1. Objectives

To assess the beneficial and adverse effects of zinc fortification of staple foods on health outcomes and zinc status in the general population

2. How studies were identified

The following databases were searched in April 2015:

• CENTRAL (The Cochrane Library 2015, Issue 3)
• MEDLINE
• EMBASE
• Web of Science
• POPLINE
• CINAHL
• OpenGrey
• AGRICOLA
• BIOSIS (ISI)
• Bibliomap and TRoPHI (Trials Register of Promoting Health Interventions)
• IBECS
• SciELO
• Global Index Medicus
• LILACS
• PAHO
• WHOLIS
• WPRO
• IMSEAR
• IndMED
• Native Health Database

Reference lists were also hand-searched and the authors contacted organizations in the field, including: the Department of Nutrition for Health and Development and the regional offices of the World Health Organization, as well as the nutrition section of the Centers for Disease Control and Prevention, the United Nations Children’s Fund, the World Food Programme, the Micronutrient Initiative, Global Alliance for Improved Nutrition, Helen Keller International, World Vision, Sight and Life, PATH, premix producers DSM and BASF, Food Fortification Initiative, and the International Zinc Nutrition Consultative Group

3. Criteria for including studies in the review

3.1 Study type

Randomized and cluster-randomized controlled trials, non-randomized trials with a concurrent control group, controlled before-and-after studies

3.2 Study participants

Members of the general population over two years of age, including pregnant and lactating women

(Studies on participants with severe disease, those who were being tube-fed or were nil-by-mouth, or older adults in long-term care facilities, were excluded)

3.3 Interventions

Staple foods (e.g., wheat or maize flour/subproducts, oils/fats, milk and dairy products, pulses, rice, sugar, sauces, fruit juices/nectar), condiments, or seasonings that had been industrially fortified with zinc versus food without added zinc or no intervention

(A minimum of two weeks until follow-up was required)

(Studies comparing zinc fortification to other forms of micronutrient interventions, such as dietary diversification, were excluded, as were trials involving point-of-use fortification with micronutrient powders, biofortification, or zinc supplements. Trials involving zinc fortification of water were not eligible for inclusion, and trials in which zinc fortification was not the only difference between treatment groups were also ineligible)

3.4 Primary outcomes

Children 24 to 59 months of age

• Zinc deficiency (as defined by trialists)
• Serum or plasma zinc (in μmol/L)
• Stunting (as defined by trialists)
• Underweight (as defined by trialists)

Children five to 11.9 years of age

• Zinc deficiency (as defined by trialists)
• Serum or plasma zinc (in μmol/L)

Adolescent boys and girls (12 to 18.9 years of age)

• Zinc deficiency (as defined by trialists)
• Serum or plasma zinc (in μmol/L)

Pregnant and lactating women

• Zinc deficiency (as defined by trialists)
• Serum or plasma zinc (in μmol/L)

Secondary outcomes for all participants included diarrhoea, pneumonia, all-cause morbidity, haemoglobin (g/dL), anaemia, and adverse effects (iron status as measured by serum ferritin in μg/L or serum transferrin receptor in mg/L, copper status as measured by serum or plasma copper in μg/dL, vomiting, other adverse effects). Additional secondary outcomes for children 24 to 59 months of age additionally included: weight, height or length, mid-upper arm circumference, cognitive and motor skill development, and all-cause death. For adult participants, additional secondary outcomes additionally included cognitive and work performance

4. Main results

4.1 Included studies

Eight controlled trials, enrolling 709 participants, were included in this review

• One trial was cluster-randomized, while the others were controlled at the individual level
• Four studies compared zinc-fortified foods to the same food without added zinc, and four trials compared zinc plus other micronutrient fortification to the same micronutrient fortification without zinc
• Three trials included 455 children aged between 24 and 59 months of age, with one of these trials enrolling only stunted and moderately anaemic children and the other enrolling healthy children
• Two trials involved 77 children between five and 11.9 years of age, with one of these trials enrolling zinc-deficient children only
• Two trials enrolled 224 adult participants, with one selecting zinc-deficient participants only
• One trial included 313 healthy pregnant women
• Seven trials used zinc-fortified cereal-based foods (bread [3 trials], biscuits, breakfast cereals, rice flour, wheat products) and one trial used zinc-fortified milk
• The dose of elemental zinc ranged from three to 40 mg per 100 g of food; types of zinc salts used were zinc oxide (3 trials), zinc sulphate (3 trials), zinc lactate and zinc amino chelate; intervention periods ranged from one to nine months

4.2 Study settings

• China, Colombia, Iran, Peru, Senegal, Sri Lanka, Turkey, and the United States of America
• Seven of the eight studies were conducted in zinc-deficient populations or in populations at high risk of zinc deficiency
• One trial country was conducted in a high-income country (the United States of America), five were conducted in upper-middle-income countries (China, Colombia, Iran, Peru, Turkey), and two were conducted in lower-middle-income countries (Senegal, Sri Lanka)

4.3 Study settings

How the data were analysed
Three comparisons were made: i) food fortified with zinc versus the same food without added zinc, ii) food fortified with zinc plus other micronutrients versus food fortified with the same micronutrients without zinc, and iii) food fortified with zinc versus no intervention. The last recorded time point during the intervention period was used for analysis. If similar outcomes, such as zinc deficiency, were reported using multiple biomarkers, the mean of these measures was used. For trials with more than two intervention groups, the data in the intervention groups were combined to produce a single pair-wise comparison with the control group. Cluster-adjusted effect estimates were extracted from cluster-randomized trials, or if adjusted estimates were not available, the intraclass correlation coefficient derived from the trial or another similar trial was used. Dichotomous data were summarized using risk ratios (RR) with corresponding 95% confidence intervals (CI), and continuous data were summarized using mean differences (MD) and 95% CI. Subgroup analysis was used to explore potential sources of heterogeneity where substantial heterogeneity was detected (I²>30%). Subgroup analyses were performed for the primary outcomes zinc deficiency and serum zinc, as follows:

• Sex: males, females, mixed
• Duration of the intervention: < six months, six months to one year, > one year
• Type of food vehicle: oils/fats, sugar, wheat flour/wheat flour subproducts, maize flour/cornmeal/maize flour subproducts, rice/rice subproducts, condiments/seasonings, milk/dairy products, fruit juices/nectars, others
• Type of zinc compound: zinc sulphate, zinc oxide, other forms of zinc, unspecified
• Dose of zinc fortification: ≤10 mg/100 g of food or >10 mg/100 g of food
• World Bank classification of the country: low- and middle-income, others

Sensitivity analyses excluding trials at risk of bias for components of methodological quality (allocation concealment, blinding, attrition) were conducted for primary outcomes where ≥ three trials contributed data to the outcome

Results
Zinc-fortified foods versus unfortified foods
Four trials were included in this comparison, including two studies in healthy children between 24 and 59 months of age, one study included children between five and 11.9 years of age who were known to be zinc deficient, and one trial was in zinc-deficient adults.

Zinc deficiency
No trials comparing zinc-fortified foods to unfortified foods reported on this outcome.

Serum or plasma zinc
Serum zinc concentrations were 2.12 μmol/L greater among participants consuming zinc-fortified foods compared to those consuming unfortified foods (95% CI [1.25 to 3.00 μmol/L], p<0.00001; 3 trials/158 participants). Subgroup analysis revealed no rise in serum zinc in the single study conducted in a high-income country with >6 months intervention and using zinc oxide to fortify maize-based cereals. No differences in effect were found for subgroup analysis by dose of zinc or in any sensitivity analyses.

Stunting
Zinc fortified food in comparison to unfortified food had no statistically significant effect on the risk of stunting among preschool children (RR 0.88, 95% CI [0.36 to 2.13], 2 trials/397 children).

Underweight
The risk of underweight was not reduced with zinc fortification in two trials including 397 preschool children (RR 3.10, 95% CI [0.52 to 18.38]).

Additional outcomes
In one trial evaluating the effects of zinc-fortified milk, the risk of respiratory infections was statistically significantly reduced in the treatment group (RR 0.49, 95% CI [0.28 to 0.86], 1 trial/301 children). In the same trial, diarrhoeal episodes did not differ between groups. All-cause morbidity, including diarrhoea, upper respiratory infections, lower respiratory infections, and pyoderma, was significantly lower among the group receiving zinc-fortified bread in comparison to those receiving unfortified bread in small trial of 24 children (MD -1.30 episodes, 95% CI [-2.34 to -0.26], p<0.05). Zinc-fortified milk statistically significantly increased weight (MD 1.04 kg, 95% CI [0.47 to 1.61]) and height (MD 3.15 cm, 95% CI [1.48 to 4.82]) in one trial of 307 children, although in another trial in a similar age group, no effect of zinc fortification on weight, height or mean arm muscle circumference were observed. No studies provided data on other pre-specified outcomes.

Adverse effects
No statistically significant adverse effect of zinc fortification on haemoglobin (0.29 g/dL, 95% CI [-0.51 to 1.10], 1 trial/24 participants), serum ferritin (0.29 μg/L [log transformed], 95% CI [-0.02 to 0.60], 1 trial/24 participants), or serum copper (-8.73 μg/dL, 95% CI [-18.03 to 0.58], 2 trials/82 participants) was observed. Vomiting and any adverse effects were not different between treatment groups in one trial of 301 children.

Zinc-fortified food in combination with other micronutrients versus the same food containing the same micronutrients without zinc
Four trials contributed data for this comparison, including: one study in moderately anaemic and stunted children less than 59 months of age, one study in healthy children five to 11.9 years of age, one study in healthy adults, and one study in healthy pregnant women.

Zinc deficiency
One factorial study comparing wheat products fortified with zinc plus iron to the same products fortified with iron alone in 40 moderately anaemic, stunted children found a non-statistically significant reduction in zinc deficiency in the zinc group (RR 0.17, 95% CI [0.01 to 3.94]).

Serum or plasma zinc
No statistically significant effect of zinc fortification on serum zinc was found in pooled analysis of four trials (MD 0.03 μmol/L, 95% CI [-0.67 to 0.72], 250 participants). Subgroup analyses revealed no differences by duration of intervention, type of zinc compound, type of food vehicle, or World Bank classification of the country. Sensitivity analyses including only those studies with a low risk of bias for blinding or attrition also had no meaningful effect on the results.

Additional outcomes
Pooled results from two trials in healthy pregnant women and moderately anaemic, stunted children found no effect of zinc fortification on the prevalence of anaemia (RR 0.89, 95% CI [0.35 to 2.28], 2 trials/137 participants). Likewise, meta-analysis of three trials in 186 participants revealed no effect of zinc fortification on haemoglobin concentrations (MD 0.13 g/dL, 95% CI [-0.33 to 0.59]). Micronutrient fortification including zinc in comparison to micronutrient fortification without zinc had no benefit on child weight (MD 0.02 kg, 95% CI [-1.09 to 1.13], 1 trial/41 children) or height (MD -0.44 cm, 95% CI [-3.40 to 2.52], 1 trial/41 children). No other pre-specified outcomes were reported on in the included trials.

Foods fortified with zinc versus no intervention
No identified trials reported on this comparison.

5. Additional author observations*

The overall methodological quality of the trials included in this review was low, with most trials at high or unclear risk of selection bias for allocation concealment or blinding. GRADE quality of evidence ratings for primary outcomes ranged from low to very low. Most of the data included in this review were from lower- or upper-middle-income countries with zinc-deficient populations or populations at high risk of zinc deficiency. While this reduced the generalizability of the findings, it also increased the applicability of the results to regions where such fortification interventions may be needed. Other factors increasing the applicability of the findings to such settings included the use of inexpensive zinc fortificants (zinc oxide and zinc sulphate) and the use of traditional phytate-rich food vehicles.

Limited evidence suggests that zinc-fortified cereal food staples increase serum zinc concentrations when zinc is used without other micronutrient fortificants such as iron. The effect of zinc fortification on children’s anthropometric status, the risk of infectious disease, and on iron and copper status remains unclear.

Further large-scale effectiveness trials are needed in this area, and these should include functional outcomes including growth, stunting and wasting of children under five, cognitive development, and work capacity. Future trials should investigate zinc plus other micronutrient fortification versus zinc fortification alone to ascertain the effects of interactions between micronutrients. Other areas requiring evaluation include the use of other food vehicles, optimal doses of zinc fortificants, and ongoing zinc food fortification programmes in terms of acceptability and adequacy of zinc intake.