This review assessed the prophylactic effects of zinc supplementation for the prevention of childhood diarrhoea and respiratory illness. The authors concluded that zinc supplementation showed modest decreases in occurrence, severity and duration of diarrhoea and frequency and severity of respiratory illness. The authors' conclusions and recommendations for further research are likely to be reliable.

To quantify the effect of prophylactic zinc supplementation (three months or more duration) on the prevention of diarrhoea and respiratory illness in children aged less than five years.

PubMed, Science Citation Index and Cochrane Central Register of Controlled Trials were searched from inception to November 2005. Search terms were reported. Reference lists of retrieved articles and previously identified literature were examined for additional studies.

Double-blinded randomised controlled trials (RCTs) of zinc supplementation for at least three months on the occurrence of diarrhoea or respiratory tract infections in children aged less than five years were eligible for inclusion in this review. Eligible studies needed to be placebo controlled and have at least 90 per cent follow-up rates. Studies that involved children who had recently recovered from a diarrhoeal episode and who were under observation were eligible for inclusion.

Studies designed to assess the effects of zinc administration for the treatment of acute or persistent diarrhoea or respiratory illness and studies with zinc administration for less than three months were not eligible for inclusion.

Participants' ages ranged from three weeks to 41 months. Health status, where stated, was generally good. Three studies included children recovering from persistent diarrhoea. The included studies evaluated various regimes (doses, frequency and duration) to administer zinc supplements with or without vitamins and minerals. Zinc compounds included zinc acetate, zinc sulphate, zinc gluconate, methionine and riboflavin. All included studies were conducted in developing countries. Reported outcomes included number of episodes of diarrhoea, severe diarrhoeal disease (severe diarrhoea or dysentric illness) and persistent diarrhoea (14 days and more), episodes of respiratory illness and severe respiratory illness (pneumonia or lower respiratory tract infection), and duration of diarrhoea and respiratory illness.

Two reviewers independently selected the studies for inclusion.

Study validity was assessed using preset quality inclusion criteria at the study selection stage. Eligible studies needed to have randomised allocation to placebo and intervention groups, double-blinded assessment of outcome measures and at least 90 per cent follow up rates.

Two reviewers independently assessed study validity. Any disagreements were resolved through joint review and discussion.

The number of episodes of diarrhoea and respiratory illness were extracted. The number of severe cases (full definitions were provided in the paper) and duration of disease episodes in the control and intervention groups were also extracted. If the number of episodes was not stated, this was calculated by multiplying the mean numbers of episodes of these illnesses per child by the total number of participants in each group. Adjusted rate ratios and 95% confidence intervals (CI) were calculated for each study. One study had two intervention arms and each was treated as a separate study. Two studies reported outcome measures separately for the same population and these were counted as one study.

The pooled rate ratios and corresponding 95% CIs were calculated using a fixed-effect meta-analyses if there was no evidence of statistical heterogeneity (p≥0.1). A random-effects model was used in the presence of statistical heterogeneity (p<0.1). Statistical heterogeneity was assessed using the Χ² and I² tests. Sub-group analyses were carried out to assess possible causes. Publication bias was investigated using funnel plot analyses, the Begg and Mazumdar rank correlation test and Egger's regression method. The Duval and Tweedie trim and fill method was used to calculate an adjusted intervention effect.

Seventeen RCTs (n=7,659) were included in the review.

Compared with placebo, zinc supplementation was associated with a statistically significant decrease in the occurrence of episodes of diarrhoea (rate ratio 0.86, 95% CI:0.79 to 0.93, p=0.0002; 15 studies), episodes of severe diarrhoea or dysentric illness (rate ratio 0.85, 95% CI 0.75 to 0.95, p=0.006; five studies) and episodes of persistent diarrhoea (rate ratio 0.75, 95% CI 0.57 to 0.98, p=0.04; three studies). There was a statistically significant decrease in the number of days with diarrhoea (rate ratio 0.86, 95% CI: 0.79 to 0.93, p=0.0004; seven studies).

Compared with placebo, zinc supplementation was associated with a statistically significant decrease in the occurrence of episodes of respiratory illness (rate ratio 0.92, 95% CI: 0.85 to 0.99, p=0.02; 12 studies), episodes of pneumonia or lower respiratory tract infection (rate ratio 0.80, 95% CI: 0.70 to 0.92, p=0.001; four studies).

Considerable statistical heterogeneity was detected between study results, with I² values greater than 50 per cent and Χ² p values of less than 0.01 for several analyses. Sub-group analyses were reported, but did not reveal reasons for the heterogeneity. Publication bias was evident from funnel plots and results of trim and fill analyses were reported.

Zinc supplementation in healthy children resulted in significant but modest decreases in the frequency of diarrhoea and respiratory illness. Reductions were also evident in the frequency of severe diarrhoea, lower respiratory tract infections and duration of diarrhoea per child, but not in the duration of respiratory illness per child.

This review had clearly stated inclusion criteria in terms of participants, study design, interventions and outcomes. The authors searched relevant databases and made efforts to identify additional studies by reviewing reference lists of relevant literature. It was unclear whether any attempt was made to identify unpublished studies and whether language restrictions were applied, therefore, relevant studies may have been missed and language bias could not be ruled out. Evidence of publication bias was found and investigated appropriately. Appropriate efforts were made to minimise reviewer bias and error in the study selection and data extraction processes. The study selection process included a formal validity assessment using strict preset minimum quality inclusion criteria. Considerable heterogeneity was evident between studies in the meta-analyses, which was investigated through appropriate sub-group analyses. The significantly high levels of heterogeneity should be borne in mind when interpreting results. The authors also recognised that effect size was dependent upon nutritional status and baseline serum zinc levels of participants, and that all included studies were conducted in the developing world. Therefore, the findings may not be generalisable to the developed world nor to certain subgroups of children. Despite the identified issues, the authors' conclusions and recommendations for further rigorous research together reflect the evidence presented and are likely to be reliable.

Practice: The authors did not state any implications for practice.

Research: The authors stated that there was a need for further large randomised controlled studies designed to assess the effect of zinc supplementation with participants stratified by nutritional status and baseline serum zinc levels. The authors also identified a need to monitor populations already participating in prophylactic zinc supplementation schemes. Further, they recommended that studies on optimal zinc supplementation doses should be carried out given the reported effects of high zinc intake.

This review was partly funded by the Government of India, Fogarty International Center, National Institutes of Health, and the Bill and Melinda Gates Foundation.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.