Introduction
About 30% of the world’s population is zinc deficient [
1], most prevalent in children under 5 years of age in developing countries [
2]. Zinc deficiency is associated with impaired immune function which results in an increase in morbidity due to infections, growth retardation, hypogonadism and cognitive dysfunction [
3,
4]. Zinc deficiency is primarily related to the consumed diet; zinc is most abundant and easily absorbable from animal proteins, whereas consumption of vegetable and cereals decreases its absorption due to binding of zinc to phytates [
5,
6].
Over the past two decades, strong evidence has come forward from multiple randomized controlled trials, in both developed and developing countries, showing an effect of zinc in decreasing morbidity and mortality in children due to gastrointestinal and respiratory infections [
7,
8]. More recent trials from sub-Saharan countries have also shown an effect on malarial morbidity [
9]. This effect of zinc against infectious diseases is therapeutic as well as preventive. Previous reviews by the Zinc Investigators’ Collaborative Group in 1999 [
10] and by Aggarwal et al. [
11] have studied the effect of preventive zinc supplementation on diarrheal and respiratory morbidity. The current systematic review presents the effect of zinc supplementation on mortality in children less than 5 years of age in developing countries. The evidence for the effect of zinc supplementation on cause-specific mortality and morbidity is assessed for diarrhea, pneumonia and malaria according to guidelines set by the Child Health Epidemiology reference Group (CHERG) for input into the Lives Saved Tool (LiST) model [
12].
Methods
Searching
A comprehensive literature search was conducted using the following search strategy: zinc AND (infection OR diarrhea OR pneumonia OR ARI OR malaria OR morbidity OR mortality OR death) in electronic bibliographic databases i.e. PubMed, the Cochrane Library and WHO regional search engines, including articles cited up to September 25th 2010 . The limits used were “Humans” and “Randomized controlled trials”. Additional studies were obtained through hand search of references from identified studies. One author reviewed the titles and abstracts to identify controlled studies conducted in developing countries in which supplemental zinc was administered and outcomes on mortality or morbidity were reported. Two authors independently assessed eligibility using the pre-defined inclusion and exclusion criteria and performed the data extraction. Any discrepancies between the reviewers in either the decision of inclusion or exclusion of studies or in data extraction were resolved by discussion aimed at reaching consensus among all.
Selection (inclusion/exclusion criteria)
Individual- or cluster-randomized controlled trials of routine (i.e. daily or weekly) zinc supplementation administered to children less than 5 years of age in developing countries.
• Studies with zinc supplementation carried out for 3 months of intervention in both intervention and comparison groups.
• Trials in which other nutrient co-interventions [e.g., vitamin A, riboflavin, iron-folic acid (IFA)] were administered to both control and zinc arms were included. Results were analyzed with and without studies in which zinc was given with IFA.
• Studies on small-for-gestational age or low birth weight infants were excluded from the review.
Validity assessment
The quality assessment of each study was carried out on key variables with regard to study design, study limitations, intervention specifics, and outcome specific and graded according to the adapted GRADE technique [
12,
13]. Any study with a final grade of very low was excluded from the review. This review is shaped in large part by the needs of the LiST model. In this model, increases in coverage of an intervention results in a reduction of one or more cause-specific deaths or in reduction of a risk factor. Therefore the reviews and the grade process used were designed to develop estimates of the effect of an intervention in reducing either a risk factor or a death due to specific cause. For more details of the review methods, the adapted grade approach or the LiST model, see the method’s paper[
12].
Data abstraction
Data from all studies that met final inclusion and exclusion criteria were abstracted into a standardised form for each outcome of interest [
12]. Data extracted from each eligible study included the following primary outcome variables: total number of deaths in both arms, deaths due to specific infection (diarrhea, pneumonia or malaria), the total number of episodes of illness (diarrhea, pneumonia and malaria) in each arm, the total amount of person-time accumulated in each arm (reported as person-days) and incidence rate ratio (IRR).
Quantitative data synthesis
Meta-analyses were generated for all-cause mortality and cause specific mortality of diarrhea, pneumonia and malaria in children under 5 years of age. Pooled estimates for incidence of diarrhea, respiratory disease and malaria were also generated. Data from cluster randomised trials were pooled with that of individual randomised trial. In this case, cluster adjusted values were used irrespective of method used by the primary authors. Generic inverse variance method of meta-analyses was used for pooling the data. The assessment of statistical heterogeneity among trials was done by visual inspection i.e. the overlap of the confidence intervals among the studies, and by the Chi square (P-value) of heterogeneity in the meta-analyses. A low P value (less than 0.10) or a large chi-squared statistic relative to its degree of freedom was considered as providing evidence of heterogeneity. The I
2 values were also looked into, and an I
2 greater than 50% was taken to represent substantial heterogeneity. In situations of substantial heterogeneity being present, causes were explored, sub-group analyses performed and random effects model was used. Results of pooled estimates are described as relative risk (RR) with 95% CI. All meta-analyses were conducted using the software RevMan version 5 [
14]. Data, where available, were taken for zinc versus placebo/no treatment (control) groups and for zinc plus other micronutrient versus other micronutrient, making zinc supplementation the only difference between the intervention and control groups.
We followed standardized guidelines of Child Health Epidemiology Reference Group (CHERG) to get estimates of effectiveness of preventive zinc supplementation in reducing diarrhea and pneumonia specific mortality [
12]. Application of CHERG rules is based on three components 1) the volume and consistency of the evidence; 2) the size of the effect, or risk ratio and 3) the strength of the statistical evidence for an association between the intervention and outcome. The detailed description and application of these rules to collective morbidity and mortality outcomes is provided in the results and discussion section.
Discussion
The benefits of preventive zinc supplementation are well established on morbidity related to infections, however; there has been no attempt to quantify the effect of preventive zinc supplementation on cause specific mortality in children which is a new aspect in our review compared to previous reviews on the subject. In their review of preventive zinc supplementation, Brown et al. [
38] pooled data on children across all age groups (infants, preschoolers and older prepubertal children) and reported a 6% non-significant reduction in all-cause mortality. They also indicated that the mortality reduction with zinc supplementation versus placebo was only significant (18%) among children older than 12 months of age, with a tendency towards increased mortality in children less than 12 months. We, on the other hand, have evaluated the effect of zinc on all cause mortality, cause specific mortality and cause specific morbidity in children less than 5 years of age from developing countries receiving preventive zinc supplementation for a minimum period of three months. The studies of shorter duration were excluded as preventive zinc supplementation may require longer duration of supplementation to replete the body stores compared to that of therapeutic supplementation for diarrhea/pneumonia where two weekly doses are enough to fulfill the acute deficiency. Our results are similar to those reported by Brown et al., even though we restricted to the studies with children less than five years of age. Our 9% reduction in all-cause mortality was also non-significant. In our review, for the LiST model, estimates are presented from zinc only studies compared with placebo/no treatment, which may not be applicable to countries where there are on-going national supplementation programs of iron-folate, for example, India.
Table
1 gives grade quality of overall evidence. Zinc supplementation alone was associated with non-significant reductions in mortality due to diarrhea, pneumonia and malaria. Table
2 summarizes the application of CHERG rules to estimate effects for diarrhea and pneumonia specific mortality [
12]. Zinc supplementation was associated with an 18% (RR = 0.82; 95% CI: 0.64, 1.05) reduction in deaths due to diarrheal diseases in under 5 children from developing countries. However, as the above values were statistically non-significant the evidence on cause specific mortality was labeled as low quality. In looking at severe morbidity estimates as surrogate measures, such as hospitalization due to diarrhea, this information was only available from one study with an overall event rate of less than 50 [
12]. As per CHERG rules, we, therefore, considered the impact of zinc supplementation on diarrheal incidence for a point estimate of impact on diarrhea mortality as the evidence was of moderate quality[
12]. With the above outcomes we took the most conservative effect between the available low quality evidence on diarrheal specific mortality and moderate quality evidence on mild morbidity. Using this sequential application of CHERG rules to the above outcomes, the observed 13% reduction (RR = 0.87; 95% CI: 0.81, 0.94) in diarrhea incidence was taken as an effect estimate for the protective effect of zinc supplementation on diarrheal disease mortality and recommended for inclusion in LiST tool [
12].
Table 1
Quality assessment of overall evidence for effect of zinc supplementation (alone) in reducing morbidity and mortality in children > 5 years of age in developing countries
Outcome: All-cause mortality: Quality of evidence: Low
|
7 | RCTs | Sequence generation and allocation concealment was unclear in few of the included studies | I2= 50% | Yes (all studies were conducted in developing countries) | The median dose of supplementation was 10 mg/day and median duration of supplementation was for 6 months. | 0.91 (0.82-1.01) |
Outcome: Diarrhea specific mortality: Quality of evidence: Low
|
4 | RCTs | Allocation concealment was unclear in two of the included studies | I2=0% | Yes (all studies were conducted in developing countries) | The median dose of supplementation was 10 mg/day and median duration of supplementation was for 6 months. | 0.82 (0.64-1.05) |
Outcome: Diarrhea specific morbidity: Quality of evidence: Moderate
|
14 | RCTs | Sequence generation and allocation concealment was unclear in few of the included studies | I2=79% | Yes (all studies were conducted in developing countries) | The median dose of supplementation was 10 mg/day and median duration of supplementation was for 6 months. | 0.87 (0.81-0.94) |
Outcome: Pneumonia specific mortality: Quality of evidence: Low
|
4 | RCTs | Allocation concealment was unclear in two of the included studies | I2= 39% | Yes (all studies were conducted in developing countries) | The median dose of supplementation was 10 mg/day and median duration of supplementation was for 6 months. | 0.85 (0.65-1.11) |
Outcome: Pneumonia specific morbidity: Quality of evidence: Moderate
|
6 | RCTs | Sequence generation and allocation concealment was unclear in few of the included studies | I2=0% | Yes (all studies were conducted in developing countries) | The median dose of supplementation was 10 mg/day and median duration of supplementation was for 6 months. | 0.81 (0.73-0.90) |
Outcome: Malaria specific mortality: Quality of evidence: Low
|
1 | RCT | None | NA | Study conducted in Zanzibar | Dose of supplementation was 10 mg/dl for children > 1 year and 5mg/day for children < 1 years. | 0.90 (0.77-1.06) |
Outcome: Malaria specific morbidity: Quality of evidence: Low
|
4 | RCTs | Allocation concealment was unclear in two of the included studies | I2=0% | Yes (all studies were conducted in developing countries) | The median dose of supplementation was 10 mg/day and median duration of supplementation was for 6 months. | 0.92 (0.82-1.04) |
Table 2
Application of standardized rules for choice of final outcome to estimate effect of zinc supplementation on mortality due to diarrhea, pneumonia and malaria in children less than 5 years of age
Cause specific mortality (diarrhea) | 4 | 18% reduction; RR = 0.82 (0.64, 1.05) | (Low quality of evidence) |
Incidence of diarrhea | 14 | 13% reduction; RR = 0.87 (0.81, 0.94) | Rule 6 is applied (Moderate quality of evidence) (Effect Recommended for LiST) |
Cause specific mortality (pneumonia) (Zinc only studies) | 4 | 15% reduction; RR = 0.85 (0.65, 1.11) | Rule 6 applied (Low quality of evidence) (Effect Recommended for LiST) |
Pneumonia morbidity | 6 | 19% reduction; RR = 0.81 (0.73, 0.90) | Rule 6 applied (Moderate quality of evidence) |
A similar effect of zinc was shown for pneumonia. Zinc only supplementation studies revealed a pooled effect size of 15% (RR = 0.85; 95% CI: 0.65, 1.11) reduction on pneumonia mortality but the overall quality of evidence was rated as low. We applied CHERG rules on studies with information on pneumonia morbidity outcomes where the overall evidence quality was rated as moderate. Applying CHERG rules we used the most conservative impact estimate from available outcomes i.e. 15% as our pooled surrogate effect size for mortality[
12]. Deaths due to malaria were included only in one study which reported a 10% (RR = 0.90; 95% CI: 0.77, 1.06) reduction in malarial mortality (low quality of evidence). In the present review there was insufficient evidence of a protective effect of zinc supplementation on malarial mortality or morbidity. Therefore no estimates can be given for effect of zinc supplementation on reduction of deaths due to malaria for inclusion in the LiST tool.
Our review did not give any conclusive evidence about any possible positive or negative interaction between zinc and iron whereby iron decreases the absorption or bioavailability of zinc. With the inclusion of IFA studies for all-cause mortality, the results remained non-significant. For diarrhea-specific mortality, with inclusion of IFA studies, the impact estimate of reduction decreased from 18% to 9% but the results in both the situations were statistically non-significant. For pneumonia-specific mortality, however, a beneficial effect was seen as the impact of all the studies came out to be significant (RR = 0.80; 95% CI: 0.67, 0.96) compared to a 15% non-significant reduction with zinc alone studies (RR = 0.85; 95% CI: 0.65, 1.11).
To conclude, the application of CHERG rules to available evidence on diarrhea and pneumonia for providing effect estimates of protective effect for zinc supplementation in children for a minimum period of 3 months indicates 13% (RR = 0.87; 95% CI: 0.81, 0.94) reduction in mortality due to diarrheal diseases and 15% (RR = 0.85; 95% CI: 0.65, 1.11) reduction in pneumonia specific mortality. Our analysis supports inclusion of preventive zinc supplementation in public health programs to improve child health and survival and the suggested point estimates provide a suitable starting point for inclusion within the LiST tool and further program evaluation in effectiveness settings.
Acknowledgment
This work was supported in part by a grant to the US Fund for UNICEF from the Bill & Melinda Gates Foundation (grant 43386) to “Promote evidence-based decision making in designing maternal, neonatal and child health interventions in low- and middle-income countries”.
This article has been published as part of
BMC Public Health Volume 11 Supplement 3, 2011: Technical inputs, enhancements and applications of the Lives Saved Tool (LiST). The full contents of the supplement are available online at
http://www.biomedcentral.com/1471-2458/11?issue=S3.
Competing interests
We do not have any financial or non-financial competing interests for this review.
Authors' contributions
Professor Zulfiqar A Bhutta developed the parameters for the review and secured requisite support. Drs Mohammad Yawar Yakoob, Evropi Theodoratou, Afshan Jabeen, Aamer Imdad and Thomas P Eisele, Joy Ferguson, Arnoupe Jhass and Igor Rudan participated in literature search, data extraction and writing of the manuscript under the overall supervision of Professor Bhutta. All authors contributed to the critical review and finalization of the manuscript.