Enteral Zinc Supplementation in Preterm or Low Birth Weight Infants: A Systematic Review and Meta-analysis Open Access

This review was registered in PROSPERO (CRD42021238971). Preferred Reporting Items for Systematic Reviews and Meta-Analyses-Protocol (PRISMA-P) guidance was followed.⁸ 

An electronic search was conducted on PubMed, Cochrane Central Register of Controlled Trials (CENTRAL), Cochrane Register of Studies Online, and Embase through March 24, 2021. There were no date or language restrictions (Appendix 1). We also searched the reference lists of the selected articles to identify additional relevant articles.

We used the Covidence systematic review software,⁹  Veritas Health Innovation, Melbourne, Australia, for the review. Two review authors independently screened the titles and abstracts to identify relevant citations. The review authors retrieved the full texts of the relevant articles and independently assessed the eligibility of the studies using predefined inclusion criteria and performed data extraction. For data extraction, a modified version of the Cochrane Effective Practice and Organization of Care Group¹⁰  data collection checklist (Cochrane EPOC Group 2017) was used, which included study identifiers and context, study design and limitations, intervention specifics, and outcome effects. We also collected data on explanatory variables, including if the babies were very low birth weight (VLBW) <1.5 kg or if the gestational age was <32 weeks. Any disagreements or discrepancies between reviewers was resolved by discussion and by consulting a third review author. Study authors were contacted where the full text was not available or to obtain any additional information.

We included randomized control trials (RCT)s and quasi-randomized trials in which individual preterm or LBW infants fed with mother’s milk or donor human milk were either allocated to receive zinc or no zinc supplementation. We assessed enteral zinc supplementation alone, or in combination with other micronutrients, if the control group also received the same micronutrients. We excluded studies where infants were formula-fed or had other specific comorbidities, such as human immunodeficiency virus.

The key outcomes studied were mortality, morbidity, hospitalization, sepsis, acute respiratory infection, diarrhea, necrotizing enterocolitis, growth (weight, length, and head circumference) at latest follow up, neurodevelopment (mental and psychomotor development). All outcomes were measured at latest follow up.

The analysis was done using Stata 16 software (Texas)¹¹  following the Cochrane Handbook for Systematic Reviews of Interventions recommendations.⁸  We reported pooled relative risks (RR) for categorical outcomes, and mean differences (MD) for continuous outcomes along with respective 95% confidence intervals (CIs) using the “meta” command in Stata.

If the relative risk was not provided, we calculated it using the available data in the article. Adjusted results were preferred over unadjusted results whenever available. Fixed-effects meta-analysis (inverse variance method) was used to pool data to estimate effects. In situations of high heterogeneity (ie, I² value greater than 50%),¹²  a random effects model using the Restricted Maximum Likelihood Method (REML) was used. Egger’s test was used to assess publication bias for outcomes with more than 10 studies.

Subgroup analyses decided a priori were based on gestational age (<32 weeks; very preterm), birth weight (<1500 g; VLBW); dose of zinc (<3, 3–5, >5 mg per day); age at enrolment; income level of the country (high or upper middle and lower middle or low income) and fortification of breastmilk.

Assessment of the risk of bias for included studies was done using revised Cochrane “Risk of bias” tool for randomized trials and risk of bias in nonrandomized studies of interventions tool for nonrandomized studies.⁸  Certainty of evidence for the pooled estimates for the outcomes was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.¹³ 

Our search on March 24, 2021 resulted in 1922 records. After title and abstract screening, we examined 315 full text articles. 14 trials were included that compared enteral zinc to no zinc supplementation in preterm or LBW infants^3,14–26  (Appendix 2). The included studies reported on 9940 preterm or LBW infants from 11 countries of which 3.1% (2 trials, 304 infants) were VLBW or very preterm.^21,23  Of the included studies, 1 was a cluster randomized trial,²⁴  1 was a quasi-randomized trial,¹⁶  and the remaining 12 were individually randomized controlled trials. Six trials^3,20–24  had low risk of bias, and others had some concerns or high risk of bias (Appendix 3). Three trials were conducted in high-income countries,^16,23,26  three in upper middle-income countries,^14,15,18  and eight in low or lower-middle income countries.^{3,17,19–22,24,25}  The median (IQR) dose of zinc used in the studies was 5 (5 to 7) mg. The time of start of intervention ranged from birth to 5.2 weeks and the median (interquartile range [IQR]) duration of supplementation was 140.7 days (98 to 182.7 days) (Appendix 4). There were no studies that compared dosing or timing.

Findings are summarized in Table 1 and Appendix 5. Six trials with a total of 8801 LBW or preterm infants reported infant deaths with a median follow-up duration of 26 (14 to 152) weeks.^{18,19,22–24}  The pooled RR of zinc supplementation on mortality was 0.73 (95% CI 0.46 to 1.16, I² = 59%, low certainty). In a sensitivity analysis, excluding studies that enrolled term LBW infants, the pooled RR of zinc supplementation for mortality was 0.68 (95% CI 0.43 to 1.09, I² = 52%, 4 trials,^19,23,24  1609 participants).

At latest follow up (median 26, IQR 20 to 26 weeks), the RR for hospitalization in infants supplemented with zinc versus no zinc was 0.70 (95% CI 0.24 to 2.00, I² = 82%, 2 trials,^18,19  277 infants, very low certainty). The mean difference in the duration of hospitalization, in infants who received zinc supplementation versus no zinc supplementation was 3.49 days (95% CI −7.11 to 14.08; I² = 49%, 2 trials, 398 infants).

At latest follow up (20 to 52 weeks), we found a mean difference of 378.6 g in weight (95% CI 275.26 to 481.88, I² = 45%, 8 trials,^{3,14,16,17,19,21,23,25}  798 participants, moderate certainty) 2.9 centimeters in length (95% CI 1.53 to 4.31, I² = 77%, 6 trials,^{3,16,17,19,21,25}  529 participants, low certainty), and 0.56 centimetres in head growth (95% CI 0.23 to 0.90, I² = 41%, 5 trials,^{14,17,19,21,25}  466 infants, low certainty).

At latest follow up (median 26, IQR 20 to 52 weeks) we found a RR of 0.81 for diarrhea (95% CI 0.68 to 0.97, I² = 28%, 6 trials,^3,17–19,22  1947 infants, moderate certainty). At latest follow up (median 13, IQR 6 to 20 weeks), the RR for acute respiratory infections was 0.32 (95% CI 0.09 to 1.17, I² = 0%, 2 trials,^17,19  172 infants, very low certainty) and for sepsis was 1.12 (95% CI 0.62 to 2.02, I² = 18%, 2 trials,^19,23  265 participants, low certainty). We found a single trial²³  (n = 193) that reported the effect of zinc supplementation on necrotizing enterocolitis (RR 0.08, 95% CI 0.00 to 1.33), bronchopulmonary dysplasia (RR 0.66, 95% CI 0.31 to 1.40), retinopathy of prematurity (RR 0.14, 95% CI 0.01 to 2.70), and fever (RR 1.66, 95% CI 1.09 to 2.53, 1 trial,²⁰  82 infants).

At latest follow up (52 weeks) we found a mean difference between the infants with zinc and no zinc supplementation^15,26  of −4.18 in Bayley Scale of Infant Development version 1 (BSID I) mental development scores (95% CI −6.51 to −1.85, I² = 9%, 2 trials, 301 infants, low certainty); and of 5.75 in BSID psychomotor development scores (95% CI −4.83 to 16.33, I² = 95%, very low certainty). In 72 infants Mathur et al¹⁹  used the Amiel-Tison scale for measuring neurodevelopment and reported a RR of 0.09 for “hyperexcitability” (95% CI 0.01 to 0.7) and a RR of 0.22 for “brisk bicipital reflex” (RR 0.22, 95% CI 0.07 to 0.7), but alertness and attention was similar across groups (RR 0.47, 95% CI 0.05 to 4.99).

No studies reported serious adverse events. Two trials^17,22  reported vomiting as an adverse event but showed no substantial difference between the zinc and no zinc groups.

Two trials provided <3 mg per day, 10 trials provided 3 to 5 mg per day, and 4 trials provided >5 mg per day of elemental zinc to the infants. There was no clear evidence of difference in effect for the mortality by dose of zinc supplementation (<3 mg, 3 to 5 mg and >5 mg). There were no data for morbidity outcomes including diarrhea. Compared with no zinc, the MD for weight in infants supplemented with 3 to 5 mg per day of zinc was 316.47 g (95% CI 161.75 to 471.19, 5 trials), whereas that was 550.98 g (95% CI 246.39 to 855.56, 3 trials) for zinc >5mg per day. Compared with no zinc, the MD for length in infants supplemented with zinc 3 to 5 mg per day was 2.18 cm (95% CI, 0.71 to 3.65, 4 trials), and that for zinc >5 mg per day was 3.90 (95% CI 1.46 to 6.34, 2 trials). Compared with no zinc, the relative risk of diarrhea in zinc supplemented infants with <3 mg per day was 1.0 (95% CI 0.72 to 1.39, 1 trial), and that with 3-5 mg per day was 0.75 (95% CI 0.62 to 0.92, 5 trials) (Appendix 5).

In very preterm and VLBW infants, a single available trial with 193 participants reported a 71% decrease in the risk of death with zinc supplementation (RR 0.29, 95% CI 0.11 to 0.76).²³  There was substantial effect of zinc supplementation on weight, length and head growth, and no effect on sepsis in very preterm or VLBW infants (Appendix 5 and 6). Subgroup analyses by country income, age of enrolment, fortification of human milk, wherever available, are described in Appendix 5.

Our systematic review included 14 trials of 9940 preterm or LBW infants from 11 countries. We found low to moderate certainty evidence that enteral zinc supplementation increased weight, length, and head circumference and decreased diarrhea at the end of latest follow up. We found very low to low certainty evidence that zinc may have little or no effect on mortality, acute respiratory infection, sepsis, and hospitalization. We also found very low to low certainty evidence from 2 trials that zinc supplementation may decrease mental development scores and may have no effect on psychomotor development scores. There were no reports of reported serious adverse events.

A systematic review in 2011 of 3 trials and 2220 participants on zinc supplementation in LBW breastfed infants did not find evidence of an effect of zinc on all-cause mortality, infectious morbidities, and growth.⁶  However, there have been 7 more relevant studies since 2011. A recently published Cochrane systematic review in 2021 assessed the effect of zinc supplementation in breastfed or formula-fed preterm infants and included 5 trials.⁷  The review reported similarities with our review on improved weight gain (standardized mean difference in z scores [SMD] 0.46, 95% CI 0.28 to 0.64, 5 RCT, 481 participants, moderate certainty), linear growth (SMD 0.75, 95% CI 0.36 to 1.14, 3 RCTs, 289 participants, low certainty), but reported little to no effect on head growth (SMD 0.21 95% CI −0.02 to 0.44). The review did not report on diarrhea but reported little or no effect on sepsis (RR 1.11, 95% CI 0.60 to 2.04). The review also reported a decrease in mortality (RR 0.55, 95% CI 0.31 to 0.97, 3 RCT, 345 participants, low certainty), but found no studies reporting neurodevelopment or long-term growth outcomes. The differences in the estimates between our review and this Cochrane review are primarily because of the difference in the study populations and inclusion criteria. We included all available studies on zinc supplementation in preterm or LBW infants who were fed with mother’s milk or donor milk and excluded studies where infants were exclusively fed on formula. The Cochrane 2021 review included infants irrespective of breastfed or formula-fed, but only included preterm infants and excluded studies with term LBW infants. Infant formula contains high levels of zinc and other micronutrients, and it is likely that effects of zinc supplementation are different in human milk fed only infants than formula fed infants.

We reported a decrease in BSID mental development scores and no effect on psychomotor development scores from 2 trials.^15,26  In 1 of these included trials (Ashworth A et al 1998¹⁵ ), there was an error in the manufacturing of zinc solution that led to 1 cohort of infants being provided 1 mg instead of 5 mg. Later, a second cohort of infants was recruited in the study that received 5 mg of zinc. Because of this error, there might be a baseline imbalance in infants allocated to treatment and control groups that may influence the neurodevelopment outcome measure in this trial. Given the high risk of bias in the included studies, the effect of zinc on neurodevelopment scores seems inconclusive and should be interpreted with caution. In another trial, zinc supplementation showed improved neurodevelopmental outcomes assessed using Amiel-Tison scale.¹⁹  The 2021 Cochrane review,⁷  did not include these trials as the population did not meet their specified inclusion criteria of preterm infants, although they were LBW infants. Moreover, the reliability of BSID score (gold standard for childhood developmental assessment) is debatable when measured before 24 months of age.^27–29  We graded this evidence as low to very low certainty because of small sample size, inconsistency, and risk of bias in the included studies. More studies as well as longer-term follow up of zinc supplemented infants is necessary to clarify this observation and generate better quality evidence.

Other systematic reviews have shown that zinc supplementation improved linear growth in under 5 children,³⁰  weight-for-age and weight-for-length Z scores in infants³¹  but had no effect on mental or motor development in infants.³²  Zinc supplementation is also seen to have preventive effects on diarrheal and respiratory illness.^33,34  Biologically, the effects on infections and growth are plausible given the role of zinc in basic metabolic and cellular functions, synthesis of proteins and enzymes, carbohydrate metabolism, cell division, immunity, and skeletal growth.^35–37  However, mechanisms for the effect on zinc on neurodevelopmental outcomes are sketchy.

The key limitations were high heterogeneity between studies and high risk of bias in 55% of the included studies. Although, we conducted subgroup analysis defined a priori, much of the heterogeneity remain unexplained. Data were limited for outcomes of acute respiratory infection, sepsis, necrotising enterocolitis, and neurodevelopment, and for the subgroup of VLBW or very preterm infants, as many of these were assessed in 1 or 2 studies. No studies compared dosing or timing of initiation. The strengths of our study include extensive literature search in multiple databases and inclusion of relevant studies with a total of 9940 preterm or LBW infants. To minimize data availability bias, we contacted study authors to get relevant data related to our study population. The statistical power for some of the critical outcomes was high. The included trials were undertaken in a variety of healthcare settings including high-, middle- and low-income settings, with the period of follow-up ranging from 42 weeks post-conceptual age to 35 months of age. These factors are to be considered in relation to generalizability of the findings for each outcome. Although, our review findings are applicable to human milk fed LBW or preterm infants, globally including LMICs, caution is warranted as the certainty of evidence ranges from moderate to very low.

It is standard practice in many neonatal units to give VLBW infants a multicomponent fortifier with human milk, which provides an additional 0.5 to 1.8 mg/kg per day of zinc until the infant reaches a weight of 1800 to 2000 g.³⁸  Few units provide zinc as an individual supplement. Our review provides moderate to very low certainty evidence suggesting that enteral zinc compared with no zinc supplementation in human milk fed preterm or LBW infants can prevent diarrhea, improve weight, length, and head growth, but may have little or no benefits in reducing deaths. Our findings on mental development require further research. Further studies on dosing and timing of zinc supplementation are needed. Large-scale, high-quality trials among infants with different gestational age groups and birth weight categories including very preterm and VLBW infants will be helpful to generate better quality evidence.

The Society for Applied Studies acknowledges the core support provided by the Department of Maternal, Newborn, Child and Adolescent Health, World Health Organization, Geneva (WHO Collaborating Centre IND-158).

BSID

Bayley Scale of Infant Development

CI

confidence interval

GRADE

Grading of Recommendations Assessment, Development and Evaluation

LBW

Low birth weight

MD

mean difference

RCT

randomized controlled trial

RR

relative risk

SMD

standardized mean difference

VLBW

very low birth weight

WHO

World Health Organization