The added benefit of zinc supplementation after zinc treatment of acute childhood diarrhoea: a randomized, double-blind field trial


Corresponding Author Charles P. Larson, BC Children’s Hospital, Room K4-104, 4480 Oak Street, Vancouver, BC, Canada V6H 3V4. Tel.: +1 604 875 2345 ext. 5642; Fax: +1 604 875 3076; E-mail:


Objectives  To determine whether continuing with zinc supplementation after zinc treatment (ZT) of an acute diarrhoea episode will result in additional clinical benefits beyond ZT alone.

Methods  Children 6–23 months of age, living in an urban slum in Dhaka, Bangladesh with acute childhood diarrhoea (ACD), were enrolled in a randomized, double-blind field trial. All children received 10 days of ZT (20 mg/day) and were then randomized to zinc (10 mg/day) or placebo supplementation for 3 months. Weekly follow-up of all children occurred over a period of 9 months.

Results  A total of 353 subjects were enrolled, with 93% of the zinc supplemented and 96% of the placebo children followed for 9 months. The incidence density of ACD among those receiving zinc supplementation compared to those receiving placebo was reduced by 28% (2.64 vs.3.66 episodes/p-y follow-up) over the 3 months while on supplementation and by 21% (2.05 vs.2.59 episodes/p-y follow-up) over the 9 months of follow-up. There was no observed effect on the incidence of acute respiratory infections (ARIs) or on growth.

Conclusions  Zinc supplementation after treatment provides additional preventive ACD benefits to children in early childhood. Larger, effectiveness trials of this strategy are warranted.


L’avantage additif de la supplémentation en zinc après traitement avec du zinc de la diarrhée aiguë de l’enfant: une étude randomisée, en double aveugle sur le terrain

Objectifs:  Déterminer si la poursuite d’une supplémentation en zinc après un traitement avec du zinc lors d’un épisode de diarrhée aiguë se traduira par des avantages cliniques supplémentaires par rapport au traitement seul avec du zinc.

Méthodes:  Des enfants de 6 à 23 mois d’âge, vivant dans un bidonville de Dhaka, au Bangladesh, avec une diarrhée aiguë ont été recrutés dans une étude randomisée, en double aveugle sur le terrain. Tous les enfants ont reçu un traitement de 10 jours avec du zinc (20 mg/jour) et ont ensuite été randomisés pour recevoir soit une supplémentation de zinc (10 mg/jour) ou de placebo pendant 3 mois. Tous les enfants ont été suivis hebdomadairement pendant 9 mois.

Résultats:  Un total de 353 sujets étaient recrutés, avec 93% de ceux recevant la supplémentation en zinc et 96% de ceux recevant le placebo, suivis pendant 9 mois. Le taux d’incidence de diarrhée aigüe de l’enfant (DAE) chez ceux recevant la supplémentation en zinc par rapport à ceux recevant le placebo a été réduit de 28% (2,64 vs 3,66 épisodes/personnes année suivies) au cours des 3 mois sous supplémentation et de 21% (2,05 vs 2,59 épisodes/personnes année suivies) sur les 9 mois. Un effet sur l’incidence des IRA ou sur la croissance n’a pas été observé.

Conclusions:  La supplémentation en zinc après traitement offre des bénéfices de prévention additionnelle contre la DAE aux enfants dans la petite enfance. Des études d’efficacité plus larges sur cette stratégie sont justifiées.


El beneficio añadido de la suplementación con zinc después de haber tratado con zinc la diarrea aguda en niños: un ensayo de campo aleatorizado y doble ciego

Objetivos:  Determinar si el suplementar con zinc después de haber tratado con zinc un episodio de diarrea aguda resulta en beneficios clínicos adicionales más allá del tratamiento único con zinc.

Métodos:  Niños con edades comprendidas entre los 6 y 23 meses, viviendo en una barriada urbana en Dhaka, Bangladesh, y con diarrea infantil aguda, fueron incluidos en un ensayo de campo aleatorizado y doble ciego. Todos los niños recibieron tratamiento con zinc durante 10 días (20 mg/día) y fueron aleatorizados para recibir suplementación con zinc (10 mg/día) o placebo durante los 3 meses siguientes a la finalización del tratamiento. Todos los niños fueron seguidos semanalmente durante 9 meses.

Resultados:  Se incluyeron un total de 353 sujetos, con un 93% de los niños suplementados con zinc y un 96% de los niños con placebo seguidos durante 9 meses. La densidad de incidencia de diarrea aguda entre aquellos recibiendo suplementación con zinc comparada con aquellos recibiendo placebo se redujo en un 28% (2.64 vs.3.66 episodios/p-año de seguimiento) durante los 3 meses que recibieron suplementación y en un 21% (2.05 vs.2.59 episodios/p-año de seguimiento) durante los 9 meses. No se observó un efecto sobre la incidencia de IRAs o en el crecimiento.

Conclusiones:  La suplementación con zinc después del tratamiento provee beneficios adicionales en la prevención de la diarrea infantil aguda durante los primeros meses de vida. Son necesarios ensayos más grandes para corroborar su efectividad.


Zinc deficiency is widespread among children in developing countries with zinc-deficient diets (Holtz & Brown 2004). Facility and population-based studies have consistently observed an association between zinc deficiency and infectious disease morbidity (Bahl et al. 1998; Srinivas et al. 1998). Children with relatively modest zinc deficiency experience higher rates of skin infections, diarrhoea, respiratory infections and delayed wound healing (Aggett & Comerford 1995; Bahl et al. 1998). Based upon estimates of the impact of zinc deficiency on diarrhoeal disease alone, it is estimated that correction of this deficiency has the potential to save nearly 450,000 under five deaths annually (Jones et al. 2003).

Randomized efficacy trials of daily zinc supplementation have consistently demonstrated reductions in the occurrence of acute and persistent childhood diarrhoea (Sazawal et al. 1996, 1997; Zinc Investigators’ Collaborative Group 2000; Bhandari et al. 2002a; Penny et al. 1999; Brown et al. 2009). Similarly, several randomized efficacy studies of zinc as a treatment for childhood diarrhoea have demonstrated both curative and preventive benefits to children under 5 years of age. This includes decreased duration or severity of an acute or persistent diarrhoeal illness episode and decreased likelihood of a repeat diarrhoea episode (Sazawal et al. 1995; Bhutta et al. 1999; Baqui et al. 2002; Aggarwal et al. 2007; Haider & Bhutta 2009). Based upon this evidence, the WHO/UNICEF guidelines for the treatment of childhood diarrhoea have been modified to include zinc treatment (ZT) of any diarrhoeal illness episode in children under 5 years of age (WHO/UNICEF 2004). Current recommendations do not include continued daily supplementation after treatment in children who have experienced a diarrhoeal illness. Is this a missed opportunity to further reduce the likelihood of recurrent infectious illnesses?

The aims of this investigation were twofold. First, to determine whether there is a reduced likelihood of repeat infection of public health importance if a child continues to receive zinc supplementation for an additional 3 months following 10 days of ZT of an acute diarrhoea episode. Potential benefits would be preventive and include further decreases in the recurrence of acute childhood diarrhoea (ACD) or acute respiratory infections (ARIs). Second, to document the immunological response to zinc among a subset of children whose diarrhoea was caused by enterotoxigenic Escherichia coli (ETEC). This article addresses the first objective.


Study design

This was a randomized, double-blind field trial conducted in an urban slum located within Mirpur District, Dhaka, Bangladesh. The study was carried out between November, 2004 and August, 2006. It was reviewed and separately approved by the ICDDR,B scientific and ethical review committees. Under the title “Immunological and clinical responses to zinc in children with diarrhoea, this study was registered with Current Controlled Trials, ID No. NCT00408356, and a data safety and monitoring board created. Informed written consent was obtained from each participating child’s mother or father.

Hypotheses tested

We hypothesized that children 6–23 months of age who receive an additional 3 months of zinc supplementation (10 mg/day) immediately after 10 days of ZT for ACD will experience a significantly fewer episodes of ACD or ARI than children receiving a placebo. This hypothesis was tested at 3 months (while on supplementation) and 9 months (6 months after terminating zinc or placebo supplementation).

Study population

Children 6–23 months of age living in urban slums in Mirpur, Dhaka, with an acute episode of diarrhoea of 24–72 h duration were eligible for entry into this study. Diarrhoea was defined as the occurrence of 3 or more loose or watery stools in the past 24 h. Patients were identified through systematic, daily door-to-door household case finding by trained field research assistants and then referred to a local ICDDR,B field clinic in Mirpur, where they were seen by staff physicians. A physician assessed each child for the following exclusion criteria: severe dehydration, suspected cholera or pneumonia, bipedal oedema, currently receiving zinc, a weight-for-height z-score below −3 or already participating in another study involving nutritional or therapeutic interventions. Informed written consent was obtained from the parent of each child fitting the study’s eligibility criteria. Following consent, a stool specimen was sent for culture to identify children positive for ETEC, as the immunological investigations were limited to children with this diarrhoea aetiology. The results of the stool cultures were reported within 48–72 h. Enrolment was continued until the desired sample size for both the immunological and clinical outcome studies were obtained.

Sample size estimation

This study was powered to detect differences in ACD incidence density. We assumed the baseline incidence of ACD would be four episodes per person-year of exposure, with a standard deviation of one episode. The level of confidence was set at 0.95 and the power at 0.90. Allowing for a minimal detectable reduction in 0.4 episodes per person-year exposure (20% reduction) in the incidence of ACD over any specified follow-up interval, the calculated sample size was 132/group. Assuming a 20% loss to follow-up, we aimed to enrol 162 subjects per group.


Zinc treatment.  All children entered into the study received once daily for 10 days 20 mg of zinc as sulphate, provided in a blister pack of 10 dispersible tablets. The first dose was given at the clinic after a demonstration of its preparation.

Zinc supplementation.  At day 10, immediately following a full course of treatment, children were randomized to receive 10 mg zinc as sulphate or placebo supplementation daily for a period of 90 days. Both were provided in a locally produced syrup formulation. The syrup was chosen because 10 mg tablets were not available at the time of the study. Caretakers were advised to give their children the syrup formulations in the morning without mixing with food.


Randomization occurred on the third day, following receipt of the culture results. The trial of placebo vs. supplementation started immediately following 10 days of treatment, as depicted in Figure 1. The supplier of the syrups prepared 10 lots of placebo syrup (100 bottles per lot) and 10 lots of zinc syrup. These lots were numbered and randomly assigned in permuted blocks of six (three placebo, three zinc).

Figure 1.

 Architecture of study, study population selection and follow-up results.

All placebo and zinc-containing bottles of syrup (1000 of each) were serially numbered in lots of 100. These lots were listed and then sequentially selected based upon random assignment of permuted blocks of six. Upon completion of the 10-day ZT, each child received a 3-month supply of syrup (five bottles) and the lot number was recorded. The randomization code was not broken until after all children had completed the trial and the data had been entered and verified.


Upon entry into the study, caretakers were interviewed to provide information on household social and demographic characteristics and the child’s illness history. Heights and weights were also recorded and a stool for culture obtained. All children were tested for serum zinc levels on days 3, 15, 90 and 180. A non-fasting specimen was drawn at the study field station by venepuncture and collected in containers specifically designated for heavy metal analyses. Height and weight measurements were repeated at 90, 180 and 270 days. At days 3 and 10, a field research assistant visited each child’s household to document clinical outcomes related to the acute diarrhoea episode and adherence to treatment instructions. Thereafter, households were visited weekly for the 9-month duration of follow-up. During these visits, field research assistants documented compliance with syrup intake. All new or continuing episodes of diarrhoea (three or more loose or watery stools in the past 24 h) or ARIs (presence of cough or sore throat, with or without fever or coryza) were documented. Adherence with treatment or supplementation instructions was based upon parent report and visual inspection of the bottles.


Data were entered and verified using spss PC version 12.0. The data set was then converted to stata version 10 for all analyses. For each variable, frequency distributions were run to identify outliers and to assess the symmetry of the data. Descriptive summary means or proportions with estimates of spread were then determined. This was followed by bivariate testing and multiple linear regression analyses for main group effects, adjusted for host and socio-demographic characteristics. Negative binomial regression models were tested for intervention group by time period or sex interaction.


As summarized in Figure 1, we identified 770 children between 6 and 23 months of age with an acute episode of diarrhoea, of whom 652 (85%) fit the study criteria and had parental consent to participate. All these children received 10 days of zinc in treatment. A total of 148 children with ETEC-positive diarrhoea were identified from stool culture results at day 3. To achieve the desired sample size, another 205 children with ETEC-negative stool cultures were systematically selected, for a total of 353 study subjects. These children were randomly assigned to either zinc supplementation or placebo after receiving the 10 days of ZT. The two groups were comparable in terms of the child’s age, mother’s education and the nutritional status of the children (Table 1). In the ZT group, 49% were female and 88% were breast feeding, while in the zinc treatment plus supplementation (ZT + S) group, 51% were female and 84% were breast feeding. In the ZT group, 74 were ETEC positive and 102 negative, and in the ZT + S group, 74 were ETEC positive and 103 negative.

Table 1.   Comparison of the study groups at time of randomization
CharacteristicMean (SD)
ZT (n = 176)ZT + S (n = 177)
  1. ZT, zinc treatment; ZT + S, zinc treatment plus supplementation.

Child’s age (months)15.5 (4.9)15.4 (5.3)
Education (years)
 Mother3.2 (3.3)2.7 (3.0)
 Father5.3 (8.3)3.9 (4.3)
 Weight (kg)/height (m)11.3 (1.1)11.1 (1.1)
Anthropometric z-scores
 Weight for age−1.91 (0.93)−2.09 (0.89)
 Height for age−1.64 (1.15)−1.80 (1.19)
 Weight for height−1.19 (0.85)−1.23 (0.88)

Compliance with the ZT and supplementation instructions was based upon tablet counts or levels of syrup remaining in the distributed bottles. Overall, it was estimated that 99% compliance was achieved. Figure 2 summarizes the mean serum zinc levels attained by the ZT only and zinc treatment +90 days supplementation (ZT + S) groups at 2, 15, 90 and 180 days after entry into the study. At day two, the zinc levels are not statistically different, but on day 15 (5 days post-treatment) and day 90, serum zinc levels were significantly lower (p < 0.001) in children not receiving supplementation. By day 180 (90 days following termination of supplementation), there was no statistical group difference in serum zinc levels.

Figure 2.

 Mean serum zinc levels and 95% confidence intervals by treatment group over time.

Table 2 summarizes the effects of additional zinc supplementation on the incidence density of ACD and ARI episodes over the initial 3 months while on zinc supplementation and over the 4- to 6- and 7- to 9-month follow-up intervals. In all intervals, the incidence of ACD was lower in the ZT + S group. The difference in incidence density, expressed as episodes per person-year of follow-up, at 3 months was 1.02 fewer episodes (95% CI 0.26, 1.79) and over 9 months 0.54 fewer episodes (95% CI .07, 1.01) of diarrhoea per person-year of follow-up than in children receiving the placebo syrup. A secondary analysis of treatment group effects by whether the child’s stool culture was ETEC positive or negative was carried out. At 3 months, the incidence density of ACD was similarly lower, whether ETEC positive or negative: if ETEC+, 1.09 fewer episodes (95% CI −0.022, 2.39) and if ETEC – 0.99 fewer episodes (95% CI 0.05, 1.92) among the children receiving zinc supplementation. Over the 9 months of follow-up, placebo compared to zinc supplemented households were 1.26 (95% CI 1.07, 1.49) more likely to seek help from a health provider for a diarrhoeal illness. There were no differences in the occurrence of ARIs over any interval tested. In addition, there were no treatment group differences identified in the growth of children over the 9 months of follow-up as measured by mean weight-for age, height-for age and weight-for-height z-scores.

Table 2.   Incidence density of acute childhood diarrhoea (ACD) and acute respiratory infection (ARI) episodes by follow-up interval
Incidence density*
Interval (months)ZTZT + SDifference(95% CI)Reduced (%)
  1. ZT, zinc treatment; ZT + S, zinc treatment plus supplementation.

  2. *Number of episodes of ACD or ARIs per person-year of exposure.

1. ACD
 0–33.662.641.02(026, 1.79)28
 4–62.512.140.37(−0.35, 1.07)15
 7–91.551.370.18(−0.41, 0.75)12
 0–92.592.050.54(0.07, 1 01)21
2. ARI
 0–36.726.560.16(−0.68, 0.96)2
 4–65.084.960.12(−0.48, 0.84)2
 7–93.764.040.28(−0.96, 0.36)−7
 0–−0.52, 0.52)0

To identify important predictors of ACD incidence during the period children were receiving zinc or placebo supplementation, adjusted for child and household characteristics, a multiple linear regression model was tested. The results are summarized in Table 3. Predictors of a decreased likelihood of reoccurring ACD included having received zinc supplementation, higher weight-for-height z-score at entry into the study, older age and higher maternal education. No treatment group by time or sex interaction was found. The impact of zinc supplementation was next assessed for effect modification on the basis of gender and age at entry into the study (6–12 months vs. 13–24 months) (refer to Table 4). The beneficial effect of zinc supplementation on the incidence of ACD over the 3 months they were receiving supplementation was more than double in female than in male (38%vs. 16% reduction). When these results are disaggregated by age, among males no treatment group effect was observed in either age group. Among female, there was a consistent strength of association observed in those under 12 months and over 12 months of age. No age by sex interaction was identified. The distribution of ACD episodes over 3 and 9 months of follow-up by treatment group is summarized in Figure 3. Children receiving zinc supplementation syrup were at half the risk of experiencing two-or-more episodes of acute diarrhoea during the 3 months they received supplementation (RR 0.45, 95% CI 0.25, 0.79).

Table 3.   Results of the multiple linear regression analysis for the prediction of acute childhood diarrhoea incidence density over the period of supplementation (0–3 months)
(95% CI)P-value
  1. ZT, zinc treatment; ZT + S, zinc treatment plus supplementation.

  2. Non-significant variables included in the model, ETEC positive or negative, child’s gender, and household income.

  3. No significant (p < 0.05) interaction among the variables in this model was found.

Group (ZT 0, ZT + S 1)−1.14(1.92, 0.364)0.0030.009
Child’s age (months)0.262(0.364, 0.104)0.0010.052
Mother’s education (years)0.156(0.312, 0.052)0.0070.003
Weight-for-height (z-score)0.6.03(−10.1, 1.40)0.0110.997
Table 4.   Gender modification by age group of the impact of zinc supplementation on the occurrence of acute childhood diarrhoea while receiving zinc supplementation
 ZTZT + SDifference(95% CI)Decrease (%)
Males (all)3.392.850.54(−0.59, 1.66)16
 ≤12 months2.632.91−0.28(−1.94, 1.38)−11
 13–24 months3.852.811.07(−0.46, 2.59)28
Females (all)3.952.431.52(0.46, 2.58)38
 ≤12 months5.082.852.23(0.31, 4.31)44
 13–24 months3.462.151.31(0.03, 2.57)38
Figure 3.

  Distribution of subjects by the number of acute diarrhoea episodes over the initial 3 and total 9 months of follow-up.


Daily zinc supplementation after ZT of an ACD episode offers additional protection against future diarrhoeal illness episodes. Among urban slum children who had recently experienced a diarrhoeal episode and had received 10 days of ZT, subsequent daily zinc supplementation for 3 months further reduced the incidence of diarrhoea by 28% while receiving supplementation. This protective impact continued for the 6 months after supplementation had stopped, with an overall 21% reduction in ACD incidence over 9 months of follow-up. It is interpreted that this overall preventive impact on diarrhoeal disease burden is in addition to that conferred by the ZT alone. The greater impact over the initial 3 months of follow-up is interpreted to reflect the period of greatest vulnerability to repeat infection.

Study limitations include the homogeneity of the children studied: all came from urban slum, low income households. Whether similar zinc supplementation clinical impacts would be observed in rural or urban, non-slum populations remain to be confirmed. It is assumed that the children enrolled in this trial had marginal zinc levels, but we did not have a serum zinc estimation until 2 days after the onset of treatment. Although one cannot be certain, it is likely the mean zinc levels obtained at 180 days, which are marginal, do reflect the pre-trial levels. A study carried out in Dhaka among children 1–3 years of age found mean serum zinc levels of 11.3 μmol/l, which is nearly equal to the mean in this study of 11.4 μmol/l in the two groups at day 180 (Rahman et al. 2002) These values are also similar to mean baseline zinc levels found in rural (Baqui et al. 2002) and urban (Roy et al. 1999) Bangladeshi clinical trials. This trial does not include a control group not receiving either ZT or supplementation. Given the publication of WHO/UNCEF guidelines recommending all under five children with diarrhoea be treated with zinc, the inclusion of a no treatment control group is not considered ethical in Bangladesh. Because this study was combined with a second objective to observe immune responses to zinc in children with ETEC diarrhoea, this may have biased the results to reflect a clinical response to this diagnostic subgroup of children. When assessed separately, the trends in clinical outcomes were similar in ETEC-positive and negative children. Strengths of this study include the high compliance with treatment instructions and the proportion of children who completed the full 9 months of follow-up.

Our finding of a reduced incidence is consistent with other supplementation trials; however, comparisons with other studies may be biased by two important differences in our study design. First, only children who had experienced an acute episode of diarrhoea were enrolled in this trial, which means we followed children who were probably more likely to experience subsequent ACD episodes. Second, all children received ZT prior to beginning zinc or placebo supplementation. It is expected that the ZT will have a preventive impact on the occurrence of diarrhoea over the next 2–3 months. Estimates of the preventive effect of ZT on reducing acute diarrhoea incidence have been fairly consistent at around 15%. (Baqui et al. 2002) (Fontaine 2001) The observed additional reduction in ACD incidence among those receiving zinc supplementation, compared to treatment only control children, is interpreted to be in addition to the benefits obtained from prior ZT of the acute diarrhoea episode.

Independently of treatment group effect, older age and improved nutritional status of the child, as well as increased years of the mother’s education, were found to be protective predictors of recurring diarrhoeal illness. Recent reports suggest that zinc may not have any treatment or preventive effects in children under 6 months of age and only marginal impact in children 6–12 months of age (Fischer-Walker et al. 2006; Walker et al. 2007). Within developing country populations, the prevalence of stunting (weight-for-height z-score of ≤2.0) in children under five years of age is known to be predictor of zinc deficiency within that population. Stunted children have consistently been observed to be at higher risk for diarrhoeal illness (Brown 2003; Brown et al. 2004).

The response to zinc, as measured by serum zinc levels, was identical in the male and female. The effect modification of gender on the incidence of ACD found in this trial stems from a secondary analysis of the data and not an a priori hypothesis of differential effect. The gender differences in impact favoring females were unexpected. Other zinc supplementation studies have found gender differences favoring males in India and Peru (Sazawal et al. 1995; Penny et al. 1999). More recently, Garenne et al. (2007) carried out a randomized field trial of zinc supplementation in rural Burkino Faso, with the aim of identifying differential effects in boys and girls enrolled between the ages of 6–30 months. Zinc supplementation reduced the prevalence of diarrhoea, with the risk reduction in the number of days of diarrhoea greater in male (21%) than female (8%). The explanation for these gender differences in zinc impact on diarrhoea incidence is not evident. Hypothetically, it could be explained by biological differences, in particular in immune system maturation. Studies have shown that male have higher levels of cytotoxic T cells in comparison with female (Lee et al. 1996; Aster et al. 1999). Garenne and others have suggested that gender differences may be explained by differential Th1 and Th2 responses to infection, with females having a higher proportion of CD4 T cells than males (Aster et al. 1999; Garenne et al. 2007). It has also been observed that males from varied geographic regions have higher levels of natural killer cells, suggesting a difference in the innate immune response between sexes (Tugume et al. 1995; Kam et al. 1996). The gender differences might also reflect social factors not measured in this study. There were no gender differences in weight-for-age or weight-for height z-scores at entry into the study or at the conclusion of follow-up; however, mean height-for-age z-scores were lower in female at entry into the study.

Several randomized trials of zinc supplementation have associated supplementation with decreased lower respiratory tract infections (Bhandari et al. 2002b; Brooks et al. 2004, 2005). A similar effect has not been consistently demonstrated for acute upper respiratory tract infections (ARIs), which was the outcome monitored in this study. The finding of no effect in this trial is consistent with other zinc supplementation studies in South Asia (Bhandari et al. 2002b). A similar trial carried out in Peru did find fewer respiratory illnesses (Penny et al. 2004). Finally, this study also found no effect of zinc supplementation on growth. Zinc supplementation if given over substantial periods of time and to nutritionally deprived children has been demonstrated to improve growth outcomes in young children (Bhandari et al. 2001; Brown et al. 2004). In this trial, the duration of supplementation was probably inadequate to influence growth outcomes.

This trial presents evidence that important preventive benefits, beyond that obtained by ZT, occur when young children are placed on a short course of zinc supplementation. Several unanswered questions remain, including the generalizability of this trial’s results and whether in a less supervised context households would value zinc supplementation enough to continue purchasing and administering it following a diarrhoeal episode. In contrast to a strategy that begins with zinc supplementation and adds ZT for an ACD episode, initiating supplementation only following the illness might be expected to result in higher levels of adherence. These parents will have experienced the benefits of zinc first hand, thus enhancing their motivation to continue with supplementation. Given the results of this trial, further effectiveness or impact studies combining ZT with supplementation in elevated risk populations are warranted.


We thank the households participating in this study and the ICDDR,B Mirpur clinic staff for their diligence in caring for the children enrolled. This study was funded through a grant from the Bill & Melinda Gates Foundation.