologies of diarrhoea and early childhood growth.
The cohort was chosen after a community census.
Siblings were not eligible to participate in the cohort
simultaneously. Details of this cohort have previously
been reported.21
Participating families were visited three times
weekly by a trained health promoter to document
the number and consistency of stools passed by the
child over the previous 24-hour period, as well as
other symptoms. This generated a continuous history
of illness over the surveillance period for each participating child. Families also reported medications prescribed uniquely for vivax malaria and distributed by
the local health post. The initiation of treatment
required a positive thick smear for malaria. As per
national guidelines, the treatment for P. vivax infection is chloroquine, with or without primaquine (not
given to children under 6 months of age). These
medications are used exclusively for the treatment
of vivax malaria allowing for the retrospective diagnosis from medication history. Drugs distributed
through the Peruvian health system have been
found to be of high quality and malarial medications
are not available in the private sector; self-medication
for suspected malaria is not possible.25 No supply
shortages of anti-malarial medications occurred
during the study period.
Ascaris and Trichuris infections were determined by
parasitology in diarrhoeal and quarterly asymptomatic
stools. Instances of helminth treatment were defined
by reported albendazol and/or mebendazol use, separated by at least 3 treatment-free days.
Length/height and weight were measured monthly,
with each child measured on the day of their birth.
Children were weighed on Salter scales (Salter
Housewares Ltd, Tonbridge, UK). Length (children
0–23 months) or height (children 24–72 months)
was measured using a marked platform with a
sliding footboard. Socio-economic and demographic
information was collected during two community censuses before and during the study period. Informed
consent was obtained and the study protocol was
approved by the institutional review boards of Johns
Hopkins Bloomberg School of Public Health
(Baltimore, MD), US Navy Medical Research Center
(Silver Springs, MD), Asociacion Benefica PRISMA
(Lima, Peru), and the Regional Health Department
of Loreto Peru.
Clinical definitions
Diarrhoea was defined by three or more semi-liquid
stools reported over a 24-hour period. Episodes of
diarrhoea or fever were defined by symptoms separated by at least 3 symptom-free days. The incidence
of vivax malaria was determined by at least 1 day
VIVAX MALARIA AND CHILD GROWTH
of reported primaquine or chloroquine treatment.
If anti-malarial therapy was repeated within 28
days, it was interpreted as a single episode.
Diarrhoea and vivax malaria were considered ‘associated with fever’ whenever fever was reported
within 2 days of the illness episode. Stunting and
underweight in this population were defined according to the 2006 WHO standards.26
Statistical analysis
The primary outcomes of interest were ponderal and
linear velocities for each child (see Supplementary
Table S1), computed over 2, 4 and 6 months to account for both short- and longer-term trends within
the confines of the data available. Incidence rates
(IRs) and 95% CIs were computed for each age category using Poisson regression. Unadjusted smoothed
plots of growth velocity among children with and
without vivax malaria, and with 475th percentile
and <25th percentile diarrhoeal incidence for their
age, were generated.
Longitudinal linear models with random effects
were used to examine the effects of diarrhoea, fever
and vivax malaria on ponderal and linear velocity. The
covariance structure of the models was determined
based on a visual inspection of the autocorrelation
function as well as model fit. Linear velocity models
explicitly accounted for the correlated nature of the
data by specifying an appropriate autoregressive covariance structure for each 2-, 4- and 6-month
model; ponderal velocity retained an unstructured covariance. Age terms were included as fractional polynomials, which is a method to adjust for curvilinear
relationships.27
Based on prior literature documenting the importance of the detected presence of helminth infection,28
seasonality,29 socio-economic status,30 parental
height31 and prior nutritional status on childhood
growth, variables considered for model inclusion
included Ascaris and Trichuris infections, sex,
mother’s height, per-capita income, seasonality (as a
categorical variable, with each category representing
a calendar month), underweight and stunting,
birth weight, maternal age, maternal education,
crowding, the presence of a household latrine,
household water connection type, and breast-feeding
status; variables were selected into the final
model based on model fit. In linear but not ponderal velocity models, incident episodes that occurred
in the 2 weeks before anthropometry were discounted (subtracted from the total incidence in the
period).
The final ponderal and linear velocity models are
shown in Equation 1. The models tested the impact
of diarrhoea, P. vivax and fever controlling for age,
sex, per-capita income, seasonality and the presence
of underweight and stunting.
533
ðYij Yin,j Þ ¼ bj þ 0 þ 1 diarrhoea þ 2 vivax
þ 3 fever þ 4 gender
þ þ 21 Age Term 1
ð1Þ
þ 22 Age Term 2 þ "ij
In Equation 1, j represents the child, i represents the
age in months, (YijYin,j) the change in total length
or weight in the previous n months (n ¼ 2, 4 or 6
representing the 2-, 4- or 6-month model), diarrhoea,
fever and vivax refer to the number of incident
episodes over the (in) period.
All analysis was performed using Stata 11
(Statacorp, College Station, TX, USA).
Results
Of the 442 children enrolled, eight were excluded
because of limited anthropometric data. The 434
remaining children yielded 838.78 total child-years
of surveillance and 11 248 monthly anthropometry
visits.
Breastfeeding was nearly universal, and the timing of
the introduction of complementary foods and weaning
fairly homogeneous, with 85.7% of 3-months-olds
exclusively breastfed, 76.6% 9-month-olds partially
breastfed and 88.8% fully weaned by 2 years. The community was also fairly homogeneous socio-economically, with 52.5% of mothers having completed primary
education and 46% having completed secondary education. Sixty-eight percent of families had a household
water connection and 69% had a private latrine
(Supplementary Table S2).
An average of 0.60 Ascaris or Trichurus infections
per child per year were identified through study
stools, and 1.25 albendazole or mebendazole treatments per child per year were reported.
Disease incidence
A total of 3987 diarrhoeal and 5613 fever episodes
were identified. Of the fever episodes, 4117 (73.3%)
were not temporally associated with diarrhoea or malaria; fever associated with 33 falciparum malaria episodes was also discounted. The mean duration of
diarrhoeal episodes was 2.6 days and the median 1
day (inter-quartile range 3 days). The mean duration
of non-diarrhoeal, non-malaria fever was 1.7 days
(median 1 day, interquartile range 2 days). Diarrhoeal
incidence was highest (9.6 episodes/year) in the 12–17
month age group, and declined in older children;
fever was highest in the youngest children (6.5 episodes/year) and declined more gradually (Table 1).
Two hundred and three vivax malaria episodes were
identified. Two hundred and ninety-seven children
never experienced vivax malaria during the study
period (68.4%), 95 experienced 1 episode (21.2%)
and 42 experienced 41 (16.9%). Infection was most
534
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Table 1 Disease incidence by age
Months
0–5
Sample size (child-years)
29
Diarrhoea (95% CI)
8.5 (8.2–8.8)
Fever (95% CI)
6.5 (6.3–6.7)
Vivax (95% CI)
0.04 (0.02–0.06)
6–11
51
8.6 (8.4–8.8)
6.5 (6.3–6.7)
0.10 (0.08–0.12)
12–17
63
9.6 (9.4–9.8)
6.1 (6.0–6.3)
0.30 (0.27–0.34)
18–23
68
7.8 (7.6–8.0)
5.5 (5.3–5.6)
0.30 (0.26–0.33)
24–35
152
5.5 (5.4–5.6)
5.1 (5.1–5.2)
0.25 (0.23–0.27)
36–47
163
3.7 (3.6–3.8)
4.8 (4.7–4.9)
0.26 (0.24–0.29)
48–59
159
2.7 (2.6–2.7)
4.3 (4.2–4.4)
0.28 (0.26–0.31)
60–72
149
2.1 (2.0–2.2)
4.1 (4.0–4.2)
0.22 (0.20–0.24)
Overall
833
4.79 (4.75–4.83)
4.95 (4.91–4.99)
0.244 (0.235–0.253)
Incidence ¼ episodes/child-year; Sample size ¼ number of child-years of data available for analysis.
frequent in May (27 episodes) and least frequent in
December (7 episodes).
Participants with vivax malaria during the study
period were similar to participants without in mother’s
education (P ¼ 0.286), per-capita income (P ¼ 0.146),
birth weight (P ¼ 0.91), birth length (P ¼ 0.68) and
mean height-for-age and weight-for-age (by two-sided
t-tests at 12, 24, 36, 48, 60, 72 months of age, data not
shown). The incidence of vivax malaria was lower
among children <24 months of age, and roughly
stable among children 24–72 months (see Table 1).
Effects of diarrhoea, fever and malaria on
growth
Smoothed plots of the unadjusted data (Supplementary Figure S1) suggested that children with relatively
higher diarrhoeal burdens experienced, on average,
decreased ponderal and linear velocities compared
with children with less disease. Similarly, the data
suggested that vivax malaria was associated with
decreased 6-month ponderal and linear velocities
(Supplementary Figure S2).
In the adjusted models, these relationships persisted. Diarrhoeal incidence predicted decreases in
ponderal velocity of 21.9, 17.2 and 13.8 g/episode
for 2-, 4- and 6-month intervals, respectively, and
decreases in linear velocity over 4- and 6-month
intervals of 0.029 and 0.028 cm/episode, respectively
(Tables 2, 3 and Figure 1).
Non-diarrhoea, non-malaria associated fever predicted decreased ponderal velocities over all intervals
of 39.0, 30.3 and 25.6 g/episode less for 2-, 4- and 6month intervals), but was consistently non-predictive
of linear velocity (Table 2, 3 and Figure 1).
Vivax malaria led to decreases in ponderal velocity
of 138.6 (95% CI:81.9, 195.4), 108.0 (62.8, 153.2), and
61.0 (20.9, 101.1) g/episode over 2, 4 and 6 months,
respectively (Table 2). Vivax malaria also predicted
a 0.070 cm/episode decrease in 4-month linear velocity
and a 0.083 cm/episode decrease over 6 months
(Table 3).
Cumulative impact of infections
Figures 2 and 3 represent the cumulative impact of diarrhoea, fever and vivax malaria on mean velocities at the
population level. Overall, the combined burden of these
three illnesses on the cohort was estimated to be
12.2% (95% CI 15.3–9.0%), 13.9% (17.4–10.4%), 11.5%
(14.3–8.7%) and 9.5% (7.2–11.5%) less weight gain at
12, 24, 36 and 48 months of age, respectively. Diarrhoea
and vivax malaria were estimated to decrease length
gain by 1.5% (95% CI 0.7–2.2%), 2.4% (1.3–3.5%),
1.9% (1.1–2.8%) and 1.5% (0.8–2.1%) at 12, 24, 36 and
48 months of age, respectively.
The effect of diarrhoea was greater than that of
vivax until 35 months of age. After 42 months,
fever became the largest cause of decreases in ponderal velocity. The three illnesses combined had the largest absolute effect at 13 months (66.7 g less weight
gain), but resulted in the greatest percent effect
at 20 months (14.2% less weight gain). Similarly,
diarrhoea and vivax malaria led to 0.11 cm less
linear gain at 17 months, but had the greatest percent
decrease (2.4%) at 24 months of age. Although a
single episode of vivax malaria was associated with
the greatest decline in linear velocity compared
with diarrhoea and fever, its low incidence in this
community meant that its contribution to growth faltering at the population level was lower than that of
diarrhoea.
Discussion
This prospective cohort study evaluated the effect of
vivax malaria, diarrhoea and fever on the growth of
children 0–72 months of age in an area of unstable
malaria transmission. An episode of vivax malaria had
effects on ponderal and linear velocity that were 4–6
and 2–3 times greater than for diarrhoea, respectively.
However, the high incidence of diarrhoeal illness
makes it a more important determinant of linear
growth deficits in this community. The cumulative
effect of these three syndromes resulted in deficits
VIVAX MALARIA AND CHILD GROWTH
535
Table 2 Ponderal velocity models
2 months
weight
gain (grams)
4 months
weight
gain (grams)
6 months
weight
gain (grams)
21.9yyy (33.3, 10.5)
17.2yyy (26.3, 8.2)
13.8yyy (22.0, 5.7)
Morbidity
Diarrhoea incidence
Vivax incidence
Fever incidence
yyy
138.6
(195.4, 81.9)
39.0yyy (51.2, 26.8)
108.0
yyy
61yy (101.1, 20.9)
(153.2, 62.8)
30.3yyy (40.2, 20.5)
25.6yyy (34.5, 16.6)
Demographic
Sex: female
11.7 (33.5, 10.1)
15.4yyy (4.7, 26.2)
Per-capita income*
0.3 (38.9, 39.4)
34.8yyy (15.3, 54.3)
0.3 (58.6, 57.9)
60.8yyy (31.5, 90.1)
Seasonal**
Anthropometric
Stunted at onset
48.4yyy (71.1, 25.7)
yyy
Underweight at onset
197.7
FP Age Term 1
56.7yyy (92.9, 20.6)
yyy
(149.2, 246.2)
FP Age Term 2
82.9
Constant
351.8yyy (312.3, 391.3)
(98.9, 66.9)
41.2yyy (75.7, 6.6)
9.8 (33.6, 53.3)
383.9yyy (317.9, 450.0)
509.2yyy (431.8, 586.6)
74.6yyy (67.0, 82.1)
46.7yyy (76.4, 17.0)
117.2
yyy
(162.1, 72.2)
624.2yyy (570.0, 678.7)
115.4yyy (135.2, 95.6)
883.1yyy (817.7, 948.6)
y
P 4 0.10 level.
P 4 0.05 level.
yyy
P 4 0.01 level.
*(variable mean)/SD.
**Included in each model as a categorical variable (by month), results not shown.
Per-capita income is expressed in terms of standard deviations from the population mean. FP age terms are fractional polynomial
age terms (age in days divided by 1000): 2-month velocity model, Age Term 1 ¼ Age1 and Age Term 2 ¼ Age1ln(Age); 4-month
velocity model, Age Term 1 ¼ Age2 and Age Term 2 ¼ Age1; 6-month model velocity model, Age Term 1 ¼ Age2; and Age Term
2 ¼ Age2ln(Age).
yy
Table 3 Linear velocity models
2 months linear gain (cm)
4 months linear gain (cm)
6 months linear gain (cm)
0.029yyy
(0.043, 0.014)
0.028yyy
(0.043, 0.013)
Morbidity
Diarrhoea incidence
0.007
(0.022, 0.008)
yy
Vivax incidence
0.016
(0.086, 0.054)
0.070
(0.137, 0.004)
0.083yy
(0.151, 0.015)
Fever incidence
0.007
(0.008, 0.023)
0.002
(0.017, 0.013)
0.004
(0.019, 0.011)
Sex: female
0.005
(0.031, 0.042)
0.059
(0.018, 0.136)
0.076
(0.035, 0.186)
Per-capita income*
0.037yyy
(0.018, 0.055)
0.088yyy
(0.049, 0.127)
0.129yyy
0.149yyy
(0.116, 0.182)
0.413yyy
(0.361, 0.465)
0.567yyy
(0.506, 0.628)
(0.186, 0.067)
yyy
0.108yyy
(0.189, 0.027)
(0.369, 0.610)
0.309yyy
(0.279, 0.340)
(0.454, 0.330)
0.807yyy
(0.661, 0.953)
yyy
(3.222, 3.424)
Demographic
(0.074, 0.185)
Seasonal**
Anthropometric
Stunted at onset
Underweight at onset
0.126
FP Age Term 1
0.186yyy
FP Age Term 2
yyy
Constant
y
yyy
0.060
yyy
1.213
(0.17, 0.193)
(0.058, 0.063)
(1.162, 1.264)
0.130
0.489yyy
yyy
0.392
yyy
2.165
(0.205, 0.056)
(2.082, 2.249)
3.323
P 4 0.10 level.
yy
P 4 0.05 level.
yyy
P 4 0.01 level.
*(variable mean)/SD.
**Included in each model as a categorical variable (by month), results not shown.
Per-capita income is expressed in terms of standard deviations from the population mean. FP age terms are fractional polynomial
age terms (age in days divided by 1000): 2-month velocity model, Age Term 1 ¼ Age2 and Age Term 2 ¼ Age2ln(Age) and
4-month velocity model, Age Term 1 ¼ Age1 and Age Term 2 ¼ Age1ln(Age) and 6-month model velocity model: Age Term
1 ¼ Age2 and Age Term 2 ¼ Age1.
536
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Figure 2 This figure shows the percent difference in
model-predicted weight gain between healthy children
and children experiencing an average disease burden.
For example, 36-month-olds have predicted 2-month
ponderal velocity declines of 1.3, 3.5 and 6.7% due to vivax
malaria, diarrhoea and fever, respectively, yielding a
cumulative decrement of 11.5%
Figure 1 This figure shows the mean and estimated 95% CI
for predicted ponderal and linear velocity deficits due to one
incident episode of diarrhoea, fever or vivax malaria, using
the adjusted models presented in Tables 2 and 3. Declines
which were associated with a P ¼ 0.05 or less are denoted
with an ‘‘asterisk’’
in the mean ponderal velocity of 7–14% and deficits in
mean linear growth of 1–2.4%.
Our study corroborates prior findings that diarrhoeal
disease is associated with poorer ponderal and linear
growth5,6 and that fever is associated with decreased
ponderal but not linear velocity.8 This finding is
undermined by our inability to ascertain the source
of fevers, in particular lower respiratory tract infections. In the context where this surveillance occurred,
diagnostic testing is limited, which is typical in underdeveloped areas. Dengue32 and other viral infections
are endemic alongside more common aetiologies such
as respiratory or urinary tract infections, and it is
quite possible that specific aetiologies of febrile illness
within this subset have significantly different effects
on child growth compared with the composite category of non-malarial, non-diarrhoeal febrile illness
described in this analysis.
This study site offers several advantages in looking
at the effects of P. vivax on childhood growth. First,
other studies have found that the majority of P. vivax
in this community is symptomatic, especially in young
Figure 3 This figure shows the impact of disease burden on 6-month linear velocity at the population level.
For example, 36-month-olds with an average disease
burden have a predicted 6-month linear velocity decline
of 0.26% (0.009 cm) and 1.67% (0.059 cm) due to vivax
malaria and diarrhoea, respectively: a cumulative
decline of 1.94%
children,23,24 although more asymptomatic vivax is
found clustered around cases.33 Our methods did
not allow for the estimation of the effect of asymptomatic infections on growth. Additionally, the public
clinic in this area is the heavily utilized, unique distributor of malaria medications, which are prescribed
only following a positive thick smear. Therefore,
symptomatic malaria in this cohort tended to be identified rapidly and treated correctly. Nevertheless, because untreated episodes could not be counted, our
methods may underestimate the incidence of vivax
VIVAX MALARIA AND CHILD GROWTH
malaria in this population. Untreated disease or disease for which treatment was delayed could also potentially lead to more severe growth impacts than we
observed. Therefore the measured deficits likely
underestimate the severity of growth deficits caused
by P. vivax in settings where treatment is less readily
available.
Because in most intervals only a single vivax malaria
episode was detected, the decreasing size of the vivax
coefficient moving from 2- to 6-month ponderal velocity models provides some suggestion of how long a
child might require to overcome the weight-related
impact of an infection: the impact of vivax malaria
over a 4-month interval was roughly 78% of what it
was over 2 months, and over 6 months, roughly 44%.
However, the impact of vivax was still large at 6
months, implying that child growth had still not recovered from the episode (Figure 1).
In linear velocity models, episodes that began in the
2 weeks prior to anthropometry were discounted. This
choice was based on findings by Checkley34 that the
effect of diarrhoea on height appears to be delayed by
at least this long. We verified this relationship in our
data by comparing the model fit.
A variety of methods have been used to model attained weight/height by age including fractional polynomial models35 and cubic spline methods.34 We
chose the automated selection of fractional polynomial to model non-linear age/velocity trends. This
allowed us to adjust for the effect of age in our analysis with relatively few terms, and the models generated fit the data well over the age range of our
study participants.
Malnutrition is known to prolong the incidence and
duration of diarrhoeal episodes.36 Evidence also
suggests that undernutrition augments severity in
P. falciparum infection, at least in cases of severe
undernutrition.1,37 In this population under longitudinal surveillence there was no evidence that children
who experienced vivax malaria were of poorer nutritional status before the episode than their counterparts based on anthropometric measures. In this
epidemiological context of low transmission, children
are non-immune or pauci-immune, and it is possible
that this association may differ in an area of higher
transmission intensity if the immunomodulatory effects of undernutrition are primarily related to the
efficacy of the acquired, rather than the innate response. There is evidence to support this.38 Whether
malnutrition predisposes to more severe vivax malaria, however, is a separate issue that cannot be addressed by our data, since no children presented with
severe disease.
Plots of observed velocities by recent vivax malaria
status (Supplementary Figure S2) suggested a more
severe effect on growth in younger children.
Therefore, models that included interaction terms between disease incidence and age were considered.
Results indicated a greater effect in younger children
537
(<24 months of age); however, they failed to reach
statistical significance. The small number of vivax
malaria episodes in our study population and limited
child-years of data among younger children limit our
ability to draw conclusions. Further investigation is
needed to clarify whether vivax malaria results in
greater decelerations in growth velocity among
younger children.
There is strong evidence that helminth infections
have a negative impact on growth.28 In this study,
stool collection protocols included the collection of
diarrhoeal and quarterly asymptomatic stools and an
average of 7.8 stools were analysed per child per year;
helminth infections detected were generally treated
promptly within 3 days of stool collection.
Additionally, treatment was frequently reported
when the helminth infection had not been detected,
likely as a result of public clinic and parental initiative. Binary variables indicating the presence of
Ascaris and Trichuris were not predictive in simple
and multi-variate ponderal and linear velocity
models. That these infections did not result in
growth deficits in this population although they
were quite common supports arguments for intensive
control programmes.
There are a number of reports documenting the
presence of severe clinical disease resulting from
P. vivax infection.39,40 The children in this study had
uncomplicated clinical disease and were all treated on
an ambulatory basis, did not present signs of severe
anaemia, respiratory distress or impaired consciousness and would not meet case definitions of severe
disease. Despite this, our findings are evidence that
typical clinical cases of P. vivax are not benign, especially in young children. Poorer linear growth and
weight gain increase a child’s risk of becoming
stunted or underweight, conditions that in turn increase a child’s risk of more severe illnesses36,41 and
of mortality.1,37 Furthermore, growth deficits in early
life are associated with lifelong deficits in work capacity3 and cognitive function,2 an enduring consequence of these common childhood infections.
In developing countries, growth faltering is frequently followed by catch-up growth. However, in
settings where infectious diseases are persistent and
diets marginal, children who become malnourished
often do not catch up.7
Our study again reinforces the need for measures to
control diarrhoeal disease. Improved water-storage
practices and maternal education that extends
beyond primary school have previously been identified
as low-cost achievable interventions with high potential impact on diarrhoeal burden in this community.21
Additionally, we found that vivax malaria has an
important and previously undescribed impact on
childhood growth, extending at least 6 months in
duration. Further studies will be required to determine if these deficits are lasting or can be modulated
by specific disease control interventions.
538
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Supplementary Data
Supplementary Data are available at IJE online.
Funding
This work was supported by the National Institutes of
Health (K01-TW05717 to M.K.) and a grant from the
Johns Hopkins Malaria Research Institue and the
Bloomberg Family Foundation and by GEIS funding
number 847705 82000 25GB0016. Gwenyth Lee was
supported by a National Institutes of Health
International Maternal and Child Health Training
Grant (T32HD046405) (PI Dr Joanne Katz) and a
Proctor
and
Gamble
Doctoral
Dissertation
Fellowship, awarded by the Johns Hopkins
Bloomberg School of Public Health.
Acknowledgements
We would like to thank Matilda Bustos Aricara,
Victora Lopez Manuyama, Marla Judith Aricari
Huanari and Lleny Amasifuen Llerena, for their hard
work and thoughtful contributions in the field. We
would also like to thank S.T. Unt for his unwavering
support of the team throughout.
Conflict of interest: None declared.
KEY MESSAGES
Vivax malaria in early childhood adversely effects ponderal and linear growth. The effect seen was
greater per disease episode than that of diarrhoea for both ponderal and linear growth and that of
fever for ponderal growth. This deficit suggests a durable adverse effect of vivax malaria on child
health that has not been recognized previously.
The increased incidence of diarrhoea relative to vivax malaria resulted in greater cumulative deficits
in growth due to diarrhoea.
Successful interventions for both diarrhoea and vivax malaria would be expected to improve child
growth in endemic settings.
References
1
2
3
4
5
6
7
8
Black RE, Allen LH, Bhutta ZA et al. Maternal and child
undernutrition: global and regional exposures and health
consequences. Lancet 2008;371:243–60.
Grantham-McGregor S. A review of studies of the effect
of severe malnutrition on mental development. J Nutr
1995;125:2233S–38S.
Haas JD, Martinez EJ, Murdoch S, Conlisk E, Rivera JA,
Martorell R. Nutritional supplementation during the
preschool years and physical work capacity in adolescent
and young adult Guatemalans. J Nutr 1995;125:
1078S–89S.
Victora CG, de Onis M, Hallal PC, Blossner M,
Shrimpton R. Worldwide timing of growth faltering:
revisiting implications for interventions. Pediatrics 2010;
125:e473–80.
Checkley W, Buckley G, Gilman RH et al. Multi-country
analysis of the effects of diarrhoea on childhood stunting.
Int J Epidemiol 2008;37:816–30.
Moore SR, Lima AAM, Conaway MR, Schorling JB,
Soares AM, Guerrant RL. Early childhood diarrhoea and
helminthiases associate with long-term linear growth faltering. Int J Epidemiol 2001;30:1457–64.
Victora CG, Barros FC, Kirkwood BR, Vaughan JP.
Pneumonia, diarrhea, and growth in the first 4 y of
life: a longitudinal study of 5914 urban Brazilian children. Am J Clin Nutr 1990;52:391–96.
Becker S, Black RE, Brown KH. Relative effects of diarrhea, fever, and dietary energy intake on weight gain in
rural Bangladeshi children. Am J Clin Nutr 1991;53:
1499–503.
9
10
11
12
13
14
15
16
17
Martorell R, Habicht JP, Yarbrough C, Lechtig A,
Klein RE, Western KA. Acute morbidity and physical
growth in rural Guatemalan children. Am J Dis Child
1975;129:1296–301.
Rowland MGM, Cole TJ, Whitehead RG. A quantitative
study into the role of infection in determining nutritional
status in Gambian village children. Br J Nutr 1977;37:
441–50.
Torres AM, Peterson KE, Souza ACT, Orav EJ, Hughes M,
Chen LC. Association of diarrhoea and upper respiratory
infections with weight and height gains in Bangladeshi
children aged 5 to 11 years. Bull World Health Organ 2000;
78:1316–23.
Price RN, Tjitra E, Guerra CA, Yeung S, White NJ,
Anstey NM. Vivax malaria: neglected and not benign.
Am J Trop Med Hyg 2007;77:79–87.
Ehrhardt S, Burchard GD, Mantel C et al. Malaria,
anemia, and malnutrition in African children – Defining
intervention priorities. J Infect Dis 2006;194:108–14.
Ronald LA, Kenny SL, Klinkenberg E et al. Malaria and
anaemia among children in two communities of Kumasi,
Ghana: a cross-sectional survey. Malar J 2006;5:105.
Nyakeriga AM, Troye-Blomberg M, Chemtai AK,
Marsh K, Williams TN. Malaria and nutritional status
in children living on the coast of Kenya. Am J Clin Nutr
2004;80:1604–10.
Williams TN, Maitland K, Phelps L. Plasmodium vivax: a
cause of malnutrition in young children. QJM 1997;90:
751–57.
Cole TJ, Parkin JM. Infection and its effect on the growth
of young children: a comparison of The Gambia and
Uganda. Trans R Soc Trop Med Hyg 1977;71:196–98.
VIVAX MALARIA AND CHILD GROWTH
18
19
20
21
22
23
24
25
26
27
28
29
Villamor E, Mbise R, Spiegelman D et al. Vitamin A
supplements ameliorate the adverse effect of HIV-1, malaria, and diarrheal infections on child growth. Pediatrics
2002;109:e6.
Molineaux L, Gramiccia G. The garki project. Geneva:
World Health Organization, 1980.
Ter Kuile FO, Terlouw DJ, Kariuki SK et al. Impact of
permethrin-treated bed nets on malaria, anemia, and
growth in infants in an area of intense perennial malaria
transmission in western Kenya. Am J Trop Med Hyg 2008;
68:68–77.
Kosek M, Yori PP, Pan WK et al. Epidemiology of highly
endemic multiply antibiotic-resistant shigellosis in children in the Peruvian Amazon. Pediatrics 2008;122:
e541–49.
Rocha C, Morrison AC, Forshey BM et al. Comparison of
two active surveillance programs for the detection of
clinical dengue cases in Iquitos, Peru. Am J Trop Med
Hyg 2009;80:656–60.
Roper MH, Torres RS, Goicochea CG et al. The epidemiology of malaria in an epidemic area of the Peruvian
Amazon. Am J Trop Med Hyg 2000;62:247–56.
Roshanravan B, Kari E, Gilman RH et al. Endemic malaria
in the Peruvian Amazon region of Iquitos. Am J Trop Med
Hyg 2003;69:45–52.
Paredes JG, Thumm M, Briggs J, Paredes P. Evaluación
de la disponibilidad y uso de medicamentos para enfermedades de la infanica a nivel de la comunidad: regiones
Loreto y la Libertad, Perú, 2003. Presentado a la Agencia
de los Estados Unidos para el Desarrollo Internacional por
el Programa Rational Pharmaceutical Management Plus,
de Arlington, Virginia: Management Sciences for Health,
2004.
WHO. Child Growth Standards. Methods and Development: Length/Height-for-age, Weight-for-age, Weight-forlength, Weight-for-height and Body Mass Index-for-age.
Geneva: WHO, 2006.
Royston P, Ambler G, Sauerbrei W. The use of fractional
polynomials to model continuous risk variables in epidemiology. Int J Epidemiol 1999;28:964–74.
Hall A, Hewitt G, Tuffrey V, De Silva N. A review and
meta-analysis of the impact of intestinal worms on child
growth and nutrition. Matern Child Nutr 2008;4:118–236.
Maleta K, Virtanen SM, Espo M, Kulmala T, Ashorn P.
Seasonality of growth and the relationship between
30
31
32
33
34
35
36
37
38
39
40
41
539
weight and height gain in children under three years of
age in rural Malawi. Acta Paediatr 2003;92:491–97.
Desai P, Standard KL, Miall WE. Socio-economic and cultural influences on child growth in rural Jamaica. J Biosoc
Sci 1970;2:133–43.
Tanner JM, Goldstein H, Whitehouse RH. Standards for
children’s height at ages 2–9 years allowing for height of
parents. Arch Dis Child 1970;45:755–62.
Morrison AC, Minnick SL, Rocha C et al. Epidemiology of
dengue virus in Iquitos, Peru 1999 to 2005: interepidemic
and epidemic patterns of transmission. PLoS Negl Trop Dis
2010;4:e670.
Branch O, Casapia WM, Gamboa DV et al. Clustered local
transmission and asymptomatic Plasmodium falciparum
and Plasmodium vivax malaria infections in a recently
emerged, hypoendemic Peruvian Amazon community.
Malar J 2005;4:27–43.
Checkley W, Epstein LD, Gilman RH, Cabrera L,
Black RE. Effects of acute diarrhea on linear growth in
Peruvian children. Am J Epidemiol 2003;157:166–75.
Stein AD, Barnhart HX, Wang M et al. Comparison of
linear growth patterns in the first three years of life
across two generations in Guatemala. Pediatrics 2004;
113:e270–75.
Guerrant RL, Schorling JB, McAuliffee JF, Auxiladora De
Souza M. Diarrhea as a cause and an effect of malnutrition: diarrhea prevents catch-up growth and malnutrition
increases diarrhea frequency and duration. Am J Trop Med
Hyg 1992;47:28–35.
Caulfield LE, de Onis M, Blössner M, Black RE.
Undernutrition as an underlying cause of child deaths
associated with pneumonia, malaria, and measles. Am J
Trop Med Hyg 2004;80:193–98.
Fillol F, Sarr JB, Boulanger D et al. Impact of child malnutrition on the specific anti-Plasmodium falciparum antibody response. Malar J 2009;8:116.
Alexandre MA, Ferreira CO, Siqueira AM et al. Severe
Plasmodium vivax malaria, Brazilian Amazon. Emerg
Infect Dis 2010;16:1611–14.
Tjitra E, Anstey NM, Sugiarto P et al. Multidrug-resistant
Plasmodium vivax associated with severe and fatal malaria: a prospective study in Papua, Indonesia. PLoS Med
2008;5:e128.
Tupasi TE, Mangubat NV, Sunico ES et al. Malnutrition
and acute respiratory tract infections in filipino children.
Rev Infect Dis 1990;12:S1047–54.