| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
 |
ABSTRACT |
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Food supplementation often has only a limited impact on child growth and nutritional status in less developed countries. Infections, by stimulating the production of pro-inflammatory cytokines, cause decreased appetite and abnormal metabolism of nutrients, making food resources of limited benefit. Previous studies have provided some evidence for a positive effect of malaria prevention on the nutritional status of young children including that achieved through chemoprophylaxis.1 In this paper, we report the impact of seasonal intermittent preventive treatment (IPT) on nutritional indices in young Senegalese children.
A randomized, double-blind, placebo-controlled intervention study of intermittent preventive treatment of malaria was conducted in a cohort of 2- to 59-month-old children living in a rural area of Senegal where transmission of malaria is highly seasonal. Treatment consisted of monthly doses of artesunate and sulfadoxine-pyrimethamine given during the months of September, October, and November, the peak malaria transmission season. The intervention was highly efficacious, reducing the risk of a clinical malaria attack by 86% as described previously.8 This study describes the nutritional benefits of malaria prevention in this cohort.
|
MATERIALS AND METHODS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The nutritional status of infants and preschool children varies strongly by season. Body weight is greatest in the dry season (April–May) and lowest at the end of the rainy season (October–November).13,14
Outcomes. This report describes changes in nutritional status and in the prevalence of malnutrition, which were secondary outcomes of the trial. Results related to the incidence of clinical attacks of malaria have been published previously.8
Subjects. Criteria for inclusion of individual children in the study were as follows: an age between 6 weeks and 59 months, no enrollment in any other malaria study, parental intention to remain in the area for the duration of the intervention (3 months), and informed consent from the parents. Oral, witnessed, informed consent of parents was sought at a specific home visit by study staff 2–3 weeks before the start of the trial. Exclusion criteria were the presence of a severe illness or severe anemia (hemoglobin [Hb] < 8 g/dL) at the time of an initial screening visit.
Another (observational) malaria study was being conducted in two of the villages in the Niakhar study zone at the time of this trial.15 Eleven of the remaining 28 villages were selected for participation in the IPT trial on the grounds of size, accessibility, and interest in the study. On the basis of a census conducted in July 2002, 1,203 children 2–59 months of age were selected from among those in the eligible age range, and 1,088 were included in the trial. Fifteen children had no anthropometric measurements taken during the first assessment in September 2002, and 10 were excluded from the analysis because of unlikely measurements (height-for-age or weight-for-age < –6 z-scores or > +2 z-scores, height increment from September–November < –2 cm or > +6 cm). Thus, 1,063 children (515 boys and 548 girls) were included in this analysis, 529 and 534 from the intervention and control groups, respectively. Among them, 486 (92%) and 495 (93%), respectively, had complete data at both time points in Year 1 and were included in analyses on growth rates during the intervention period. In Year 2, 954 and 934 children, respectively, were included in the anthropometric surveys conducted in July and December.
Intervention. Details of the study intervention and how it was administered have been given previously.8 In brief, artesunate (4 mg/kg body weight) and sulfadoxine-pyrimethamine (sulfadoxine: 25 mg/kg; pyrimethamine: 1.25 mg/kg) or matching placebo were given monthly on three occasions at the height of the rainy season during the first year of the study (from mid-September to mid-November). No chemoprevention was given during the second year of observation. Compliance with the protocol was high: in both groups, 99% of subjects received the first dose and 93% received all three doses of either preventive treatment or placebo.
Case management. During the study, children with clinical malaria benefited from rapid treatment, following the national malaria public health program in Senegal: chloroquine as first-line treatment, quinine or sulfadoxine-pyrimethamine as second-line treatment, and injectable quinine for cases with persistent vomiting or severe malaria. Children who presented with pale mucosae or an illness suggestive of anemia were provided with iron supplementation. Clinically malnourished children were offered free treatment at an ad hoc nutrition rehabilitation center.
Data collection. Anthropometric data were collected at health centers by two trained measurers during drug administration in September and November 2002 (Year 1) and during home visits in July and December 2003 (Year 2), in accordance with internationally recommended procedures.16 Weight measurements were taken using baby scales (SECA, Hamburg, Germany), precise to the nearest 10 g, for children weighing < 16 kg, and an electronic scale (Téfal, Paris, France), precise to the nearest 100 g, for older children. Recumbent length measurements were taken for children < 2 years of age, whereas standing height was measured in children beyond that age. Measurements were precise to the nearest millimeter. Left upper-arm circumference (MUAC) measurements were taken using a non-extensible tape to the nearest millimeter, and triceps and subscapular skinfold thickness were measured using a Holtain adipometer (Siber-Hegner, Miribel, France), precise to the nearest 0.2 mm. All measurements except for weight were taken twice, and the average was used for analysis. Because of error, MUAC measurements were omitted in September 2002.
Socio-demographic data (child age, sex, and village of residency, maternal age, religion, education, and professional activity) were taken from the Niakhar study area database.
Ethics. All children benefited from participation in the trial, because both intervention and control children received rapid treatment if they had malaria. Case detection was active and passive, because mothers do not necessarily seek health care for their ill children. The study protocol was approved by the ethical review committees of the Senegalese Ministry of Health and the London School of Hygiene and Tropical Medicine.
Statistical analysis.
The nutritional indicators height-forage (HAZ), weight-for-height (WHZ), and weight-for-age (WAZ) were computed in z-scores of the WHO/NCHS reference,17 using Epi Info software V.6. Stunting, wasting, and underweight were defined as being < –2 z-scores for HAZ, WHZ, and WAZ, respectively. Arm muscle circumference (AMC) was computed using MUAC and triceps skinfold (TSF) as AMC = MUAC – (
x TSF). Increments in weight, height, skinfold thicknesses, and arm circumference during the transmission seasons of Year 1 and Year 2 were computed as the difference between measures at extremities of the intervals, divided by the precise duration (in months).
All analyses were done using intention-to-treat; children who had only received one or two preventive treatments (of three possible) were kept in the analysis. Five age groups at the first anthropometric assessment (September 2002) were created (0–11.9, 12–23.9, 24–35.9, 36–47.9, and 48–59.9 months) and maternal school education was made binary (any versus none). Mean nutritional status in September and November of Year 1 and July and December of Year 2 and growth between September and November of Year 1 were compared between intervention and control children using t tests, whereas the prevalence of stunting, wasting, and underweight were compared using 
2 test, without any adjustment for confounders as children had been randomly allocated to intervention and control groups.
In sub-analyses, growth rates between groups were compared separately for those who had suffered clinical malaria from September to November in Year 1 (49 intervention and 218 control children) and those who had not (437 intervention and 277 control children), using general linear models (GLMs) to adjust for potential confounders (5 groups of child age and 11 villages of residency) because the risk of clinical malaria and rate of growth differed for each of these variables. Clinical malaria was defined, using active case detection data, as an illness with a body temperature of 37.5°C or greater, or a history of fever or vomiting within the previous 24 hours, or both; no other obvious cause for the fever or vomiting; and the presence of P. falciparum asexual stage parasitemia at any density.
GLM was also used for testing the change in HAZ, WHZ, and WHZ over time. The change in the prevalence of malnutrition over time was tested using the Mantel-Haenszel procedure for repeated measures, i.e., by means of a stratified analysis where each subject constituted a stratum.18 SAS (version 8.2) was used for all analyses.
|
RESULTS |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
).
|
Growth during the malaria transmission season in Year 1.
There was a significant difference in weight gain between children in intervention and control groups between September and November in Year 1; children in the intervention group gained almost three times more weight than those in the control group (P < 0.0001; Table 2). Triceps and subscapular skinfold thickness fell in both groups, but this loss was significantly greater in the control group than in the intervention group (P < 0.0001 for both). These differences were also significant when the analysis was restricted to children who had not experienced any clinical malaria episode during the period of follow-up (Table 2). Sex and maternal age did not modify significantly the effect of the intervention on growth in weight or skinfolds and neither did age of the child (Table 3). There was no difference in height gain between the two groups (Table 2).
|
|
).
|
), whereas the prevalence of stunting did not vary over time in either of the two groups.
Nutritional status at the end of the Year 1 transmission season.
Mean MUAC and triceps and subscapular skinfold thicknesses were significantly greater in the intervention group (P < 0.05, < 0.01, and < 0.001, respectively) at the end of the Year 1 malaria transmission season, but the differences between groups were modest (Table 5). There was no significant difference between groups in terms of the mean HAZ, WAZ, WHZ, and arm muscle circumference (Table 5).
|
Change in nutritional status over time. The prevalence of stunting and underweight decreased significantly between November of Year 1 and December of Year 2 in both groups of children. Stunting decreased mainly between July and December of Year 2 (from 21.0% [95% CI: 18.4, 23.6] to 12.2% [95% CI: 10.1, 14.3], P < 0.0001), whereas underweight became less prevalent between November of Year 1 (26.2%; 95% CI, 23.4, 29.0) and July of Year 2 (14.3%; 95% CI, 12.1, 16.5) and remained constant thereafter (13.4%; 95% CI, 11.2, 15.6 in December of Year 1). There was no significant change in the prevalence of wasting.
Because the prevalence of stunting may change with age, and our cohort aged by 15 months during the course of the study, additional analyses were conducted to compare the change in mean HAZ over time within age categories. There was a significant linear trend of increase over time for all age groups (results not shown) and for almost all age- and intervention-specific groups (Table 6).
|
|
DISCUSSION |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Season is a strong determinant of many aspects of life in rural areas of West Africa, particularly in the Sahelian and Sudanese-Sahelian climatic zones, which are defined by average yearly rainfalls of 250–500 and 500–900 mm, respectively, and include Burkina Faso, The Gambia, Mali, Mauritania, Niger, Senegal, Sudan, and Tchad (in total or partly). Body weight varies strongly with season in newborns,19 9-month-old infants,13 preschool children, and adults.20–22 In preschool children, reduced food intake because of lack of resources and an increased work load in women during the rainy season are considered important risk factors for malnutrition, in addition to increased morbidity from diarrhea and malaria, but evidence of the relative importance of each of the factors in constraining growth is scarce.
This study had several strengths. First, randomization of study children to the intervention and placebo groups and the double-blind design prevented a number of potential problems in interpretation of the findings. Second, anthropometric measurements were taken by well-trained, experienced field workers providing powerful comparisons even for skinfold measurements.
The observed differences in growth and nutritional status between intervention and placebo groups probably underestimated the true negative impact of malaria in unprotected populations, because control children benefited from frequent home visits and early treatment of a clinical malaria episode with an effective anti-malarial drug, conditions that do not always occur outside a trial situation. Also, the impact of the intervention on growth and nutritional status was assessed in early November at the time of administration of the last dose of intermittent preventive treatment, and the study could therefore not capture the likely nutritional benefit of this third dose.8
Conversely, intermittent treatment with artesunate could have had an impact on the prevalence of Schistoma haematobium infection,23 and this may, in turn, have had an impact on nutrition that was unrelated to malaria prevention.
It seems that the intervention had an effect mainly on fat mass, because differences in skinfold thickness were highly significant, whereas those in arm muscle circumference were not. Also, the difference in change in triceps skinfold between intervention and control groups was significantly greater in villages with more intense exposure to malaria, suggesting a stronger beneficial effect of the intervention in these villages (data not shown). The same tendency existed for weight gain, but it was not significant.
There was no interaction between child age and the effect of the intervention on growth in weight and skinfold thickness. However, the difference in weight gain between groups was greatest among the 4-year-old children, considerably greater than among their 1-year-old counterparts (146 versus 83 g/mo).
Malaria prevention has been shown to have a positive effect on child weight and arm circumference in a number of studies. A community-randomized impregnated bed net trial among preschool children in The Gambia resulted in significant increases in WAZ and WHZ (P < 0.01 for both) but not in HAZ.24 Eighteen months after the implementation of an impregnated bed net intervention among preschool children with perennial exposure to malaria, the mean arm circumference was greater in the intervention group (–1.05 versus –1.25 z-scores, P = 0.008), and there was a non-significant tendency towards a greater mean HAZ (–1.06 versus –1.21, P = 0.14).25,26 In Tanzania, combined chemoprophylaxis and use of impregnated bed nets improved weight gain during the rainy season in preschool children, particularly in those 6–17 months of age,27 whereas a randomized IPT trial providing sulfadoxine-pyrimethamine to infants at 2, 3, 4, and 9 months of age did not have any impact on body weight at 18 months.28
There was no positive effect of the IPT on linear growth in these Senegalese preschool children. The time period chosen for assessment of impact on growth (September–November) might have been too short to detect any effect on linear growth. Also, many authors consider that linear growth is sensitive to environmental stress (food intake and infections) mainly up to 2 years of age.29 Bradley-Moore and others30 reported on a significantly greater mean HAZ in children receiving malaria chemoprophylaxis compared with controls, within the age range of 12–23 months (93.2% versus 91.7% of the NCHS mean, P < 0.05). However, we found no impact of the intervention even among children < 2 years of age (results not shown).
Unexpectedly, we found an impressive 50% decrease in the prevalence of stunting in both groups (from 23.7% in September of Year 1 to 12.2% in December of Year 2). This occurred especially during the rainy season in Year 2, although no specific malaria prevention program was provided at that time. This change was apparent already in April of Year 2 during a nutritional survey conducted on a subgroup of 543 study children 6–42 months of age: compared with September of Year 1, HAZ and the prevalence of stunting had decreased from –1.48 to –1.02 z-scores and from 31.8% to 17.4%, respectively (P < 0.0001 for both).31 As would be expected, the prevalence of underweight also decreased sharply (because weight is closely correlated with height), contrasting with a constant prevalence of wasting.
This positive trend in height-for-age over time contrasts with results of prior studies in preschool children in this area. In one large-scale study (~6,000 children), mean height-forage decreased from birth up to 18–24 months of age and remained constant from 2 to 5 years.7 No seasonal variations occurred for this indicator. In a cohort study of 443 children, the prevalence of stunting was 35.9%, 30.7%, 31.1%, and 30.5%, respectively, and mean HAZ was –1.7, –1.5, –1.5, and –1.4, respectively, at four 6-month rounds conducted at mean ages of 18.0, 23.4, 28.7, and 34.2 months.14
Few controlled nutrition or health interventions have achieved similar effect sizes on height status. However, the difficulty in interpreting the results of this study is that both intervention and control groups had improved height-for-age and study children benefited from several medical services, independently of malaria prevention and treatment. Management of acute, clinical malnutrition in an ad hoc nutritional rehabilitation center during the study is not likely to have had any notable impact on linear growth, because few children were enrolled and stayed for an average of 7 days only (A. Diallo, unpublished observations). However, weekly contacts between health workers and study children for active malaria case detection may have had a non-specific positive effect on child health and growth, the so-called "health worker effect."32,33
If malaria-specific interventions are responsible for this accelerated linear growth rate, the most likely explanation is that early treatment of clinical malaria prevented prolonged carriage of P. falciparum and its associated growth-depressing immune response. Snow and others1 suggested that clinical malaria, if adequately managed, has limited or no effects on linear growth in young children. In contrast, long-term carriage, resulting from either sub-clinical infections or non-treated clinical malaria with spontaneous recovery, may have a strong negative effect through the production of growth-depressing cytokines. Tumor necrosis factor-
(TNF) is produced in response to malaria34 and is known to inhibit the production of insulin-like growth factor (IGF), which mediates the effect of growth hormone. However, although biologically plausible, little evidence is available to support this hypothesis at the present time. In African adolescents, malaria-induced TNF production was associated with significantly lower BMI, but linear growth was not considered.34
Thus, other studies are necessary to test our hypothesis of a long-term negative effect of parasite carriage on linear growth in preschool children through the production of pro-inflammatory cytokines,34 but the design is challenging because it is ethically impossible to monitor growth in children without treating cases of clinical malaria.
In conclusion, nearly complete prevention of clinical malaria by seasonal intermittent preventive treatment improved growth in weight and skinfolds of preschool children over a 2-month period, but did not affect linear growth compared with controls benefiting from active malaria case detection and prompt treatment.
Received January 25, 2007. Accepted for publication May 13, 2007.
Acknowledgments: The authors thank Amady Ndiaye, Tofène Ndiaye, and Bassirou Fall for taking the anthropometric measurements, Adama Marra for identification of eligible subjects and data management, Jacqueline Milet for help with anthropometric data management, Ernest Faye for supervision in the field, and the entire study field team for their work.
Financial support: The study was funded by the Gates Malaria Partnership, which receives support from the Bill and Melinda Gates Foundation, the LSHTM DFID Malaria Knowledge Programme, and the IRD.
* Address correspondence to Kirsten B. Simondon, Centre IRD, BP 64501, 34394 Montpellier, France. E-mail: Kirsten{at}ird.fr 
Authors’ addresses: Balthazar Ntab, François Simondon, and Kirsten B. Simondon, Centre IRD, BP 64501, 34394 Montpellier, France. Badara Cissé, UCAD, Service de Parasitologie, BP 5005, Dakar, Senegal. Denis Boulanger, Cheikh Sokhna, and Jean-François Trape, Centre IRD, BP 1386, CP18524, Dakar, Senegal. Geoffrey Targett, Jo Lines, Neal Alexander, and Brian Greenwood, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK.
|
REFERENCES |
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
