Introduction
Low birth weight, small-for-gestational age (SGA), preterm birth, stillbirths, perinatal and neonatal mortality are important adverse outcomes of pregnancy [
1]. The incidence of low birth weight in developing countries varies from 6 - 30%, and at least one-third of these are small for gestational age, especially in settings with high rates of maternal undernutrition. Small for gestational age (SGA) babies are those whose birth weight lies below the 10
th percentile for a particular gestational age [
2]. Vast majority of these are due to fetal growth problems that occur during pregnancy, including intrauterine growth restriction (IUGR) [
3]. Full term SGA infants may not have complications related to organ immaturity like those of pre-term infants of similar size, but are at an increased risk of stillbirth and perinatal/neonatal mortality due to perinatal asphyxia, meconium aspiration and hypoglycaemia [
4,
5]. Women of reproductive age, especially pregnant women, in developing countries are recognized to be at risk of multiple micronutrient deficiencies, such as iron, folic acid, iodine, zinc, vitamins A and D, riboflavin, B6 and B12, with the likelihood of adverse effects on the mother and pregnancy outcomes [
1,
6]. Pregnancy represents a state of increased metabolic requirements, and intake of key micronutrients by pregnant women especially in developing countries is usually inadequate. This inadequate intake and increased requirement further exacerbates the pre-existing maternal deficiency [
7].
Iron deficiency contributes to one of the largest prevalence of micronutrient deficiencies among pregnant women. For example, anaemia affects approximately 41.8% of all pregnancies globally [
8], with iron deficiency accounting for half of the cases [
9]. Retrospective and observational studies have demonstrated a higher risk of maternal mortality in severely anaemic pregnant women, predisposing to death from haemorrhage and infections. Maternal iron deficiency anaemia also has adverse effects on birth outcomes including a greater risk of birth asphyxia, low birth weight, preterm delivery and lower Apgar scores [
10,
11]. The association of maternal anemia with adverse outcomes is largely based on observational studies and risk assessment [
12] and objective evidence that iron-folate supplementation in pregnancy improves outcomes other than anemia is not well established.
Given the significant impact of deficiencies of key micronutrients during pregnancy [
6]
, supplementation with multiple micronutrients during pregnancy may be a feasible public health strategy. One potential advantage of multiple micronutrients could be that they might have comparable benefits to iron-folate in reducing anemia, and could also have additional benefits on intrauterine growth and outcomes in the neonatal period and infancy [
7]. Many workers had attempted augmentation of iron-folate supplementation in pregnancy with additional micronutrients, but the first systematic efforts to undertake this were almost a decade ago [
7]. In 1999 the UNICEF/WHO/UN University proposed a prenatal supplement UNIMAPP containing fifteen micronutrients, including iron and folic acid which could provide one recommended daily allowance of each and potentially replace standard iron-folate supplements for pregnant women in low and middle income countries [
13]. It was speculated that a combination of more than one micronutrient may have additive and/or synergistic benefit on maternal and child outcomes. Systematic information on the benefits if any, of such supplements in improving maternal and pregnancy outcomes is, however, limited. A Cochrane review on the subject in 2006 [
14] indicated that compared to iron-folate, multiple micronutrient supplementation (defined as administration of three or more micronutrients), did not have any significant benefits on maternal anemia (data from one study) and prevalence of SGA births (data included from 2 studies). Another systematic review was undertaken by a team commissioned by UNICEF/WHO/SCN which analysed data from 12 trials using the UNIMAPP formulation and evaluated effects on maternal and pregnancy outcomes [
15] There have been concerns regarding increased risk of perinatal and neonatal mortality with multiple micronutrient supplementation through increased birth asphyxia in heavier babies [
9] In this paper, we included additional studies since the Cochrane review in 2006 and have evaluated the evidence of the potential impact of multiple micronutrient supplementation in pregnancy on the maternal and pregnancy outcomes using Child Health Epidemiology Reference Group (CHERG) rules [
16]. Given that the Lives Saved Tool (LiST) largely models SGA outcomes for potential maternal nutrition interventions, we specifically focused our analysis to evaluate the impact of maternal multiple micronutrient interventions on risk of delivering SGA infants and also evaluated the impact on neonatal mortality.
Discussion
Our evaluation of multiple micronutrient supplements during pregnancy did not show a significant benefit of the supplement on maternal anaemia in third trimester as compared to iron-folate. These findings corroborates those from earlier reviews [
14,
36] which also revealed no significant benefit of multiple micronutrients over iron-folate. However, our review does indicate that multiple micronutrient administration is associated with a significant reduction in SGA births in comparison with iron-folate administration (9% and 12% reduction in the fixed and random effects models, respectively). These findings are comparable to the results of the systematic review of UNIMAPP trials [
37] (pooled OR = 0.90; 95% CI: 0.82 – 0.99) based on 12 studies, which are also included in our analysis. These findings are however, different from our previous Cochrane review which was based on a limited set of studies [
14] and reported a non-significant effect on SGA (RR = 1.04; 95% CI: 0.93 – 1.17). Our current data are also at variance with a recent meta-analysis by Shah et al. [
38] which analyzed results from only 5 RCTs and reported a statistically insignificant impact on SGA (RR = 0.89; 95% CI: 0.77 – 1.01). In our current review, 13 out of 14 studies showed consistent results with respect to SGA babies i.e. a risk ratio of less than 1. The study by Christian et al. reported a risk ratio of greater than 1, but was statistically not significant (RR = 1.04; 95% CI: 0.94 – 1.15) [
23]. The overall quality of evidence for this outcome was high, with a strong statistical association with the P-value of 0.0002 in the fixed model. Given that the LiST model uses the impact pathway via SGA, we are recommending a point estimate of 9% reduction in SGA for inclusion in the LiST tool for situations where countries may decide to replace routinely used iron-folate supplements during pregnancy with multiple micronutrients. The sub-group analysis with respect to maternal BMI showed that there was a significant reduction in SGA only in mothers whose mean baseline BMI was ≥ 22 kg/m
2, and the result was non-significant for BMI less than this. It is difficult to explain these findings as it may in part be an artefact caused by the way mothers’ BMI status were determined. It should to be noted that weight gain in pregnancy is not linear, and women who entered early in the study would be classified as thin compared to those who entered the study in the later stages. Another possible explanations could be hypothesized like decreased ability of malnourished mother to utilize multiple nutrients [
39]. However a trial conducted in India on malnourished women did not support this hypothesis [
33]. Certain other factors that need to be considered are the age of the mother, composition of multiple micronutrient supplement and co-supplementation of macronutrient like balanced protein energy supplementation [
40].
There is considerable debate on the potential adverse effects of providing multiple micronutrient supplements during pregnancy with a concern about an increase in neonatal mortality in less developed health systems that have suboptimal maternal care [
41,
42]. There are a few studies that reported mortality beyond the neonatal period with maternal multiple micronutrient supplementation, the data of which has not been pooled. Christian et al. [
26] had reported infant deaths (0 – 3 months) in the multiple micronutrient recipients versus controls (receiving vitamin A only) with a non-significant higher risk in the micronutrient group (RR = 1.07; 95% CI: 0.75 – 1.58). In contrast, Shankar et al. [
21] in the considerably larger SUMMIT trial, reported a statistically significant 18% reduction in early infant mortality (0 – 3 months) in the multiple micronutrient group compared to iron-folate (RR = 0.82; 95% CI: 0.70 – 0.95; P = 0.010), with comparable results for post-neonatal mortality – from 29 days to 90 days after birth (RR = 0.70; 95% CI: 0.55 – 0.89). Our analysis of data from nine studies showed a non-significant effect on neonatal mortality (fixed effects: RR = 1.05; 95% CI: 0.92 – 1.19; random effects: RR = 1.17; 95% CI: 0.95 – 1.44). There was evidence that this risk might be related to the provision of skilled deliveries and standard of care in the health system. The 4 intervention trials conducted in populations where the majority of births occurred at home were associated with significant increase in the risk of neonatal mortality. However, this effect was not seen in the other 5 settings where the majority of births were in facilities. These results are generally consistent with a UNIMAPP trials analysis [
43] with a notable, though non-significant increase in early neonatal mortality (OR = 1.23; 95% CI: 0.96 – 1.59), while there was a 6% non-significant reduction in late neonatal mortality (OR = 0.94; 95% CI: 0.73 – 1.23). These findings suggest that the use of multiple micronutrient supplements in populations to address maternal anemia and reduce the incidence of SGA, 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 [
44,
45]. It must be noted that in most studies evaluated, despite reduction in rates of SGA, there was no reduction in neonatal mortality. A review of the relationship between IUGR and neonatal mortality largely based on observational studies did not distinguish if the effects on neonatal mortality were mediated through IUGR or other concurrent factors [
5]. Applying the CHERG Rules for Evidence Review, we ranked the studies with neonatal mortality outcomes as high grade of evidence. Included studies had some minor limitations, especially insufficient power to detect differences in mortality between the groups. The studies, in general, were representative of the low-income populations. However, caution must be exercised in adapting the findings of multiple micronutrient supplements to programs without further effectiveness trials and robust evaluation in health systems.
It is surprising to note that the participants remained anemic in the included studies, even though they were taking seemingly adequate amounts of supplemental iron. There is no exact reason for this but this has been seen in many effectiveness trials over the years [
40]. An important aspect to consider is the timing of initiation of multiple micronutrient supplements. Starting the supplementation during pregnancy may be too late for many women, especially those with pre-existing anemia [
40]. Berger et al. from Vietnam had reported that the pre-pregnancy use of weekly iron–folic acid is associated with better iron status in the first and second trimesters and with reduced prevalence of LBW compared with pregnant women who only received daily iron–folic acid supplementation during pregnancy [
46]. Another reason could be the amount of iron in the supplement, as higher dosages may be associated with more side effects of diarrhea and thus less absorption [
40,
47]. Similarly one possible reason could be infections, since these were not systematically treated across the trials. World health organization recommends the treatment of infections (especially deworming) for control of anemia along with supplementation of iron [
48].
The observed benefit of multiple micronutrients in the SUMMIT trial [
21] on early infant and post-neonatal mortality opens the interesting possibility that such intrauterine supplements may benefit infants beyond the neonatal period. This is further corroborated by the improved growth beyond infancy noted among recipients of maternal micronutrient supplements in Nepal [
49]. Given the multiple effects of micronutrients in the regulation of body metabolism and growth [
21], these effects are biologically plausible but need further evaluation with a range of additional outcomes, including child development indicators.
While there have also been calls for greater evaluation of potential interactions between various micronutrients [
50], we feel that the field is ripe for well conducted effectiveness trials in health systems. The overall results of maternal micronutrient supplements call for further carefully conducted demonstration projects in health systems with adequate monitoring of pregnancy outcomes. Such studies may also overcome several of the limitations of size and power in the currently available literature. It must be underscored that intake of multivitamins and micronutrients are almost universal during pregnancy in developed countries and among the rich in developing countries [
51]. Policy makers in health systems with adequate provision of skilled care may choose to introduce multiple micronutrient supplements in lieu of routine iron-folate for pregnant women with adequate monitoring and evaluation.
Competing interests
We do not have any financial or non-financial competing interests for this review.
Authors' contributions
Professor Zulfiqar A Bhutta developed the parameters for the review and secured support. Dr Batool Azra Haider and Dr Mohammad Yawar Yakoob undertook the literature search, data extraction and wrote the manuscript under Professor Bhutta’s supervision.