点击显示 收起
1 From the Projecto de Saúde de Bandim, Bissau, Guinea-Bissau; the Department of Epidemiology Research, Danish Epidemiology Science Centre, Statens Serum Institut, Copenhagen; the Center for International Health, University of Bergen, Norway; All India Institute of Medical Sciences, New Delhi; and Laboratorio Nacional de Saúde Publica, Bissau, Guinea-Bissau.
2 Supported by the STD3 Programme, the Commission of the European Union (TS3*-CT94-0311), the Danish Council for Development Research (104.Dan.8/717 and 9501931), DANIDA, The Danish Medical Research Council, The Danish National Research Foundation, University of Bergen (grant to HS), L Meltzers Høyskolefond, and The Norwegian Research Council. 3 Address reprint requests to P Valentiner-Branth, Department of Epidemiology Research, Statens Serum Institut, Artillerivej 5, DK-2300 Copenhagen S, Denmark. E-mail: pvb{at}ssi.dk.
ABSTRACT
Background: Uncontrolled hospital-based studies in developing countries have reported promising results of dietary rehabilitation of children with persistent diarrhea.
Objective: The objective was to determine the immediate and long-term effects of a dietary supplement and micronutrients given to children with persistent diarrhea during the episode and for 1 wk during convalescence.
Design: The study was open, controlled, and community-based and was conducted in a periurban area in Guinea-Bissau. Children <3 y of age with persistent diarrhea were identified during weekly household visits. The children randomly assigned to the treatment and control groups were examined by a physician and all medical conditions were treated. The children in the treatment group were offered home-based dietary treatment consisting of locally available foods and micronutrient supplements.
Results: There were 141 episodes of persistent diarrhea during the study: 70 in the treatment group (in 58 children) and 71 in the control group (in 62 children). During the intervention period (median: 17 d), weight gain in the treatment group exceeded that of the control group by 61.5 g/wk (95% CI: 49.2, 73.8), whereas there was no significant difference in linear growth on the basis of knee-heel length. At a median follow-up period of 6.6 mo after the intervention was stopped, weight gain in the treatment group exceeded that of the control group by 12.5 g/wk (95% CI: 7.7, 17.3); knee-heel length was 7.5 mm/y (4.8, 10.2) greater and total length was 0.65 cm/y (0.11, 1.19) greater in the treatment group.
Conclusion: Therapeutic feeding and micronutrient supplementation had an immediate and sustained beneficial effect on growth in children with persistent diarrhea.
Key Words: Persistent diarrhea • dietary treatment • micronutrients • growth • children • community-based study • controlled study • Guinea-Bissau
INTRODUCTION
Persistent diarrhea is defined as a diarrheal episode of presumed infectious etiology lasting 14 d (1). This operational definition identifies children with an elevated diarrheal burden (2, 3) and an elevated risk of nutritional insult (4). The management of children with persistent diarrhea is frustrating for both health professionals and family members, leading to a substantial overuse of drugs and other resources (5); effective management protocols are needed. Zinc supplementation during acute diarrhea reduces the duration of the episode (6, 7). The effect of zinc supplementation during an episode of persistent diarrhea remains to be established because studies have not shown a significant reduction in the duration of the episode (8, 9) or showed a shorter duration in certain subgroups only (10, 11). Recent studies of the long-term effect of short-course zinc supplementation, initiated during an episode of acute or persistent diarrhea, showed reduced diarrheal morbidity (12) and increased linear growth (13) during follow-up. In recent years, centers in developing countries reported promising results in managing children with persistent diarrhea by using dietary treatment consisting of locally available inexpensive foods and micronutrient supplementation. A multicenter, hospital-based study indicated that a protocol with a brief period of nutritional rehabilitation led to cessation of the diarrheal episode and weight gain in 88% of the children (14). However, because the study did not include a control group, it was not possible to separate the effect of the nutritional supplementation from the catch-up growth that occurs in most children after recovery from diarrhea (15, 16). Furthermore, because there was no follow-up after discharge from the hospital, no conclusions could be drawn about the long-term effect of the treatment algorithm on growth. To address these issues we implemented a modified protocol for the management of persistent diarrhea in a controlled study that included follow-up after the intervention.
The purpose of this dual intervention was to restore a positive nutritional balance via dietary supplementation to achieve catch-up growth and to restore intestinal mucosal integrity by providing energy and important bowel nutrients such as vitamin A and zinc. Because most children with persistent diarrhea are unlikely to receive hospital care, this study was carried out in the community and, except for the initial instruction at the local health center on how to prepare and use the diet, the dietary treatment was home-based.
SUBJECTS AND METHODS
Subjects
The present study, which was approved by the Ministry of Public Health in Guinea-Bissau and the Danish Central Scientific Ethical Committee, was conducted in a periurban area of the capital of Guinea-Bissau. It was designed as an open, controlled study nested in a prospective community-based surveillance study of diarrheal disease. A total of 603 houses were selected at random in the districts Bandim II and Belem and were randomly assigned to either a treatment or a control group. All children residing in these houses who were born between 1 July 1994 and 15 January 1997 were included. Diarrhea was defined according to the mother's perception of diarrhea and persistent diarrhea as diarrhea lasting for 14 d with <2 consecutive intervening days without diarrhea. The duration of diarrhea after inclusion was defined as ending with the last day of diarrhea followed by a 48-h diarrhea-free period.
Diet
On the basis of the results from a previously conducted dietary survey (17), a millet gruel was chosen for the dietary treatment. This traditional complementary food consists of millet, egg, banana, margarine, and sugar. It is accepted by a broad age range of children, and mothers will feed the gruel to children who have diarrhea. Furthermore, it is logistically simple to distribute and to prepare because it is supplied as a flour mix containing all ingredients except water. The dietary requirements for catch-up growth are an energy density of 3.8 MJ/L and 8.7% of energy from protein (18). The gruel was modified to meet these criteria by adding to it full-strength milk powder, sugar, and margarine (17).
Procedures
All children were visited weekly and stool samples were collected from those with acute diarrhea. Children who had had diarrhea for 7 d were revisited 14 d after the onset of that episode. If the diarrheal episode was found to persist, informed oral consent was obtained from the mothers to treat the children at the Bandim health center. These children were examined by a physician and offered fluid therapy and antibiotic treatment according to World Health Organization guidelines (19). Exclusion criteria included suspicion of acute surgical conditions and danger signs; however, no children were excluded on the basis of these criteria. The importance of continued breast-feeding and of feeding their children plenty of nutritious foods was stressed to the mothers. In addition, the children in the treatment group were offered dietary treatment and micronutrient supplementation. The amount of micronutrients offered was approximately twice the recommended dietary allowance (RDA; 20), provided as 2 tablets/d for children >4 mo of age and as 1 tablet/d for children 3–4 mo of age (see Table 1 for daily dosages). The tablets were chewable but could be easily crushed to be mixed in the gruel if desired. Children <3 mo of age received no dietary treatment. Children aged 3–4 mo were provided dietary treatment only if they had already started receiving complementary foods. Thus, a total of 3 children who were randomly assigned to the treatment group received micronutrients as the sole treatment. From 4 mo of age, dietary treatment was instituted regardless of nutritional history.
View this table:
TABLE 1.. Daily dosages of micronutrients supplied to the children in a controlled community-based study of persistent diarrhea in Guinea-Bissau (1996–1997)1
The first dietary treatment was provided at the health center. After the gruel was prepared in the presence of the mothers, the children were given the tablets and a portion of the gruel. The remaining treatments were provided at home. The mothers were supplied with the flour mix and micronutrient tablets and were instructed to prepare the gruel 3 times/d and to feed it to their children 6 times/d. The micronutrient supplement was to be given once a day. The daily energy intake was targeted to be 628 kJ/kg body wt. The family was visited daily by a fieldworker, except on weekends and public holidays, to interview the mothers about the children's intake of the gruel and tablets. The children in the treatment group were visited a total of 1138 times (median: 11/episode; interquartile range: 9–20/episode) to monitor compliance. Dietary intervention continued until the end of a 7-d interval without diarrhea. The intake of other foods was not quantified. To standardize the conditions for the control and treatment groups, the mothers in both groups were asked about their children's diarrhea only during the weekly diarrhea-recall visits. The children received a median of 2 diarrhea-recall visits during the intervention period (interquartile range: 2–4). Each mother participated with one child only.
Growth monitoring
Knee-heel length (21, 22) and body weight were measured at inclusion, ie, when the children had had diarrhea for 14 d, and weekly thereafter during the intervention period. During follow-up, knee-heel length and body weight were measured monthly. Total length was measured at the time of inclusion and at 3-mo intervals thereafter. In addition, all children had their length and weight measured at 6, 9, and 18 mo of age and a subgroup also had measurements taken at 7.5 and 10.5 mo of age. The children were weighed while nude or dressed in light clothing with an electronic Seca scale (Seca, Hamburg, Germany) at the health center and with a portable balance (Super Samson Scale; Salter, West Bromwich, United Kingdom) in the field. Total length was measured with a measuring board as height in children from 2 y of age and as recumbent length in children <2 y of age. The duration of anthropometric follow-up was calculated from the end of the intervention period until the end of follow-up—defined as the last anthropometric measurement before the beginning of a new episode of persistent diarrhea or before 31 December 1997.
Microbiological analyses
From 15 January 1996 to 1 January 1998, microbiological examination of stool samples from 138 diarrheal episodes progressing to persistent diarrhea was undertaken by using methods described previously (23), except that in the present study a semisolid, blood-free, selective motility medium was used to identify Campylobacter spp. (24). Furthermore, DNA-DNA colony hybridization with 9 different DNA probes was used to identify diarrheagenic Escherichia coli, including enteroaggregative E. coli (25), enteroinvasive E. coli (26), diffusely adherent E. coli (27), verotoxin-producing E. coli (28), eae-Positive E. coli (29), and enterotoxigenic E. coli (30). In addition, isolates with the characteristics of Salmonella or Shigella were examined with the F1217 (31) and the ipaH DNA probes (26), respectively. The samples (one per episode) were selected as shortly after the beginning of the diarrheal episode as possible and no later than 31 d after the onset of the episode, leading to the exclusion of 3 episodes. Rectal swabs were taken in 11 samples from the treatment group and in 11 samples from the control group, ie, in 15.9% (22/138) of the samples.
Data analyses
We analyzed the anthropometric data by using a generalized linear mixed model tailored for the analysis of repeated measurements (32–34). Maximum likelihood methods were used for estimation, and hypothesis testing was done by using log likelihood ratio tests.
Because the effect of the treatment algorithm on ponderal and linear growth was likely to depend on age, sex, and nutritional status, we estimated indexes separately for each sex and for the age groups 0–11, 12–23, and 24 mo. Nutritional status was grouped as weight-for-age or height-for-age z scores < or -2, depending on the model. On the basis of pooled anthropometric measurements from the study area, the age dependence of ponderal growth was modeled by a piecewise linear curve, allowing the coefficient to be different depending on whether the child was >or <3 mo of age; the age dependence of linear growth was modeled by a piecewise linear curve, allowing the coefficient to be different depending on whether the child was > or <5 mo of age.
The PROC MIXED procedure of the SAS software package (version 6.12; SAS Institute, Cary, NC) was used for the regression analysis and the EPINUT program (Epiinfo; Centers for Disease Control and Prevention, Atlanta) was used for computing anthropometric z scores.
The prevalence of diarrhea during follow-up, defined as the number of days with diarrhea divided by the total number of days with information about diarrheal status for each child, was analyzed by logistic regression with generalized estimating equations to correct for repeated entries. Morbidity was recorded during the weekly visits. Information on morbidity recorded for >7 d retrospectively was recorded as missing. To compare the tendency of the subjects in the treatment and control groups to develop a new diarrheal episode, the period from the end of the persistent diarrheal episode to the subsequent diarrheal episode was determined by using the log-rank test. Energy intakes from gruel and the duration of diarrheal episodes were analyzed by using the Kruskal-Wallis test.
RESULTS
From 1 January 1996 to 31 December 1997, 141 episodes of persistent diarrhea were identified and included for analysis in the study. Of the 120 children taking part in the study, 101 children participated once, 17 children participated twice, and 2 children participated 3 times. A total of 71 episodes (in 62 children) from the control group and 70 episodes (in 58 children) from the treatment group were included during the whole study period. Restricting the analyses to the first episode in each of the 120 children did not change the ponderal or linear growth estimates. Accordingly, all 141 episodes were included in the analyses.
Baseline characteristics
The baseline characteristics of the children at the time of diagnosis with persistent diarrhea are shown in Table 2. The only significant difference between the treatment and control groups was in height-for-age z scores. The prevalence of stunting (defined as a height-for-age z score < -2) was higher in the control group (n = 15) than in the treatment group (n = 9); however, this difference was not significant [odds ratio (OR): 1.82; 95% CI: 0.67, 4.97]. Nutritional status and age were effect modifiers of the growth estimates; therefore, age, sex, and nutritional status were included in the models. Neither malarial parasitemia, dysentery, or pneumonia confounded the estimates of growth and thus were not included in the final models.
View this table:
TABLE 2.. Characteristics at entry of the children with persistent diarrhea in a controlled community-based study of persistent diarrhea in Guinea-Bissau (1996–1997)
Isolation frequency of enteric pathogens
The stool samples were collected at a mean of 6.1 d (median: 5.0 d) after the beginning of the diarrheal episode. Seventy-two percent of the samples were collected during the first week of diarrhea (interquartile range: 2–9 d). No pathogens were found in 38% of the control and treatment groups, one pathogen was found in 45% of the control group and 38% of the treatment group, and more than one pathogen was found in 17% of the control group and 25% of the treatment group. There were no significant differences in enteropathogen isolation frequencies between the 2 groups (Table 3).
View this table:
TABLE 3.. Isolation of established and potential enteropathogens from stool samples of 138 episodes of diarrhea progressing to persistent diarrhea in a controlled community-based study of persistent diarrhea in Guinea-Bissau (1996–1997)1
Adherence to the treatment protocol
The mean daily intake of millet gruel was 104 mL/kg body wt (median: 102 mL/kg body wt), corresponding to a median daily energy intake of 385 kJ/kg body wt and a median daily lactose intake of 1.25 g/kg body wt (interquartile range: 0.96, 1.58). The energy intake from the gruel was significantly higher in the underweight children (weight-for-age z score < -2) than in the nonunderweight children (weight-for-age z score
Information on the duration of a diarrheal episode was missing for 3 episodes in the treatment group and for 8 episodes in the control group (6 because of traveling, 3 because of death, and 2 because of relocation outside the study area); therefore, these episodes were excluded from the analyses of diarrheal duration. We found no significant differences in the duration of diarrhea (median: 7 and 6 d, respectively) or in the duration of the intervention period (median: 17 and 16 d, respectively) after inclusion in the study. The study had a power of only 17% to detect a clinically relevant difference of 3 d in the duration of diarrhea. Two children in the treatment group and 2 in the control group were hospitalized during the intervention period because of systemic chronic infection, dehydration, septicemia, or middle ear infection. Two children in the control group died during the intervention period, one from septicemia and one from pneumonia; one severely malnourished child in the treatment group died, likely from a systemic chronic infection.
Diarrheal morbidity
The prevalence of diarrhea in the period from the end of the intervention period until 31 December 1997 was compared between the control and the treatment groups for the children participating in the study for the first time with an episode of persistent diarrhea. Information about the duration of the diarrheal episode at entry was missing for 11 children and morbidity data were missing for another 2 children after the end of the intervention period. Thus, data for 107 of 120 children were used in the analysis. The children had diarrhea on 2578 of the 24936 study days and the prevalence of diarrhea was thus 10.3%. The mean follow-up period, defined as the number of days with information about diarrheal status, was 6.87 mo (interquartile range: 3.32–9.49 mo; n = 53) in the control group and 8.43 mo (4.86–12.78 mo; n = 54) in the treatment group. In a multivariate analysis in which age was controlled for, the prevalence of diarrhea in the treatment group during follow-up was not significantly lower than in the control group (OR: 0.85; 95% CI: 0.60, 1.20). There was no significant difference in the period from the end of the persistent diarrheal episode until the first subsequent diarrheal episode between the children in the 2 groups (Plog rank = 0.91).
Effect of the intervention on weight gain
The final model for weight gain included a total of 1829 measurements. The effects of the dietary treatment, combined for age and nutritional status, are shown in Table 4. The treatment group had a weight gain exceeding that of the control group by 61.5 g/wk during the intervention period and by 12.5 g/wk during follow-up (median: 6.6 mo). In children >3 mo of age, weight gain during the intervention period was 98 g/wk in the control group and 159 g/wk in the treatment group; during follow-up, the weight gain was 38 g/wk in the control group and 50 g/wk in the treatment group. During the intervention period, the weight gain in both groups was significantly larger than the weight gain during follow-up. There was substantial catch-up growth in both the control and treatment groups during the intervention period, but to a larger extent in the treatment group. Nutritional status had a significant effect on the additional weight gain in the treatment group in the intervention period (P < 0.0001). Thus, in children <2 y of age the additional weight gain in the treatment group was higher in underweight children than in nonunderweight children (Table 5). Few children 2 y were underweight and consequently the estimates for this age group could not be determined accurately. During follow-up, the additional weight gain in the treatment group was independent of age, sex, and nutritional status.
View this table:
TABLE 4.. Weight gain and increases in knee-heel length and height in the treatment group (n = 70) exceeding that of the control group (n = 71) in the intervention period and during follow-up in a controlled community-based study of persistent diarrhea in Guinea-Bissau (1996–1997)1
View this table:
TABLE 5.. Weight change in the treatment group (n = 70) relative to that of the control group (n = 71) in the intervention period in a controlled community-based study of persistent diarrhea in Guinea-Bissau (1996–1997)1
Effect of the intervention on linear growth
Knee-heel length
A total of 1165 knee-heel length measurements were used in the model. During the intervention period, there was no significant increase in knee-heel growth in the treatment group (2.7 mm/y; Table 4). During follow-up (median: 5.4 mo), however, knee-heel growth in the treatment group exceeded that of the control group by 7.5 mm/y. The additional knee-heel growth during follow-up was independent of age, sex, and nutritional status on the basis of height-for-age z scores. Knee-heel growth during the intervention period was 30.4 and 33.1 mm/y and during follow-up was 30.4 and 37.9 mm/y for children >5 mo of age in the control and treatment groups, respectively. In the control group, there was no significant difference in knee-heel growth between the intervention and follow-up periods.
Total length (height)
A total of 623 length measurements were used in this model. The increase in length from the time of entry to the end of follow-up (median: 6.0 mo) in the treatment group exceeded that in the control group by 0.65 cm/y (Table 4). Linear growth was 9.2 and 9.9 cm/y in children >5 mo of age in the control and treatment groups, respectively. As in the knee-heel-length model, the additional linear growth was independent of age, sex, and nutritional status on the basis of height-for-age z scores.
DISCUSSION
The present study, conducted in a West African community, documents short-term and long-term weight gains and a long-term improvement in linear growth after dietary treatment and micronutrient supplementation in children with persistent diarrhea. The duration of the presenting diarrheal episode did not differ between the control and treatment groups. Although the criteria for stopping dietary treatment was not explained to the mothers, we cannot exclude the possibility that some of the mothers gave incorrect information about the diarrheal status of their children to prolong dietary supplementation.
The information required to calculate sample sizes was not available before initiation of the study, which did not have sufficient power to determine a clinically relevant difference in the duration of the diarrheal episodes. Other studies suggested a reduction in the duration of diarrheal episode in children with persistent diarrhea who were offered dietary treatment (35, 36) but none of these studies included a control group. Some randomized controlled trials in children with acute diarrhea showed that zinc supplementation reduces the duration of diarrheal episodes (6, 7), whereas supplementation during established persistent diarrheal episodes does not seem to have a significant effect (8–11).
There was a tendency to a reduced prevalence of diarrhea in the treatment group during the period of 8 mo of follow-up (OR: 0.85; 95% CI: 0.60, 1.20). A 1.56-mo shorter mean follow-up period for the control group was mainly due to a difference between the groups in the time from the end of the intervention to the end of the study. Assuming that the effect of the intervention decreased with time, the relatively longer follow-up for the treatment group may have led to an underestimation of the difference between diarrheal morbidity in the 2 groups. This OR points in the same direction as that of a recent pooled analysis of randomized controlled trials of short-term (2 wk) zinc supplementation, which showed reduced diarrheal prevalence (OR: 0.66; 95% CI: 0.52, 0.83) during follow-up (12). Note that the present study provided only 1–2 times the RDA of zinc whereas the cited trials provided 2–4 times the RDA.
In the present study, the mothers' definition of diarrhea was used as opposed to the clinical definition of 3 loose stools/24 h. The correspondence between these definitions was not examined in the present study but was shown to be acceptable in other settings (37, 38).
The present study showed a substantial catch-up in weight during the intervention period in both the control and treatment groups. This indicates that the increased weight gain reported during the dietary treatment of persistent diarrhea in earlier studies (14, 35, 36) and in the present study might be attributed partly to factors related to the recovery from the diarrheal episode and partly to an additional catch-up growth resulting from the intervention. Episodes of persistent diarrhea were shown to be associated with a decline in ponderal and linear growth (39–41). Some controversy exists as to whether diarrhea-induced growth faltering is transient or sustained (15, 16, 42). Our findings corroborate the notion that catch-up growth partially covers the weight loss associated with persistent diarrhea. However, the negative effect of persistent diarrhea on linear growth was sustained during the period of follow-up, ie, was not covered by catch-up growth.
During follow-up we found higher rates of linear growth in the treatment group than in the control group, measured both by knee-heel length and total length. The knee-heel length is considered to constitute 25–30% of the total length of the child (22). However, it cannot be assumed that an increase in knee-heel length, mainly reflecting growth of long bones, is directly proportional to the overall linear growth, which is mainly influenced by the growth of short bones in the spine (22, 43).
The treatment protocol consisted of 2 interventions, either of which may have contributed to the increased growth observed. Compliance with tablet ingestion was excellent, whereas compliance with the study diet was somewhat lower than recommended. The estimate of additional weight gain or of linear growth (data not shown) was not significantly different when based on a high compared with a low intake of the study diet [> or <80 mL•kg body wt-1•d-1, respectively (25th percentile)]. The intake of other foods and of breast milk could not be quantified, so the magnitude of the effect of energy intake on growth could not be estimated in the present study. According to a study in Bangladesh, zinc supplementation during acute diarrhea resulted in significantly increased weight gain during the supplementation period (44), indicating that micronutrient supplementation independent of the dietary treatment may improve catch-up growth.
Children with persistent diarrhea have lower concentrations of vitamin A and zinc than do age-matched control children (45). These deficiencies may be exacerbated further if persistent diarrhea is combined with malnutrition (46). In a study of malnourished children in Bangladesh who were treated with a rice diet, those receiving vitamin supplements containing iron and zinc had greater weight gains than did those receiving vitamin supplements containing iron but no zinc (47). Thus, in the present study, underlying micronutrient deficiencies combined with increased dietary intakes may explain why the underweight children benefited most from the intervention.
In the present study, the higher rates of ponderal and linear growth in the treatment than in the control group were sustained beyond the intervention period. According to a study in Bangladesh, short-term zinc supplementation during acute diarrhea increased the subsequent linear growth and reduced the incidence of diarrhea in both stunted and underweight children during follow-up (13). Furthermore, a pooled analysis of short-term zinc supplementation showed reduced diarrheal prevalence during follow-up (12). This indicates that a period of adequate zinc consumption may have more prolonged functional benefits. It is conceivable that children consuming a diet marginal in zinc, for example, could maintain adequate zinc status longer if their micronutrient stores are replenished after a period of supplementation. Although no significant reduction in the prevalence of diarrhea was observed in the present study, the better nutritional status of the treatment group may explain the reduced susceptibility of this group to the detrimental effects of diarrhea on growth during the follow-up period (48). Another possible reason for the increased growth was long-term changes in maternal feeding behavior as a result of the intervention.
Although the present study was designed as a controlled clinical trial, it could not be blinded. Furthermore, the fieldworkers were required to have closer contact with the treatment households than with the control households in the intervention period. However, because the mothers from both groups were asked about diarrhea during the weekly diarrhea-recall visits only, this design was not likely to have influenced the recorded outcome of the diarrheal episode.
In conclusion, we report that children with persistent diarrhea can be managed with home-based dietary treatment with a single low-lactose, cereal, milk-based diet consisting of locally available foods, micronutrient supplementation, and antibiotic treatment when indicated. Although no significant effect on diarrheal morbidity was shown, the intervention did lead to short- and long-term beneficial effects on ponderal growth and a long-term beneficial effect on linear growth in children. The study was not designed to show whether this effect was due to the micronutrient supplement alone, the dietary treatment, or both. Controlled trials are important for the establishment of evidence-based treatment regimens for this important health problem in developing countries. The described treatment protocol may well be modified for implementation in a wide range of health institutions, eg, in outpatient clinics of local hospitals and in health centers. Future research is warranted to examine the sustainability and effectiveness of the intervention when implemented on a larger scale.
ACKNOWLEDGMENTS
We thank Queba Djana, Francisco da Silva, Carlos Sá, Adelino Bassam, Alexandre Nhaga, Leontino da Silva, and Domingos Pereira, all of whom participated in the fieldwork; Kim Fleischer Michaelsen, The Royal Veterinary and Agricultural University, for his advice regarding dietary treatment and the use of knemometry; and Flemming Scheutz, Statens Serum Institut, for his help in establishing the microbiological analyses.
REFERENCES