Zulfiqar A. Bhutta, Noordin Noormahomed Sharief Professor & Founding Chair, Division of Women & Child Health, The Aga Khan University, Karachi 74800, Pakistan. E-mail: firstname.lastname@example.org
Anaemia is one of the most resilient global public health problems and affects a staggering 1.62 billion of the world's population, largely concentrated in Asia and Africa. Anaemia contributes to almost 120 000 maternal deaths globally and indirectly to almost a fifth (18%) of the burden of maternal mortality. Given the widespread prevalence of micronutrient deficiencies in developing countries, supplementation with multiple micronutrients rather than iron folate alone, could be of potential benefit to the mother and the fetus. This review evaluates the evidence of the impact of multiple micronutrient supplements during pregnancy, in comparison with standard iron folate supplements, on maternal anaemia. A systematic review of randomised controlled trials was conducted using search engines like PubMed, the Cochrane Library and World Health Organization Regional Databases. Primary outcomes were ‘maternal anaemia’ and ‘haemoglobin level’. We included seven studies for detailed data abstraction. There was no differential benefit of multiple micronutrients as compared with iron folate on maternal anaemia in the third trimester (relative risk = 1.03 [95% confidence interval 0.94, 1.12]). Results were similar for haemoglobin levels. In summary, multiple micronutrients have a similar effect on maternal anaemia compared with iron folate supplementation. These findings have to be interpreted in the context of other benefits of multiple micronutrient supplementation such as promoting better fetal growth and the possible increased risk of neonatal and perinatal mortality that are best resolved through large-scale effectiveness trials.
Anaemia is one of the most intractable global public health problems and affects a staggering 1.62 billion of the world's population, largely concentrated in Asia and Africa.1 Anaemia contributes to almost 120 000 maternal deaths globally and indirectly to almost a fifth (18%) of the burden of maternal mortality.2 In addition to maternal deaths, there are several adverse health outcomes associated with anaemia including maternal morbidity, stillbirths and neonatal mortality, low birthweight (LBW) and poor cognitive development in the offspring.3 In terms of risks, women of reproductive age are more vulnerable to anaemia because of recurrent menstrual loss and the demands of pregnancy and repeated childbearing but corresponding estimates from adolescents and young women are lacking. The most recent global estimates suggest that the prevalence of anaemia is 41.8% among pregnant women and 30.2% among non-pregnant women.1–4
Anaemia, as defined by low haemoglobin or haematocrit, is commonly used to assess the severity of iron deficiency in populations without high rates of malaria. The aetiology of anaemia is multi-factorial including genetic factors, dietary deficiencies, repeated pregnancies and high burden of infectious diseases.5 A multi-pronged approach is needed for its prevention and treatment. Although iron deficiency and infections are the most common aetiological factors, other conditions may also be important. These include nutritional deficiencies of other micronutrients like vitamin A, vitamin B12, folate and riboflavin as well as thalassaemias and haemoglobinopathies.3–5 Pregnancy greatly increases the demand for macronutrients and micronutrients. To illustrate, iron requirements increase almost tenfold by the third trimester.6,7 The high physiological requirement for iron in pregnancy is difficult to meet with most local diets even if fortified. Therefore, it is recommended that pregnant women routinely receive iron supplements, especially in developing countries with high rates of iron deficiency anaemia.8
Most of the focus in public health programmes on the prevention of anaemia in pregnancy has been on iron/folic acid (IFA) supplementation in pregnancy, as an integral part of antenatal care. However, global coverage data indicate that this is far from satisfactory. To illustrate, in 2005/06 only two-thirds of Indian women report taking IFA during their last pregnancy with only 17% taking IFA for the 100 days or more recommended by India's health policy.9 Adherence to iron supplements during pregnancy in a national programme is dependent on a variety of factors, including adequate programme support, sufficient delivery of services, and patient related factors such as promotion and support, particularly important given gastrointestinal side-effects of iron supplementation in some women.10,11 The latter have been widely regarded as a major impediment to achieving sufficient compliance and hence programme coverage. However, a large multi-county qualitative study evaluating this issue suggested that perceived side-effects were not a major barrier and that other factors could be contributing to poor coverage.12 Other studies from India evaluating the National Nutritional Anemia Prophylaxis Programme at national or state level indicated that a combination of factors were responsible for poor coverage including irregular and inadequate supply of iron tablets, insufficient supply of services, inadequate consumption by beneficiaries and side-effects.13,14
In recent years in addition to the aforementioned issues of poor uptake and reach of IFA supplementation programmes, two factors have emerged, which suggest that the policy for isolated IFA supplementation in pregnancy ought to be reviewed. The first is the recognition that many women of reproductive age and pregnant women have multiple micronutrient (MMN) deficiencies15–18 and the other is that in addition to anaemia, important birth outcomes such as LBW should be considered.
Although an earlier meta-analysis of iron folate supplementation in pregnancy did not suggest a significant improvement in birthweight,19 there was evidence to indicate that there may be benefits [mean difference (MD) 36.05 g [95% confidence interval (CI) −4.84, 76.95 g]] and 21% reduction in LBW [relative risk (RR) = 0.79 [95% CI 0.61, 1.03]].19 In an updated analysis, we have shown20 that IFA supplementation during pregnancy decreases incidence of LBW (RR = 0.81 [95% CI 0.74, 0.89] and increases mean birthweight (MD = 42.18 g [95% CI 9.27, 75.09]. Further evidence of potential benefits of iron folate on birthweight and mortality outcomes in the newborn is evident from population-based observational studies. To illustrate, analysis of a nationally representative survey in Zimbabwe suggested that maternal iron supplementation was associated with an increase in birthweight of 103 g [95% CI 42 g, 164 g].21 Others have indicated that IFA supplementation was associated with a significant reduction in early neonatal mortality by 47% in Indonesia22 and that the combination of iron folate supplementation and IFA and IPTp (intermittent preventive treatment of plasmodium malaria) was associated with 24% reduction on neonatal mortality in 19 sub-Saharan African countries.23 In a recent analysis, Roberfroid and colleagues have shown that fetal growth with regards to iron supplements is associated with cumulative micronutrient intake and duration of supplementation.24 Both cumulative micronutrient intake and duration were associated with decrease in incidence of small for gestation babies and increase in birthweight.
Studies of MMN supplementation in pregnancy have suggested benefits on reduction of maternal anaemia and improved birthweight.25,26 In general these studies provided a lower daily dose of iron on the assumption that other micronutrients such as vitamin A and ascorbic acid would facilitate bioavailability and uptake of iron with lower gastrointestinal side-effects.27 Some MMN supplementation studies have also suggested benefits on infant28 and newborn mortality29 and growth in later infancy.30 Others have cautioned against the use of these supplements on the basis of an increased risk of perinatal mortality observed in some settings31–33 although this has been disputed given the overall variation in context.18,26
In the current conundrum an evaluation of the potential benefits and risks of replacing iron folate supplements with MMN is necessary and must in the first instance be based on a factual evaluation of the scientific benefits and risks. The logic of replacing iron folate supplements (usually at 60 mg daily in developing countries) with MMN (with a lower dose of iron) is that if they had similar benefits to iron folate in reducing anaemia yet additional benefits on intrauterine growth and outcomes in childhood, they would offer a huge advantage for public health programmes.
We undertook an updated systematic review to compare and synthesise the most up-to-date evidence comparing the effects of MMN supplements during pregnancy vs. iron folate for the prevention of maternal anaemia. Effects on birth outcomes have been assessed elsewhere in this supplement.33
A literature search was conducted on PubMed, the Cochrane Systematic Reviews, the World Health Organization Regional Databases and hand search of bibliographies of relevant reviews. The last date of search was 15 September 2011. Studies were included, irrespective of language and status of publication. Experts were also contacted in the field for unpublished data. The basic search strategy used was:
(‘Mothers’[Mesh] OR ‘Pregnancy’[Mesh] OR mother* OR maternal OR pregnancy) AND (‘Micronutrients’[Mesh] OR ‘multiple micronutrient*’ OR multivitamin OR micronutrient*) AND (supplement*).
1Only prospective randomised controlled trials evaluating MMN supplementation in women during pregnancy were included.
2Multiple micronutrients were defined as supplementation with at least five micronutrients including the UNICEF (The United Nations Children's Fund)/WHO (World Health Organization)/UNU (United Nation University) international multiple micronutrient preparation (UNIMMAP) formulation27 or those with similar composition.
3The comparison group received less than three micronutrients (mainly iron folate) during pregnancy.
4There were no limits on gestational age at the time of enrolment in the study and the duration of supplementation.
1Animal studies were excluded.
2Non-intervention studies, review articles, cross-sectional, case–control, cohort studies, commentary letter or editorial were excluded.
3Studies with non-healthy pregnant women (hospitalised patients, symptomatic HIV-positive women) were excluded.
4Studies with fewer than five micronutrients studies were excluded.
5Studies of fortified food, and Sprinkles powder were excluded.
6Studies of preconceptual or periconceptual interventions were excluded.
Data abstraction and validity assessment
Each study that satisfied the eligibility criteria was included in the review. Two authors abstracted the data in a standardised excel sheet for variables like sample size, study location, setting and study methods like blinding, allocation concealment and description of intervention and control groups (in terms of dosage and time of enrolment). Each study was graded based on: (1) study design, (2) study quality, and (3) relevance to the objectives of the review.34 The following four grades were given to individual studies: high, moderate, low or very low. A study received an initial score of high if it was a randomised or cluster randomised trial. The grade was decreased by one for each study design limitation. In addition, studies reporting an intent-to-treat analysis or with statistically significant strong levels of association (>80% reduction) received one- to two-grade increases. Any study with a final grade of very low was excluded on the basis of inadequate study quality.
The overall quality of evidence of an outcome was also assessed and graded according to the Child Health Epidemiology Group adaptation of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) technique34,35 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, as reflected by the P-value.
Meta-analysis was performed where data were available from more than one study. Primary outcomes were haemoglobin levels and risk of anaemia. Birth outcomes have been assessed in another paper in this supplement.33 Continuous data were pooled to get a standard mean difference. Dichotomous data were pooled to get a relative risk, which was reported along with 95% CI. Fixed models were used for all the analyses. The assessment of statistical heterogeneity among the pooled data was done by visual inspection, that is, the overlap of the CI 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 I2 values were also looked into, and roughly an I2 greater than 50% was taken to represent substantial and high heterogeneity. In situations of substantial or high heterogeneity being present, causes were explored by sensitivity analysis and subgroup analyses. For all cluster randomised trials, cluster adjusted estimates were used. For the purpose of the analyses of the studies with factorial designs, we assumed that there was no interaction between other interventions and the effect of supplementation, as all interventions were randomised. All analyses were performed using Review Manager (RevMan) [Computer program]. Version 5.1 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2011).
The literature search revealed 4770 hits (Figure 1) on PubMed. After screening the titles and abstracts, 43 studies were initially considered eligible and, finally, seven studies were selected for inclusion in this review.29,36–40 We evaluated the impact of the intervention on the following outcomes: maternal anaemia and maternal haemoglobin levels.
Web Table S1 shows characteristics of included studies. All studies were from low-/middle-income settings. Four studies used MMN supplement formula called UNIMMAP, which consisted of 30 mg iron, 400 µg folic acid, 15 mg zinc, 2 mg copper, 65 µg selenium, 800 µg RE vitamin A, 1.4 mg vitamin B1, 1.4 mg vitamin B2, 18 mg niacin, 1.9 mg vitamin B6, 2.6 µg vitamin B12, 70 mg vitamin C, 5 µg vitamin D, 10 mg vitamin E and 150 µg iodine. The formulas used in rest of the studies were similar to UNIMMAP except in small variation in dose of IFA used. Among all studies, supplementations were started at 28 weeks at the latest. Some of the studies had limitations based on study design and execution. Web Table S2 shows risk of bias for included studies.
Quantitative data synthesis
Table 1 reports the overall quality grading of the outcomes and results of the corresponding meta-analyses. MMN supplementation had no significant effect on incidence of maternal anaemia in the third trimester compared with iron folate based on data from seven trials [RR = 1.03, 95% CI 0.94, 1.12] (Figure 2). Similarly, when data were pooled for haemoglobin level, there was no significant effect between MMN and IFA (standard mean difference = −0.01 [95% CI −0.08, 0.06]) (Figure 3). A standardised mean difference was calculated for this outcome as the units of measurements were different among the studies.
Table 1. Quality assessment of trials of multiple micronutrient supplementation for maternal anaemia (multiple micronutrients vs. iron folate)
This review demonstrates that MMN supplements have similar beneficial effects to iron folate supplementation for prevention of anaemia in pregnancy. The pooled studies had non-significant statistical heterogeneity and there was no study that showed statistically significant different results between the two study groups. These results corroborate earlier reviews that also revealed similar findings and support the overall findings by others.26 This review was limited to outcomes for maternal anaemia but benefits of MMN supplements over IFA alone in terms of birthweight and LBW have been presented elsewhere in this supplement.33 Briefly, the second review included 16 trials and the combined results showed that compared with control supplementation that was usually iron plus folic acid in most studies, MMN supplementation resulted in a significant reduction in the incidence of LBW (RR = 0.86 [95% CI 0.81, 0.92]) and small-for-gestational age (RR = 0.83 [95% CI 0.73, 0.95]) and an increase in mean birthweight (Weighed Mean Difference 54.5 g [95% CI 45.4 g, 63.5 g]). There was no significant difference in the overall risk of preterm birth, stillbirth or neonatal mortality. None of the studies evaluated maternal morbidity or mortality.
Our findings should be considered in the context of the overall effects of MMN on pregnancy outcomes. MMN has similar beneficial effects on maternal anaemia as that of iron folate supplementation and have additional advantages of prevention of Intrauterine Growth Restriction and LBW.20,33 In addition to our group,36,41 other investigators have also reported similar findings from previous reviews of trials of MMN using a specific formulation (UNIMMAP) and indicated a significant benefit in comparison with IF supplements on reducing small-for-gestational age [pooled odds ratio (OR) = 0.90 [95% CI 0.82, 0.99] based on a pooled analysis of data from 12 studies].42
There are two additional considerations that impact the potential move to replace IF with MMN. One is the potential cost of such a move and the other the potential adverse effects. There have been concerns over the potential adverse effects of an increase in neonatal mortality in less developed settings that have suboptimal maternal care.31,32 A non-significant increase in the risk of neonatal mortality was noted in UNIMMAP trials analysis (OR = 1.23 [95% CI 0.96, 1.59]).36 We also evaluated this possibility in a recent pooled analysis from nine studies and demonstrated an overall non-significant effect on neonatal mortality (RR = 1.05 [95% CI 0.92, 1.19]). There was however increase risk of neonatal mortality where the majority of births occurred at home compared with that where the majority of births were in facilities.18 Also, in the review by Ramakrishnan et al. included in this supplement, an increased risk of neonatal death with MMN supplementation was seen in the subgroup of trials that began the intervention after the first trimester.33 These findings suggest that the use of MMN supplements should be started in the first trimester or preconceptionally and must be accompanied by the provision of skilled care at delivery and facility births to offset any potential increase in the risk of obstructed labour and birth asphyxia.36,41
Oral iron supplementation during pregnancy is known to cause side-effects especially that of gastrointestinal system. A Cochrane review that evaluated effectiveness and safety of iron supplementation during pregnancy has shown that side-effects (any) were more common among women who received daily iron or IFA supplementation than among those who received no treatment or placebo (RR = 3.92 [95% CI 1.21, 12.64]).19 It remains to be seen in detail if MMN supplements have similar adverse effects to routine IF supplements. A study from Mali43 compared acceptability and adherence to a daily MMN supplementation to iron folate supplementation during pregnancy. The results showed that there were no significant differences between comparison groups with respect to women's perceptions about supplement size, colour, taste or flavour. Adherence to the MMN supplementation scheme was better (257.5 ± 20.9 tablets; average adherence 95.4%) than that to the IFA supplementation scheme (238.5 ± 32.7 tablets; average adherence 92.2%; P = 0.008). A study from China44 showed that reported side-effects did not differ between the IFA group and the group receiving MMN, even though the IFA group received twice the amount of iron as the group receiving MMN. The most commonly reported side-effects reported were nausea (47%), severe gastrointestinal symptoms (34%) and vomiting (16%), with no differences between the supplement groups. This preliminary data showed that side-effect profile does not differ between MMN and iron folate group. It has also been shown that side-effect profile is not the leading cause of current low adherence to oral supplements (either MMN or IFA). Instead it depends on adequate supply of supplements and community education within programmes to achieve high level of compliance.45,46
There have been relatively few studies that examine costs and cost-effectiveness of using MMN in pregnancy.44 MMN supplements are more expensive than iron folate supplements, but bulk production could bring costs down.47 However, the costs of distribution of either MMN or IFA are similar and given the potential benefits of addressing other micronutrient deficiencies concurrently, MMN supplements remain an attractive option. Moreover, if benefits for women of reproductive age and adolescent girls can be established, there are also other mechanisms to ensure a cost-effective supply chain. The recently constituted UN Commission on Commodities is one step in the direction of establishing a steady and reliable supply at low cost, and could be used to ensure a steady source for procurement in low and middle income countries.
In summary, MMN and iron folate supplements have similar effects on reducing maternal anaemia. A plea to scale up iron folate supplements has been made48 in view of the beneficial effect for prevention of Intrauterine Growth Restriction and that of LBW from other studies. However, MMN supplementation improves these outcomes beyond what iron folate supplements achieve.49,50 The concern about the potential risk of increased perinatal and neonatal mortality with the use of MMN in pregnancy however remains, and the quantum of risk in relation to early access to prenatal care and availability of obstetric services should be and is best evaluated in appropriate effectiveness trials.
We thank Beth Imhoff- Kunsch for the valuable inputs during the review process.
Conflicts of interest
The authors have not declared any conflicts of interest.