Types of trials: All randomized or qausi–randomized trials with individual or cluster allocation and using concurrent controls. Trials employing a factorial design with multiple intervention groups were eligible for inclusion.

Participants: Low birth weight infants (birth weight less than 2500 grams) irrespective of gestational status who were exclusively or predominantly breastfed at the initiation of intervention. Exclusive breast-feeding was defined as no feed other than breastmilk. Predominant breastfeeding was defined as taking only water or multivitamins or medicines other than breastfeeds.

Intervention: Zinc salts given as tablets or syrups orally to provide at least 1 RDA of elemental zinc [6, 7] for at least a period of 14 days and introduced within one month of birth. Trials providing additional supplements (for example, vitamin A, micronutrient mixtures, iron) were considered if the only difference between the two comparison arms was zinc supplement. The comparison groups included no intervention or placebo.

Primary: We examined all cause mortality in the child at two time points: during infancy, in the period between initiation of intervention and the last follow-up until the age of one year; and during the neonatal period between initiation of intervention and the last follow-up until the age of one month.

Secondary: In the period between initiation of intervention and the last follow-up until the age of one year we measured cause specific mortality due to diarrhea, acute respiratory infections and causes other than these (as defined by the authors, irrespective of ascribing a single or multiple causes of death), severity of morbidities as assessed by clinic or hospital visits or hospitalizations (as defined by the authors of the trials) and morbidities because of sepsis including probable or culture proven bacterial sepsis, acute respiratory infection or respiratory difficulty, diarrhea, meningitis, ear infections, cough or running nose, fever or severe malnutrition (severe wasting, and pedal edema or kwashiorkor). We also measured weight and height at the end of intervention and adverse effects including vomiting.

The trials were identified by simultaneous searches of medical databases (till August 26, 2009) including PubMed (since 1966), EMBASE (since 1980), Cochrane Controlled Trials Register, Web of Science (WoS), Allied and Complementary Medicine (AMED) (since 1985), British Nursing Index (BNI) (since 1994) and CAB abstracts (since 1973) with no language restrictions. For PubMed the search strategy employed was: (newborn OR neonat* OR infant OR neonates OR postnatal OR post-natal) AND (low birth weight OR preterm OR small for gestational age OR premature) AND ("zinc" OR micronutrient* OR vitamin*) AND  ((clinical[Title/Abstract] AND trial[Title/Abstract]) OR clinical trials[MeSH Terms] OR clinical trial[Publication Type] OR random*[Title/Abstract] OR random allocation[MeSH Terms] OR systematic OR review OR metaanalysis OR meta-analysis).

A lateral search using the related articles link in PubMed was done for selected articles. We reviewed the reference lists of identified articles and hand searched reviews, and abstracts of international micronutrient conferences of past three years. To avoid publication bias, we included published and unpublished trials. Requests for information were sent to experts and major development and aid agencies including United Nations Children’s Fund (UNICEF), WHO, United States Agency for International Development, and Bill and Melinda Gates Foundation.

Data extraction and management: We extracted data in duplicate on study design, participant characteristics, interventions, and outcomes and contacted authors for additional information if required. Any discrepancies in extracted data were resolved by discussion.

Assessment of risk of bias in included studies: This was assessed in relation to sequence generation, allocation sequence concealment, blinding, incomplete outcome data assessment, selective reporting and any other bias [8].

In factorial trials and in multi-arm designs yielding two or more intervention groups (different dosage or administration regimens) and a single control group, the data in the intervention groups, including the variation in the intervention characteristic, was to be pooled and compared against the single control group to prevent unit-of-analysis error.

Using the search strategy, 12 potentially eligible reports were identified [15-26]. Amongst these, 9 reports [15-23] were excluded for various reasons (Web Fig. 1). The remaining 3 reports [24-26] provided data on 3 trials satisfying the inclusion criteria. The baseline characteristics of the included trials, all from developing countries [27] and the notable individual study specific features are summarized in Web Table I and Web Annexure. Web Table II summarizes the assessed risk of bias in the included studies. In 2 trials, no risk of bias was evident; while in the third trial; intention to treat analysis was not performed. Data was insufficient to calculate the routine measures of publication bias (Begg-Mazumdar bias and Horbold-Egger bias statistics).

Fig. 1 Flow chart depicting the trial flow for selection of randomized control trials included in the systematic review.

Infant mortality: Only one trial reported on the number of infant deaths [26]. Of the 1026 infants randomized in each group, there were 21 deaths in the supplemented and 19 deaths in the placebo groups. The relative risk of mortality with zinc supplementation group was 1.11 (95% CI 0.57 to 2.18) at one year.

Neonatal mortality: No trial reported on the number of neonatal deaths.

Cause specific mortality: The only trial [26] with information on deaths did not report on cause specific mortality.

Severity of morbidities: Only one trial [26] reported on hospitalization rate as number/1000 child-years of follow up (189.9 in zinc supplemented and 175.3 in placebo groups, respectively). The odds ratio for hospitalization with zinc supplementation was 1.10 (95% CI 0.87 to 1.39). No trial had reported upon clinic or hospital visits.

Other morbidities: No trial reported morbidities other than diarrhea and ARI.

Adverse effects: Only one trial had reported on adverse effects [26]. There was no evidence of an increase in vomiting following zinc supplementation. The pooled relative risk (random effects model) of vomiting with zinc supplementation was 1.06 (95% CI 0.88 to 1.28) till one year of age and 1.17 (95% CI 0.91 to 1.51) till 6 months of age.

Exploratory subgroup and meta-regression analyses were not feasible because outcomes were reported in a small number (one to three) of trials.

There was no convincing evidence of benefit following zinc supplementation in breastfed LBW infants. Limited data from 1 to 3 trials did not indicate a reduced risk of mortality (1 trial), hospitalization rate (1 trial), acute respiratory infection (1 trial), or diarrhea (2 trials). However, the trial reporting on mortality was not adequately powered for evaluating this outcome. There was no evidence of an increase in weight (3 trials) or height (2 trials) at either 6 months or one year of age or of an increased risk of vomiting following zinc supplementation. Serum zinc levels at the end of intervention were significantly higher in the supplemented group (2 trials) indicating successful absorption of the micronutrient.

Following limitations merit consideration. First, only three trials satisfied the inclusion criteria and inferences on some outcomes were based on 1 or 2 trials only. The only trial reporting on mortality was not adequately powered to evaluate this outcome. Second, all trials were conducted in developing countries and two were from India, which limits the generalization of findings to other regions in developing countries and developed nations. Third, the possibility of publication bias and predictors for heterogeneity could not be explored due to the small number of trials. Fourth, information on exclusive or predominant breast feeding status was available only at the time of initiation of the study but not in a longitudinal manner till the end of the intervention.

The following criterion for data inclusion deserves elucidation: (i) The intervention was evaluated in LBW breastfed infants to guide policy formulation by the WHO for optimal feeding of LBW infants in developing countries. LBW newborns comprise a heterogeneous population of preterm and small for gestational age (SGA) newborns that are physiologically different. Preterm infants are likely to have higher zinc deficit and dietary requirements as nearly 60% fetal zinc is acquired during third trimester of pregnancy. It is therefore conceivable that the response to zinc supplementation may be variable amongst preterm and SGA babies. It would therefore have been ideal to evaluate the effect of intervention in relation to gestation and growth retardation. However, most of the relevant trials have not segregated the data into preterm and/or growth retarded babies. Also in the context of developing guidelines in resource starved countries, the only practical option is categorization as LBW rather than SGA and/or preterm because gestation is difficult to determine with accuracy. Further, nearly 70% of LBW babies in developing countries are SGA unlike in the developed world where the bulk of LBW babies are preterms [1]. Thus we feel that our study criteria are apt to feed the policy requirement. (ii) The type of intervention was defined as zinc salts given as tablets or syrups orally to provide at least 1 RDA of elemental zinc for at least a period of 14 days introduced within one month of birth. The RDAs for zinc in low birth weight breast fed infants have not been defined [6, 7]. However, the suggested average dietary intakes (ADI, which are invariably lower than RDA) for infants in the age group of 0 to 6 months are 2 mg(6) and it is believed that the requirements would be higher in LBW infants. We therefore used the conservative cut-off of 2 mg elemental zinc intake to define the RDA in LBW newborns.

It could be postulated that beneficial effects were not evident because the intervention may not have led to an increase in zinc nutriture of the participants. However, a significant increase of serum zinc in sub-groups of two trials refutes this hypothesis. Further, as per the inclusion criteria the participants had received at least one recommended ADI per day.

All the reviewed evidence pertains to developing countries, and is primarily from populations at risk of developing zinc deficiency. In view of no convincing evidence of benefits from the limited data available currently, there is no justification for recommending routine zinc supplementation for breastfed low birth weight newborns in these populations. Future research and trials on this subject should examine: outcomes in more settings in Asia and Africa; and should be adequately powered to estimate mortality. These trials should also record feeding status of participants and record causes of death and morbidities other than diarrhea and ARI.

Contributors: AG prepared the protocol, applied the search strategy, retrieved the articles, and extracted data. HPSS developed the idea for review, finalized the protocol and search strategy, confirmed the extracted data, and did the statistical analysis. SB prepared and finalized the protocol. All authors contributed to the drafting of the final version of the paper. AG and HPSS will act as joint guarantors.

Funding: Department of Child and Adolescent Health and Development, World Health Organization and Sitaram Bhartia Institute of Science and Research, B-16 Qutab Institutional Area, New Delhi, India. The funding sources had no involvement in the study or the decision to publish the manuscript.

Competing interest: None stated.

1. Sachdev HPS, Shah D. Epidemiology of maternal and fetal malnutrition in South Asia. In: Bhutta ZA, editor. Oxford: Perinatal and Newborn Care in South Asia. Ed.: Oxford University Press. 2007. p. 75-105.

2. Singh PP, Khushlani K, Veerwal PC, Gupta RC. Maternal hypozincemia and low-birth-weight infants. Clin Chem. 1987;33:1950.

3. Goel R, Misra PK. Study of plasma zinc in neonates and their mothers. Indian Pediatr. 1982;19:611-4.

4. Jeswani RM, Vani SN. A study of serum zinc levels in cord blood of neonates and their mothers. Indian J Pediatr. 1991;58:683-6.

5. Bhandari N, Taneja S, Mazumder S, Bahl R, Fontaine O, Bhan MK; Zinc Study Group. Adding zinc to supplemental iron and folic acid does not affect mortality and severe morbidity in young children. J Nutr. 2007;137: 112-7.

6. Dietary reference intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc. A Report of the Panel on Micronutrients, Subcommittees on Upper Reference Levels of Nutrients and of Interpretation and Uses of Dietary Reference Intakes, and the Standing Committee on the Scientific Evaluation of Dietary Reference Intakes, Food and Nutrition Board, Institute of Medicine. Zinc. National Academy Press, Washington, DC, 2001 p 442-501. Available from: www.nap.edu/openbook.php?record_id= 10026 & page=442. Accessed on December 2, 2009.

7. Hotz C. Dietary indicators for assessing the adequacy of population zinc intakes. Food Nutr Bull. 2007;28:S430-48.

8. Higgins JPT, Altman DG on behalf of the Cochrane Statistical Methods Group and the Cochrane Bias Methods Group. Assessing risk of bias in included studies. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. New Jersey: John Wiley & Sons Inc. 2008. p.187-241.

9. Sterne JAC, Egger M, Smith GD. Investigating and dealing with publication and other biases. In: Egger M, Smith GD, Altman DG, eds. Systematic Reviews in Health Care: Meta-analysis in Context. London: BMJ Books. 2001. p. 189-208.

10. Sterne JAC, Bradburn MJ, Egger M. Meta-analysis in STATA TM. In: Egger M, Smith GD, Altman DG editors. Systematic Reviews in Health Care: Meta-analysis in Context. London: BMJ Books. 2001. p.347-369.

11. Steichen TJ, Egger M, Sterne JAC.  sbe19.1: Tests for publication bias in meta-analysis.  Stata Technical Bulletin. 1998;44:3-4.

12. Harris R, Bradburn MJ, Deeks J, Harbord R, Altman D, Steichen T, et al. Stata Version 9 Update (Distribution Date February 19, 2007) for the Stata User Written Programme sbe24 (Bradburn MJ, Deeks JJ, Altman D.  sbe24: metan - an alternative meta-analysis command.  Stata Technical Bulletin 1999; 44: 4-15). Available from: http://fmwww. bc.edu/RePEc/bocode/m. Accessed on March 24, 2007.

13. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21:1539-58.

14. Harbord R, Steichen T, Sharp SJ. 2004. "metareg: update to Stata module to perform meta-analysis regression" (Sharp SJ.  sbe23: Meta-analysis regression.  Stata Technical Bulletin 1998; 42: 16-22). Statistical Software Components S4446201, Boston College Department of Economics, revised 02 Feb 2005. Available from: http: //ideas.repec.org/c/boc/bocode/s446201.html. Accessed March 15, 2007.

15. Friel JK, Andrews WL, Matthew JD, Long DR, Cornel AM, Cox M, et al. Zinc supplementation in very-low-birth-weight infants. J Pediatr Gastroenterol Nutr. 1993;17:97-104.

16. Ashworth A, Morris SS, Lira PI, Grantham-McGregor SM. Zinc supplementation, mental development and behaviour in low birth weight term infants in northeast Brazil. Eur J Clin Nutr. 1998;52:223-7.

17. Sazawal S, Black RE, Menon VP, Dinghra P, Caulfield LE, Dhingra U, et al. Zinc supplementation in infants born small for gestational age reduces mortality: a prospective, randomized, controlled trial. Pediatrics. 2001;108:1280-6.

18. Castillo-Durán C, Perales CG, Hertrampf ED, Marín VB, Rivera FA, Icaza G. Effect of zinc supplementation on development and growth of Chilean infants. J Pediatr. 2001;138:229-35.

19. Osendarp SJM, Santosham M, Black RE, Wahed MA, van Raaij JMA, Fuchs GJ. Effect of zinc supplementation between 1 and 6 mo of life on growth and morbidity of Bangladeshi infants in urban slums. Am J Clin Nutr. 2002;76:1401-8.

20. Díaz-Gómez NM, Doménech E, Barroso F, Castells S, Cortabarria C, Jiménez A. The effect of zinc supplemen-tation on linear growth, body composition, and growth factors in preterm infants. Pediatrics. 2003;111:1002-9.

21. Black MM, Sazawal S, Black RE, Khosla S, Kumar J, Menon V. Cognitive and motor development among small-for-gestational-age infants: impact of zinc supplementation, birth weight, and caregiving practices. Pediatrics. 2004;113:1297-1305.

22. Loui A, Raab A, Wagner M, Weigel H, Grüters-Kieslich A, Brätter P, et al. Nutrition of very low birth weight infants fed human milk with or without supplemental trace elements: a randomized controlled trial. J Pediatr Gastroenterol Nutr. 2004;39:346-53.

23. Jimenez R, Martinez M, Penalver R. Effect of zinc on growth and development of children with low birth weight. Colombia Medica. 2007;38:6-13.

24. Castillo-Duran C, Rodriguez A, Venegas G, Alvarez P, Icaza G. Zinc supplementation and growth of infants born small for gestational age. J Pediatr. 1995;127:206-11.

25. Sur D, Gupta DN, Mondal SK, Ghosh S, Manna B, Rajendran K, et al. Impact of zinc supplementation on diarrheal morbidity and growth pattern of low birth weight infants in Kolkata, India: a randomized, double-blind, placebo-controlled, community-based study. Pediatrics. 2003;112:1327-32.

26. Taneja S, Bhandari N, Rongsen-Chandola T, Mahalanabis D, Fontaine O, Bhan MK. Effect of zinc supplementation on morbidity and growth in hospital-born, low-birth-weight infants. Am J Clin Nutr. 2009;90:385-91.