1. Objectives

To assess zinc supplementation in children aged six months to 12 years of age for the prevention of mortality and morbidity, and for promoting growth

2. How studies were identified

The following databases were searched in December 2012/January 2013:

• Cochrane Infectious Diseases Group Specialized Register
• CENTRAL (The Cochrane Library 2012)
• MEDLINE
• MEDLINE In-Process & Other Non-Indexed Citations
• EMBASE
• African Index Medicus
• IndMED
• Global Health
• Latin American Caribbean Health Sciences Literature (LILACS)
• Conference Proceedings Citation Index -Science (Web of Science)
• WHO Library & Information Networks for Knowledge Database (WHOLIS)
• metaRegister of Controlled Trials (mRCT)
• ProQuest Dissertations & Theses Database

Reference lists were also searched, and authors of included studies were contacted to identify any ongoing trials

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials and cluster-randomized trials; quasi-randomized trials were excluded

3.2 Study participants

Children aged six months to 12 years (inclusive)

(Children aged <6 months were excluded, as were hospitalized children, and children with severe protein-energy malnutrition, HIV infection, or chronic diseases or conditions that could affect growth. Studies were excluded if ≤50% of participants met inclusion criteria and disaggregated data were unavailable)

3.3 Interventions

Oral zinc supplementation, regardless of compound, formulation, dose, duration or frequency, compared to placebo, no treatment or waiting list controls

(Studies evaluating food fortification or intake, mixed micronutrient supplementation, or zinc supplementation for therapeutic purposes, such as zinc administration during episodes of diarrhoea, were excluded. Studies with co-interventions, such as vitamin A, were included if they were administered to both treatment and control groups)

3.4 Primary outcomes

• All-cause mortality
• Cause-specific mortality
• Mortality due to all-cause diarrhoea
• Mortality due to lower respiratory tract infection, including pneumonia
• Mortality due to malaria

Secondary outcomes included all-cause hospitalization, diarrhoea (incidence, prevalence and hospitalization due to all-cause diarrhoea, incidence and prevalence of severe diarrhoea, incidence, prevalence and hospitalization due to persistent diarrhoea), lower respiratory tract infection (incidence, prevalence and hospitalization due to lower respiratory tract infection, including pneumonia), malaria (incidence, prevalence and hospitalization due to malaria), growth (height, weight, weight-to-height ratio, prevalence of stunting), and zinc status (serum or plasma zinc concentration, prevalence of zinc deficiency). Adverse event outcomes included side effects (study withdrawal, participants with one or more side effect, vomiting episodes, participants with one or more vomiting episode), haemoglobin status (blood haemoglobin concentration, prevalence of anaemia), iron status (serum or plasma ferritin concentration, prevalence of iron deficiency), and copper status (serum or plasma copper concentration, prevalence of copper deficiency)

4. Main results

4.1 Included studies

Eighty randomized controlled trials, enrolling 205,923 eligible participants, were included in this review

• Nine trials were cluster-randomized, two were crossover trials, and the remainder were individually randomized; seven identified studies did not contribute data to meta-analyses
• Of the 76 trials which could be classified into an age subgroup, 52 were in the under five category, and 24 were in the five to 13 years category
• Both stunted and non-stunted children were included in 42 studies; five studies included only stunted children, while five included only non-stunted children
• Baseline zinc concentrations were reported in 46 studies, the overall median of which was 72.5 microg/L

4.2 Study settings

• Bangladesh (7 trials), Belize, Brazil (5), Burkina Faso (2), Canada, Chile (4), China (4), Colombia, Ecuador (2), Ethiopia, Guatemala (5), India (8), Indonesia (3), Iran (6), Jamaica (2), Japan, Mexico (2), Nepal, Pakistan, Papua New Guinea, Peru (4), Senegal, South Africa, Sri Lanka, Thailand, Turkey (3), Uganda, the United Kingdom of Great Britain and Northern Ireland, the United Republic of Tanzania (2 trials), the United States of America (5), Viet Nam, and Zimbabwe
• Of the 72 studies that described the study setting, 21 were conducted in rural areas, 46 took place in urban or peri-urban areas and five were conducted in both rural and urban settings

4.3 Study settings

How the data were analysed
Two comparisons were made: i) zinc supplementation compared to no zinc supplementation, and ii) zinc plus iron supplementation versus iron alone (a post hoc comparison). Dichotomous data were summarized using risk ratios (RR), and for incidence data, RR (events per child) were combined with rate ratios (events per child year). Standardized mean differences (SMD) were generated for continuous data, and all outcomes were reported with corresponding 95% confidence intervals (CI). Fixed-effects meta-analysis was used, with random-effects models used in sensitivity analyses. All comparisons from factorial studies were included in meta-analyses separately, i.e., arms were not combined according to the absence or presence of zinc to create a single comparison from each trial. Data from the first period only of crossover trials were analysed, and data from cluster-randomized trials were adjusted for clustering. For outcome measures with at least 10 studies contributing data, the following subgroup analyses were planned to investigate potential heterogeneity:

• By country-level income: low- and middle-income countries versus high-income countries
• By age: six months to under one year, one year to under five years, five years to under 13 years
• By presence of stunting: height-for-age Z-score of <-2 versus height-for-age Z-score ≥-2
• By daily zinc dose: <5 mg, 5 to <10 mg, 10 to <15 mg, 15 to <20 mg, ≥20 mg
• By duration: < five months, six to 11 months, ≥12 months
• By co-intervention with iron: iron plus zinc versus iron alone, zinc versus no supplementation
• By formulation: solution, pill/tablet, capsule, powder

Results
Zinc versus no zinc
All-cause mortality
No reduction in all-cause mortality was observed with zinc supplementation in comparison to no zinc supplementation (RR 0.95, 95% CI [0.86 to 1.05], p=0.32; 13 studies/138,302 participants). Analysis by dose, duration, iron co-intervention and formulation did not reveal any differences in effect between subgroups.

Cause-specific mortality
Death due to diarrhoea did not differ between treatment groups (RR 0.95, 95% CI [0.69 to 1.31], p=0.95; 4 trials/132,321 participants), nor did death due to lower respiratory tract infection (RR 0.86, 95% CI [0.64 to 1.15], p=0.31; 3 trials/132,063 participants), or mortality due to malaria (RR 0.90, 95% CI [0.77 to 1.06], p=0.20; 2 comparisons/42,818 children).

Additional outcomes: all-cause hospitalization
No effect was found for zinc supplementation on the number of hospitalizations or the number of participants hospitalized.

Additional outcomes: diarrhoea
The incidence of all-cause diarrhoea was lowered by 13% among children receiving zinc supplementation in comparison to controls (RR 0.87, 95% CI [0.85 to 0.89], p<0.00001; 26 trials/15,042 participants), although substantial heterogeneity was detected (I²=0.88). While dose subgroups were significantly different (p<0.00001), no clear pattern across doses was observed, and the only non-significant dose range was <5 mg/day. Formulation subgroups were also significantly different (p<0.00001), with capsules and powders having no statistically significant effect on the incidence of diarrhoea (both p≥0.059). Iron co-intervention appeared to remove the benefit of zinc supplementation on the incidence of diarrhoea, with an RR of 1.00 (95% CI [0.96 to 1.05]) for iron co-intervention versus an RR of 0.82 (95% CI [0.80 to 0.84]) for no iron co-intervention. Similar results were found for the prevalence of all-cause diarrhoea, with a 12% reduction in prevalence with zinc supplementation (0.88, 95% CI [0.86 to 0.90], p<0.00001; 13 trials/8519 children; I²=0.88). In subgroup analyses, it appeared that those aged one to five years benefited the most from zinc supplementation, and that higher doses were more effective, with doses <10 mg/day showing no beneficial effect (both p<0.00001 for subgroup differences). As for the incidence of diarrhoea, powder formulation did not have a beneficial effect on diarrhoea prevalence, and iron co-intervention reduced the benefit of zinc supplementation (RR 0.96, 95% CI [0.88 to 1.08], versus RR 0.88, 95% CI [0.86 to 0.90] for no iron co-intervention). The incidence of severe diarrhoea was reduced by 11% among those receiving zinc (RR 0.89, 95% CI [0.84 to 0.95], p=0.00023; 6 trials/4982 children). The incidence and prevalence of persistent diarrhoea were reduced by 27% and 30%, respectively (incidence: RR 0.73, 95% CI [0.62 to 0.85], 7 trials/ 6216 children, and prevalence: 0.70, 95% CI [0.64 to 0.76], 1 trial/665 children; both p≤0.0001). Hospitalization due to diarrhoea was not reduced with zinc supplementation.

Additional outcomes: lower respiratory tract infection
Overall, no effect of zinc supplementation on the incidence of lower respiratory tract infection was observed. However, the prevalence of lower respiratory tract infection was found to be 20% higher among those receiving zinc (RR 1.20, 95% CI [1.10 to 1.03], p=0.000034; 3 trials/1955 children), although this result was highly heterogeneous (I²=97%) and became non-significant in a random-effects model. There was no statistically significant effect of zinc supplementation the risk of hospitalization due to lower respiratory tract infection.

Additional outcomes: malaria
No effect of zinc supplementation on the incidence or prevalence of malaria was found.

Additional outcomes: growth
A statistically significant increase in height was observed with zinc supplementation (SMD -0.09, 95% CI [-0.13 to -0.06], p<0.0001; 59 comparisons/13,669 children), translating to an improvement in height-for-age Z-score of 0.1 (95% CI [0 to 0.2]). A greater benefit was found in older age groups (one to five years and five to 13 years; p<0.00001 for subgroup differences), and although significant differences were found between subgroups by duration and by dose (both p<0.00001 for subgroup differences), there was no clear pattern of increasing or decreasing effect. Co-intervention with iron removed the beneficial effect of zinc supplementation, and with regard to formulation, the largest benefits were seen among the solution and capsule groups. Weight was statistically significantly increased with zinc supplementation (SMD -0.10, 95% CI [-0.18 to -0.02], p<0.00001; 44 trials/12,305 children), with greater benefits observed in older age categories and for solution and capsule formulations when analysed by subgroup. One study appeared to cause the reduction in weight with zinc supplementation observed in the youngest age group (SMD 0.31, 95% CI [0.25 to 0.38]). No difference in effect was found in subgroup analyses by country-level income, stunting or iron co-intervention. Weight-to-height ratio was statistically significantly improved among those receiving zinc (SMD -0.05, 95% CI [-0.10 to -0.01], p=0.024; 29 studies/7901 individuals). There was no difference in the prevalence of stunting between treatment and control groups.

Additional outcomes: zinc status
Serum or plasma zinc concentrations improved with zinc supplementation (SMD -0.62, 95% CI [-0.67 to -0.58], p<0.00001; 46 trials/9810 participants), with greater benefits seen in the one to five year age group, in low- and middle-income countries, with larger doses, with shorter supplementation durations, in capsule formulations, and when given without iron co-intervention (all p≤0.00001 for subgroup differences). The prevalence of zinc deficiency was more than halved with zinc supplementation (RR 0.49, 95% CI [0.45 to 0.53], p<0.00001; 15 trials/5434 children), with larger benefits observed in older age categories, for capsule and solution formulations, and when administered without iron co-intervention (all p≤0.00001 for subgroup differences). Subgroup analyses by dose and by duration were also significantly different (both p≤0.00001 for subgroup differences), although no clear pattern was observed.

Adverse effects
No clear evidence of a difference in study withdrawal was noted between treatment and control groups, although there was some evidence of a difference in the risk of having more than one side effect (RR 1.13, 95% CI [1.00 to 1.27], p=0.043; 3 trials/850 participants). Vomiting episodes (RR 1.68, 95% CI [1.61 to 1.75], p<0.00001; 5 trials/4095 children), and more than one vomiting episode (RR 1.29, 95% CI [1.14 to 1.46], p=0.000044; 5 trials/35,192 children) were more common with zinc supplementation. No statistically significant effect was found on haemoglobin concentrations or the prevalence of anaemia overall; however, due to the influence of one trial, there appeared to be a reduction in risk of anaemia with higher doses ≥20 mg (RR 0.17, 95% CI [0.06 to 0.46]) and with treatment of less than six months duration (RR 0.18, 95% CI [0.06 to 0.48]). Serum or plasma ferritin concentrations were slightly increased among those receiving zinc overall (SMD -0.07, 95% CI [-0.13 to -0.00], p=0.038; 20 trials/4474 participants). In subgroup analyses, this finding appeared to be reversed with increasing duration of treatment, with an SMD in ferritin of 0.34 (95% CI [0.24 to 0.45]) for supplementation lasting 12 months or more. A harmful effect was also found for subgroup analysis by iron co-intervention, with those in the subgroup without iron having lower ferritin levels (SMD 0.27, 95% CI [0.17 to 0.38]). Although differences were detected in subgroup analysis by dose, there was no clear directionality in change of effect. The single study from a high-income country showed some evidence of harm, with a decrease in ferritin in those receiving zinc supplementation (SMD 0.88, 95% CI [0.29 to 1.47]). The prevalence of iron deficiency was not significantly affected by zinc supplementation (RR 0.99, 95% CI [0.89 to 1.10], p=0.79; 10 trials/3149 children). Serum or plasma copper concentrations were decreased among those receiving zinc supplementation (SMD 0.22, 95% CI [0.14 to 0.29], p<0.00001; 13 trials/3071 children). Although results differed by dose, there was no pattern to the changes in effect size; however, the decrease in copper levels became non-significant with duration of supplementation lasting six months or more. A greater reduction in copper levels was observed in the pill/tablet subgroup (SMD 0.83, 95% CI [0.65 to 1.01]). The prevalence of copper deficiency was statistically significantly increased among those treated with zinc relative to controls (RR 2.64, 95% CI [1.28 to 5.42], p=0.0083; 3 trials/1337 children).

Zinc plus iron versus zinc alone
All-cause mortality
No statistically significant difference in all-cause mortality was observed between those treated with zinc plus iron versus those receiving zinc alone (RR 0.33, 95% CI [0.01 to 8.39], p=0.50; 1 trial/323 children).

Additional outcomes
All-cause hospitalization and hospitalization due to diarrhoea did not differ between treatment groups. The incidence of all-cause diarrhoea was higher among those treated with zinc and iron compared to those treated with zinc alone (RR 1.10, 95% CI [1.03 to 1.18], p=0.007; 5 trials/1530 individuals), although this displayed substantial heterogeneity (I²=76%) and was non-significant when analysed using a random-effects model. The prevalence of all-cause diarrhoea and the incidence of severe diarrhoea did not differ between groups, nor did the incidence of lower respiratory tract infection or malaria. Growth, as measured by height, weight, and weight-to-height ratio, did not differ between groups; however, in two trials reporting on the prevalence of stunting, zinc plus iron appeared to be beneficial in comparison to zinc alone (RR 0.92, 95% CI [0.85 to 0.99], p=0.037; 2 trials/462 participants). Zinc alone was associated with a greater benefit in raising serum zinc concentrations (SMD 0.16, 95% CI [0.05 to 0.27], p=0.0039; 8 trials/1337 participants), but this was not significant in random-effects meta-analysis, and the prevalence of zinc deficiency did not differ between treatment groups. Haemoglobin concentrations were increased with zinc plus iron supplementation (SMD -0.23, 95% CI [-0.34 to -0.12], p<0.0001; 8 trials/1341 children), as were serum ferritin concentrations (SMD -1.78, 95% CI [-1.99 to -1.56], p<0.00001; 6 trials/945 children), and zinc plus iron also lowered the prevalence of iron deficiency (RR 0.12, 95% CI [0.04 to 0.32], p=0.000034; 2 trials/248 children). Serum copper concentrations did not differ between treatment groups.

5. Additional author observations*

The quality of the evidence for all-cause and cause-specific mortality outcomes was rated moderate to high, indicating that further trials evaluating the effects of zinc supplementation on mortality in the populations represented in this review may not be necessary. The 80 included trials were conducted in diverse populations with 32 countries represented; thereby the results are likely to have high external validity. However, most trials were conducted in low- and middle-income countries with a high underlying prevalence of zinc deficiency, and thus the findings may be most applicable in these settings. A paucity of data disaggregated by stunting status may have underpowered this review to detect differences in treatment effect between stunted and non-stunted children. For secondary outcomes and adverse event outcomes, the quality of the evidence was mixed, and for many outcomes substantial heterogeneity was detected which remained unexplained following subgroup analyses.

Preventive zinc supplementation was demonstrated to reduce diarrhoea morbidity and improve zinc status in children aged six months to 12 years, and small but statistically significant improvements in growth outcomes were observed. In contrast, vomiting was increased and copper status was negatively impacted by zinc supplementation, but the benefits of zinc supplementation are likely to outweigh the harms in low- and middle-income settings with an elevated risk of zinc deficiency.

Future research on the optimal dose, duration, frequency, timing and formulation of zinc supplementation is warranted, as is further research on ways to improve the delivery of zinc interventions in difficult to reach populations. In addition, future studies could evaluate zinc fortification and dietary interventions in comparison to zinc supplementation.