Multiple‐micronutrient supplementation for women during pregnancy
Abstract
Background
Multiple‐micronutrient (MMN) deficiencies often coexist among women of reproductive age in low‐ and middle‐income countries. They are exacerbated in pregnancy due to the increased demands of the developing fetus, leading to potentially adverse effects on the mother and baby. A consensus is yet to be reached regarding the replacement of iron and folic acid supplementation with MMNs. Since the last update of this Cochrane Review in 2017, evidence from several trials has become available. The findings of this review will be critical to inform policy on micronutrient supplementation in pregnancy.
Objectives
To evaluate the benefits of oral multiple‐micronutrient supplementation during pregnancy on maternal, fetal and infant health outcomes.
Search methods
For this 2018 update, on 23 February 2018 we searched Cochrane Pregnancy and Childbirth’s Trials Register, ClinicalTrials.gov, the WHO International Clinical Trials Registry Platform (ICTRP), and reference lists of retrieved studies. We also contacted experts in the field for additional and ongoing trials.
Selection criteria
All prospective randomised controlled trials evaluating MMN supplementation with iron and folic acid during pregnancy and its effects on pregnancy outcomes were eligible, irrespective of language or the publication status of the trials. We included cluster‐randomised trials, but excluded quasi‐randomised trials. Trial reports that were published as abstracts were eligible.
Data collection and analysis
Two review authors independently assessed trials for inclusion and risk of bias, extracted data and checked them for accuracy. We assessed the quality of the evidence using the GRADE approach.
Main results
We identified 21 trials (involving 142,496 women) as eligible for inclusion in this review, but only 20 trials (involving 141,849 women) contributed data. Of these 20 trials, 19 were conducted in low‐ and middle‐income countries and compared MMN supplements with iron and folic acid to iron, with or without folic acid. One trial conducted in the UK compared MMN supplementation with placebo. In total, eight trials were cluster‐randomised.
MMN with iron and folic acid versus iron, with or without folic acid (19 trials)
MMN supplementation probably led to a slight reduction in preterm births (average risk ratio (RR) 0.95, 95% confidence interval (CI) 0.90 to 1.01; 18 trials, 91,425 participants; moderate‐quality evidence), and babies considered small‐for‐gestational age (SGA) (average RR 0.92, 95% CI 0.88 to 0.97; 17 trials; 57,348 participants; moderate‐quality evidence), though the CI for the pooled effect for preterm births just crossed the line of no effect. MMN reduced the number of newborn infants identified as low birthweight (LBW) (average RR 0.88, 95% CI 0.85 to 0.91; 18 trials, 68,801 participants; high‐quality evidence). We did not observe any differences between groups for perinatal mortality (average RR 1.00, 95% CI 0.90 to 1.11; 15 trials, 63,922 participants; high‐quality evidence). MMN supplementation led to slightly fewer stillbirths (average RR 0.95, 95% CI 0.86 to 1.04; 17 trials, 97,927 participants; high‐quality evidence) but, again, the CI for the pooled effect just crossed the line of no effect. MMN supplementation did not have an important effect on neonatal mortality (average RR 1.00, 95% CI 0.89 to 1.12; 14 trials, 80,964 participants; high‐quality evidence). We observed little or no difference between groups for the other maternal and pregnancy outcomes: maternal anaemia in the third trimester (average RR 1.04, 95% CI 0.94 to 1.15; 9 trials, 5912 participants), maternal mortality (average RR 1.06, 95% CI 0.72 to 1.54; 6 trials, 106,275 participants), miscarriage (average RR 0.99, 95% CI 0.94 to 1.04; 12 trials, 100,565 participants), delivery via a caesarean section (average RR 1.13, 95% CI 0.99 to 1.29; 5 trials, 12,836 participants), and congenital anomalies (average RR 1.34, 95% CI 0.25 to 7.12; 2 trials, 1958 participants). However, MMN supplementation probably led to a reduction in very preterm births (average RR 0.81, 95% CI 0.71 to 0.93; 4 trials, 37,701 participants). We were unable to assess a number of prespecified, clinically important outcomes due to insufficient or non‐available data.
When we assessed primary outcomes according to GRADE criteria, the quality of evidence for the review overall was moderate to high. We graded the following outcomes as high quality: LBW, perinatal mortality, stillbirth, and neonatal mortality. The outcomes of preterm birth and SGA we graded as moderate quality; both were downgraded for funnel plot asymmetry, indicating possible publication bias.
We carried out sensitivity analyses excluding trials with high levels of sample attrition (> 20%). We found that results were consistent with the main analyses for all outcomes. We explored heterogeneity through subgroup analyses by maternal height, maternal body mass index (BMI), timing of supplementation, dose of iron, and MMN supplement formulation (UNIMMAP versus non‐UNIMMAP). There was a greater reduction in preterm births for women with low BMI and among those who took non‐UNIMMAP supplements. We also observed subgroup differences for maternal BMI and maternal height for SGA, indicating greater impact among women with greater BMI and height. Though we found that MMN supplementation made little or no difference to perinatal mortality, the analysis demonstrated substantial statistical heterogeneity. We explored this heterogeneity using subgroup analysis and found differences for timing of supplementation, whereby higher impact was observed with later initiation of supplementation. For all other subgroup analyses, the findings were inconclusive.
MMN versus placebo (1 trial)
A single trial in the UK found little or no important effect of MMN supplementation on preterm births, SGA, or LBW but did find a reduction in maternal anaemia in the third trimester (RR 0.66, 95% CI 0.51 to 0.85), when compared to placebo. This trial did not measure our other outcomes.
Authors' conclusions
Our findings suggest a positive impact of MMN supplementation with iron and folic acid on several birth outcomes. MMN supplementation in pregnancy led to a reduction in babies considered LBW, and probably led to a reduction in babies considered SGA. In addition, MMN probably reduced preterm births. No important benefits or harms of MMN supplementation were found for mortality outcomes (stillbirths, perinatal and neonatal mortality). These findings may provide some basis to guide the replacement of iron and folic acid supplements with MMN supplements for pregnant women residing in low‐ and middle‐income countries.
PICOs
Plain language summary
Vitamin and mineral supplements for women during pregnancy
What is the issue?
In low‐ and middle‐income countries, many women have poor diets and are deficient in nutrients and micronutrients that are required for good health. Micronutrients are vitamins and minerals that are needed by the body in very small quantities, but are important for normal functioning, growth and development. During pregnancy, these women often become more deficient because of the need to provide nutrition for the baby too, and this can negatively affect their health, along with the health of the baby.
Why is this important?
Combining multiple micronutrients into one supplement has been suggested as a cost‐effective way to achieve multiple benefits for women during pregnancy. Micronutrient deficiencies are known to interact, and a greater effect may be achieved by multiple supplementation rather than single‐nutrient supplementation. However, interactions could also lead to poor absorption of some of the nutrients. High doses of some nutrients may also cause harm to the mother or her baby.
What evidence did we find?
We searched Cochrane Pregnancy and Childbirth's Trials Register (23 February 2018). This systematic review included 21 trials (involving 142,496 women), but only 20 trials (involving 141,849 women) contributed data. The included trials compared pregnant women who supplemented their diets with multiple micronutrients (including iron and folic acid) with pregnant women who received iron (with or without folic acid) or a placebo. Overall, we found that pregnant women who received multiple‐micronutrient supplementation had fewer babies that were born too small (weighing less than 2500 g), fewer babies who were smaller in size than normal for their gestational age, and fewer births that occurred before week 37 of pregnancy. The evidence for the main outcomes of low birthweight and small‐for‐gestational age was found to be of high quality and moderate quality, respectively.
What does this mean?
These findings, which have been observed elsewhere, may provide a basis to guide the replacement of iron and folic acid supplements with multiple‐micronutrient supplements for pregnant women in low‐ and middle‐income countries.
Authors' conclusions
Summary of findings
| Multiple micronutrients compared to control (iron with or without folic acid) for women during pregnancy | ||||||
| Patient or population: women during pregnancy | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Quality of the evidence | Comments | |
| Risk with control (iron with or without folic acid) | Risk with multiple micronutrients | |||||
| Preterm births | Study population | RR 0.95 | 91,425 | ⊕⊕⊕⊝ | ||
| 197 per 1000 | 188 per 1000 | |||||
| Small‐for‐gestational age | Study population | RR 0.92 | 57,348 | ⊕⊕⊕⊝ | ||
| 337 per 1000 | 310 per 1000 | |||||
| Low birthweight | Study population | RR 0.88 | 68,801 | ⊕⊕⊕⊕ | ||
| 212 per 1000 | 187 per 1000 | |||||
| Perinatal mortality | Study population | RR 1.00 | 63,922 | ⊕⊕⊕⊕ | Raw event and participant data were unavailable for Ramakrishnan 2003 and West 2014, so have not been included in No of participants column. | |
| 39 per 1000 | 39 per 1000 | |||||
| Stillbirths | Study population | RR 0.95 | 97,927 | ⊕⊕⊕⊕ | ||
| 30 per 1000 | 28 per 1000 | |||||
| Neonatal mortality | Study population | RR 1.00 | 80,964 | ⊕⊕⊕⊕ | Raw event and participant data were unavailable for Bhutta 2009a and Fawzi 2007 so have not been included in No of participants column. | |
| 29 per 1000 | 29 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| aStrong evidence of funnel plot asymmetry, indicating possible publication bias. | ||||||
Background
Description of the condition
Micronutrient deficiencies are common among women of reproductive age (15 to 49 years of age), especially those residing in low‐ and middle‐income countries where diets often lack diversity and fortified foods are less available (Black 2013; FAO/WHO 2004). Infections and chronic illness can also contribute to micronutrient deficiencies by directly inhibiting nutrient absorption (Bailey 2015). Micronutrient deficiencies are exacerbated during pregnancy due to increased requirements of the growing fetus, placenta and maternal tissues. An inability to fulfil the increased demands results in potentially adverse effects on the mother and the fetus (Berti 2011). Additionally, there can be sustained intergenerational effects. Maternal malnutrition has been shown to affect both short‐term and long‐term outcomes for the offspring, including growth, neurodevelopment and cognition, and cardiometabolic, pulmonary, and immune function (Gernand 2016).
Up‐to‐date population‐level estimates are largely lacking for individual micronutrients due to measurement and cost challenges associated with collecting these indicators (Gernand 2016). In addition, there are few data that have been disaggregated by age, parity, wealth status and other factors that can influence nutrition throughout pregnancy (Gernand 2016). However, we do know that anaemia due to iron deficiency is one of the most prevalent micronutrient deficiencies globally. According to 2013 estimates, the worldwide prevalence of prenatal iron‐deficiency anaemia was 19.2% (95% confidence interval (CI) 17.1% to 21.5%; Black 2013). Anaemia during pregnancy has been found to be associated with increased risk of maternal mortality, perinatal mortality, and infants with low birthweight (LBW) (Allen 2001; Christian 2010; Haider 2013; Murray‐Kolb 2013). Vitamin A is another important nutrient that, when deficient, can lead to night blindness. According to global estimates for the time period between 1995 and 2005, vitamin A deficiency, measured using night blindness and low serum retinol levels, affected 9.8 million (95% CI 8.7 to 10.8 million), and 19.1 million pregnant women (95% CI 9.30 to 29.0 million), respectively. This corresponds to 15.3% (95% CI 6.0% to 24.6%), of pregnant women being deficient in vitamin A (Black 2013). Deficiency of vitamin A has been associated with poor birth and mortality outcomes; however, supplementation with vitamin A during pregnancy has demonstrated no beneficial effect on these outcomes (McCauley 2015), but has been shown to reduce the risk of maternal anaemia, infection, and night blindness (McCauley 2015).
In the past decade, deficiency of vitamin D has also emerged as an important nutritional problem with high prevalence being reported in high‐income, as well as low‐income, populations (Datta 2002; Ginde 2010; Sachan 2005). Iodine deficiency, often measured by urinary iodine, is also common among pregnant women. The median urinary iodine level in a nationally representative sample of pregnant women in Nepal was reported to be 134 mcg/L (Benoist 2008), indicating insufficient iodine intake (Andersson 2007). Severe iodine deficiency during pregnancy results in pregnancy loss, mental retardation and cretinism (Dunn 1993). Although severe deficiency is now rare, mild to moderate deficiency continues to be a problem (Andersson 2007).
Deficiencies of other micronutrients are also common among pregnant women. According to the 2012 estimates, around 17% of the world’s population have reduced dietary intake of zinc (Wessells 2012). Zinc deficiency has been associated with complications of pregnancy and delivery such as pre‐eclampsia, premature rupture of membranes, and congenital abnormalities (Black 2001; Caulfield 1998). However, a review of trials of zinc supplementation showed a reduction in the risk of preterm birth only (Hess 2009; Ota 2015). Folic acid deficiency can lead to haematological consequences and congenital malformations; however, association with other birth outcomes is equivocal (Black 2001; De‐Regil 2010). Concurrent deficiencies that could include vitamins A, D, E, riboflavin, B6, B12, folic acid, iron, and zinc have also been reported in studies conducted among pregnant women (Jiang 2005; Pathak 2004). Deficiencies in other minerals such as magnesium, selenium, copper and calcium may also potentially be associated with complications of pregnancy, childbirth or fetal development (Black 2001).
Description of the intervention
The World Health Organization (WHO) currently recommends iron and folic acid supplementation for women during pregnancy as part of routine antenatal care (WHO 2012). The recommended dose of iron ranges from 30 mg to 60 mg. In areas where anaemia is a severe public health problem, defined as a prevalence of 40% or higher, a daily dose of 60 mg of iron is preferred. The standard dose of 60 mg of iron was first recommended in 1959 and was based on maternal requirements during pregnancy (WHO 1959). Despite its provision as part of national antenatal care programmes for the last few decades in most low‐ and middle‐income countries, compliance with the supplement is low. Gastrointestinal side‐effects including constipation, nausea, vomiting, and diarrhoea are the most common complaints among women consuming high doses of iron (Oriji 2011; Seck 2008).
Supplementation with iron and folic acid during pregnancy has been found to be associated with reduction in the risk of maternal anaemia and infants with LBW (Haider 2013; Pena‐Rosas 2015). To overcome other possible maternal micronutrient deficiencies, the United Nations Children's Fund (UNICEF), United Nations University (UNU) and the WHO, in 1999, agreed on the composition of a proposed multiple‐micronutrient (MMN) tablet (UNICEF 1999). This UNIMMAP tablet provides one recommended daily allowance of vitamin A, vitamin B1, vitamin B2, niacin, vitamin B6, vitamin B12, folic acid, vitamin C, vitamin D, vitamin E, copper, selenium and iodine with 30 mg of iron and 15 mg of zinc for pregnant women. In contrast to the WHO recommendation, a lower dose of iron was recommended as the absorption of iron was expected to be enhanced due to vitamin C, vitamin A, and riboflavin, and given that the majority of pregnant women suffer from mild anaemia and the potential side‐effects associated with higher doses of iron.
How the intervention might work
Vitamins and minerals play critical roles in cellular metabolism, growth and maintenance of normal functioning of the human body. They are also important in many enzymatic processes, signal transduction and transcription pathways (McArdle 1999; WHO 2004). Recent studies have suggested a possible benefit of multiple micronutrient supplementation for improving pregnancy outcomes through placental function, including modulation of inflammation and oxidative stress and vascular function (Owens 2015; Richard 2017). Deficiencies of these micronutrients rarely exist in isolation. Additionally, because of their role at various levels in the biological pathways, it is difficult to assign a clinical or pre‐clinical condition to the deficiency of a single micronutrient (McArdle 1999). Micronutrient deficiencies are also known to interact. Combining MMN in a single delivery mechanism has been suggested as a cost‐effective way to achieve multiple benefits.
Why it is important to do this review
The interest of the global research community in eliminating micronutrient deficiencies stems from their significant negative impact on the health of women and infants. The health effects during the fetal life may also have consequences later as an adult. Several trials have demonstrated that supplementation with MMN during pregnancy reduces the risk of micronutrient deficiencies (Haider 2012). Findings from individual trials regarding the benefit on other maternal and pregnancy outcomes are inconsistent, as individual studies may not have statistical power to evaluate effects on these outcomes. Several meta‐analyses have systematically reviewed and synthesised the evidence of the effect of supplementation with multiple micronutrients, with the first such synthesis of evidence being an earlier version of this Cochrane Review (Bhutta 2012; Haider 2006; Haider 2011; Haider 2012; Kawai 2011; Ramakrishnan 2012). On the basis of the evidence, supplementation with MMN during pregnancy has been recommended (Bhutta 2008; Bhutta 2013). However, a consensus is yet to be reached regarding the replacement of iron and folic acid supplementation with MMN. Since the last update of this Cochrane Review (Haider 2017), evidence from a few large trials has recently been made available, inclusion of which is critical to inform global policy.
This review updates a previously published Cochrane Review on MMN supplementation during pregnancy that had demonstrated positive effect of supplementation on birth outcomes (Haider 2017). The effects of supplementation with individual micronutrients during pregnancy have been evaluated in other Cochrane Reviews. The effect of MMN supplementation in HIV‐infected pregnant women has been evaluated in another Cochrane Review (Siegfried 2012).
Objectives
To evaluate the benefits of oral multiple‐micronutrient supplementation during pregnancy on maternal, fetal and infant health outcomes.
Methods
Criteria for considering studies for this review
Types of studies
All prospective randomised controlled trials evaluating multiple micronutrient (MMN) supplementation during pregnancy and its effects on pregnancy outcomes were eligible, irrespective of language or publication status of the trials. Trial reports that were published as abstracts were eligible for inclusion. We included cluster‐randomised trials, but excluded quasi‐randomised trials.
Types of participants
Pregnant women. There was no limit on the length of gestation at the time of enrolment in the study. We excluded HIV‐infected pregnant women from the review, as this population is at a greater risk of nutritional disorders compared to uninfected women. We also excluded studies recruiting women at high risk of nutritional disorders for other reasons. Another Cochrane Review has evaluated the effect of MMN supplementation in HIV‐infected pregnant women (Siegfried 2012).
Types of interventions
Since WHO recommends use of iron and folic acid supplementation in women during pregnancy as a part of routine antenatal care, we evaluated the effect of MMN supplementation with iron and folic acid in pregnant women versus supplementation with iron, with or without folic acid. We also included trials comparing the outcomes of providing pregnant women with MMN supplements with iron and folic acid compared to placebo. The composition of MMN supplement by trial can be found in Table 1.
| Study ID | Iron | Folic acid | Vit A | Beta‐carotene | Vit C | Vit D | Vit E | Vit B1 (thiamine) | Vit B2 (riboflavin) | Vit B3 (niacin) | Vit B5 (pantothenic acid) | Vit B6 | Vit B12 | Vit K | Copper | Selenium | Zinc | Iodine | Magnesium | Calcium | Phosphorus | Biotin | Potassium | Manganese |
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | |||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | |||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ||||||||||||||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | |||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
We excluded trials that used fewer than three micronutrients in the intervention group, regardless of their outcomes. There were no limits on the duration of supplementation.
We included the following specific comparisons in the review.
-
Multiple micronutrients with iron and folic acid versus control (iron with or without folic acid)
-
Multiple micronutrients with iron and folic acid versus control (placebo)
The review focuses on daily oral supplements. One trial (Biggs 2010), provided MMN twice weekly; we excluded data from this trial from the main results, but included them in the subgroup analysis for dose of iron to examine whether there are any notable differences when women are provided with a lower dose of iron throughout pregnancy (Biggs 2010 provided women with 120 mg per week compared to 210 mg or 420 mg per week in all other trials). We excluded trials examining parenteral MMN or food fortification with MMN.
Types of outcome measures
Primary outcomes
-
Preterm births (births before 37 weeks of gestation)
-
Small‐for‐gestational age (SGA) (as defined by the authors of the trials)
-
Low birthweight (LBW) (birthweight less than 2500 g)
-
Perinatal mortality
-
Stillbirths
-
Neonatal mortality
Secondary outcomes
-
Maternal anaemia (third trimester haemoglobin (Hb) < 110 g/L)
-
Maternal mortality
-
Miscarriage (loss of pregnancy before 28 weeks of gestation)
-
Premature rupture of membranes
-
Pre‐eclampsia
-
Mode of delivery: caesarean section (not prespecified)
-
Macrosomia (not prespecified)
-
Placental abruption
-
Very preterm births (births before 34 weeks of gestation)
-
Neurodevelopmental delay (assessed using Bayley Scale of Infant Development (BSID) at six and 12 months of age)
-
Nutritional status of children (stunting, wasting and underweight at six, 12 and 24 months of age)
-
Cost of supplementation
-
Side‐effects of MMN supplements
-
Congenital anomalies (including neural tube defects)
-
Maternal well‐being or satisfaction
Search methods for identification of studies
The following search methods section of this review is based on a standard template used by Cochrane Pregnancy and Childbirth.
Electronic searches
For this update we searched Cochrane Pregnancy and Childbirth’s Trials Register by contacting their Information Specialist (23 February 2018).
The Register is a database containing over 25,000 reports of controlled trials in the field of pregnancy and childbirth. It represents over 30 years of searching. For full current search methods used to populate Pregnancy and Childbirth’s Trials Register including the detailed search strategies for CENTRAL, MEDLINE, Embase and CINAHL; the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service, please follow this link
Briefly, Cochrane Pregnancy and Childbirth’s Trials Register is maintained by their Information Specialist and contains trials identified from:
-
monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);
-
weekly searches of MEDLINE (Ovid);
-
weekly searches of Embase (Ovid);
-
monthly searches of CINAHL (EBSCO);
-
handsearches of 30 journals and the proceedings of major conferences;
-
weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.
Search results are screened by two people and the full text of all relevant trial reports identified through the searching activities described above is reviewed. Based on the intervention described, each trial report is assigned a number that corresponds to a specific Pregnancy and Childbirth review topic (or topics), and is then added to the Register. The Information Specialist searches the Register for each review using this topic number rather than keywords. This results in a more specific search set that has been fully accounted for in the relevant review section (Included studies; Excluded studies; Studies awaiting classification; Ongoing studies).
In addition, we searched ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP) for unpublished, planned and ongoing trial reports (24 February 2018) using the search methods detailed in Appendix 1.
Searching other resources
We searched reference lists of retrieved articles and key reviews. We contacted experts in the field for additional and ongoing trials.
We did not apply any language or date restrictions.
Data collection and analysis
For methods used in the previous version of this review, see Haider 2017.
For this update, we used the following methods for assessing the reports that were identified as a result of the updated search.
The following methods section of this review is based on a standard template used by Cochrane Pregnancy and Childbirth.
Selection of studies
Two review authors independently assessed for inclusion all the potential studies identified as a result of the search strategy. We resolved any disagreement through discussion.
Data extraction and management
We designed a form to extract data. For eligible studies, two review authors extracted the data using the agreed form. We resolved discrepancies through discussion or, if required, we consulted a third review author. We entered data into Review Manager 5 software (Review Manager 2014), and checked them for accuracy.
When information regarding any of the above was unclear, we planned to contact authors of the original reports to provide further details.
Given some of the changes and data edits in the previous versions of this review, for this update, we re‐extracted data for all primary and secondary outcomes for all included studies from the outset, not just those found from the most recent search.
Assessment of risk of bias in included studies
Two review authors independently assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). We resolved any disagreement by discussion or by involving a third assessor.
(1) Random sequence generation (checking for possible selection bias)
We described for each included study the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.
We assessed the method as:
-
low risk of bias (any truly random process, e.g. random number table; computer random number generator);
-
high risk of bias (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number);
-
unclear risk of bias.
(2) Allocation concealment (checking for possible selection bias)
We described for each included study the method used to conceal allocation to interventions prior to assignment and assessed whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment.
We assessed the methods as:
-
low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
-
high risk of bias (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);
-
unclear risk of bias.
(3.1) Blinding of participants and personnel (checking for possible performance bias)
We described for each included study the methods used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. We considered that studies were at low risk of bias if they were blinded, or if we judged that the lack of blinding unlikely to affect results. We assessed blinding separately for different outcomes or classes of outcomes.
We assessed the methods as:
-
low, high or unclear risk of bias for participants;
-
low, high or unclear risk of bias for personnel.
(3.2) Blinding of outcome assessment (checking for possible detection bias)
We described for each included study the methods used, if any, to blind outcome assessors from knowledge of which intervention a participant received. We assessed blinding separately for different outcomes or classes of outcomes.
We assessed the methods used to blind outcome assessment as:
-
low, high or unclear risk of bias.
(4) Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)
We described for each included study, and for each outcome or class of outcomes, the completeness of data including attrition and exclusions from the analysis. We stated whether attrition and exclusions were reported and the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes. Where sufficient information was reported, or could be supplied by the trial authors, we planned to re‐include missing data in the analyses which we undertook.
We assessed methods as:
-
low risk of bias (e.g. no missing outcome data; missing outcome data balanced across groups);
-
high risk of bias (e.g. numbers or reasons for missing data imbalanced across groups; ‘as treated’ analysis done with substantial departure of intervention received from that assigned at randomisation);
-
unclear risk of bias.
(5) Selective reporting (checking for reporting bias)
We described for each included study how we investigated the possibility of selective outcome reporting bias and what we found.
We assessed the methods as:
-
low risk of bias (where it is clear that all of the study’s pre‐specified outcomes and all expected outcomes of interest to the review have been reported);
-
high risk of bias (where not all the study’s pre‐specified outcomes have been reported; one or more reported primary outcomes were not pre‐specified; outcomes of interest are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported);
-
unclear risk of bias.
(6) Other bias (checking for bias due to problems not covered by (1) to (5) above)
We described for each included study any important concerns we had about other possible sources of bias.
(7) Overall risk of bias
We made explicit judgements about whether studies were at high risk of bias, according to the criteria given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017). With reference to (1) to (6) above, we planned to assess the likely magnitude and direction of the bias and whether we considered its likelihood to impact the findings. For cluster‐randomised trials, we carefully considered recruitment bias, baseline imbalance, loss of clusters, incorrect analysis, and comparability with individually randomised trials (Higgins 2011a). We explored the impact of attrition bias through undertaking sensitivity analyses ‐ see Sensitivity analysis.
Assessment of the quality of the evidence using the GRADE approach
For this update, we assessed the quality of the evidence using the GRADE approach as outlined in the GRADE Handbook (Schünemann 2013). We assessed the quality of the body of evidence relating to the following outcomes for the comparison of MMN versus iron and folic acid supplements:
-
Preterm births
-
SGA
-
LBW
-
Perinatal mortality
-
Stillbirths
-
Neonatal mortality
We used the GRADEpro Guideline Development Tool (GRADEpro GDT 2015), to import data from Review Manager 5.3 (Review Manager 2014), in order to create a 'Summary of findings' table. We produced a summary of the intervention effect and a measure of quality for each of the above outcomes using the GRADE approach. The GRADE approach uses five considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of the body of evidence for each outcome. The evidence was downgraded from 'high quality' by one level for serious limitations (or by two levels for very serious), depending on assessments for risk of bias, indirectness of evidence, serious inconsistency, imprecision of effect estimates or potential publication bias. We did not downgrade evidence for heterogeneity with an I² statistic value of less than 50% (Deeks 2017; Higgins 2003).
Measures of treatment effect
Dichotomous data
For dichotomous data, we presented results as summary risk ratio (RR) with 95% confidence intervals (CI).
Continuous data
We used the mean difference (MD) if outcomes were measured in the same way between trials. In future updates as appropriate, we plan to use the standardised mean difference to combine trials that measure the same outcome, but use different methods.
If both adjusted and unadjusted effect estimates were provided in the primary report, we utilised the adjusted estimate. If only event and participant raw data were provided, we used these to calculate a relevant effect estimate (RR or MD) using the Review Manager 5 calculator function (Review Manager 2014). Where there were significant discrepancies between our re‐extracted data and what had been reported in previous versions of this review, we contacted trial investigators for clarification.
Unit of analysis issues
Cluster‐randomised trials
We included cluster‐randomised trials (Christian 2003; Bhutta 2009a; Biggs 2010; SUMMIT 2008; Sunawang 2009; West 2014; Zagre 2007; Zeng 2008), in the analyses along with individually‐randomised trials. We extracted cluster‐adjusted effect estimates with their confidence intervals, which we analysed along with individually‐randomised trials using the generic inverse variance method. If effect estimates were not cluster‐adjusted, or if number of events was used to calculate the effect estimate for a given outcome, then we used the reported intracluster correlation coefficient (ICC) and average cluster size (M) to determine the design effect for a trial (1+(M‐1)*ICC) (Higgins 2011a). Some trials reported the design effect, which we then applied to the number of events and sample size for dichotomous outcomes, and sample size only for continuous outcomes, to reduce cluster‐randomised trial data to their effective sample size.
We had to adjust all reported estimates from each cluster‐randomised trial, with the exception of Bhutta 2009a (all outcomes), Christian 2003 (all outcomes), SUMMIT 2008 (SGA, LBW, perinatal mortality, neonatal mortality, stillbirth, and maternal mortality), and West 2014 (preterm birth, SGA, LBW, perinatal mortality, neonatal mortality, and stillbirth). Details on reported ICCs and design effect by trial can be found in Table 2.
| Trial | Outcome(s) | Reported design effect | Reported ICC | Calculated M | Calculated design effecta |
| Preterm birth, miscarriage | 1.2 | ‐ | ‐ | ‐ | |
| All | 1.2 | ‐ | ‐ | ‐ | |
| Maternal anaemia, miscarriage, maternal mortality, very preterm birth | 1.15 | ‐ | ‐ | ‐ | |
| All | 1.2 | ‐ | ‐ | ‐ | |
| Preterm birth, very preterm birth | ‐ | 0.02 | 11 | 1.2 | |
| LBW | ‐ | 0.03 | 11 | 1.3 | |
| All other outcomes | ‐ | 0.03b | 11 | 1.3 | |
| Maternal anaemia | ‐ | 0.03 | 12.1 | 1.3 | |
| All other outcomes | ‐ | 0 | 12.1 | 1.0 | |
| ICC: intracluster correlation coefficient; M: average cluster size | |||||
aDesign effect = 1 + (M‐1)*ICC.
bZeng 2008 reported ICCs specific to gestational age and birthweight. For all other outcomes, we used the more conservative of the two ICCs (0.03) to calculate the design effect.
Trials with multiple intervention groups
For the majority of trials with multiple intervention groups, we selected one pair of interventions and excluded the others. This is one approach recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a).
We included trials with more than two intervention groups in the analysis after selecting the comparison groups (intervention and control groups) that satisfied the 'types of intervention' criterion and were relevant to the review. For Christian 2003, we included data from group 4 (MMN group with iron and folic acid) versus group 2 (control group iron with or without folic acid). We did not include groups 1 (folic acid with vitamin A) and 5 (vitamin A only), since they did not satisfy the inclusion criterion of the review. Further, group 3 (iron, folic acid, vitamin A, and zinc) did not include the majority of micronutrients being considered for inclusion in a MMN supplement for pregnant women and was also not comparable to the UNIMMAP formulation proposed by UNICEF, UNU, and WHO. For Kaestel 2005, we included group 1 (MMN with iron and folic acid) and group 3 (iron and folic acid) in the review. We selected the group with one recommended daily allowance, since the MMN supplement in group 1 was comparable to the UNIMMAP formulation. For Lui 2013, data from group 3 (MMN with iron and folic acid) versus group 2 (iron and folic acid) fitted the types of intervention criterion of the review and we included them in the analyses. Similarly, we included data for the comparison of groups 3 (MMN with iron and folic acid) versus 2 (iron and folic acid) for Zeng 2008. Group 1 in both Lui 2013 and Zeng 2008 had received folic acid only and did not satisfy the control definition of the review.
For Moore 2009, we included data from group 1 (iron and folic acid) versus group 2 (MMN with iron and folic acid). We excluded data from groups 3 (protein‐energy plus iron and folic acid) and 4 (protein‐energy plus MMN with iron and folic acid) due to the provision of a food‐based supplement in both groups. For Biggs 2010, we included data from groups 1 (daily iron and folic acid) versus 3 (twice weekly MMN with iron and folic acid), though only for the purpose of the subgroup analysis looking at dose of iron. We excluded group 2 because they provided iron and folic acid twice weekly as opposed to daily.
If more than two intervention groups had met the eligibility criteria, we would have combined groups to create a single pair‐wise comparison, as recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a). We used this approach for Tofail 2008, where we combined early invitation and usual invitation to food groups to compare MMN with iron and folic acid to iron (30 mg) and folic acid.
Dealing with missing data
For included studies, we noted levels of attrition. We assessed the impact of including studies with high levels of missing data (> 20%) in the overall assessment of treatment effect by using sensitivity analysis. For all outcomes, we carried out analyses, as far as possible, on an intention‐to‐treat basis, that is, we attempted to include all participants randomised to each group in the analyses. The denominator for each outcome in each trial was the number randomised minus any participants whose outcomes were known to be missing.
Assessment of heterogeneity
We assessed statistical heterogeneity in each meta‐analysis using the Tau², I² (Higgins 2003), and Chi² statistics. We regarded heterogeneity as substantial if an I² statistic was greater than 30% and either a Tau² was greater than zero, or there was a low P value (less than 0.10) in the Chi² test for heterogeneity. If we identified substantial heterogeneity (above 30%), we explored it by pre‐specified subgroup analysis (Deeks 2017).
Assessment of reporting biases
Where there are 10 or more studies in the meta‐analysis, we investigated reporting biases (such as publication bias) using funnel plots. We assessed funnel plot asymmetry visually (Sterne 2017).
Data synthesis
We carried out statistical analysis using Review Manager 5 software (Review Manager 2014). We pooled together data from individually randomised and cluster‐randomised trials during meta‐analysis, using the generic inverse variance method, after appropriate adjustment of estimates from cluster‐randomised trials, as noted above. In response to feedback received for the previous version of this review, we conducted all analyses using a random‐effects model, given the clinical heterogeneity amongst the included trials. We treated the random‐effects summary as the average of the range of possible treatment effects, and we discussed the clinical implications of treatment effects differing between trials. If the average treatment effect was not clinically meaningful, we did not combine trials. We have presented the results as the average treatment effect with 95% confidence intervals, and the estimates of Tau² and I² statistic.
Subgroup analysis and investigation of heterogeneity
If we identified substantial heterogeneity at the outcome‐level, we investigated it using subgroup analyses. We considered whether an overall summary was meaningful, and if it was, we used random‐effects analysis to produce it.
We carried out the following subgroup analyses.
-
Timing of supplementation (categorised as either before or after 20 weeks of gestation)
-
Dose of iron in the MMN and control supplements (*15 to 20 mg versus 30 mg versus 60 mg of iron)
-
Baseline nutritional status of the mother, including body mass index (BMI < 20 kg/m2 versus ≥ 20 kg/m2) and height (< 154.9 cm versus ≥ 154.9 cm)
-
UNIMMAP versus **non‐UNIMMAP MMN supplement formulation
*We used the 15 to 20 mg dosage range to allow us to incorporate the trial that provided 60 mg of iron to women twice weekly, which equates to about 17 mg iron daily.
**Variations on the UNIMMAP formulation (e.g. varying concentrations of micronutrients) would be included in the non‐UNIMMAP group.
We assessed subgroup differences by interaction tests available within Review Manager 5 (Review Manager 2014). We reported the results of subgroup analyses quoting the Chi² statistic and P value, and the interaction test I² value (Deeks 2017; Higgins 2003).
Sensitivity analysis
We carried out sensitivity analyses for all outcomes to explore the effect of risk of bias, as assessed by high attrition rates. We excluded trials at high risk of attrition bias (> 20% loss to follow up) from each analysis in order to assess whether this exclusion affected the overall result.
Results
Description of studies
Results of the search
See: Figure 1
Study flow diagram
For this 2018 updated search, we retrieved and assessed 130 trial reports. We also reassessed the two studies (three reports) awaiting classification in the previous version of the review, and we checked the progress of the seven ongoing studies (14 reports).
We included four new trials in this update (Ashorn 2010; Biggs 2010; Dewey 2009; Moore 2009). We also added 30 new reports to trials already included.
We identified a total of 21 trials (involving 142,496 women) as eligible, of which 20 trials (involving 141,849 number of women) contributed data to the review.
We excluded 23 new studies (39 reports) and added 24 reports to studies that we had previously excluded. We also excluded two trials that we had included in the previous update (Hininger 2004; Theobald 1937), because the MMN supplement did not contain iron.
One study is awaiting classification (Gathwala 2012), due to missing group denominators. Two new studies (three reports) are ongoing (NCT03287882; Sumarmi 2015), in addition to NCT02190565 and Tu 2013 that have remained ongoing since the previous update. See Characteristics of ongoing studies for more information.
Included studies
We identified a total of 21 trials (involving 142,496 women) as eligible for inclusion in this review. Of these, one study (Sood 1975), presented data in a format that precluded its inclusion. Hence, this study did not contribute data to the analyses. A total of 141,849 women participated in the remaining 20 included trials (Ashorn 2010; Bhutta 2009a; Biggs 2010; Brough 2010; Christian 2003; Dewey 2009; Fawzi 2007; Friis 2004; Kaestel 2005; Lui 2013; Moore 2009; Osrin 2005; Ramakrishnan 2003; Roberfroid 2008; SUMMIT 2008; Sunawang 2009; Tofail 2008; West 2014; Zagre 2007; Zeng 2008), of which eight were cluster‐randomised (Bhutta 2009a; Biggs 2010; Christian 2003; SUMMIT 2008; Sunawang 2009; West 2014; Zagre 2007; Zeng 2008). One trial was conducted in a high‐income country (Brough 2010). Most of the outcomes were defined in the same way across different trials except for one trial that used different cut‐offs for stillbirth and very preterm birth (West 2014), and three trials that used different cut‐offs for miscarriage (Osrin 2005; West 2014; Zagre 2007). See Characteristics of included studies table for further details of included studies.
All trials reported sources of funding, with the exception of Sood 1975.
Seventeen trials included a statement of disclosure as to whether or not there was a potential conflict of interest related to the study (Ashorn 2010; Biggs 2010; Brough 2010; Christian 2003; Dewey 2009; Fawzi 2007; Friis 2004; Kaestel 2005; Lui 2013; Moore 2009; Osrin 2005; Ramakrishnan 2003; Roberfroid 2008; SUMMIT 2008; Tofail 2008; West 2014; Zeng 2008). Of these, there were 12 trials where all authors declared no conflict of interest (Biggs 2010; Brough 2010; Christian 2003; Fawzi 2007; Friis 2004; Kaestel 2005; Lui 2013; Moore 2009; Ramakrishnan 2003; Roberfroid 2008; SUMMIT 2008; Tofail 2008), and five studies where an author or several authors had indicated a conflict of interest (Ashorn 2010; Dewey 2009; Osrin 2005; West 2014; Zeng 2008). The remaining four studies did not provide a statement of disclosure, and therefore we could not assess potential conflicts of interest (Bhutta 2009a; Sood 1975; Sunawang 2009; Zagre 2007).
Participants
The 20 trials contributing data to the analyses included 141,849 pregnant women at varying gestational stages, ranging from early pregnancy to 36 weeks of gestation. Pregnant women with a haemoglobin (Hb) concentration of less than 80 g/L, with a serious medical condition or a complication of pregnancy such as cardiac disease, pneumonia and threatened abortion were not eligible for inclusion in the trials. Two trials (Ashorn 2010; Friis 2004), included a subgroup of pregnant women who were HIV‐1‐infected, but the data for these subgroups were excluded from the review. Baseline characteristics of the participants in the intervention and the control groups were comparable in the included trials except for minor differences in five trials (Christian 2003; Friis 2004; Ramakrishnan 2003; Roberfroid 2008; Zagre 2007). In Friis 2004, a higher proportion of primigravidae was found in the placebo group. In Ramakrishnan 2003, there was a higher proportion of single mothers and a lower mean BMI in the intervention group. In Christian 2003, more participants in the control group belonged to a specific ethnic background and owned land. In Roberfroid 2008, the Hb level was lower in the intervention group and the BMI was lower in the control group. In Zagre 2007, the intervention group had more households and preventive measures against malaria, whereas the placebo group had less education and more poverty.
Intervention
Seventeen trials assessed MMN supplementation versus control; iron with or without folic acid (Ashorn 2010; Bhutta 2009a; Biggs 2010; Christian 2003; Dewey 2009; Kaestel 2005; Lui 2013; Moore 2009; Osrin 2005; Ramakrishnan 2003; Roberfroid 2008; SUMMIT 2008; Sunawang 2009; Tofail 2008; West 2014; Zagre 2007; Zeng 2008). Two trials had a component of nutritional education along with MMN supplementation (Bhutta 2009a; Zagre 2007), and one trial had a food supplement co‐intervention along with MMN supplementation (Tofail 2008). Three trials assessed MMN supplementation against a placebo (Brough 2010; Fawzi 2007; Friis 2004); however, in Fawzi 2007 and Friis 2004, all participants received iron and folic acid supplements. In Brough 2010, participants not taking folic acid received recommendations to take it daily.
The composition of the MMN supplement was different in all included trials. Eighteen trials included iron and folic acid in the MMN supplement (Ashorn 2010; Bhutta 2009a; Biggs 2010; Brough 2010; Christian 2003; Dewey 2009; Kaestel 2005; Lui 2013; Moore 2009; Osrin 2005; Ramakrishnan 2003; Roberfroid 2008; SUMMIT 2008; Sunawang 2009; Tofail 2008; West 2014; Zagre 2007; Zeng 2008). All supplements were given orally to the pregnant women throughout pregnancy from the time of enrolment, except for one trial where supplementation was started when the participants reached 14 weeks of gestation (Tofail 2008). The duration of supplementation varied because the time of enrolment differed across the trials. Six trials enrolled participants in the first trimester of pregnancy (Brough 2010; Christian 2003; Ramakrishnan 2003; Tofail 2008; West 2014; Zagre 2007). Two trials enrolled participants with a gestation of less than 16 weeks (Bhutta 2009a; Biggs 2010), four trials less than 20 weeks (Ashorn 2010; Dewey 2009; Lui 2013; Moore 2009), and one trial less than 28 weeks (Zeng 2008). Two trials enrolled participants in the second trimester (Osrin 2005; Sunawang 2009), one trial enrolled women in both the second and third trimester (Friis 2004), and two trials enrolled women who were less than 37 weeks of gestation (Fawzi 2007; Kaestel 2005). Two trials enrolled pregnant women irrespective of gestational age (Roberfroid 2008; SUMMIT 2008). Supplementation was given until delivery in 11 of the included trials (Bhutta 2009a; Brough 2010; Friis 2004; Kaestel 2005; Lui 2013; Moore 2009; Osrin 2005; Ramakrishnan 2003; Tofail 2008; Zagre 2007; Zeng 2008). Supplementation continued until four weeks after delivery in Sunawang 2009, six weeks after delivery in Fawzi 2007, 12 weeks after delivery in five trials (Biggs 2010; Christian 2003; Roberfroid 2008; SUMMIT 2008; West 2014), and for five weeks after a stillbirth or miscarriage in Christian 2003, and 24 weeks after delivery in two trials (Ashorn 2010; Biggs 2010). The frequency of MMN supplementation in all included trials was once daily, except for one trial where supplementation was provided six days a week (Ramakrishnan 2003), and another trial where supplementation was twice weekly (Biggs 2010).
Excluded studies
We excluded 117 trials from the review. Briefly, 32 trials evaluated the effects of a single or two micronutrients or compounds (Beazley 2002; Bergmann 2006; Carrasco 1962; Caulfield 1999; Chames 2002; Goldenberg 1995; Gopalan 2004; Hillman 1963; Holly 1955; Hossain 2014; Hunt 1983; Hunt 1985; Iannotti 2008; Lucia 2007; Ma 2008; Malvasi 2014; Marya 1987; Mathan 1979; Merialdi 1999; Muslimatun 2001; NCT01795131; Ochoa‐Brust 2007; Robertson 1991; Sachdeva 1993; Sagaonkar 2009; Salzano 2001; Schmidt 2001; Semba 2000; Suharno 1993; Suprapto 2002; Tanumihardjo 2002; Zavaleta 2000). Fourteen trials did not satisfy the trial design criteria (ACTRN12616001449426; Aguayo 2005; Biswas 1984; Kubik 2004; Kynast 1986; Itam 2003; Menon 1962; Patimah 2013; Park 1999; People's League 1946; Pezzack 2014; Sun 2010; Thauvin 1992; Wijaya‐Erhardt 2014), and six trials studied HIV‐positive women (Arsenault 2010; Fawzi 1998; Khavari 2014; Merchant 2005; Olofin 2014; Webb 2009), and hence we excluded them from the review. Six trials gave MMN supplements to both groups of participants (Ahn 2006; Asemi 2014; Dawson 1987; Dawson 1998; Magon 2014; Taghizadeh 2014). Czeizel 1996, ICMR 2000, Cooper 2012, Gunaratna 2015, Khulan 2012 and Otoluwa 2017 evaluated supplementation in the periconceptional period, and Nguyen 2012 evaluated the effect of pre‐conception supplementation. An 2001, Guldholt 1991, Graham 2007, Fleming 1986, and Wibowo 2012 assessed different doses of micronutrients; Agarwal 2012 evaluated different durations of the same micronutrients, while Feyi‐Waboso 2005 and Nwagha 2010 evaluated parenteral infusion or injection. Ramirez‐Velez 2011 did not contain an adequate comparison arm (they provided calcium in addition to ferrous sulphate and folic acid). Godfrey 2017 evaluated a drink enriched with micronutrients, probiotics and myo‐inositol, Callaghan‐Gillespie 2017 evaluated a food supplement (corn‐soy blend) enriched with micronutrients, Fall 2006 evaluated a micronutrient‐rich snack, Huang 2017 evaluated different maternal milk preparations, Nakano 2010 assessed the effect of chlorella tablets and Ling 1996 evaluated a herbal tonic. Li 2014 evaluated the effect of supplementation with folic acid and milk. Six excluded trials assessed the effect of fortification with MMN (Dieckmann 1944; Janmohamed 2016; Jarvenpaa 2007; Kureishy 2017; Tatala 2002; Vadillo‐Ortega 2011). Twelve trials included high‐risk women (Asemi 2015; Azami 2016; Christian 2003; Devi 2017; Gupta 2007; IRCT2015041321736N1; IRCT201704225623N109; ISRCTN83599025; NCT02802566; NCT02959125; Nossier 2015; Rumiris 2006). We excluded eight trials because they evaluated different forms of supplementation such as powder, tablet or spread (Choudhury 2012; Hambidge 2014; Huynh 2017; Lanou 2014; Young 2010); balanced energy protein supplementation (Huybregts 2009); weekly food provision (Wijaya‐Erhardt 2011); or polyunsaturated fatty acids fortification in milk fortified with MMN (Mardones 2007). The cohort of an included trial (Tofail 2008), was later randomised to breastfeeding counselling or standard care groups measuring the impact on postnatal growth in children (Kabir 2009), and hence we excluded it. We excluded Leroy 2010 because it compared a traditional food‐assisted maternal and child health and nutrition (MCHN) programme versus a newly designed approach to prevent malnutrition in children; Nguyen 2017 because it compared a nutrition‐focused MNCH programme with a standard MNCH (antenatal care with standard nutrition counselling) programme. We excluded one abstract of a trial because it was a trial in women with alcohol consumption during pregnancy (Kable 2012). We excluded Coles 2015 because of the trial's quasi‐randomised method of allocation to intervention and control arms. We excluded Dewey 2012 and Fernald 2016 because they evaluated the effects of lipid‐based nutrient supplements.
We reclassified Hininger 2004 and Theobald 1937 from included to excluded for the 2018 update because the MMN supplement did not contain iron.
See the Characteristics of excluded studies table for more details.
One report comparing MMN supplementation versus iron folic acid remains awaiting assessment due to missing group denominators (Gathwala 2012).
Risk of bias in included studies
The risk of bias of the included studies was generally low with at least 50% of the judgements at 'low risk' for two domains (allocation concealment and incomplete outcome data) and at least 75% of judgements at 'low risk' for the remaining five domains. The domain with the highest risk of bias was incomplete outcome data (attrition bias), for which we have conducted a sensitivity analysis. It is unlikely that the evidence presented in this review is affected by the biases evaluated.
See Figure 2; Figure 3 and Characteristics of included studies table for further details on the risk of bias of the included studies.
'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study
Allocation
The included trials were of variable risk of bias. Seventeen trials adequately randomised participants to the treatment groups (Ashorn 2010; Bhutta 2009a; Biggs 2010; Christian 2003; Dewey 2009; Fawzi 2007; Friis 2004; Kaestel 2005; Lui 2013; Moore 2009; Osrin 2005; Ramakrishnan 2003; Roberfroid 2008; Sood 1975; SUMMIT 2008; West 2014; Zeng 2008), whereas the remaining trials did not describe the method they used for generating the randomisation sequence in sufficient detail to permit judgement.
Fourteen trials concealed the allocation of participants to the intervention and control groups (Ashorn 2010; Bhutta 2009a; Biggs 2010; Dewey 2009; Fawzi 2007; Friis 2004; Lui 2013; Moore 2009; Osrin 2005; Ramakrishnan 2003; Roberfroid 2008; SUMMIT 2008; West 2014; Zeng 2008); it was unclear in six trials (Brough 2010; Christian 2003; Kaestel 2005; Sunawang 2009; Tofail 2008; Zagre 2007); whereas the remaining one trial probably did not conceal allocation (Sood 1975).
Blinding
Eighteen trials blinded the participants, personnel and outcome assessors to the treatment allocation (Ashorn 2010; Bhutta 2009a; Brough 2010; Christian 2003; Dewey 2009; Fawzi 2007; Friis 2004; Kaestel 2005; Lui 2013; Osrin 2005; Ramakrishnan 2003; Roberfroid 2008; Sood 1975; SUMMIT 2008; Tofail 2008; West 2014; Zagre 2007; Zeng 2008). However, Sunawang 2009 showed blinding of participants only and Moore 2009 indicated blinding of outcome assessors only. In Biggs 2010, it was not possible to blind the participants and personnel to the daily supplementation arm.
Incomplete outcome data
Loss to follow‐up was less than 5% in four trials (Dewey 2009; Moore 2009; West 2014; Zeng 2008); between 5% to 9.9% in eight trials (Ashorn 2010; Biggs 2010; Christian 2003; Fawzi 2007; Lui 2013; Osrin 2005; Roberfroid 2008; SUMMIT 2008); and between 10% to 19.9% in four trials (Bhutta 2009a; Brough 2010; Sunawang 2009; Zagre 2007). In one trial (Zagre 2007), although attrition was less than 20% and they reported the reasons for attrition, the proportion of women who dropped out was significantly higher in the MMN versus the iron‐folic acid group. In addition, they did not report exclusion data, so we have deemed this trial as 'unclear risk' of bias. Loss to follow‐up was more than 20% in five trials (Friis 2004; Kaestel 2005; Ramakrishnan 2003; Sood 1975; Tofail 2008). In Osrin 2005, although attrition was 5% and they reported reasons for it, exclusion was 39.5% and without reported reasons, and so we assessed it as being at 'high risk'. Similarly, we assessed Moore 2009 to be 'high risk' because, although attrition was 4.8% and reasons for it were not reported, exclusion was 25.6% and reasons were reported. All of the trials used intention‐to‐treat analysis except for two trials that used modified intention‐to‐treat analysis (Ashorn 2010; Biggs 2010).
Selective reporting
There was no indication of selective reporting in most of the included trials. One trial, however, did not present growth outcomes mentioned in the methods section in the results section of the paper, including weight‐for‐age and weight‐for‐length (Biggs 2010).
Other potential sources of bias
We did not identify any other potential sources of bias, including those related specifically to cluster design, in the included trials.
Effects of interventions
Comparison 1: multiple micronutrients (MMN) versus control (all trials)
Twenty trials contributed data to this comparison. Nineteen out of 20 of these trials were carried out in low‐ and middle‐income countries and compared MMN supplements containing iron and folic acid to iron, with or without folic acid. One trial carried out in the UK compared MMN with placebo and contributed data to only four outcomes. In view of the differences in the settings where trials were conducted, and considering the control group conditions, we have presented results separately in the forest plots and in the text below.
Multiple micronutrients (MMN) with iron and folic acid versus iron, with or without folic acid
In this comparison, we included 19 trials conducted in low‐ and middle‐income countries that evaluated UNIMMAP or similar formulations of MMN supplement (Ashorn 2010; Bhutta 2009a; Biggs 2010; Christian 2003; Dewey 2009; Fawzi 2007; Friis 2004; Kaestel 2005; Lui 2013; Moore 2009; Osrin 2005; Ramakrishnan 2003; Roberfroid 2008; SUMMIT 2008; Sunawang 2009; Tofail 2008; West 2014; Zagre 2007; Zeng 2008). In two trials (Fawzi 2007; Friis 2004), women received iron and folic acid as separate supplements, and in one trial (Ramakrishnan 2003), women in the control group received iron only.
Primary outcomes
When we compared MMN supplementation against supplementation with iron with or without folic acid, there was probably a slight reduction in preterm births (average risk ratio (RR) 0.95, 95% confidence interval (CI) 0.90 to 1.01; studies = 18; random‐effects, Tau² = 0.00, I² = 49%; moderate‐quality evidence; Analysis 1.1), although the confidence interval for the pooled effect estimate just crossed the line of no effect. MMN supplementation also probably reduced births that were considered small‐for‐gestational age (SGA) (average RR 0.92, 95% CI 0.88 to 0.97; studies = 17; random‐effects, Tau² = 0.00, I² = 39%; moderate‐quality evidence; Analysis 1.2). MMN reduced births that were considered low birthweight (LBW) (average RR 0.88, 95% CI 0.85 to 0.91; studies = 18; random‐effects, Tau² = 0.00, I² = 0%; high‐quality evidence; Analysis 1.3), and made little or no difference to perinatal mortality (average RR 1.00, 95% CI 0.90 to 1.11; studies = 15; random‐effects, Tau² = 0.01, I² = 42%; high‐quality evidence; Analysis 1.4). Similar to preterm births, there was a slight reduction in stillbirths (average RR 0.95, 95% CI 0.86 to 1.04; studies = 17; random‐effects, Tau² = 0.00, I² = 12%; high‐quality evidence; Analysis 1.5), though the confidence interval for the pooled effect estimate just crossed the line of no effect. MMN supplementation did not have an important effect on neonatal mortality (average RR 1.00, 95% CI 0.89 to 1.12; studies = 14; random‐effects, Tau² = 0.01, I² = 22%; high‐quality evidence; Analysis 1.6). Visual inspection of funnel plots for all primary outcomes revealed no obvious funnel plot asymmetry (Figure 4; Figure 5; Figure 6; Figure 7), with the exception of preterm births (Figure 8), and SGA (Figure 9), where smaller studies appeared to report slightly more pronounced treatment effects.
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.3 Low birthweight
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.4 Perinatal mortality
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.5 Stillbirths
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.6 Neonatal mortality
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.1 Preterm births
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.2 Small‐for‐gestational age
It should be noted that where none of the individual trial reports reported an outcome, then we obtained data from a separate supplement (Food and Nutrition Bulletin 2009), where possible. We took data for SGA estimates for the following trials from the Food and Nutrition Bulletin: Friis 2004; Kaestel 2005; Osrin 2005; Ramakrishnan 2003; Sunawang 2009; Tofail 2008; Zagre 2007. Similarly, we took data for preterm birth for the following trials: Kaestel 2005; Sunawang 2009; Tofail 2008; Zagre 2007; data for LBW for Tofail 2008; data for perinatal mortality for Sunawang 2009; and data for stillbirth for Kaestel 2005, from the same report.
Secondary outcomes
When we compared MMN supplementation to iron supplementation with or without folic acid, there was little or no difference between groups in: maternal anaemia in the third trimester (average RR 1.04, 95% CI 0.94 to 1.15; studies = 9; random‐effects, Tau² = 0.01, I² = 50%; Analysis 1.7); maternal mortality (average RR 1.06, 95% CI 0.72 to 1.54; studies = 6; random‐effects, Tau² = 0.00, I² = 0%; Analysis 1.8); miscarriage (average RR 0.99, 95% CI 0.94 to 1.04; studies = 12; random‐effects, Tau² = 0.00, I² = 0%; Analysis 1.9); delivery via a caesarean section (average RR 1.13, 95% CI 0.99 to 1.29; studies = 5; random‐effects, Tau² = 0.00, I² = 0%; Analysis 1.10); and congenital anomalies (average RR 1.34, 95% CI 0.25 to 7.12; studies = 2; random‐effects, Tau² = 0.00, I² = 0%; Analysis 1.11). However, there was probably a reduction in very preterm births (average RR 0.81, 95% CI 0.71 to 0.93; studies = 4; random‐effects, Tau² = 0.00, I² = 10%; Analysis 1.12). Of the secondary outcomes examined, only miscarriage had a sufficient amount of studies to create a funnel plot. Visual inspection of the funnel plot for this outcome (Figure 10), suggested that smaller studies reported more substantial treatment effects.
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.9 Miscarriage (loss before 28 weeks)
There were a number of prespecified clinically important outcomes that we could not assess due to insufficient data from the included trials. These included the following outcomes, which only one or no trials measured, or which the trials presented in a format that precluded their inclusion in the analysis: premature rupture of membranes, pre‐eclampsia, macrosomia (Roberfroid 2008), placental abruption, neurodevelopmental delay of infants (Zeng 2008), nutritional status of the children (Dewey 2009; Roberfroid 2008), cost of supplementation, side‐effects of MMN supplementation (Lui 2013; Tofail 2008), and maternal well‐being or satisfaction.
Multiple micronutrients (MMN) versus placebo
One trial conducted in the UK (Brough 2010), contributed data to this analysis. In this trial, women in the control group were advised to take folic acid. Brough 2010 randomised 402 women. Women receiving supplements were at reduced risk of anaemia in the third trimester (average RR 0.66, 95% CI 0.51 to 0.85; Analysis 1.7), but there were little or no differences between women receiving supplements and those in the placebo group for any of the other outcomes reported: preterm birth (average RR 1.09 95% CI 0.43 to 2.77; Analysis 1.1); SGA (average RR 0.94, 95% CI 0.60 to 1.48; Analysis 1.2); or LBW (average RR 1.58, 95% CI 0.67 to 3.72; Analysis 1.3).
Subgroup analysis (data shown in comparison 2) multiple micronutrients (MMN) with iron and folic acid versus iron with or without folic acid)
For the trials comparing MMN with iron and folic acid versus iron with or without folic acid (19 trials), we found substantial heterogeneity in the analyses for preterm birth, SGA, and perinatal mortality, and explored its presence through subgroup analyses.
For the outcome preterm birth, MMN supplementation probably led to fewer preterm births for women in a subgroup of trials with mean maternal BMI of less than 20 kg/m² (average RR 0.85, 95% CI 0.81 to 0.90; studies = 3), compared to women with a BMI of at least 20 kg/m² (average RR 0.99, 95% CI 0.96 to 1.03; studies = 15; the test for subgroup differences P < 0.00001, I² = 95.2%; Analysis 2.1). There were little or no differences among subgroups based on mean maternal height (Analysis 2.2), the timing of supplementation (Analysis 2.3), or the dose of iron (Analysis 2.4) (all P > 0.05). However, MMN supplementation probably led to fewer preterm births among women who took non‐UNIMMAP supplements (average RR 0.89, 95% CI 0.82 to 0.98; studies = 8) compared to those who took UNIMMAP supplements (average RR 1.00, 95% CI 0.96 to 1.03; studies = 10; the test for subgroup differences P = 0.02, I2 = 80.4%; Analysis 2.5).
For the outcome SGA, MMN supplementation probably led to fewer SGA births for women in a subgroup of trials with mean maternal BMI of at least 20 kg/m² (average RR 0.88, 95% CI 0.83 to 0.93; studies = 14), but not for women with a mean maternal BMI of less than 20 kg/m² (average RR 1.00, 95% CI 0.92 to 1.08; studies = 3; test for subgroup differences P = 0.01, I² = 84.6%; Analysis 2.6). Similarly, we observed some differences between the subgroups of studies based on maternal height (Analysis 2.7). MMN supplementation probably reduced SGA births for women with a mean maternal height of at least 154.9 cm (average RR 0.85, 95% CI 0.79 to 0.91; studies = 9), and probably slightly reduced SGA births for women with a mean maternal height of less than 154.9 cm (average RR 0.98, 95% CI 0.96 to 1.00; studies = 8; test for subgroup differences P < 0.0001, I² = 93.7%; Analysis 2.7). There were little or no differences in SGA among subgroups based on timing of supplementation (Analysis 2.8), dose of iron (Analysis 2.9), or MMN supplement formulation (Analysis 2.10), when comparing MMN supplementation to iron with or without folic acid.
While we found that MMN supplementation made no difference to perinatal mortality as an outcome, the analysis demonstrated substantial statistical heterogeneity. However, subgroup analyses did not show clear differences based on mean maternal BMI (Analysis 2.11), mean maternal height (Analysis 2.12), dose of iron (Analysis 2.14) and MMN supplement formulation (Analysis 2.15) (all P > 0.05). However, we observed differences between subgroups based on the timing of supplementation. The reduction in perinatal mortality was probably higher in the subgroup with supplementation after 20 weeks (average RR 0.89, 95% CI 0.80 to 0.98; studies = 3) compared to the subgroup where women began supplementation before 20 weeks (average RR 1.09, 95% CI 0.92, 1.27; studies = 12; test for subgroup differences P = 0.04, I² = 76.5%; Analysis 2.13).
Sensitivity analysis (data shown in comparison 3) multiple micronutrients (MMN) with iron and folic acid versus iron, with or without folic acid
We undertook sensitivity analysis to study the effect of MMN supplementation on each outcome by excluding trials with loss to follow‐up of more than 20% from the analyses (Friis 2004; Kaestel 2005; Ramakrishnan 2003; Tofail 2008). This exclusion did not significantly affect the findings for the outcomes.
Discussion
Summary of main results
We identified 21 trials (involving 142,496 women) as eligible for inclusion in this review but only 20 trials (involving 141,849 women) contributed data to the review. This updated Cochrane Review summarises the current evidence on the effect of MMN supplementation during pregnancy on fetal, infant, and maternal outcomes. Overall, MMN supplementation with iron and folic acid versus supplementation with iron (with or without folic acid) showed an 8% reduction in the risk of SGA births (average risk ratio (RR) 0.92, 95% CI 0.88 to 0.97; moderate‐quality evidence), and a 12% reduction in the risk of LBW (average RR 0.88, 95% CI 0.85 to 0.91; high‐quality evidence). MMN supplementation slightly reduced the risk of both preterm births (average RR 0.95, 95% CI 0.90 to 1.01; 18 trials, moderate‐quality evidence), and stillbirths (average RR 0.95, 95% CI 0.86 to 1.04; high‐quality evidence), with the caveat that the confidence intervals for the pooled effects for both of these outcomes just crossed the line of no effect. We found that MMN supplementation did not have an important effect on perinatal mortality (average RR 1.00, 95% CI 0.90 to 1.11; high‐quality evidence), and neonatal mortality (average RR 1.00, 95% CI 0.89 to 1.12; high‐quality evidence). Of the secondary outcomes examined, we found that MMN supplementation reduced the risk of very preterm births by 19% (average RR 0.81, 95% CI 0.71 to 0.93). A summary of the main findings for trials comparing MMN with iron and folic acid versus iron with or without folic acid is presented in summary of findings Table for the main comparison.
Overall completeness and applicability of evidence
This review included a total of 21 trials that evaluated the impact of MMN supplementation. The earliest trial was published in 1975 (Sood 1975), though this trial did not contribute data to the review because it reported its outcomes in a format that we could not use in the meta‐analyses. An additional trial (Biggs 2010), did not contribute data to the main analyses because women received supplements twice‐weekly as opposed to daily. All trials evaluating the UNIMMAP supplement (Bhutta 2009a; Kaestel 2005; Lui 2013; Osrin 2005; Roberfroid 2008; SUMMIT 2008; Sunawang 2009; Tofail 2008; Zagre 2007; Zeng 2008), proposed in 1999 by UNICEF, UNU, and WHO, started recruitment of participants as early as 2001 and we included them in the analysis along with trials that evaluated adapted‐UNIMMAP formulations. Inclusion of these and older trials, identified as a result of an extensive search of literature published over the last several decades, represents overall completeness of evidence.
As iron/folic acid is recommended for pregnant women in low‐ and middle‐income countries, we primarily evaluated the effect of adding additional micronutrients to the iron and folic acid supplement. The vast majority (95%) of the included trials were conducted in low‐ and middle‐income countries, and included pregnant women at varying gestational stages, ranging from early pregnancy to 36 weeks of gestation.
In many low‐ and middle‐income countries, fertility rates are high, alongside a high prevalence of maternal iron‐deficiency anaemia and subclinical micronutrient deficiencies (Bhutta 2008). Studies have shown that a significant proportion of pregnant women suffer from multiple concurrent micronutrient deficiencies, especially throughout pregnancy when nutritional demands are increased. These deficiencies have been associated with a host of poor pregnancy outcomes including LBW, preterm, SGA, perinatal mortality, and maternal mortality (Allen 2001; Allen 2005; Christian 2010; De‐Regil 2016; Dror 2011; Haider 2013; Keen 2003; MacKay 2001; Ota 2015). Increased risk of infection has also been postulated to be associated with anaemia, especially as a result of iron deficiency (Oppenheimer 2001). Our findings of improved LBW, SGA, and very preterm births, then, could be reflective of improved nutritional status of mothers and hence better resistance to maternal infections following MMN supplementation.
Maternal anthropometry pre‐pregnancy and weight gain during pregnancy have also been associated with various neonatal and child outcomes. Maternal height is a relatively stable and easily measurable variable in the setting of low‐ and middle‐income countries. Reviews have identified short maternal stature (short height) as an important determinant of intrauterine growth retardation and LBW (Kramer 2003; WHO 1995). Short maternal stature has been associated with an increased risk of child mortality, underweight infants, and stunting (Ozaltin 2010; Voigt 2010). Our subgroup analyses indicated that MMN supplementation did not reduce SGA in women with poor nutritional status at baseline, defined as maternal height less than 154.9 cm and BMI less than 20 kg/m². However, MMN supplementation did reduce the risk of SGA babies among women with a mean maternal height of at least 154.9 cm. Similarly, MMN supplementation reduced the risk of SGA babies among women with a mean BMI of at least 20 kg/m². These findings should be interpreted with caution, but suggest a possible role of MMN in preventing SGA only among women with good nutritional status at baseline, and underscore an absence of similar effect in women with poor nutritional status at the time of conception. On the contrary, the findings for the subgroup analysis for preterm birth indicate a reduction among women with low BMI, but not among those with higher BMI. These differences further highlight the complex contribution of maternal malnutrition to fetal anthropometry and infant growth that relates to the intergenerational transfer of malnutrition. Though micronutrient supplementation has been implicated in improved child growth and survival, there is currently no evidence to suggest that maternal MMN supplementation during pregnancy improves child growth or survival when compared with iron/folic acid supplementation. Additional studies with long‐term follow‐up are required.
Quality of the evidence
We evaluated the quality of the available evidence by using the GRADE methodology as outlined in the GRADE Handbook. We created a 'Summary of findings' table for the primary outcomes of preterm birth, SGA, LBW, stillbirths, perinatal and neonatal mortality for Comparison 1: MMN with iron and folic acid versus iron with or without folic acid supplement.
When assessed according to GRADE criteria, the quality of evidence for the review's primary outcomes overall was moderate to high. We based pooled results for all primary outcomes on multiple randomised controlled trials with large sample sizes and precise estimates. For the comparison of MMN versus control (iron with or without folic acid) we graded the following outcomes as high quality: LBW, stillbirth, neonatal mortality, and perinatal mortality. The outcomes of preterm birth and SGA we graded as moderate quality; we downgraded both by one for funnel plot asymmetry, indicating possible publication bias.
Potential biases in the review process
This update of the review includes additional data published since the last update (2017). We conducted an extensive literature search to identify any additional studies since the last search. Two review authors independently conducted the screening of the updated search results, selection of eligible studies and data extraction. Two review authors also assessed the risk of bias. Given the application of the above Cochrane methodology, it is unlikely that the findings of this review are affected by biases in the review process.
Agreements and disagreements with other studies or reviews
The findings of reduction in SGA, LBW and very preterm risk as a result of MMN supplementation in the current review corroborate those of other systematic reviews and meta‐analyses conducted since the first version of this Cochrane Review (Bhutta 2012; Kawai 2011; Margetts 2009; Ramakrishnan 2012). Another systematic review and meta‐analysis (Christian 2015), also reports reduction in the risk of preterm births, LBW, and SGA; however, the effect on SGA is reported to be marginal (RR 0.91, 95% CI 0.84 to 1.00). This review included a smaller number of studies in the SGA analysis (studies = 7), as opposed to the current review (studies = 17), thereby explaining the difference between the estimates reported in the two reviews.
Another recent exercise (Smith 2017), used individual participant data from 17 randomised trials in low‐ and middle‐income countries to perform a two‐stage meta‐analysis comparing MMN and iron‐folic acid supplementation among pregnant women and identify potential modifiers of effect. Smith 2017 also found a reduction in LBW (random‐effects RR 0.86, 95% CI 0.81 to 0.92), a slight reduction in SGA (random‐effects RR 0.94, 95% CI 0.90 to 0.98), and some reduction in preterm births (random‐effects RR 0.93, 95% CI 0.87 to 0.98). Similar to this review, there were no important survival benefits (or harms) noted. Smith 2017 also demonstrated a greater effect of MMN supplementation on preterm births among underweight women when compared to normal weight women and, although tests for heterogeneity were not significant, there appeared to be a trend towards higher impact on SGA for women with greater BMI and height, as we have also shown. Though we did not stratify data by anaemia status, authors found that MMN supplementation caused greater reductions in LBW, SGA, and six‐month mortality among anaemic pregnant women when compared to those who were not anaemic. Taken together with the results of our subgroup analyses, this points to maternal nutritional status at baseline as a modifier of the effect of MMN supplementation for some important outcomes, and should instigate further studies to better elucidate these effects and the mechanisms behind them.
We found that MMN supplementation did not have an important effect on perinatal mortality, stillbirths, and neonatal mortality, which are similar to findings from earlier versions of this Cochrane Review and other systematic reviews and meta‐analyses (Christian 2005; Haider 2011; Ronsmans 2009; Smith 2017). With the additional studies included in this 2018 update, RR estimates for perinatal mortality, stillbirths, and neonatal mortality declined from 1.01 (0.91 to 1.13), 0.97 (0.87 to 1.09), and 1.06 (0.92 to 1.22) to 1.00 (0.90 to 1.11), 0.95 (0.86 to 1.04), and 1.00 (0.89 to 1.12), respectively. Previously, concerns were raised regarding the possibly harmful effects of MMN supplements relating to risk of perinatal and neonatal mortality through increased birth asphyxia in heavier babies (Christian 2005). Indeed, the Smith 2017 analysis found a slightly increased risk of large‐for‐gestational‐age births (defined by the Intergrowth standard) following MNN supplementation, but no indication of heightened risk of stillbirth or mortality. Two earlier trials conducted in Nepal (Christian 2003; Osrin 2005), both found no important increase in the risk of neonatal and perinatal mortality, though their pooled effect estimate showed an increase in the risk of these outcomes. This concern was questioned by other researchers in the field and has not been observed in other studies (Bhutta 2009b; Huffman 2005; Shrimpton 2005). Importantly, our current findings of no impact on neonatal mortality are supported by those of two MMN supplementation trials that were individually powered to evaluate an effect on early infant mortality (SUMMIT 2008; West 2014). The recent large trial in Bangladesh did not show an increase in neonatal or early infant mortality risk in the MMN supplementation group verus iron and folic acid (West 2014). The trial authors, however, in post‐hoc analysis, report higher neonatal mortality among boys due to birth asphyxia (West 2014). This finding should be interpreted with caution as the cause of death was ascertained by verbal autopsy with parents, which may result in misclassification of the underlying cause of death. Moreover, the trial was not powered to detect a statistically significant difference in cause‐specific mortality by gender. Interestingly, Smith 2017 found about a 15% reduction in neonatal mortality for females babies but no effect for male babies, indicating that there may be a biological difference in the impact of MMN supplementation relating to sex. Previous work has suggested that this may be a function of birth complications relating to size, with greater length, weight, and head circumference typically being true of male babies (West 2014). However, Smith 2017 noted no sex‐specific differences in stillbirths, indicating that other, potentially context‐specific, mechanisms may be responsible for these differences. These findings warrants further research.
'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies
'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.3 Low birthweight
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.4 Perinatal mortality
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.5 Stillbirths
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.6 Neonatal mortality
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.1 Preterm births
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.2 Small‐for‐gestational age
Funnel plot of comparison 1. Multiple micronutrients vs control, outcome: 1.9 Miscarriage (loss before 28 weeks)
Comparison 1 Multiple micronutrients vs control, Outcome 1 Preterm births.
Comparison 1 Multiple micronutrients vs control, Outcome 2 Small‐for‐gestational age.
Comparison 1 Multiple micronutrients vs control, Outcome 3 Low birthweight.
Comparison 1 Multiple micronutrients vs control, Outcome 4 Perinatal mortality.
Comparison 1 Multiple micronutrients vs control, Outcome 5 Stillbirths.
Comparison 1 Multiple micronutrients vs control, Outcome 6 Neonatal mortality.
Comparison 1 Multiple micronutrients vs control, Outcome 7 Maternal anaemia (third trimester Hb < 110 g/L).
Comparison 1 Multiple micronutrients vs control, Outcome 8 Maternal mortality.
Comparison 1 Multiple micronutrients vs control, Outcome 9 Miscarriage (loss before 28 weeks).
Comparison 1 Multiple micronutrients vs control, Outcome 10 Mode of delivery: caesarean section.
Comparison 1 Multiple micronutrients vs control, Outcome 11 Congenital anomalies.
Comparison 1 Multiple micronutrients vs control, Outcome 12 Very preterm birth (before 34 weeks of gestation).
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 1 Preterm births: mean maternal BMI.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 2 Preterm births: mean maternal height.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 3 Preterm births: timing of supplementation.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 4 Preterm births: dose of iron.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 5 Preterm births: MMN supplement formulation.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 6 Small‐for‐gestational age: mean maternal BMI.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 7 Small‐for‐gestational age: mean maternal height.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 8 Small‐for‐gestational age: timing of supplementation.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 9 Small‐for‐gestational age: dose of iron.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 10 Small‐for‐gestational age: MMN supplement formulation.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 11 Perinatal mortality: mean maternal BMI.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 12 Perinatal mortality: mean maternal height.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 13 Perinatal mortality: timing of supplementation.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 14 Perinatal mortality: dose of iron.
Comparison 2 Subgroup analysis for primary outcomes (MMN with iron and folic acid vs iron with or without folic acid)), Outcome 15 Perinatal mortality: MMN supplement formulation.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 1 Preterm births.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 2 Small‐for‐gestational age.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 3 Low birthweight.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 4 Perinatal mortality.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 5 Stillbirths.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 6 Neonatal mortality.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 7 Maternal anaemia (third trimester Hb < 110 g/L).
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 8 Miscarriage (loss before 28 weeks).
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 9 Maternal mortality.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 10 Very preterm birth (before 34 weeks of gestation).
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 11 Congenital anomalies.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 12 Neurodevelopmental outcome: BSID scores.
Comparison 3 Sensitivity analysis (all trials) excluding trials with > 20% loss to follow up, Outcome 13 Mode of delivery: caesarean section.
| Multiple micronutrients compared to control (iron with or without folic acid) for women during pregnancy | ||||||
| Patient or population: women during pregnancy | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Quality of the evidence | Comments | |
| Risk with control (iron with or without folic acid) | Risk with multiple micronutrients | |||||
| Preterm births | Study population | RR 0.95 | 91,425 | ⊕⊕⊕⊝ | ||
| 197 per 1000 | 188 per 1000 | |||||
| Small‐for‐gestational age | Study population | RR 0.92 | 57,348 | ⊕⊕⊕⊝ | ||
| 337 per 1000 | 310 per 1000 | |||||
| Low birthweight | Study population | RR 0.88 | 68,801 | ⊕⊕⊕⊕ | ||
| 212 per 1000 | 187 per 1000 | |||||
| Perinatal mortality | Study population | RR 1.00 | 63,922 | ⊕⊕⊕⊕ | Raw event and participant data were unavailable for Ramakrishnan 2003 and West 2014, so have not been included in No of participants column. | |
| 39 per 1000 | 39 per 1000 | |||||
| Stillbirths | Study population | RR 0.95 | 97,927 | ⊕⊕⊕⊕ | ||
| 30 per 1000 | 28 per 1000 | |||||
| Neonatal mortality | Study population | RR 1.00 | 80,964 | ⊕⊕⊕⊕ | Raw event and participant data were unavailable for Bhutta 2009a and Fawzi 2007 so have not been included in No of participants column. | |
| 29 per 1000 | 29 per 1000 | |||||
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). | ||||||
| GRADE Working Group grades of evidence | ||||||
| aStrong evidence of funnel plot asymmetry, indicating possible publication bias. | ||||||
| Study ID | Iron | Folic acid | Vit A | Beta‐carotene | Vit C | Vit D | Vit E | Vit B1 (thiamine) | Vit B2 (riboflavin) | Vit B3 (niacin) | Vit B5 (pantothenic acid) | Vit B6 | Vit B12 | Vit K | Copper | Selenium | Zinc | Iodine | Magnesium | Calcium | Phosphorus | Biotin | Potassium | Manganese |
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | |||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | |||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ||||||||||||||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | |||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ||||||||||
| ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Trial | Outcome(s) | Reported design effect | Reported ICC | Calculated M | Calculated design effecta |
| Preterm birth, miscarriage | 1.2 | ‐ | ‐ | ‐ | |
| All | 1.2 | ‐ | ‐ | ‐ | |
| Maternal anaemia, miscarriage, maternal mortality, very preterm birth | 1.15 | ‐ | ‐ | ‐ | |
| All | 1.2 | ‐ | ‐ | ‐ | |
| Preterm birth, very preterm birth | ‐ | 0.02 | 11 | 1.2 | |
| LBW | ‐ | 0.03 | 11 | 1.3 | |
| All other outcomes | ‐ | 0.03b | 11 | 1.3 | |
| Maternal anaemia | ‐ | 0.03 | 12.1 | 1.3 | |
| All other outcomes | ‐ | 0 | 12.1 | 1.0 | |
| ICC: intracluster correlation coefficient; M: average cluster size | |||||
| aDesign effect = 1 + (M‐1)*ICC. | |||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Preterm births Show forest plot | 19 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 1.1 MMN with iron and folic acid vs iron with or without folic acid | 18 | Risk Ratio (Random, 95% CI) | 0.95 [0.90, 1.01] | |
| 1.2 MMN with iron and folic acid vs placebo | 1 | Risk Ratio (Random, 95% CI) | 1.09 [0.43, 2.77] | |
| 2 Small‐for‐gestational age Show forest plot | 18 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 2.1 MMN with iron and folic acid vs iron with or without folic acid | 17 | Risk Ratio (Random, 95% CI) | 0.92 [0.88, 0.97] | |
| 2.2 MMN with iron and folic acid vs placebo | 1 | Risk Ratio (Random, 95% CI) | 0.94 [0.60, 1.48] | |
| 3 Low birthweight Show forest plot | 19 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 3.1 MMN with iron and folic acid vs iron with or without folic acid | 18 | Risk Ratio (Random, 95% CI) | 0.88 [0.85, 0.91] | |
| 3.2 MMN with iron and folic acid vs placebo | 1 | Risk Ratio (Random, 95% CI) | 1.58 [0.67, 3.72] | |
| 4 Perinatal mortality Show forest plot | 15 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 4.1 MMN with iron and folic acid vs iron with or without folic acid | 15 | Risk Ratio (Random, 95% CI) | 1.00 [0.90, 1.11] | |
| 5 Stillbirths Show forest plot | 17 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 5.1 MMN with iron and folic acid vs iron with or without folic acid | 17 | Risk Ratio (Random, 95% CI) | 0.95 [0.86, 1.04] | |
| 6 Neonatal mortality Show forest plot | 14 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 6.1 MMN with iron and folic acid vs iron with or without folic acid | 14 | Risk Ratio (Random, 95% CI) | 1.00 [0.89, 1.12] | |
| 7 Maternal anaemia (third trimester Hb < 110 g/L) Show forest plot | 10 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 7.1 MMN with iron and folic acid vs iron with or without folic acid | 9 | Risk Ratio (Random, 95% CI) | 1.04 [0.94, 1.15] | |
| 7.2 MMN with iron and folic acid vs placebo | 1 | Risk Ratio (Random, 95% CI) | 0.66 [0.51, 0.85] | |
| 8 Maternal mortality Show forest plot | 6 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 8.1 MMN with iron and folic acid vs iron with or without folic acid | 6 | Risk Ratio (Random, 95% CI) | 1.06 [0.72, 1.54] | |
| 9 Miscarriage (loss before 28 weeks) Show forest plot | 12 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 9.1 MMN with iron and folic acid vs iron with or without folic acid | 12 | Risk Ratio (Random, 95% CI) | 0.99 [0.94, 1.04] | |
| 10 Mode of delivery: caesarean section Show forest plot | 5 | Risk Ratio (Random, 95% CI) | 1.13 [0.99, 1.29] | |
| 10.1 MMN with iron and folic acid vs iron with or without folic acid | 5 | Risk Ratio (Random, 95% CI) | 1.13 [0.99, 1.29] | |
| 11 Congenital anomalies Show forest plot | 2 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 11.1 MMN with iron and folic acid vs iron with or without folic acid | 2 | Risk Ratio (Random, 95% CI) | 1.34 [0.25, 7.12] | |
| 12 Very preterm birth (before 34 weeks of gestation) Show forest plot | 4 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 12.1 MMN with iron and FA vs iron with or without folic acid | 4 | Risk Ratio (Random, 95% CI) | 0.81 [0.71, 0.93] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Preterm births: mean maternal BMI Show forest plot | 18 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 1.1 BMI < 20 kg/m2 | 3 | Risk Ratio (Random, 95% CI) | 0.85 [0.81, 0.90] | |
| 1.2 BMI ≥ 20 kg/m2 | 15 | Risk Ratio (Random, 95% CI) | 0.99 [0.96, 1.03] | |
| 2 Preterm births: mean maternal height Show forest plot | 18 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 2.1 Maternal height < 154.9 cm | 8 | Risk Ratio (Random, 95% CI) | 0.93 [0.85, 1.03] | |
| 2.2 Maternal height ≥ 154.9 cm | 10 | Risk Ratio (Random, 95% CI) | 0.99 [0.93, 1.05] | |
| 3 Preterm births: timing of supplementation Show forest plot | 18 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 3.1 Supplementation started before 20 weeks | 14 | Risk Ratio (Random, 95% CI) | 0.93 [0.87, 1.00] | |
| 3.2 Supplementation after 20 weeks | 4 | Risk Ratio (Random, 95% CI) | 1.00 [0.96, 1.04] | |
| 4 Preterm births: dose of iron Show forest plot | 19 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 4.1 MMN with 30 mg iron vs. supplement with 60 mg iron | 6 | Risk Ratio (Random, 95% CI) | 1.05 [0.94, 1.17] | |
| 4.2 MMN with 60 mg iron vs. supplement with 60 mg iron | 5 | Risk Ratio (Random, 95% CI) | 0.96 [0.88, 1.05] | |
| 4.3 MMN with 30 mg iron vs. supplement with 30 mg iron | 5 | Risk Ratio (Random, 95% CI) | 0.92 [0.84, 1.02] | |
| 4.4 MMN with 15‐20 mg iron vs. supplement with 60 mg iron | 3 | Risk Ratio (Random, 95% CI) | 0.90 [0.64, 1.27] | |
| 5 Preterm births: MMN supplement formulation Show forest plot | 18 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 5.1 UNIMMAP formulation | 10 | Risk Ratio (Random, 95% CI) | 1.00 [0.96, 1.03] | |
| 5.2 Non‐UNIMMAP formulation | 8 | Risk Ratio (Random, 95% CI) | 0.89 [0.82, 0.98] | |
| 6 Small‐for‐gestational age: mean maternal BMI Show forest plot | 17 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 6.1 BMI < 20 kg/m2 | 3 | Risk Ratio (Random, 95% CI) | 1.00 [0.92, 1.08] | |
| 6.2 BMI ≥ 20 kg/m2 | 14 | Risk Ratio (Random, 95% CI) | 0.88 [0.83, 0.93] | |
| 7 Small‐for‐gestational age: mean maternal height Show forest plot | 17 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 7.1 Maternal height < 154.9 cm | 8 | Risk Ratio (Random, 95% CI) | 0.98 [0.96, 1.00] | |
| 7.2 Maternal height ≥ 154.9 cm | 9 | Risk Ratio (Random, 95% CI) | 0.85 [0.79, 0.91] | |
| 8 Small‐for‐gestational age: timing of supplementation Show forest plot | 17 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 8.1 Supplementation started before 20 weeks | 13 | Risk Ratio (Random, 95% CI) | 0.98 [0.96, 1.00] | |
| 8.2 Supplementation after 20 weeks | 4 | Risk Ratio (Random, 95% CI) | 0.85 [0.72, 0.99] | |
| 9 Small‐for‐gestational age: dose of iron Show forest plot | 17 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 9.1 MMN with 30 mg iron vs. supplement with 60 mg iron | 7 | Risk Ratio (Random, 95% CI) | 0.89 [0.81, 0.97] | |
| 9.2 MMN with 60 mg iron vs. supplement with 60 mg iron | 5 | Risk Ratio (Random, 95% CI) | 0.88 [0.72, 1.08] | |
| 9.3 MMN with 30 mg iron vs. supplement with 30 mg iron | 3 | Risk Ratio (Random, 95% CI) | 0.98 [0.96, 1.00] | |
| 9.4 MMN with 20 mg iron vs. supplement with 60 mg iron | 2 | Risk Ratio (Random, 95% CI) | 0.91 [0.75, 1.09] | |
| 10 Small‐for‐gestational age: MMN supplement formulation Show forest plot | 17 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 10.1 UNIMMAP formulation | 9 | Risk Ratio (Random, 95% CI) | 0.91 [0.85, 0.98] | |
| 10.2 Non‐UNIMMAP formulation | 8 | Risk Ratio (Random, 95% CI) | 0.92 [0.84, 1.01] | |
| 11 Perinatal mortality: mean maternal BMI Show forest plot | 15 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 11.1 BMI < 20 kg/m² | 3 | Risk Ratio (Random, 95% CI) | 1.11 [0.82, 1.50] | |
| 11.2 BMI ≥ 20 kg/m² | 12 | Risk Ratio (Random, 95% CI) | 0.98 [0.86, 1.13] | |
| 12 Perinatal mortality: mean maternal height Show forest plot | 15 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 12.1 Maternal height < 154.9 cm | 8 | Risk Ratio (Random, 95% CI) | 1.00 [0.89, 1.13] | |
| 12.2 Maternal height ≥ 154.9 cm | 7 | Risk Ratio (Random, 95% CI) | 0.99 [0.78, 1.25] | |
| 13 Perinatal mortality: timing of supplementation Show forest plot | 15 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 13.1 Supplementation before 20 weeks | 12 | Risk Ratio (Random, 95% CI) | 1.09 [0.92, 1.27] | |
| 13.2 Supplementation after 20 weeks | 3 | Risk Ratio (Random, 95% CI) | 0.89 [0.80, 0.98] | |
| 14 Perinatal mortality: dose of iron Show forest plot | 15 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 14.1 MMN with 30 mg iron vs. supplement with 60 mg iron | 6 | Risk Ratio (Random, 95% CI) | 1.20 [0.95, 1.51] | |
| 14.2 MMN with 60 mg iron vs. supplement with 60 mg iron | 3 | Risk Ratio (Random, 95% CI) | 1.09 [0.74, 1.60] | |
| 14.3 MMN with 30 mg iron vs. supplement with 30 mg iron | 4 | Risk Ratio (Random, 95% CI) | 0.92 [0.86, 0.98] | |
| 14.4 MMN with 20 mg iron vs. supplement with 60 mg iron | 2 | Risk Ratio (Random, 95% CI) | 0.64 [0.36, 1.13] | |
| 15 Perinatal mortality: MMN supplement formulation Show forest plot | 15 | Risk Ratio (Random, 95% CI) | Subtotals only | |
| 15.1 UNIMMAP formulation | 9 | Risk Ratio (Random, 95% CI) | 1.06 [0.89, 1.25] | |
| 15.2 Non‐UNIMMAP formulation | 6 | Risk Ratio (Random, 95% CI) | 0.94 [0.82, 1.09] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Preterm births Show forest plot | 14 | Risk Ratio (Random, 95% CI) | 0.96 [0.90, 1.02] | |
| 2 Small‐for‐gestational age Show forest plot | 10 | Risk Ratio (Random, 95% CI) | 0.92 [0.85, 0.99] | |
| 3 Low birthweight Show forest plot | 14 | Risk Ratio (Random, 95% CI) | 0.88 [0.85, 0.91] | |
| 4 Perinatal mortality Show forest plot | 12 | Risk Ratio (Random, 95% CI) | 1.02 [0.91, 1.15] | |
| 5 Stillbirths Show forest plot | 13 | Risk Ratio (Random, 95% CI) | 0.97 [0.86, 1.08] | |
| 6 Neonatal mortality Show forest plot | 11 | Risk Ratio (Random, 95% CI) | 1.03 [0.88, 1.20] | |
| 7 Maternal anaemia (third trimester Hb < 110 g/L) Show forest plot | 8 | Risk Ratio (Random, 95% CI) | 1.03 [0.92, 1.15] | |
| 8 Miscarriage (loss before 28 weeks) Show forest plot | 9 | Risk Ratio (Random, 95% CI) | 0.99 [0.94, 1.04] | |
| 9 Maternal mortality Show forest plot | 5 | Risk Ratio (Random, 95% CI) | 1.13 [0.76, 1.68] | |
| 10 Very preterm birth (before 34 weeks of gestation) Show forest plot | 4 | Risk Ratio (Random, 95% CI) | 0.81 [0.71, 0.93] | |
| 11 Congenital anomalies Show forest plot | 2 | Risk Ratio (Random, 95% CI) | 1.34 [0.25, 7.12] | |
| 12 Neurodevelopmental outcome: BSID scores Show forest plot | 1 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 12.1 Mental development scores at 6 months of age: new subgroup | 1 | 592 | Mean Difference (IV, Random, 95% CI) | ‐0.02 [‐6.75, 6.71] |
| 12.2 Mental development scores at 12 months of age | 1 | 572 | Mean Difference (IV, Random, 95% CI) | 1.21 [‐5.04, 7.46] |
| 12.3 Psychomotor development scores ar 6 months of age | 1 | 592 | Mean Difference (IV, Random, 95% CI) | ‐0.16 [‐3.89, 3.57] |
| 12.4 Psychomotor development scores at 12 months of age | 1 | 572 | Mean Difference (IV, Random, 95% CI) | 0.34 [‐2.72, 3.40] |
| 13 Mode of delivery: caesarean section Show forest plot | 4 | Risk Ratio (Random, 95% CI) | 1.13 [0.99, 1.30] | |
