Treatments for iron‐deficiency anaemia in pregnancy

Abstract

Background

Iron deficiency, the most common cause of anaemia in pregnancy worldwide, can be mild, moderate or severe. Severe anaemia can have very serious consequences for mothers and babies, but there is controversy about whether treating mild or moderate anaemia provides more benefit than harm.

Objectives

To assess the effects of different treatments for anaemia in pregnancy attributed to iron deficiency (defined as haemoglobin less than 11 g/dL or other equivalent parameters) on maternal and neonatal morbidity and mortality.

Search methods

We searched the Cochrane Pregnancy and Childbirth Group's Trials Register (7 June 2011), CENTRAL (2011, Issue 5), PubMed (1966 to June 2011), the International Clinical Trials Registry Platform (ICTRP) (2 May 2011), Health Technology Assessment Program (HTA) (2 May 2011) and LATINREC (Colombia) (2 May 2011).

Selection criteria

Randomised controlled trials comparing treatments for anaemia in pregnancy attributed to iron deficiency.

Data collection and analysis

We identified 23 trials, involving 3.198 women. We assessed their risk of bias. Three further studies identified are awaiting classification.

Main results

Many of the trials were from low‐income countries; they were generally small and frequently methodologically poor. They covered a very wide range of differing drugs, doses and routes of administration, making it difficult to pool data. Oral iron in pregnancy showed a reduction in the incidence of anaemia (risk ratio 0.38, 95% confidence interval 0.26 to 0.55, one trial, 125 women) and better haematological indices than placebo (two trials). It was not possible to assess the effects of treatment by severity of anaemia. A trend was found between dose and reported adverse effects. Most trials reported no clinically relevant outcomes nor adverse effects. Although the intramuscular and intravenous routes produced better haematological indices in women than the oral route, no clinical outcomes were assessed and there were insufficient data on adverse effects, for example, on venous thrombosis and severe allergic reactions. Daily low‐dose iron supplements may be effective at treating anaemia in pregnancy with less gastrointestinal side effects compared with higher doses.

Authors' conclusions

Despite the high incidence and burden of disease associated with this condition, there is a paucity of good quality trials assessing clinical maternal and neonatal effects of iron administration in women with anaemia. Daily oral iron treatment improves haematological indices but causes frequent gastrointestinal adverse effects. Parenteral (intramuscular and intravenous) iron enhances haematological response, compared with oral iron, but there are concerns about possible important adverse effects (for intravenous treatment venous thrombosis and allergic reactions and for intramuscular treatment important pain, discolouration and allergic reactions). Large, good quality trials, assessing clinical outcomes (including adverse effects) as well as the effects of treatment by severity of anaemia are required.

Plain language summary

Treatments for anaemia in pregnancy thought to be due to iron deficiency

When the blood has insufficient red cells, or the red cells carry insufficient haemoglobin to deliver adequate oxygen to the tissues, this is called anaemia. There is normally a reduction in the haemoglobin concentrations in the mother's blood during pregnancy, and this allows a better blood flow around the womb (uterus) and to the baby. This is sometime called physiological anaemia and needs no treatment. True anaemia, however, can be mild, moderate or severe and can cause weakness, tiredness and dizziness. Severe anaemia makes women at risk of cardiac failure and is very common in low‐income countries Anaemia has many causes including a shortage or iron, folic acid or vitamin B₁₂. These are all required for making red cells and are available in a good diet. Iron shortage, however, is the most common cause of anaemia during pregnancy. Iron treatment can be given by mouth (oral), by injection into the muscle (intramuscular) or injection into the vein (intravenous). Blood transfusion or giving something which stimulates the body to produce more red cells (erythropoietin) are also possible treatments.

In this review, we identified 23 trials involving 3198 pregnant women. Many of the trials were in low‐income countries and many treatment variations were studied. Oral iron reduced the incidence of anaemia but is known to sometimes cause constipation and nausea. Although the intramuscular and intravenous routes produced better levels of red cells and iron stores than the oral route, no clinical outcomes (such as pre‐eclampsia, preterm births, postpartum haemorrhage) were assessed and there were insufficient data on adverse effects. Intravenous treatment can cause venous thrombosis (blockages in the veins) and intramuscular treatment causes important pain and discolouration at the injection site. It was unclear if women and babies were healthier when women were given iron for mild or moderate anaemia during pregnancy. There were no studies on using blood transfusions.

Overall, there was insufficient evidence to say when or how anaemia in pregnancy needs to or should be treated.

Background

Description of the condition

Iron deficiency, and particularly iron deficiency anaemia, is one of the most severe and important nutritional deficiencies worldwide. Iron deficiency is defined as "a condition in which there are no mobilizable iron stores and in which signs of a compromised supply of iron to tissues, including the erythron, are noted. The more severe stages of iron deficiency are associated with anaemia" (WHO 2001). Anaemia is a reduction in the normal number of circulating red blood cells and in the quantity of haemoglobin (Hb) in the blood. More than half of the pregnant women in low‐income countries suffer from anaemia and iron deficiency is the most common cause of anaemia in pregnancy (WHO 2001). Using data from the World Health Organization (WHO) Vitamin and Mineral Nutrition Information System for 1993 to 2005, global anaemia prevalence was estimated to be 41.8% (95% confidence interval (CI) 39.9 to 43.8 %) in pregnant women (McLean 2009). The high prevalence of anaemia has been associated with low socio‐economic status (odds ratio (OR) 1.419, 95% CI 1.05 to 1.90) (Noronha 2010).

More than half a million maternal deaths occur each year, approximately 90% of which are in low‐income countries, making evident a large discrepancy between high‐ and low‐income countries. The common causes of anaemia include iron deficiency, folate deficiency, vitamin B12 deficiency, bone marrow suppression, haemolytic diseases (sickle cell disease and malaria), chronic blood loss (for example, hook worm infestation) and underlying malignancies (WHO 1992), with iron‐deficiency being the most common cause of anaemia in pregnant women worldwide (Goroll 1997; Lops 1995; Williams 1992). The diverse main preventable factors relating to maternal mortality have been described, and include chronic anaemia, infections, bleeding, hypertensive disorders, obstructed labour and unsafe abortions (WHO 2000).

Haemoglobin is the protein in the red blood cell which carries oxygen to the tissues and iron is part of the Hb molecule. There are a variety of indicators of iron status such as bone marrow iron, serum ferritin, transferrin saturation, erythrocyte protoporphyrin, transferrin receptors and also red cell mean corpuscular volume (MCV), red cell mean corpuscular haemoglobin (MCH) and red cell mean corpuscular haemoglobin concentration (MCHC). However, neither blood Hb concentration nor serum iron are thought to be good indicators of anaemia because there can be depletion of body iron stores in the presence of normal Hb levels and serum iron fluctuates depending on recent iron intake. Serum ferritin may be a better indicator of iron status as the examination of iron stores in the bone marrow is impractical. Historically, blood Hb levels have been used as indicators of anaemia, the test being simple and inexpensive to undertake. Anaemia is associated with general weakness, tiredness and dizziness but the level of Hb associated with these symptoms in pregnancy is unknown. These symptoms are nonspecific and frequently appear in cases of severe anaemia too and in association with other medical problems. Anaemia in pregnancy is defined by the WHO as a Hb value below 11 g/dL (WHO 1992; WHO 2001). Although anaemia is frequently graded as "mild", "moderate", or "severe", the Hb values at which the division into these three categories is made vary and are arbitrary. Standardised cut‐off values are difficult to define because populations, geographic settings and needs are different according to specific areas. Some authors suggest that Hb values at sea level should be categorised as follows (WHO 1989): (1) mild anaemia (Hb 10 to 10.9 g/dL); (2) moderate anaemia (Hb 7 to 9.9 g/dL); (3) severe anaemia (Hb less than 7 g/dL). However, other criteria have been widely used in the literature to define anaemia cut‐off values: (1) mild (Hb 9 to 10.9 g/dL), (2) moderate (Hb 7 to 8.9 g/dL) and (3) severe (Hb below 7 g/dL) (Adam 2005); and (1) mild anaemia (Hb 7 to 11 g/dL), moderate anaemia (5 to 7 g/dL) and severe anaemia (below 5 g/dL) (Brabin 2001). The interpretation of Hb concentrations in pregnancy is sometimes difficult because of the normal physiological decrease in blood Hb concentration in pregnancy and in certain conditions where water balance is affected (i.e. pre‐eclampsia) Hb concentrations may be overestimated as a result of dehydration (Letsky 1991).

It is suggested that the iron stores of the woman's body become reduced during pregnancy (as a result of the increased red cell mass and the demands of the fetus exceeding iron intake), and that this can take place in the presence of normal blood Hb levels. Some will argue that this is a well‐designed physiological mechanism to continue to deliver oxygen to the tissues in the presence of lowered blood Hb levels in pregnancy. An observational study undertaken in London, UK, found that low levels of Hb, commonly considered as mild anaemia, were associated with a better prognosis for the fetus, although figures did not appear to be corrected for women with pre‐eclampsia (Steer 1995). However, others argue that reduced iron stores are a health problem for pregnant women and their babies (Letsky 2001). Several studies considered anaemia (Hb levels between 7 g/dL and 10 g/dL) as a risk factor for fetal death, premature delivery, low birthweight and other adverse outcomes (Williams 1992). Some suggest a link between maternal anaemia in pregnancy on the later developmental problems of the children (Letsky 2001; Williams 1992). There is evidence indicating that maternal Hb levels under 7 g/dL are associated with a higher risk in the mother of developing cardiac heart failure, which has adverse consequences on the mother and fetus (Lops 1995; WHO 1992; Williams 1992). A cohort study done in Pakistan found that the risk of low birthweight and preterm delivery among anaemic women (Hb under 11 g/dL) was 1.9 and four times higher, respectively, than non‐anaemic women. In addition, the neonates of anaemic women had a 3.7 greater risk of intrauterine fetal death and 1.8 times increased risk having low Apgar scores at one minute when compared to non‐anaemic women (Lone 2004). There is a strong case for studying separately physiological anaemia, mild anaemia and severe anaemia in pregnancy.

In low‐income countries, anaemia in pregnancy is frequent and has been attributed to poor nutrition and a high incidence of concurrent diseases, and can potentially complicate conditions such as postpartum haemorrhage which is a major contributor to maternal mortality in many developing countries (WHO 1992). A case control study, conducted in India, evaluated maternal mortality risk factors from 25,926 households in 411 villages. The authors reported that the major related causes of maternal deaths were haemorrhage, severe anaemia, puerperal sepsis, and abortion (Gupta 2010). However, anaemia may only be a marker of various social and nutritional conditions, and raising Hb levels could have little, if any, effect on morbidity or mortality if those conditions are not improved (Goroll 1997).

Description of the intervention

There are various considerations that need to be taken into account when treating iron‐deficiency anaemia:

Iron can be given by mouth, by intramuscular (IM) injection or intravenous (IV) injection. It is also possible to deliver iron by giving a blood transfusion, and recombinant erythropoietin in conjunction with iron is a further possibility. Some evidence suggests that oral iron given to anaemic pregnant and non‐pregnant women is associated with gastrointestinal side effects such as nausea and constipation (WHO 2000 ). IM or IV iron are thought to be associated with allergic reactions and anaphylactic shock, as well as venous thrombosis and occasionally cardiac arrest and death. Blood transfusion carries the risk of transmitting parasitic or viral infections including HIV, hepatitis, and Chagas disease (trypanosomiasis), despite preventive blood screening. There is also the possibility of bovine spongiform encephalitis, and as yet unknown viral infections. Oral iron is often the preferred route of administration for mild anaemia, while IM and IV routes are more frequently used in people with severe anaemia when the risks of cardiac failure due to severe anaemia are perceived to outweigh the risks of potential adverse effects.

There are also variations in the doses used (e.g. low or high doses), differing regimens (e.g. single, daily, monthly or intermittent doses), differing forms of iron (e.g. dextran, fumarate) and sometimes adjuncts given (e.g. vitamin A, erythropoietin).

How the intervention might work

Physiological changes during pregnancy may affect laboratory parameters such as the haematocrit and the Hb (Klajnbard 2010). Transfer of iron from the mother to the fetus is regulated by the placenta (Allen 2000). During pregnancy, there is an increase in maternal iron absorption and also an increase in both red cell mass and plasma volume to accommodate the needs of the growing uterus and fetus. However, plasma volume increases more than the red cell mass leading to a fall in the concentration of Hb in the blood, despite the increase in the total number of red cells (Letsky 1991). This drop in Hb concentration decreases the blood viscosity and it is thought this enhances the placental perfusion providing a better maternal‐fetal gas and nutrient exchange (Mani 1995). There is controversy around the significance for women and their babies of this physiological haemodilution of pregnancy and at what level of Hb women and babies would benefit from iron treatment. As discussed below, some studies suggest that the physiological decrease in Hb is associated with improved outcomes for the baby (Mahomed 1989; Steer 1995), whilst others have identified adverse long‐term outcomes for the baby (Walter 1994).

For pathological changes, treatment for iron deficiency focuses on increasing the iron stores and blood Hb levels, so they reach normal levels and can provide a normal oxygen supply to the tissues. In anaemia in pregnancy, the treatment should not only normalise iron stores and blood Hb levels, but should also be shown to improves clinical outcomes for both mother and baby.

Why it is important to do this review

Iron is the most common strategy currently used to control iron deficiency and anaemia in low‐income countries (WHO 2001). Recommendations for the treatment of anaemia are frequently based on the expectation that they may be benevolent but are seldom supported by reproducible robust studies, especially randomised controlled trials. Furthermore, they may not take into account important adverse effects such as allergic reactions, viral or parasitic transmission from blood transfusions, gastrointestinal complications, and discomfort generated by common side effects of iron (Table 1). Therefore, it is difficult to balance the benefits and harms of treatments, let alone determine if there is a case to recommend a particular anaemia treatment for all women with anaemia in pregnancy.

1. FDA iron adverse effects description.

Drug substance	Adverse effect
Iron	Oral preparations: produces gastrointestinal irritation and abdominal pain with nausea and vomiting, when administered orally. The effect is usually dose related to the amount of elemental iron, rather than the preparation. Diarrhoea and constipation. Better to administer with foods, and to increase doses gradually. Oral liquid preparations may stain teeth. Oral preparations should not be given concomitantly with parenteral preparations. Parenteral preparations: anaphylactoid reactions, peripheral vascular flushing with intravenous administration, tachycardia, hypotension and syncope, thrombophlebitis (higher if given with glucose 5% versus sodium chloride 0.9%), nausea, vomiting, taste disturbance. Delayed reactions may include arthralgia, myalgia, regional lymphadenopathy, chills, fever, paraesthesia, dizziness, malaise, headache, nausea, vomiting and haematuria. Intramuscular use in animals has resulted in the development of sarcomas at the injection site. It interacts with enalapril (potentiates adverse systemic reactions) and chloramphenicol.
Iron sulphate (oral)
Ferrous gluconate (oral)
Iron fumarate (oral)
Iron sucrose (intravenous)	In addition to the adverse effects for parenteral preparations: bronchospasm, dyspnoea, myalgia, pruritus, urticaria, rash, reactions in the injection site.
Iron sorbitol	In addition to the adverse effects for parenteral preparations: severe systemic reactions with potentially fatal cardiac complications. Dark urine. Do not administer intravenously.
Iron dextran (intramuscular)
Iron polymaltose complex ‐iron dextrin‐
Adjuvant recombinant human erythropoietin	Epoetin: recombinant human erythropoietin. Darbepoetin: derivative of epoetin. Headache, hypertension and seizures, specially in people with poor renal function. Thrombosis at vascular access sites, flu‐like symptoms, hyperkalaemia, skin rashes, and rare reports of anaphylactoid reactions.
Iron polymaltose complex (oral)
†Martindale, The Complete Drug reference. Pharmaceutical Press. 32 ed. London 1999. ‡British National Formulary. Royal Pharmaceutical Society. 49 ed. London, March 2005 ‡British National Formulary. Royal Pharmaceutical Society. 49 ed. London, March 2005 †Martindale, The Complete Drug reference. Pharmaceutical Press. 32 ed. London 1999.

The aim of this review was to use a systematic approach to identify and synthesise the evidence of randomised controlled trials evaluating the effects of treatments for anaemia in pregnancy, and provide robust valid and useful evidence to inform clinical practice.

Objectives

The principal objective was to determine the overall effects of different forms of iron therapy given to pregnant women diagnosed with anaemia attributed to iron‐deficiency, measuring neonatal and maternal morbidity and mortality, haematological parameters and side effects, especially adverse effects of treatment. The review aimed to assess the effects of iron treatments when delivered to women categorised in three groups (mild, moderate or severe anaemia, as defined by trialists) at inception into the randomised controlled trial.

The review did not address the need for iron supplementation of non‐anaemic women; this question has been addressed in several other reviews and evidence summaries. Similarly, it did not focus vitamin B12, micronutrients, folate deficiency, infectious or genetic anaemia, which will be or are covered by other reviews (Pena‐Rosas 2009; Van den Broek 2010).

Methods

Criteria for considering studies for this review

Types of studies

This review considered randomised controlled trials assessing the effects of treatments for anaemia in pregnancy attributed to iron deficiency. When information in the abstract was unclear or incomplete, we reviewed the 'materials and methods' of the reports. Quasi‐random studies were not eligible for this review.

As previously mentioned, the use of Hb <11g/dL (at sea level) only helps to diagnose anaemia but not iron deficiency anaemia. Iron‐deficiency anaemia definitions may be problematic due to the controversy about which diagnostic tests are sufficient and reliable enough to rule out other causes of anaemia, and that anaemia causes are frequently combined. Therefore, for this review we accepted the diagnosis of anaemia defined by the authors of the studies.

Types of participants

Pregnant women with a diagnosis of anaemia (Hb levels under 11 g/dL, or other tests for anaemia as described by trialists) attributed to iron deficiency.

Types of interventions

We included the following interventions.

1. Oral iron.

2. Oral iron plus adjuncts.

3. Intramuscular (IM) iron.

4. Intravenous (IV) iron.

5. Blood transfusion.

6. Recombinent erythropoietin.

We looked for studies that compared these intervention against placebo and against each other. We also looked at different doses, routes of administration and regimens.

* For the purpose of this review, regular oral iron will include preparations different from controlled‐release oral iron.

Types of outcome measures

Primary outcomes

Women

Clinical outcomes

Mortality 
 Morbidity 
 Puerperal sepsis 
 Systemic bacterial infection after delivery 
 Days in intensive care unit 
 Days hospitalised during pregnancy

Newborn

Clinical outcomes

Mortality 
 Morbidity 
 Days hospitalised 
 Admission to neonatal intensive care unit

Secondary outcomes

Women

Clinical outcomes

Preterm labour 
 Premature delivery 
 Pneumonia 
 Postpartum haemorrhage (equal to or more than 500 ml) 
 Heart failure

Haematological outcomes

Maternal serum ferritin 
 Maternal serum iron 
 Hb levels 
 Long‐term haematological outcomes (not pre‐specified in original protocol)

Maternal side effects

General symptoms (e.g. weakness, tiredness, dizziness)Gastrointestinal effects (e.g. nausea, vomiting, diarrhoea, epigastric pain, constipation) 
 Local symptoms (e.g. pain or tenderness, erythema) 
 Systemic symptoms (e.g. myalgia, arthralgia, abscess formation at injection site, allergic reactions etc)

Newborn

Clinical outcomes

Low birthweight (less than 2500 g) 
 Respiratory disease requiring ventilation 
 Small‐for‐gestational age

Haematological outcomes

Cord serum ferritin 
 Cord Hb 
 Other long‐term outcomes

Search methods for identification of studies

Electronic searches

We searched the Cochrane Pregnancy and Childbirth Group's Trials Register by contacting the Trials Search Co‐ordinator (7 June 2011).

The Cochrane Pregnancy and Childbirth Group’s Trials Register is maintained by the Trials Search Co‐ordinator and contains trials identified from: 

1. quarterly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);

2. weekly searches of MEDLINE;

3. weekly searches of EMBASE;

4. handsearches of 30 journals and the proceedings of major conferences;

5. weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.

Details of the search strategies for CENTRAL and MEDLINE, the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service can be found in the ‘Specialized Register’ section within the editorial information about the Cochrane Pregnancy and Childbirth Group. 

Trials identified through the searching activities described above are each assigned to a review topic (or topics). The Trials Search Co‐ordinator searches the register for each review using the topic list rather than keywords. 

In addition, we searched CENTRAL (The Cochrane Library 2011, Issue 2 of 4) and PubMed (1966 to June 2011) using the search strategies given in Appendix 1.

In order to further identify ongoing RCTs and unpublished studies we searched the following trials registers (2 May 2011) using the terms and phrases detailed in Appendix 1.

• Search Portal of the International Clinical Trials Registry Platform (ICTRP; www.who.int/ictrp/) for appropriate ongoing and recently completed studies

• Health Technology Assessment Program (HTA) (www.hta.ac.uk/) (United Kingdom)

• LATINREC [www.latinrec.net], Colombia

For details of searching carried out in the previous version of the review, see:Appendix 2

Searching other resources

We searched the bibliographies of all papers identified by these strategies.

We did not apply any language restrictions.

Data collection and analysis

For the methods used when assessing the trials identified in the previous version of this review, seeAppendix 3.

For this update, we used the following methods when assessing the reports identified by the updated search.

Selection of studies

Two review authors (Ludovic Reveiz (LR), Alexandra Casasbuenas (AC)) independently assessed for inclusion all the potential studies we identified as a result of the search strategy. We resolved any disagreement through discussion or, if required, we consulted a third person (Luis Gabriel Cuervo (LGC)).

Data extraction and management

We designed a form to extract data. For eligible studies, two review authors (LR, AC) extracted the data using the agreed form. We resolved discrepancies through discussion or, if required, we consulted another review author (LGC). We entered data into Review Manager software (RevMan 2011) and checked for accuracy.

When information regarding any of the above was unclear, we attempted to contact authors of the original reports to provide further details.

Assessment of risk of bias in included studies

Two review authors (LR, AC) independently assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved any disagreement by discussion.

(1) 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 (any truly random process, e.g. random number table; computer random number generator);

• high risk (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number);

• unclear risk.   

 (2) Allocation concealment (checking for possible selection bias)

We described for each included study the method used to conceal the allocation sequence and determine whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment.

We assessed the methods as:

• low risk (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);

• high risk (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);

• unclear risk.   

(3) Blinding (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 could not have affected the results. We assessed blinding separately for different outcomes or classes of outcomes.

We assessed the methods as:

• low risk, high risk or unclear risk for participants;

• low risk, high risk or unclear risk for personnel;

• low risk, high risk or unclear risk for outcome assessors.

(4) Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations)

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, 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 re‐included missing data in the analyses which we undertook. We assessed methods as:

• low risk;

• high risk;

• unclear risk.

(5) Selective 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 (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 (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 were reported incompletely and so could not be used; study failed to include results of a key outcome that would have been expected to have been reported);

• unclear risk.

(6) Other sources of bias

We described for each included study any important concerns we have about other possible sources of bias.

We assessed whether each study was free of other problems that could put it at risk of bias:

• low risk;

• high risk;

• unclear risk.

(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 2011). With reference to (1) to (6) above, we assessed the likely magnitude and direction of the bias and whether we considered it was likely to impact on the findings.  We explored the impact of the level of bias through undertaking sensitivity analyses ‐ seeSensitivity analysis.

Measures of treatment effect

Dichotomous data

For dichotomous data, we presented results as summary risk ratio (RR) with 95% confidence intervals (CIs). 

Continuous data

For continuous data, we used the mean difference (MD) if outcomes were measured in the same way between trials. We used the standardised mean difference (SMD) to combine trials that measured the same outcome, but used different methods.  

Dealing with missing data

For included studies, we noted levels of attrition. We explored the impact of including studies with high levels of missing data in the overall assessment of treatment effect by using sensitivity analysis.

For all outcomes, we carried out analyses, as far as possible, on information about retaining participants in their original assigned groups, i.e. we attempted to include all participants randomised to each group in the analyses, and all participants were analysed in the group to which they were allocated, regardless of whether or not they received the allocated intervention. The denominator for each outcome in each trial was the number randomised minus any participants whose outcomes are known to be missing.

Assessment of heterogeneity

We assessed statistical heterogeneity in each meta‐analysis using the T², I² and Chi² statistics. We regarded heterogeneity as substantial if T² was greater than zero and either I² was greater than 30% or there is a low P value (less than 0.10) in the Chi² test for heterogeneity.

Assessment of reporting biases

If there had been 10 or more studies in the meta‐analysis we planned to investigate reporting biases (such as publication bias) using funnel plots. We would have assessed funnel plot asymmetry visually, and used formal tests for funnel plot asymmetry. For continuous outcomes we would have used the test proposed by Egger 1997, and for dichotomous outcomes, the test proposed by Harbord 2006. If asymmetry had been detected in any of these tests, or was suggested by a visual assessment, we would have performed exploratory analyses to investigate it.

Data synthesis

We carried out statistical analysis using the Review Manager software (RevMan 2011). We used fixed‐effect meta‐analysis for combining data where it was reasonable to assume that studies were estimating the same underlying treatment effect: i.e. where trials were examining the same intervention, and the trials’ populations and methods were judged sufficiently similar. If there was clinical heterogeneity sufficient to expect that the underlying treatment effects differed between trials, or if substantial statistical heterogeneity was detected, we used random‐effects meta‐analysis to produce an overall summary if an average treatment effect across trials was considered clinically meaningful. The random‐effects summary was treated as the average 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.

Where we used random‐effects analyses, the results were presented as the average treatment effect with its 95% CI, and the estimates of  T² and I².

Subgroup analysis and investigation of heterogeneity

Had we identify substantial heterogeneity, we would have investigated it using subgroup analyses and sensitivity analyses. We would have considered whether an overall summary was meaningful, and if it was, used random‐effects analysis to produce it.

We had planned to carry out the following subgroup analyses:

1. level of anaemia (mild, moderate, severe).     

The following outcomes would have been used in subgroup analysis:

1. maternal and newborn mortality;

2. maternal and newborn morbidity (puerperal sepsis, days hospitalised during pregnancy, low birthweight, newborn hospitalisations days);

3. Hb levels.

For fixed‐effect inverse variance meta‐analyses, we would have assessed differences between subgroups by interaction tests. For random‐effects and fixed‐effect meta‐analyses using methods other than inverse variance, we would have assessed differences between subgroups by inspection of the subgroups’ CIs; non‐overlapping CIs indicate a statistically significant difference in treatment effect between the subgroups.

Sensitivity analysis

We had planned to carry out the following sensitivity analyses but there were insufficient data:

1. according to risk of bias assessment.

Results

Description of studies

Results of the search

The search identified 117 references: two unpublished trials, six congress abstracts, and 109 published trials. An initial trawl through this list (LR excluded 17 references of non‐randomised controlled trials (RCTs). This left 100 trials for a more detailed evaluation. Seventy three studies were further excluded after first review because they were not RCTs; included mostly non‐anaemic women; evaluated postpartum iron treatments; focused on non iron‐deficiency anaemia; or had methodological flaws that seriously compromised their validity or resulted in insufficient useful reliable information (Characteristics of excluded studies). Two review authors (LR and AC) independently checked the trials against the inclusion criteria, and a third author (LGC) acted as the arbiter. We actively tried to contact the authors using contact information provided in their articles and on the Internet. We contacted and received responses from the authors listed in the following articles: Breymann 2001; De Souza 2004; Mumtaz 2000; Singh 1998; Suharno 1993; Visca 1996. We did not receive a response to our communications from the authors of the articles listed as Al Momen 1996; Khalafallah 2010; Sarkate 2007; Ruangvutilet 2009. We were unable to contact the authors for two articles (Stein 1991; Wu 1998).

Included studies

We included 23 RCTs in the review involving 3198 women (Al 2005; Bayoumeu 2002; Breymann 2001; Dawson 1965; De Souza 2004; Digumarthi 2008; Kaisi 1988; Khalafallah 2010; Komolafe 2003; Kumar 2005; Mumtaz 2000; Nappi 2009; Ogunbode 1980; Oluboyede 1980; Saha 2007; Singh 1998; Sood 1979; Suharno 1993;Sun 2010; Symonds 1969; Wali 2002; Zhou 2007; Zutschi 2004). Most focused on laboratory results rather than clinical outcomes. Clinical outcomes were assessed in seven RCTs (Al 2005; Bayoumeu 2002; Breymann 2001; Oluboyede 1980; Singh 1998; Zhou 2007; Zutschi 2004).Although Breymann and Singh's data were unpublished, these data were provided by the main author of Singh 1998 and have been incorporated into the review. LR and AC independently extracted data from the articles. LGC was expected to act as arbiter if differences arose in the data extraction, but this did not happen. LR performed data entry, and G Gyte (GG) and AC checked data entries for accuracy. Six RCTs were included in this update (Digumarthi 2008; Khalafallah 2010; Nappi 2009; Saha 2007; Sun 2010; Zhou 2007). In addition, we have added three reports to Studies awaiting classification because authors did not provide useful data for analysis (Paleru 2011; Pandit 2009; Sarkate 2007) and one study was ongoing (Ruangvutilet 2009). We contacted authors by email but only the authors of the ongoing RCT replied informing us that they were still at the stage of analysis (Ruangvutilet 2009). 
 
 Seven groups of RCTs were described according to the type of intervention. However, groups were further divided according to co‐interventions, dose, regimen, route, or type of chemical components of the intervention (i.e. iron sucrose, dextran), as follows.

(1) Oral iron

• Oral iron versus placebo (Suharno 1993;Sun 2010; Symonds 1969)

• Oral iron plus vitamin A versus placebo (Suharno 1993)

• Oral iron plus vitamin A versus oral iron (Suharno 1993)

• Oral iron versus bovine lactoferrin (Nappi 2009)

(2) Different regimens of oral iron treatment

• Daily oral iron versus twice‐weekly (De Souza 2004; Mumtaz 2000)

• Daily oral iron versus once a week (De Souza 2004)

• Twice‐weekly iron versus once weekly iron (De Souza 2004)

• 600 mg oral iron versus 1200 mg oral iron (Ogunbode 1980)

• Controlled‐release oral iron versus regular oral iron (Symonds 1969)

• Iron polymaltose versus ferrous sulphate (Saha 2007)

• Elemental iron 20; 40 or 80 mg (Zhou 2007)

(3) Intramuscular (IM) iron

• IM iron sorbitol versus IM dextran (Dawson 1965)

• IM iron sorbitol versus intravenous (IV) iron dextran (Oluboyede 1980)

(4) Intravenous (IV) iron

• IV route versus placebo (Symonds 1969)

(5) Parenteral route (IM or IV) versus oral route

• IM versus oral iron treatment (Komolafe 2003; Kumar 2005; Ogunbode 1980; Zutschi 2004)

• IV versus oral iron treatment (Al 2005; Bayoumeu 2002; Digumarthi 2008; Khalafallah 2010; Singh 1998; Sood 1979; Symonds 1969)

(6) IV iron versus IM iron with different regimens of parenteral iron treatment

• IV iron versus IM iron (Oluboyede 1980)

• Different IM preparations (Dawson 1965)

• IV iron versus IM iron (Dawson 1965)

• IV iron with hydrocortisone versus IV iron (Dawson 1965)

• Two differing IV doses (Kaisi 1988)

• IV iron versus IM iron (Wali 2002)

(7) IV administered iron sucrose with and without adjuvant recombinant human erythropoietin (Breymann 2001)

For details of included studies see:Characteristics of included studies table. We have two studies awaiting classification, see Characteristics of studies awaiting classification). There is one ongoing study (Ruangvutilet 2009).

Excluded studies

The reasons for excluding studies are reported in Characteristics of excluded studies.

Risk of bias in included studies

We (LR, AC) assessed the risk of bias of the included studies independently as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved differences in interpretations by consensus among three review authors (LR, AC, LGC), after checking the criteria agreed in the original review protocol. When RCTs had potential validity or interpretation problems and just part of the data were deemed useful, we would only use such data. For example, when RCTs had high withdrawal rates, and therefore, incomplete data on outcomes at the end of follow‐up, but still offered complete data at a given time that fulfilled our pre‐defined inclusion criteria, we used the later data. A summary of the assessment can be viewed in Figure 1 and Figure 2.

1.

Methodological quality graph: review authors' judgements about each methodological quality item presented as percentages across all included studies.

2.

Methodological quality summary: review authors' judgements about each methodological quality item for each included study.

Allocation

Twelve out of 23 RCTs reported on how the randomisation sequence was generated (Al 2005; Bayoumeu 2002; Breymann 2001; Dawson 1965; Kaisi 1988; Khalafallah 2010; Komolafe 2003; Mumtaz 2000; Nappi 2009 ; Singh 1998; Suharno 1993; Zhou 2007); no information was available for the remaining RCTs.

Seven out of 23 studies reported adequate allocation concealment (Al 2005; Breymann 2001; Khalafallah 2010; Mumtaz 2000; Singh 1998; Suharno 1993; Zhou 2007). Published details of the randomisation were insufficient in the Singh 1998 and Breymann 2001 articles, but additional details were provided by the authors upon request.

The allocation strategy and concealment were considered adequate in six of the 23 studies (Al 2005; Breymann 2001; Khalafallah 2010; Mumtaz 2000; Suharno 1993; Zhou 2007).

Blinding

In most RCTs, blinding was not used; these were open RCTs. Six RCTs described blinding (masking), three of which were adequately reported (Nappi 2009; Saha 2007; Zhou 2007). However, in two RCTs it was unclear whether the participants or healthcare providers were blinded to the interventions (Mumtaz 2000; Sun 2010; Suharno 1993); both RCTs assessed oral administration.

Incomplete outcome data

Withdrawal rates (drop outs and losses to follow up) were reported in eleven RCTs (Al 2005; Breymann 2001;Khalafallah 2010; Komolafe 2003; Nappi 2009; Ogunbode 1980; Singh 1998;Sun 2010; Symonds 1969; Zhou 2007; Zutschi 2004).

• Less than 5%: withdrawal rates were lower than 5% in five RCTs (Nappi 2009; Oluboyede 1980; Sood 1979; Sun 2010; Zhou 2007).

• 5% to 9.9%: an RCT from Pakistan (Wali 2002) had five withdrawals (8.3%) due to intolerance in the IM iron group. An RCT from France had three withdrawals (6%) (Bayoumeu 2002) and the RCT from Australia had 17 withdrawals (8.5%) (Khalafallah 2010).

• 10% to 19.9%: the West Java RCT (Suharno 1993) had complete data available on 251 (83%) women: reasons for withdrawals are further described in the article.

• More than 20%: an RCT from Pakistan (Mumtaz 2000) recruited 191 women; of these, 160 were successfully followed for at least four weeks and supplemented for an average of 10.9 weeks. Fifty‐five per cent completed the entire duration of follow‐up; 15% of the women recruited did not return for a single visit and were excluded from the analysis. The remaining 30% did not complete the entire 12 weeks of planned follow‐up. No significant differences were found for population characteristics (age, socio‐economic status score, parity, time since last pregnancy, body mass index, initial Hb, dependants or family and the duration of follow‐up) between women who withdrew and those who completed the study. The RCT from Brazil (De Souza 2004) had 41 (21.5%) women who withdrew or were lost to follow up; the reasons were described in the article. The analysis was done using data at 16 weeks of treatment. The UK RCT (Dawson 1965) had high rates of losses to follow up. The RCT focused on assessing adverse effects. The RCT from Tasmania (Kaisi 1988) had high withdrawal rates for most outcomes (loss to follow up for Hb result was 47%) and only data on adverse effects were used for this review. The RCT from India (Kumar 2005) recruited 220 women of whom 150 (68%) completed the study. No significant differences were found on data of initial haematological parameters, gestation, parity or literacy between women who completed the study and women who withdrew. However, withdrawals were different for women receiving oral treatment (13.5%) and those receiving IM treatment (38.5%).

A number of studies did not adequately report the outcomes (i.e. standard deviations were not reported) (seeRisk of bias in included studies).

Selective reporting

Relevant neonatal and maternal outcomes were not frequently reported by most studies (seeRisk of bias in included studies). The protocol was not available in any study (Figure 1; Figure 2). Only two studies (Khalafallah 2010; Zhou 2007) were registered in an international clinical trial registry.

Effects of interventions

Twenty‐three RCTs, involving 3198 women, met the inclusion criteria. Overall, we found insufficient assessment of the outcomes relevant to the focus of this review, especially of clinical outcomes. Most results were provided by one or two small RCTs with methodological limitations. The effect size for these are represented in this review using the risk ratio (RR) and mean difference (MD). Uncertainty levels are quantified using 95% confidence intervals (CI).

(1) Oral iron

Oral iron versus placebo (comparison 01)

We found three RCTs involving 266 women (Suharno 1993; Sun 2010; Symonds 1969). Data from the first RCT showed that women receiving iron (ferrous sulphate) had a lower risk of being anaemic during the second trimester (RR 0.38; 95% CI 0.26 to 0.55, one RCT, 125 women, Analysis 1.1). In the group receiving iron, the mean Hb level was higher (MD 1.34; 95% CI 0.0.27 to 2.41; two RCT, 215 women;I² 98% Analysis 1.2) (Suharno 1993; Sun 2010); both trials had individual significant differences favouring oral iron treatments. Similarly, the mean serum ferritin was higher for women receiving iron (MD 0.70; 95% CI 0.52 to 0.88; one RCT, 125 women; Analysis 1.3). A trend towards increased adverse effects (for example, nausea, vomiting, constipation and abdominal cramps) was also noticed in the second RCT (ferrous gluconate), but figures were small to allow worthy comparisons (11/51 women with adverse effects). No other assessments were found for clinical outcomes. Hence, it is difficult to establish the clinical effects of treatments in women and newborns. Conclusions need to be approached with care because they are drawn from a small sample of participants. Furthermore, one RCT assessed outcomes at the second trimester (Suharno 1993) and it is unclear if those women sustained similar Hb levels during the rest of their pregnancy, and no assessment of haematological results was done at delivery.

1.1. Analysis.

Comparison 1 Oral iron versus placebo, Outcome 1 Anaemic during 2nd trimester.

1.2. Analysis.

Comparison 1 Oral iron versus placebo, Outcome 2 Haemoglobin levels (g/dL).

1.3. Analysis.

Comparison 1 Oral iron versus placebo, Outcome 3 Ferritin levels (ug/l).

Oral iron plus vitamin A versus placebo (comparison 02)

We found one RCT involving 125 women (Suharno 1993). It included anaemic women with a high risk of suffering vitamin A deficiency. Adding vitamin A to regular iron (ferrous sulphate), resulted in improved Hb levels. Anaemia during the second trimester was lower with oral iron plus vitamin A, compared with placebo (RR 0.04; 95% CI 0.01 to 0.15; one RCT, 125 women; Analysis 2.1). The difference was not as large when the comparator was iron therapy only (see below). The applicability of these results may be limited to women in populations with vitamin A deficiency.

2.1. Analysis.

Comparison 2 Oral iron + vitamin A versus placebo, Outcome 1 Anaemic during 2nd trimester.

Oral iron versus bovine lactoferrin (comparison 25)

One RCT involving 97 women (Nappi 2009) found a significant reduction in the mean level of anaemia at one month with oral ferrous sulphate compared with bovine lactoferrin (MD ‐0.30 95% CI ‐0.52 to ‐0.08; one RCT, 97 women, Analysis 25.1).

25.1. Analysis.

Comparison 25 Oral bovine lactoferrin versus ferrous sulphate, Outcome 1 Mean haemoglobin levels at 1 month.

(2) Different regimens of oral iron treatment (comparisons 15, 16, 17 and 21)

Daily oral iron versus twice‐weekly (comparison 15)

An RCT from Pakistan had a loss to follow up greater than 20% for most of the outcomes and so validity of the findings is unclear (Mumtaz 2000). The study found that daily oral iron (ferrous sulphate) significantly increased Hb levels at four weeks, eight weeks, and 12 weeks, compared with twice‐weekly oral iron. At 12 weeks, the mean Hb level was 11.36 g/dL compared with 10.09 g/dL, respectively (MD 1.27; 95% CI 0.68 to 1.86; one RCT, 105 women; Analysis 15.3). In women receiving daily versus twice‐weekly oral iron therapy (ferrous sulphate), an RCT from Brazil (De Souza 2004) found no significant difference in Hb levels (MD 0.30; 95% CI ‐0.01 to 0.61; one RCT, 102 women; Analysis 15.4) or anaemia (RR 1.38; 95% CI 0.86 to 2.23; one RCT, 102 women; Analysis 15.5) at 16 weeks of treatment. A trend was found between higher doses of iron and reported adverse effects (19/48 (40%) for 1/week, 24/53 (45%) for twice/week and 35/49 (71%) for daily treatment). No further description of adverse effects was provided.

15.3. Analysis.

Comparison 15 Oral iron daily versus oral iron twice weekly, Outcome 3 Haemoglobin level at 12 weeks.

15.4. Analysis.

Comparison 15 Oral iron daily versus oral iron twice weekly, Outcome 4 Haemoglobin level at 16 weeks.

15.5. Analysis.

Comparison 15 Oral iron daily versus oral iron twice weekly, Outcome 5 Haemoglobin level > 11 g/dL at 16 weeks of treatment.

Daily oral iron versus once‐weekly (comparison 16)

One RCT from Brazil also had a loss to follow up greater than 20% and so the findings also have uncertain validity (De Souza 2004). The study found that daily oral treatment (ferrous sulphate) increased Hb levels after 16 weeks of treatment, compared with weekly oral iron (MD 0.70; 95% CI 0.36 to 1.04; one RCT, 97 women; Analysis 16.1), the proportion of women non‐anaemic at the end of the follow‐up (RR 1.73; 95% CI 1.00 to 3.01; one RCT, 97 women; Analysis 16.2) and reduced treatment failure (RR 0.05; 95% CI 0.01 to 0.35; one RCT, 97 women; Analysis 16.3).

16.1. Analysis.

Comparison 16 Oral iron daily versus oral iron once week, Outcome 1 Haemoglobin level at 16 weeks.

16.2. Analysis.

Comparison 16 Oral iron daily versus oral iron once week, Outcome 2 Haemoglobin level > 11 g/dL at 16 weeks of treatment.

16.3. Analysis.

Comparison 16 Oral iron daily versus oral iron once week, Outcome 3 Treatment failure (haemoglobin < 10 g/dL) at 16 weeks.

Twice‐weekly iron versus once‐weekly iron (comparison 17)

The same RCT from Brazil (De Souza 2004) found that a twice‐weekly regimen of ferrous sulphate resulted in a modest increase in Hb levels, compared with a weekly iron regimen (MD 0.40; 95% CI 0.03 to 0.77; one RCT, 101 women; Analysis 17.1) and reduced treatment failure (RR 0.32; 95% CI 0.15 to 0.68; one RCT, 101 women; Analysis 17.3). However, no significant differences were found in the proportion of women non‐anaemic at 16 weeks of treatment (RR 1.25; 95% CI 0.69 to 2.28; one RCT, 101 women; Analysis 17.2). There is again a high risk of bias due to a loss to follow up of greater than 20%.

17.1. Analysis.

Comparison 17 Oral iron twice week versus oral iron once week, Outcome 1 Haemoglobin level at 16 weeks.

17.3. Analysis.

Comparison 17 Oral iron twice week versus oral iron once week, Outcome 3 Treatment Failure (haemoglobin < 10 g/dL) at 16 weeks.

17.2. Analysis.

Comparison 17 Oral iron twice week versus oral iron once week, Outcome 2 Haemoglobin level > 11 g/dL at 16 weeks of treatment.

600 mg oral iron versus 1200 mg oral iron (comparison 21)

An RCT conducted in Nigeria (Ogunbode 1980) found no significant differences in Hb levels at four weeks ( MD 0.37; 95% CI ‐0.77 to 1.51; one RCT, 56 women; Analysis 21.1) and eight weeks (MD 0.02; 95% CI ‐1.03 to 1.07; one RCT, 56 women; Analysis 21.2) of treatment between women receiving 600 mg versus 1200 mg of oral ferrous sulphate. All women received daily 5 mg of folic acid and 25 mg of pyrimethamine daily, in addition to ferrous sulphate.

21.1. Analysis.

Comparison 21 Oral ferrous sulphate iron 1200 mg/day versus 600 mg/day, Outcome 1 Haematocrit (%) at 4 weeks of treatment.

21.2. Analysis.

Comparison 21 Oral ferrous sulphate iron 1200 mg/day versus 600 mg/day, Outcome 2 Haematocrit (%) at 8 weeks of treatment.

Controlled‐release oral iron versus regular oral iron (comparison 03)

One RCT conducted in Australia (Symonds 1969) compared controlled‐release oral iron versus other iron preparations. It provided information on adverse effects, but data on effectiveness were not included because it had a very high withdrawal rate. It found no differences in nausea and vomiting, constipation and abdominal cramps at one month between controlled‐release iron and regular oral iron. The small sample size and broad CIs illustrate that the sample size is clearly insufficient to rule out any difference (Analysis 3.1 to Analysis 3.4).

3.1. Analysis.

Comparison 3 Controlled‐release oral iron versus regular oral iron, Outcome 1 Side effects.

3.4. Analysis.

Comparison 3 Controlled‐release oral iron versus regular oral iron, Outcome 4 Abdominal cramps.

Iron polymaltose versus ferrous sulphate (comparison 24)

One RCT from India (Saha 2007) that included 100 women compared iron polymaltose (100 mg elemental iron + folic acid 500 mcg for eight weeks versus ferrous sulphate (60 mg elemental iron + folic acid 500 mcg for eight weeks). No significant differences were found in the mean level of Hb and in the proportion of women with Hb greater than 11 g/dL between groups. Concerning adverse events, a significant difference was found in constipation (RR 0.30; 95% CI 0.14 to 0.64; one RCT, 100 women, Analysis 24.2), gastrointestinal intolerance (RR 0.41; 95% CI 0.25 to 0.69; one RCT, 100 women, Analysis 24.3) that favoured the iron polymaltose groups. However, no significant differences were found for other adverse events (metallic taste, diarrhoea, rashes), compliance, the cost due to hospital visits and the cost due to loss of work.

24.2. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 2 Constipation at 8 weeks.

24.3. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 3 Haemoglobin > 11 g/dL at 8 weeks of treatment.

Ferrous sulphate (elemental iron 20 versus 40 versus 80 mg) (comparisons 26, 27 and 28)

One RCT conducted in Australia (Zhou 2007) that included 180 women compared three different doses of ferrous sulphate elemental iron.

(3) Intramuscular (IM) iron

We found no RCTs comparing IM iron versus placebo.

IM iron sorbitol versus IM dextran (comparison 04)

We found one RCT that recruited 74 women (Dawson 1965). It did not provide effectiveness figures and had high withdrawal rates, so the findings are at high risk of bias. It found that iron sorbitol produced less skin discolouration (RR 0.21; 95% CI 0.07 to 0.65; one RCT, 48 women; Analysis 4.2) and fewer headaches (RR 0.13; 95% CI 0.02 to 0.99; one RCT, 48 women; Analysis 4.5) than IM dextran. These findings are inconclusive given the limitations of this single study and the high loss to follow up.

4.2. Analysis.

Comparison 4 Intramuscular iron sorbito‐citric acid versus intramuscular dextran, Outcome 2 Skin discolouration at injection site.

4.5. Analysis.

Comparison 4 Intramuscular iron sorbito‐citric acid versus intramuscular dextran, Outcome 5 Headaches.

IM iron sorbitol versus intravenous (IV) iron dextran (comparison 23)

We found one RCT involving 63 women conducted in Nigeria (Oluboyede 1980). It found that IM iron sorbitol increased haematocrit after four weeks of treatment (MD 2.18; 95% CI 0.77 to 3.59; one RCT, 59 women; Analysis 23.1) and after eight weeks of treatment (MD 1.48; 95% CI 0.15 to 2.81; one RCT, 43 women; Analysis 23.2), compared with IV iron dextran.

23.1. Analysis.

Comparison 23 Intramuscular iron sorbitol‐glu acid versus intravenous iron dextran, Outcome 1 Haematocrit (%) at 4 weeks of treatment.

23.2. Analysis.

Comparison 23 Intramuscular iron sorbitol‐glu acid versus intravenous iron dextran, Outcome 2 Haematocrit (%) at 8 weeks of treatment.

(4) Intravenous (IV) iron

IV route versus placebo (comparison 07)

We found one small RCT involving 54 women and conducted in Australia (Symonds 1969). The RCT provided data on adverse effects. Data on effectiveness were not included. It found no significant differences between IV iron and placebo for: nausea and vomiting (RR 0.33; 95% CI 0.01 to 7.84; one RCT, 54 women; Analysis 7.2), abdominal cramps (not estimable), and constipation (RR 0.25; 95% CI 0.03 to 2.09; one RCT, 54 women; Analysis 7.3). However, the small sample size and broad CIs illustrate that the sample size is clearly insufficient to rule out any such adverse effects.

7.2. Analysis.

Comparison 7 Intravenous iron versus placebo, Outcome 2 Nausea or vomiting.

7.3. Analysis.

Comparison 7 Intravenous iron versus placebo, Outcome 3 Constipation.

(5) Parenteral route (IM or IV) versus oral route

IM versus oral iron treatment (comparisons 12, 13 and 14)

We found four RCTs (571 women) comparing IM and oral administration of iron (Komolafe 2003; Kumar 2005; Ogunbode 1980; Zutschi 2004).

The first RCT, from India, (Zutschi 2004) evaluated 150 mg IM iron sorbitol (via three injections a day) at four‐weekly intervals versus 100 mg of elemental oral iron for at least 100 days. IM iron significantly increased Hb (MD 0.54; 95% CI 0.30 to 0.78; one RCT, 200 women; Analysis 12.2), and haematocrit levels (MD 1.40; 95% CI 0.67 to 2.13; one RCT, 200 women; Analysis 12.3), compared with oral iron. A higher proportion of women were found to be non‐anaemic at labour (RR 1.23; 95% CI 1.01 to 1.48; one RCT, 200 women; Analysis 12.1). Adverse effects were not included in the reports of the article.

12.2. Analysis.

Comparison 12 Intramuscular iron sorbitol citric acid versus oral iron, Outcome 2 Mean maternal haemoglobin at birth.

12.3. Analysis.

Comparison 12 Intramuscular iron sorbitol citric acid versus oral iron, Outcome 3 Mean maternal hematocrit level at birth.

12.1. Analysis.

Comparison 12 Intramuscular iron sorbitol citric acid versus oral iron, Outcome 1 Not anaemic at term.

The second RCT, from India, compared IM sorbitol citric acid dose versus oral ferrous sulphate (100 mg of elemental iron) plus 5 mg of folic acid at 36 weeks of pregnancy (Kumar 2005). Women receiving oral iron plus folic acid had a higher Hb level (MD 0.26; 95% CI 0.04 to 0.48; one RCT, 150 women; Analysis 14.1). No significant differences were found for caesarean section rates or mean birthweight. Adverse effects were reported by 40 women receiving IM treatment versus 16 receiving oral treatment at 36 weeks of treatment (Analysis 14.3 to Analysis 14.15). Gastrointestinal side effects (dyspepsia, constipation, diarrhoea, vomiting) were observed predominately in the oral group, while systemic reactions (local pain, staining, fever, systemic ache, arthralgia, itching and rash, immediate headache, malaise and vaso‐vagal due to apprehension (it is unclear what the authors meant by this and what diagnostic criteria they used) were more frequently found in women receiving IM iron. No anaphylactic reaction or abscess formation were observed, but too few women participated in the RCT to assess these and other important adverse effects. 
 
 The third RCT, from Nigeria, (Ogunbode 1980) was a three‐arm RCT comparing iron sorbitol versus 600 mg oral ferrous sulphate versus 1200 mg oral ferrous sulphate. All women received a daily supplement of 5 mg of folic acid and 25 mg of pyrimethamine. After eight weeks, IM iron sorbitol had significantly improved haematocrit levels compared with 600 mg of oral iron (MD 2.62; 95% CI 1.26 to 3.98; one RCT, 59 women; Analysis 12.6), and compared with 1200 mg of oral iron (MD 2.60; 95% CI 1.02 to 4.18; one RCT, 59 women; Analysis 12.8). Adverse effects were not assessed.

14.1. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 1 Mean haemoglobin at 36 weeks.

14.3. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 3 Caesarean section.

14.15. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 15 Systemic ache.

12.6. Analysis.

Comparison 12 Intramuscular iron sorbitol citric acid versus oral iron, Outcome 6 Haematocrit (%) at 8 weeks of treatment.

12.8. Analysis.

Comparison 12 Intramuscular iron sorbitol citric acid versus oral iron, Outcome 8 Haematocrit (%) at 8 weeks of treatment.

The fourth RCT, conducted in Nigeria, compared IM iron dextran (250 mg iron dextran thrice‐weekly until total calculated dose was given) versus 600 mg of oral ferrous sulphate plus vitamin C and folic acid (Komolafe 2003). It found that iron dextran significantly improved haematocrit levels after six weeks (MD 4.47; 95% CI 3.67 to 5.27; one RCT, 60 women; Analysis 13.1) and the proportion of non‐anaemic women after six weeks (RR 11.00; 95% CI 1.51 to 79.96; one RCT, 60 women; Analysis 13.2).

13.1. Analysis.

Comparison 13 Intramuscular iron dextran versus oral iron + vitamin C + folic acid, Outcome 1 Haematocrit.

13.2. Analysis.

Comparison 13 Intramuscular iron dextran versus oral iron + vitamin C + folic acid, Outcome 2 Not anaemic at 6 weeks (packed cell volume > 33%).

IV versus oral iron treatment (comparisons 08 and 09)

Pooled estimates (Al 2005; Bayoumeu 2002; Digumarthi 2008) for Hb levels at four weeks favoured IV iron (average MD 0.44; 95% CI 0.05 to 0.82; random‐effects [X² = 3.43; T² = 0.05; P = 0.18; I² 42%]; three RCTs, 167 women; Analysis 8.11). Diarrhoea was less frequent in women receiving IV iron (RR 0.16; 95% CI 0.03 to 0.86; three RCTs, 237 women; Analysis 8.5). 
 
 A French RCT compared IV iron sucrose given in six slow IV injections on days one, four, eight, 12, 15 and 21 according to a formula described in the article, with 240 mg of elemental iron sulphate tablets (Bayoumeu 2002); all women received 15 mg of folic acid in addition to iron. No significant differences were found in maternal Hb levels at four weeks of treatment, Hb levels in excess of 12 g/dL, neonatal Hb, ferritin levels, and birthweight. Similarly, no significant differences were found in the incidence of diarrhoea, postpartum haemorrhage, blood transfusion required, or neonatal mortality. The RCTs were underpowered to assess these outcomes properly.

8.11. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 11 Mean maternal haemoglobin at 4 weeks ( g/dL).

8.5. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 5 Diarrhoea.

An RCT conducted in Turkey (Al 2005) compared IV iron sucrose calculated according to a formula described in the article (total dose was administered over five days and maximum daily dose administered was 400 mg elemental iron) versus 300 mg of elemental iron (polymaltose complex); all women were given 5 mg of folic acid daily. It found that IV iron significantly increased maternal Hb at four weeks (MD 0.68; 95% CI 0.39 to 0.97; one RCT, 90 women; Analysis 8.11) and at birth (MD 0.75; 95% CI 0.34 to 1.16; one RCT, 90 women; Analysis 8.8) and increased the proportion of non‐anaemic women ‐ those with Hb levels equal or greater than 11 g/dL (RR 1.54; 95% CI 1.21 to 1.94; one RCT, 90 women; Analysis 8.25). No significant differences were found for caesarean section rates, neonatal birthweight, gestational hypertension, gestational diabetes and arthralgias (Analysis 8.2; Analysis 8.14; Analysis 8.18; Analysis 8.23; Analysis 8.24; Analysis 8.27).

8.8. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 8 Maternal haemoglobin at birth.

8.25. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 25 Haemoglobin level > 11 g/dL at birth.

8.2. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 2 Nausea or vomiting or epigastric discomfort.

8.14. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 14 Caesarean section.

8.18. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 18 Neonatal birthweight.

8.23. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 23 Gestational hypertension.

8.24. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 24 Gestational diabetes.

8.27. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 27 Arthralgia.

A comparison of oral ferrous fumarate 200 mg three times a day versus IV iron dextrin (calculated according to described formula) found that oral treatments increased constipation, compared with IV treatments (Singh 1998) (RR 0.04; 95% CI 0.00 to 0.61; one RCT, 100 women; Analysis 8.3). No significant differences were found for constipation when IV iron was compared with controlled‐release iron. However, only one small RCT (Symonds 1969) assessed this and it seemed to be underpowered to rule out clinically important effects (RR 0.22; 95% CI 0.03 to 1.85; one RCT, 51 women; Analysis 8.3). One RCT, recruiting mostly Malayan and Chinese women, found that higher Hb levels were found at the end of gestation with IV versus oral treatments (Singh 1998). However, the standard deviations are 50 to 100 times narrower than those found in other studies, raising questions about their validity (Al 2005; Suharno 1993). We excluded data from the analysis pending a response from the trial's authors. No maternal or neonatal deaths were recorded in this RCT, which was the only one specifically assessing these outcomes in women receiving oral or IV treatments.

8.3. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 3 Constipation.

Three RCTs (Al 2005; Singh 1998; Symonds 1969), including one that assessed controlled‐release iron (Symonds 1969), found that oral iron was more frequently associated with complaints of nausea than IV preparations, and the magnitude of the effects was consistent across all three RCTs (RR 0.33; 95% CI 0.15 to 0.74; three trials, 244 women; Analysis 8.2).

Two women were reported as suffering severe allergic reactions with IV dextran in an RCT comparing the latter with oral ferrous sulphate (Sood 1979). Data on other relevant outcomes were not available for comparison.

IV + oral iron versus oral iron (comparisons 29)

One RCT (Khalafallah 2010) assessed daily oral ferrous sulphate 250mg (elemental iron 80 mg) with or without a single intravenous iron polymaltose infusion. The combined treatment significantly increased the mean Hb (MD 0.48 95% CI 0.21 to 0.75, one trial, 183 women; Analysis 29.1) and in the proportion of women with mild or moderate anaemia predelivery (RR 0.55 95% CI 0.31 to 0.98) when compared with the oral treatment. Following delivery, the combined treatment also significantly increased the mean Hb (MD 0.39 95% CI 0.02 to 0.76, one trial, 183 women, Analysis 29.2). No significant differences were found in the newborn weights, gestational age at delivery, placental cord Hb or iron status between the two treatment groups. Two patients developed urticarial reactions after commencement of IV iron infusion. Trialist reported that there was a significant benefit in the IV plus oral iron group in terms of amelioration of symptoms attributed to anaemia (energy level, well being and physical activity) within a shorter period of time

29.1. Analysis.

Comparison 29 Intravenous iron + oral iron versus oral iron, Outcome 1 Mean predelivery maternal haemoglobin.

29.2. Analysis.

Comparison 29 Intravenous iron + oral iron versus oral iron, Outcome 2 Mean maternal haemoglobin after delivery.

(6) IV iron versus IM iron with different regimens of parenteral iron treatment (comparisons 04, 05, 06, 18, 20 and 23)

IV iron versus IM iron (comparison 23)

An RCT from Nigeria (Oluboyede 1980) found that IM sorbitol increased haematocrit levels compared with the IV dextran group at four weeks (MD 2.18; 95% CI 0.77 to 3.59, one RCT, 59 women, Analysis 23.1) and eight weeks (MD 1.48; 95% CI 0.15 to 2.81; one RCT, 43 women; Analysis 23.2). One women receiving IV iron suffered a severe allergic reaction whereas one participant of the IM group had viral hepatitis three months later. Authors reported that no significant differences in newborn weight and Apgar score at birth were found between the groups (no data were provided). Neonates were assessed for any complication at birth and within the first week of life; one neonate in each treatment group developed neonatal jaundice. Maternal outcomes were not reported for each group of treatment.

Different IM preparations (comparison 04)

One RCT compared two IM preparations (Dawson 1965). It found that women receiving IM iron‐sorbitol complex had a lower incidence of skin discolouration at injection sites at eight weeks (RR 0.21; 95% CI 0.07 to 0.65; one RCT, 48 women; Analysis 4.2) and fewer headaches (RR 0.13; 95% CI 0.02 to 0.99; one RCT, 48 women; Analysis 4.5) compared with IM iron dextran. Results should be interpreted with care as they come from a single, small RCT. However, this particular RCT had a robust randomisation and concealment strategy.

IV iron versus IM iron (comparison 05)

One factorial RCT conducted in the UK compared IM treatments with IV treatment (Dawson 1965). It found that IM iron was more frequently associated with pain in the injection site. This factorial design had some problems that were not addressed during the analysis: active treatments were compared with a single control group and no adjustments for multiple comparisons were done. This increases the possibilities of finding spurious associations. The RCT found a higher risk of skin discolouration in women receiving IM iron dextran compared with IV iron. Findings suggested a trend towards a higher risk of venous thrombosis with IV iron versus IM iron, but no statistical differences were found (RR 0.13; 95% CI 0.01 to 2.20; 4/26 with IV iron dextran (15%) versus 0/23 with IM iron; one RCT, 49 women; Analysis 5.3). However, this raises concern and an association can not be ruled out; the RCTs were underpowered to assess these outcomes properly, and these are very serious adverse effects.

5.3. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 3 Venous thrombosis.

The RCT found that IM iron dextran was not associated with higher complaints of headaches, compared with IV infusion of iron dextran (RR 3.96; 95% CI 0.91 to 17.17; one RCT, 49 women; Analysis 5.5). The RCT was too small to rule out important clinical differences in measured adverse effects outcomes such as shivering, itching, metallic taste in mouth, severe delayed allergic reaction (Analysis 5.4 to Analysis 5.9).

5.5. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 5 Headaches.

5.4. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 4 Nausea or vomiting.

5.9. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 9 Severe delayed allergic reaction.

Two differing IV doses of iron dextran (comparison 11)

An RCT conducted in Tanzania compared two doses of IV iron dextran by total dose infusion (Kaisi 1988). All participants were given the full dose recommended by the manufacturer; group A received an additional 10 ml whereas group B was given two‐thirds of that total dose. It found that allergic reactions after the infusion had finished were reduced with the lower dose (RR 0.62; 95% CI 0.45 to 0.86; one RCT, 623 women; Analysis 10.2). No significant differences were found for life threatening allergic reactions. This RCT was not used to assess effectiveness as it failed to fulfil our quality criteria.

10.2. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 2 Allergic reaction after infusion.

IV iron versus IM iron (comparisons 19 and 20)

An RCT conducted in Pakistan (Wali 2002) evaluated two doses of IV iron sucrose (500 mg versus 200 mg) and IM iron sorbitol. The participants were divided into three groups. In group A (n = 15), IV iron sucrose was administered intravenously according to the following formula: total iron deficit = body weight x (target Hb ‐ actual Hb) x 0.24 + 500; in group B (n = 20) IV iron sucrose was administered using the same formula but 200 mg of iron being given for storage instead of 500 mg; in group C, iron was administered IM daily or alternate days; after parenteral administration, oral iron therapy (ferrous gluconate 250 mg) was continued until the time of giving birth. No significant differences were found regarding Hb levels and the proportion of non‐anaemic women (with Hb levels greater than 11 g/dL at delivery) when the two different doses of IV iron were compared. Abdominal pain was reported by one woman in each group. Administration of IV iron sucrose (500 mg) significantly increased Hb levels (MD 1.60; 95% CI 0.87 to 2.33; one RCT, 40 women; Analysis 19.1) and the proportion of women with Hb greater than 11 g/dL at delivery (RR 2.86; 95% CI 1.45 to 5.63; one RCT, 40 women; Analysis 20.2) compared with IM iron sorbitol. Similarly, significant results favouring IV treatment were found when comparing IV iron sucrose (200 mg) and IM iron sorbitol for the same outcomes (MD 1.10; 95% CI 0.49 to 1.71; Analysis 20.1), and (RR 2.50; 95% CI 1.25 to 4.99; one RCT, 45 women; Analysis 20.2). In the IV groups 2/35 (5.7%) women had shivering and feeling of weakness within a few hours, and 3/35 (8.6%) had phlebitis at the site where IV canula was retained. In the IM group, 5/25 (20%) withdrew from the study due to intolerance (no further description was provided) and the majority complained of pain at the injection site.

19.1. Analysis.

Comparison 19 Intravenous iron sucrose 500 mg versus intramuscular iron sorbitol, Outcome 1 Maternal haemoglobin level at birth.

20.2. Analysis.

Comparison 20 Intravenous iron sucrose 200 mg versus intramuscular iron sorbitol, Outcome 2 Haemoglobin level > 11 g/dL at delivery.

20.1. Analysis.

Comparison 20 Intravenous iron sucrose 200 mg versus intramuscular iron sorbitol, Outcome 1 Haemoglobin level at delivery.

(7) IV administered iron sucrose with and without adjuvant recombinant human erythropoietin (comparison 11)

One small size study evaluated adjuvant recombinant human erythropoietin when iron sucrose was administered intravenously (Breymann 2001). No statistically significant differences were found in the number of women with a rise of Hb greater than 11 g/dL or caesarean delivery. The author provided unpublished data concerning birthweight and mean maternal blood pressure at the end of therapy; no significant differences were found (Analysis 11.6 and Analysis 11.8) for these outcomes. None of the women required additional antepartum or postpartum blood transfusion.

11.6. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 6 Birthweight.

11.8. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 8 Maternal mean blood pressure.

Discussion

Summary of main results

Although iron treatments consistently increase maternal haematological indices in women diagnosed with anaemia in pregnancy attributed to iron deficiency, we found no evidence that these laboratory improvements reflected in clinical improvements such as reduced preterm delivery, reduced infant low birthweight, lower rates of pre‐eclampsia, sepsis or postpartum haemorrhage and its complications (Scholl 1992; Scholl 2000). We found very few randomised controlled trials (RCTs) assessing clinical outcomes, and these RCTs were too small to estimate important clinical effects. Moreover, the studied populations turned out to be too small to deliver clear‐cut answers to this review's questions.

The findings suggest that gastrointestinal adverse effects are more frequent with oral iron treatments, compared with other routes of iron administration, although these gastrointestinal effects may be considered less problematic than the adverse effects of intravenous and intramuscular iron (see below). The results of one RCT suggest that daily iron treatment is better than intermittent iron supplementation in increasing Hb at delivery in pregnant women based in developing countries. Higher doses of iron were not associated with improved haematological values. In one RCT (Zhou 2007), daily low‐dose iron supplements were effective at treating anaemia in pregnancy with less gastrointestinal side effects compared with higher doses. Although women with higher iron doses had a higher mean Hb concentration at the end of intervention, all groups of treatment (low, intermediate and high doses) were within the normal reference range for women. The neonatal and maternal outcomes were similar among groups of intervention. The assessment of the effects of controlled versus regular oral iron were mostly inconclusive; there seems to be a reduced incidence of constipation. Most oral iron studies were marred by high withdrawal rates, highlighting the importance of assessing adverse effects and compliance issues with these frequently‐prescribed treatments.

The findings of this review suggest that adding vitamin A to regular iron (ferrous sulphate) resulted in improved Hb levels for women with vitamin A deficiency. Another Cochrane review that focused on vitamin A supplementation during pregnancy suggested beneficial effects for women in areas of poor nutritional intake (Van den Broek 2010).

Compared with oral iron, intramuscular (IM) iron sorbitol and iron dextran improved haematological values, reduced the proportion of women without anaemia, and resulted in lower gastrointestinal side effects. But these preparations were associated with higher rates of systemic reactions.

The findings of this review also suggest that intravenous (IV) iron sucrose is effective, but there is uncertainty whether it may increase the incidence of serious adverse effects such as thrombosis, which was frequent (9/41; 22%). Similarly, there are worrying trends towards an increased risk of severe allergic reaction with IV dextran iron, but data were few. One study suggests that the risk of venous thrombosis may be lowered by adding hydrocortisone to the infusion, but it is unclear what the real impact of this might be and whether it has any other effects. Evidence of a relationship between doses of IV iron and risk of adverse allergic reactions is inconclusive. No effectiveness assessments were done for the compared doses of IV drugs. Compared with IM iron sucrose, IV iron sucrose significantly increased haematological indices but it is unclear what the effects are on maternal and neonatal morbidity. RCTs were insufficient to determine the clinical effects of treatments in women with iron‐deficiency anaemia during pregnancy.

Overall completeness and applicability of evidence

Although there were a reasonable number of RCTs and women in this review, the data are incomplete for a number of clinically important outcomes. In addition, there were multiple comparisons with few RCTs leading to limited opportunities to pool data. The applicability of the evidence outside the research setting is reasonable as all these studies were conducted in the clinical settings that are quite similar. The comparisons in the review are commonly undertaken and not difficult to apply. The 23 RCTs came from many different countries.

Quality of the evidence

The objective of this review was to address the effects of iron anaemia treatments on maternal and neonatal morbidity and mortality. The review included 23 RCTs, most of which were small and with significant methodological flaws. These RCTs assessed many different questions and a broad range of treatments resulting in very limited opportunities to pool useful data. The paucity of robust studies assessing clinical effects of treatments makes it impossible to balance the benefits and harms of differing treatments for different levels of anaemia in pregnancy, in a meaningful and useful way. Many questions remain open. We cannot determine if women with mild anaemia, but otherwise healthy, will benefit from anaemia treatment; adverse effects can potentially outweigh benefits. It also remains unclear which treatments are safer and more effective in women with moderate or severe anaemia with and without associated illness.

Potential biases in the review process

We have produced updated coverage of randomised controlled trials of treatments for anaemia in pregnancy attributed to iron‐deficiency by summarising the best available evidence using quantitative methods. We have endeavoured to provide information to help clinicians and stakeholders choose the most appropriate treatment. Several sources were searched to identify RCTs. However, we cannot ensure that all available RCTs were located. We had to impute data (i.e. standard deviation) from a number of RCTs. Using this technique may lead to introduce bias. In addition, a number of pooled outcomes may have been measured in different ways within RCTs (i.e. adverse events).

Agreements and disagreements with other studies or reviews

Another Cochrane review that evaluated the effects and safety of preventive oral iron or iron + folic acid supplementation for women during pregnancy found that routine supplementation raised Hb levels although the authors found no evidence of the significant reduction in substantive maternal and neonatal adverse clinical outcomes (Pena‐Rosas 2009). Those findings are similar to those presented in this review.

Authors' conclusions

Implications for practice.

Avoidable limitations in the included randomised controlled trials (RCTs) resulted in these failing to provide sound evidence that currently available treatments for iron‐deficiency anaemia in pregnant women are beneficial for women or their children. We found no scientific basis to suggest that in otherwise healthy women, the benefits of treatments for mild anaemia in pregnancy will outweigh the adverse effects associated with them. We found no evidence that in women with iron‐deficiency anaemia in pregnancy, improvement in women's haematological indices translate into clinical improvements for them or their children. However, treatments are associated with frequent adverse effects such as gastrointestinal disturbances and poor compliance. Compared with oral iron, intramuscular (IM) iron improves haematological indices. But again, the support from clinical research seems to be missing and adverse effects remain poorly evaluated despite indications that treatments can result in important adverse outcomes. Intravenous (IV) iron sorbitol improves haematological values compared with IM or oral iron, but serious adverse effects are possible and remain poorly studied; knowledge of their magnitude and mitigation strategies is missing. Potential adverse effects may include venous thrombosis and severe allergic reactions. Treatment of mild anaemia in pregnancy remains controversial and unsupported by scientific proof. It is also unclear what treatments work better for severe anaemia in pregnancy. In one RCT (Zhou 2007) daily low‐dose iron supplements were effective at treating anaemia in pregnancy with less gastrointestinal side effects compared with higher doses.

Iron‐deficiency anaemia in pregnancy is frequently diagnosed and treated, but the effects of these treatments remain largely unknown. Severe iron‐deficiency anaemia affects many pregnant women in developing countries and may have considerable impact on maternal and neonatal health.

Implications for research.

Considerable resources are being used globally to diagnose and treat anaemia in pregnant women, but it remains unclear if these efforts are worthy and beneficial to individuals or populations. Also, it is unclear if there is a positive return for this investment, and if it improves people's lives. This review is an invitation for researchers, especially those working towards the improvement of health of communities in under‐resourced settings, to implement high‐quality RCTs addressing knowledge gaps (such as those flagged up by this review), for this common condition. In particular, determining when treatments are worthwhile, and providing sufficient information to allow better balancing of the benefits and harms of treatments. The authors of this systematic review consider that a solution to this would be to conduct a large multicenter RCT assessing the clinical effects of a selection of commonly‐used treatments in different regions of the world. The sample and duration of the follow‐up in such an RCT should be estimated to allow the identification of important, frequent, and long‐term effects in women and babies. Large RCTs such as the MAGPIE trial (Magpie 2002) or the CRASH trial (Edwards 2005) illustrate how gaps in knowledge can be effectively addressed through research, and how this can reduce harmful practices and inappropriate use of resources. We found a compelling case for a similar approach to be taken on iron‐deficiency anaemia in pregnancy.

Some important considerations for future research are as follows.

1. There is an urgent need to determine what treatments improve maternal and neonatal prognosis in women with severe and moderate anaemia in poorly‐resourced settings.

2. The effects of different doses, regimens and routes of administrations for commonly‐used treatments remain to be determined. The suitability of the route of administration may be influenced by the setting or cultural background.

3. Stratification according to anaemia severity can help address questions of the effects in different populations, and balance differently the benefits and harms.

4. Women with additional factors contributing to their anaemia, such as vitamin A deficiency, need to be studied as a different population.

5. Clinical outcomes, including adverse effects and quality of life, need to be better addressed and considered for study sample size calculations.

6. Offspring outcomes are particularly important given the possibility that iron has been associated with adverse effects in some observational studies.

7. RCTs need to have sample sizes big enough to allow assessing adverse effects such as venous thrombosis, allergic reactions, infections, and rare but serious adverse effects, as well as long‐term outcomes.

8. For women with mild iron‐deficiency anaemia, it would be helpful to assess whether oral iron is overall beneficial compared with placebo or no treatment. Researchers need to remain aware about the clinical effects of high iron on Hb levels, and possible overdosing.

9. We found no studies on oral erythropoietin or transfusions; these need to be evaluated in populations where they remain likely to be used. But providing scientific support for commonly‐used treatments seems to be the priority; we do not know if more harm then good is being done and yet these interventions remain widely prescribed and used.

10. Studies are needed to determine the effects in specific populations such as pregnant women who are anaemic and also infected with human immunodeficiency virus.

11. To use the CONSORT statement to improve the quality of reports of randomised trials (Schulz 2010: http://www.consort‐statement.org/).

[Note: The three citations in the awaiting classification section of the review may alter the conclusions of the review once assessed.]

What's new

Date	Event	Description
5 December 2011	Amended	Corrected typographical error.

History

Protocol first published: Issue 4, 1999
 Review first published: Issue 2, 2001

Date	Event	Description
7 June 2011	New search has been performed	This update is based on a search run in June 2011, which identified six new randomised controlled trials (RCTs). There are now a total of 23 RCTs included in the review. The inclusion of the 6 new included studies did not change the overall conclusions (Digumarthi 2008; Khalafallah 2010; Nappi 2009; Saha 2007; Sun 2010; Zhou 2007). In addition, a 'risk of bias' assessment was performed for all RCTs.
7 June 2011	New citation required but conclusions have not changed	New author joined the review team.
9 February 2010	Amended	Search updated. Fifteen reports added to Studies awaiting classification.
12 May 2009	Amended	Contact details updated.
20 September 2008	Amended	Converted to new review format.
13 February 2007	New search has been performed	An updated search of the Pregnancy and Childbirth Group's Trials Register on 31 January 2007 identified seven new trial reports which have been added to the awaiting assessment section for assessment in the next update. This update is based on a search run in December 2005, which identified twelve new trials (Al 2005; Bayoumeu 2002; Breymann 2001; De Souza 2004; Komolafe 2003; Kumar 2005; Mumtaz 2000; Ogunbode 1980; Oluboyede 1980; Sood 1979; Wali 2002; Zutschi 2004).
1 February 2007	New citation required but conclusions have not changed	There are now a total of 17 trials included in the review. The inclusion of these trials have generally not changed the conclusions although there are now concerns about possible important adverse effects. Ludovic Reveiz is now the guarantor of the review.

Appendices

Appendix 1. Search strategy

PubMed (Jan 2005 to June 2011)

(randomized controlled trial [pt] OR controlled clinical trial [pt] OR randomized [tiab] OR placebo [tiab] OR drug therapy [sh]  OR randomly [tiab]  OR trial [tiab]  OR groups [tiab]) NOT (animals [mh] NOT (humans [mh] AND animals [mh])) AND (Pregnant Women [mh] OR Pregnancy [mh] OR Pregnan* OR Prenatal Care [mh]) AND (Iron OR ferrous OR ferric OR iron [mh]) AND (Anemia* OR anaemia* OR anemic OR anaemic OR anemia [mh])

CENTRAL (2011, Issue 2 of 4)

#1 mesh descriptor Pregnant Women explode all trees 
 #2 mesh descriptor Pregnancy explode all trees 
 #3 mesh descriptor Prenatal care explode all trees 
 #4 Iron OR ferrous OR ferric OR mesh descriptor Iron explode all trees 
 #5 Anemia* OR anaemia* OR anemic OR anaemic OR mesh descriptor Anemia explode all trees 
 #6 #1 OR #2 OR #3 
 #7 #6 AND #5 AND #4

Health Technology Assessment Program (HTA) (www.hta.ac.uk/) [United Kingdom]; and LATINREC [www.latinrec.org, Colombia] (searched 2 May 2011):

(anemia OR anaemia OR anemic OR anaemic) AND (pregnant OR pregnancy) AND (iron OR ferrous) AND (random OR randomized OR randomised OR aleatory).

Search Portal of the International Clinical Trials Registry Platform (ICTRP; www.who.int/ictrp/) (searched 2 May 2011)

anemia AND pregnant AND ferrous 
 anaemia AND pregnant AND ferrous 
 anemic AND pregnant AND ferrous 
 anemia AND pregnancy AND ferrous 
 anaemia AND pregnancy AND ferrous 
 anemic AND pregnancy AND ferrous 
 anemia AND pregnant AND iron 
 anaemia AND pregnant AND iron 
 anemic AND pregnant AND iron 
 anemia AND pregnancy AND iron 
 anaemia AND pregnancy AND iron 
 anemic AND pregnancy AND iron

Appendix 2. Search strategy used in the previous version of the review

Authors searched the Cochrane Central Register of Controlled Trials (The Cochrane Library 2005, Issue 4), MEDLINE (1966 to December 2005), EMBASE (1976 to December 2005), LILACS (1982 to 40 edition), and BIOSIS Previews (from 1980 to June 2002) using the following strategy (adapted for each database):

(Randomized‐controlled‐trial:PT OR Randomized‐clinical‐trials:PT) 
 AND 
 (Pregnancy in Mesh OR Prenatal care in Mesh) 
 (Anemia, Hypochromic/drug therapy in MESH OR 
 Anemia, Hypochromic/prevention and control in MESH OR 
 Anemia, Hypochromic/therapy in MESH OR 
 Anemia, Iron deficiency/drug therapy in MESH OR 
 Anemia, Iron deficiency/prevention and control in MESH OR 
 Anemia, Iron deficiency/therapy in MESH) 
 Pregnancy complications/prevention and control

Iron/therapeutic use

Haematinics/adverse effects.

Appendix 3. Methods used to assess trials included in previous versions of this review

(1) Study selection

Two review authors (L Reveiz (LR) and LG Cuervo (LGC)) checked the titles and abstracts identified from the searches. If it was clear that the study did not refer to a randomised controlled trial on iron‐deficiency anaemia in pregnancy, it was excluded. If it was unclear, then we obtained the full text of the study for independent assessment by LR and G Gyte (GG). LR and GG assessed each trial for inclusion and resolved any disagreements through discussion, with referral to a third author (LGC) when necessary. Excluded studies and reasons for exclusion are described in the 'Characteristics of excluded studies' table.

(2) Assessment of methodological quality

We assessed trials under consideration for methodological quality and for appropriateness for inclusion without consideration of their results. We processed data from included trials as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2005). We undertook quality assessment by evaluating the following components for each included study, since there was some evidence that these are associated with biased estimates of treatment effect: 
 (a) the method of generation of the randomisation sequence; if it delivered a known chance allocation to each given group but individual allocation could not be anticipated; 
 (b) the method of allocation concealment, which was considered 'adequate' when the assignment could not be foreseen; 
 (c) who was blinded or not blinded (participants, clinicians, outcome assessors); 
 (d) participants lost to follow up in each arm of the study (split into postrandomisation exclusions and later losses if possible), and whether participants were analysed in the groups to which they were originally randomised (intention to treat).

The information was recorded in a table of quality criteria and a description of the quality of each study was given based on a summary of these components.

(3) Data extraction

Data extraction was carried out independently by one author (LR) using a data extraction form. Data were extracted for all outcomes for all relevant drugs, paying particular attention to the dosage and periodicity of treatment. GG checked the data extraction. We resolved disagreements by discussion until we reached consensus. We obtained missing data from the trial authors, when possible.

(4) Analysis

To estimate differences between treatments, we pooled the results of randomised controlled trials (RCTs) that evaluated similar interventions (and controls), and calculated a weighted treatment effect across RCTs using a fixed‐effect model. The results were expressed as relative risk, and 95% confidence intervals (CI)) for dichotomous outcomes, and weighted mean difference (and 95% CI) for continuous outcomes. Results were expressed as number needed to treat where appropriate. We summarised the information we found available. Quasi‐randomised and non‐randomised controlled studies were identified and listed, but were not further discussed. A qualitative description was provided for adverse effects when this was available.

Data and analyses

Comparison 1. Oral iron versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Anaemic during 2nd trimester	1	125	Risk Ratio (M‐H, Fixed, 95% CI)	0.38 [0.26, 0.55]
2 Haemoglobin levels (g/dL)	2	215	Mean Difference (IV, Random, 95% CI)	1.34 [0.27, 2.42]
3 Ferritin levels (ug/l)	1	125	Mean Difference (IV, Fixed, 95% CI)	0.70 [0.52, 0.88]
4 Serum iron (mg/l)	1	125	Mean Difference (IV, Fixed, 95% CI)	0.04 [0.03, 0.05]
5 Side effects	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	1.97 [0.66, 5.91]
6 Nausea and vomiting	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	4.5 [0.54, 37.54]
7 Constipation	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.32, 4.01]
8 Abdominal cramps	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	5.6 [0.28, 111.15]

1.4. Analysis.

Comparison 1 Oral iron versus placebo, Outcome 4 Serum iron (mg/l).

1.5. Analysis.

Comparison 1 Oral iron versus placebo, Outcome 5 Side effects.

1.6. Analysis.

Comparison 1 Oral iron versus placebo, Outcome 6 Nausea and vomiting.

1.7. Analysis.

Comparison 1 Oral iron versus placebo, Outcome 7 Constipation.

1.8. Analysis.

Comparison 1 Oral iron versus placebo, Outcome 8 Abdominal cramps.

Comparison 2. Oral iron + vitamin A versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Anaemic during 2nd trimester	1	125	Risk Ratio (M‐H, Fixed, 95% CI)	0.04 [0.01, 0.15]
2 Haemoglobin levels (g/dL)	1	125	Mean Difference (IV, Fixed, 95% CI)	1.30 [1.11, 1.49]
3 Ferritin levels (ug/l)	1	125	Mean Difference (IV, Fixed, 95% CI)	0.70 [0.52, 0.88]
4 Serum iron (mg/l)	1	125	Mean Difference (IV, Fixed, 95% CI)	0.08 [0.07, 0.09]

2.2. Analysis.

Comparison 2 Oral iron + vitamin A versus placebo, Outcome 2 Haemoglobin levels (g/dL).

2.3. Analysis.

Comparison 2 Oral iron + vitamin A versus placebo, Outcome 3 Ferritin levels (ug/l).

2.4. Analysis.

Comparison 2 Oral iron + vitamin A versus placebo, Outcome 4 Serum iron (mg/l).

Comparison 3. Controlled‐release oral iron versus regular oral iron.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Side effects	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.40, 2.33]
2 Nausea and vomiting	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.27, 3.41]
3 Constipation	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	0.24 [0.03, 2.00]
4 Abdominal cramps	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	0.48 [0.05, 4.95]

3.2. Analysis.

Comparison 3 Controlled‐release oral iron versus regular oral iron, Outcome 2 Nausea and vomiting.

3.3. Analysis.

Comparison 3 Controlled‐release oral iron versus regular oral iron, Outcome 3 Constipation.

Comparison 4. Intramuscular iron sorbito‐citric acid versus intramuscular dextran.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pain at injection site	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.58, 1.72]
2 Skin discolouration at injection site	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	0.21 [0.07, 0.65]
3 Venous thrombosis	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
4 Nausea or vomiting	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.06, 13.87]
5 Headaches	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	0.13 [0.02, 0.99]
6 Shivering	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	0.31 [0.01, 7.20]
7 Itching	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	0.92 [0.26, 3.26]
8 Metallic taste in mouth	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	3.68 [0.44, 30.56]

4.1. Analysis.

Comparison 4 Intramuscular iron sorbito‐citric acid versus intramuscular dextran, Outcome 1 Pain at injection site.

4.3. Analysis.

Comparison 4 Intramuscular iron sorbito‐citric acid versus intramuscular dextran, Outcome 3 Venous thrombosis.

4.4. Analysis.

Comparison 4 Intramuscular iron sorbito‐citric acid versus intramuscular dextran, Outcome 4 Nausea or vomiting.

4.6. Analysis.

Comparison 4 Intramuscular iron sorbito‐citric acid versus intramuscular dextran, Outcome 6 Shivering.

4.7. Analysis.

Comparison 4 Intramuscular iron sorbito‐citric acid versus intramuscular dextran, Outcome 7 Itching.

4.8. Analysis.

Comparison 4 Intramuscular iron sorbito‐citric acid versus intramuscular dextran, Outcome 8 Metallic taste in mouth.

Comparison 5. Intramuscular iron dextran versus intravenous iron dextran.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pain at injection site	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	4.52 [1.45, 14.05]
2 Skin discolouration at injection site	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	14.70 [2.08, 103.81]
3 Venous thrombosis	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	0.13 [0.01, 2.20]
4 Nausea or vomiting	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	0.57 [0.05, 5.83]
5 Headaches	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	3.96 [0.91, 17.17]
6 Shivering	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	0.57 [0.05, 5.83]
7 Itching	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	1.51 [0.38, 6.04]
8 Metallic taste in mouth	1	49	Risk Ratio (M‐H, Fixed, 95% CI)	1.13 [0.07, 17.07]
9 Severe delayed allergic reaction	1	62	Risk Ratio (M‐H, Fixed, 95% CI)	0.21 [0.01, 4.26]

5.1. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 1 Pain at injection site.

5.2. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 2 Skin discolouration at injection site.

5.6. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 6 Shivering.

5.7. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 7 Itching.

5.8. Analysis.

Comparison 5 Intramuscular iron dextran versus intravenous iron dextran, Outcome 8 Metallic taste in mouth.

Comparison 6. Intramuscular iron sorbitol citric acid versus intravenous iron dextran.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Pain at injection site	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	4.51 [1.46, 13.94]
2 Skin discolouration at injection site	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	3.12 [0.35, 28.03]
3 Venous thrombosis	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	0.12 [0.01, 2.04]
4 Nausea or vomiting	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	0.52 [0.05, 5.38]
5 Headaches	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	0.52 [0.05, 5.38]
6 Shivering	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	0.21 [0.01, 4.12]
7 Itching	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	1.39 [0.34, 5.58]
8 Metallic taste in mouth	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	4.16 [0.50, 34.71]

6.1. Analysis.

Comparison 6 Intramuscular iron sorbitol citric acid versus intravenous iron dextran, Outcome 1 Pain at injection site.

6.2. Analysis.

Comparison 6 Intramuscular iron sorbitol citric acid versus intravenous iron dextran, Outcome 2 Skin discolouration at injection site.

6.3. Analysis.

Comparison 6 Intramuscular iron sorbitol citric acid versus intravenous iron dextran, Outcome 3 Venous thrombosis.

6.4. Analysis.

Comparison 6 Intramuscular iron sorbitol citric acid versus intravenous iron dextran, Outcome 4 Nausea or vomiting.

6.5. Analysis.

Comparison 6 Intramuscular iron sorbitol citric acid versus intravenous iron dextran, Outcome 5 Headaches.

6.6. Analysis.

Comparison 6 Intramuscular iron sorbitol citric acid versus intravenous iron dextran, Outcome 6 Shivering.

6.7. Analysis.

Comparison 6 Intramuscular iron sorbitol citric acid versus intravenous iron dextran, Outcome 7 Itching.

6.8. Analysis.

Comparison 6 Intramuscular iron sorbitol citric acid versus intravenous iron dextran, Outcome 8 Metallic taste in mouth.

Comparison 7. Intravenous iron versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Side effects	1	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.75 [0.19, 3.04]
2 Nausea or vomiting	1	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 7.84]
3 Constipation	1	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.25 [0.03, 2.09]
4 Abdominal cramps	1	54	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

7.1. Analysis.

Comparison 7 Intravenous iron versus placebo, Outcome 1 Side effects.

7.4. Analysis.

Comparison 7 Intravenous iron versus placebo, Outcome 4 Abdominal cramps.

Comparison 8. Intravenous iron versus regular oral iron.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Side effects	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	0.38 [0.11, 1.31]
2 Nausea or vomiting or epigastric discomfort	3	244	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.15, 0.74]
3 Constipation	2	151	Risk Ratio (M‐H, Random, 95% CI)	0.11 [0.02, 0.71]
4 Abdominal cramps	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	0.18 [0.01, 3.54]
5 Diarrhoea	3	237	Risk Ratio (M‐H, Fixed, 95% CI)	0.16 [0.03, 0.86]
6 Blood transfusion required	3	167	Risk Ratio (M‐H, Fixed, 95% CI)	0.27 [0.05, 1.59]
7 Neonates mean hemoglobin	1	47	Mean Difference (IV, Fixed, 95% CI)	‐0.15 [‐1.37, 1.07]
8 Maternal haemoglobin at birth	1	90	Mean Difference (IV, Fixed, 95% CI)	0.75 [0.34, 1.16]
9 Maternal haemoglobin at 6 weeks	0	0	Mean Difference (IV, Fixed, 95% CI)	0.0 [0.0, 0.0]
10 Neonates ferritin level	1	47	Mean Difference (IV, Fixed, 95% CI)	‐2.0 [‐62.36, 58.36]
11 Mean maternal haemoglobin at 4 weeks ( g/dL)	3	167	Mean Difference (IV, Random, 95% CI)	0.44 [0.05, 0.82]
12 Maternal mortality	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
13 Preterm labour	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
14 Caesarean section	2	190	Risk Ratio (M‐H, Fixed, 95% CI)	0.88 [0.46, 1.67]
15 Operative vaginal birth	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	1.5 [0.26, 8.60]
16 Postpartum haemorrhage	2	147	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.34, 2.26]
17 Low birthweight (under 2500 g)	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
18 Neonatal birthweight	3	237	Mean Difference (IV, Random, 95% CI)	54.29 [‐170.11, 278.68]
19 Small‐for‐gestational age	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	1.6 [0.56, 4.56]
20 Five minute Apgar score under seven	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.06, 15.55]
21 Neonatal mortality	2	147	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
22 Haemoglobin level > 12 g/dL at 30 days	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	0.72 [0.18, 2.87]
23 Gestational hypertension	1	90	Risk Ratio (M‐H, Fixed, 95% CI)	5.0 [0.25, 101.31]
24 Gestational diabetes	1	90	Risk Ratio (M‐H, Fixed, 95% CI)	0.2 [0.01, 4.05]
25 Haemoglobin level > 11 g/dL at birth	1	90	Risk Ratio (M‐H, Fixed, 95% CI)	1.54 [1.21, 1.94]
26 Severe delayed allergic reaction	1	67	Risk Ratio (M‐H, Fixed, 95% CI)	5.45 [0.27, 109.49]
27 Arthralgia	1	90	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.06, 15.50]

8.1. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 1 Side effects.

8.4. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 4 Abdominal cramps.

8.6. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 6 Blood transfusion required.

8.7. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 7 Neonates mean hemoglobin.

8.10. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 10 Neonates ferritin level.

8.12. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 12 Maternal mortality.

8.13. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 13 Preterm labour.

8.15. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 15 Operative vaginal birth.

8.16. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 16 Postpartum haemorrhage.

8.17. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 17 Low birthweight (under 2500 g).

8.19. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 19 Small‐for‐gestational age.

8.20. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 20 Five minute Apgar score under seven.

8.21. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 21 Neonatal mortality.

8.22. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 22 Haemoglobin level > 12 g/dL at 30 days.

8.26. Analysis.

Comparison 8 Intravenous iron versus regular oral iron, Outcome 26 Severe delayed allergic reaction.

Comparison 9. Intravenous iron versus controlled‐release oral iron.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Side effects	1	52	Risk Ratio (M‐H, Fixed, 95% CI)	0.40 [0.12, 1.37]
2 Nausea or vomiting	1	52	Risk Ratio (M‐H, Fixed, 95% CI)	0.10 [0.01, 1.82]
3 Constipation	1	52	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.06, 14.03]
4 Abdominal cramps	1	52	Risk Ratio (M‐H, Fixed, 95% CI)	0.31 [0.01, 7.26]

9.1. Analysis.

Comparison 9 Intravenous iron versus controlled‐release oral iron, Outcome 1 Side effects.

9.2. Analysis.

Comparison 9 Intravenous iron versus controlled‐release oral iron, Outcome 2 Nausea or vomiting.

9.3. Analysis.

Comparison 9 Intravenous iron versus controlled‐release oral iron, Outcome 3 Constipation.

9.4. Analysis.

Comparison 9 Intravenous iron versus controlled‐release oral iron, Outcome 4 Abdominal cramps.

Comparison 10. 2/3 dose intravenous iron versus full dose intravenous iron.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Allergic reaction during infusion	1	623	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.35, 1.25]
2 Allergic reaction after infusion	1	623	Risk Ratio (M‐H, Fixed, 95% CI)	0.62 [0.45, 0.86]
3 Life‐threatening allergic reaction during infusion	1	623	Risk Ratio (M‐H, Fixed, 95% CI)	2.54 [0.50, 13.00]
4 Discomfort needing analgesics after infusion	1	623	Risk Ratio (M‐H, Fixed, 95% CI)	0.49 [0.27, 0.89]
5 Immobilised by painful joints	1	623	Risk Ratio (M‐H, Fixed, 95% CI)	0.79 [0.30, 2.10]
6 Non‐live births	1	507	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.36, 2.03]
7 Neonatal death	1	507	Risk Ratio (M‐H, Fixed, 95% CI)	0.52 [0.13, 2.07]
8 Stillbirth	1	507	Risk Ratio (M‐H, Fixed, 95% CI)	0.70 [0.25, 1.93]
9 Spontaneous abortion	1	507	Risk Ratio (M‐H, Fixed, 95% CI)	3.13 [0.33, 29.92]

10.1. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 1 Allergic reaction during infusion.

10.3. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 3 Life‐threatening allergic reaction during infusion.

10.4. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 4 Discomfort needing analgesics after infusion.

10.5. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 5 Immobilised by painful joints.

10.6. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 6 Non‐live births.

10.7. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 7 Neonatal death.

10.8. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 8 Stillbirth.

10.9. Analysis.

Comparison 10 2/3 dose intravenous iron versus full dose intravenous iron, Outcome 9 Spontaneous abortion.

Comparison 11. Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Hb < 11 g/dL at 4 weeks	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	0.2 [0.03, 1.56]
2 Mean corpuscular volume	1	40	Mean Difference (IV, Fixed, 95% CI)	6.30 [2.96, 9.64]
3 Caesarean section	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.39, 2.58]
4 Metallic taste	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	0.5 [0.05, 5.08]
5 Warm feeling	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	1.0 [0.07, 14.90]
6 Birthweight	1	40	Mean Difference (IV, Fixed, 95% CI)	‐130.0 [‐380.44, 120.44]
7 Birth < 37 weeks	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	0.33 [0.01, 7.72]
8 Maternal mean blood pressure	1	40	Mean Difference (IV, Fixed, 95% CI)	‐0.20 [‐5.02, 4.62]
9 Need transfusion	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]

11.1. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 1 Hb < 11 g/dL at 4 weeks.

11.2. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 2 Mean corpuscular volume.

11.3. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 3 Caesarean section.

11.4. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 4 Metallic taste.

11.5. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 5 Warm feeling.

11.7. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 7 Birth < 37 weeks.

11.9. Analysis.

Comparison 11 Intravenous iron sucrose with adjuvant recombinant human erythropoietin versus intravenous iron sucrose, Outcome 9 Need transfusion.

Comparison 12. Intramuscular iron sorbitol citric acid versus oral iron.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Not anaemic at term	1	200	Risk Ratio (M‐H, Fixed, 95% CI)	1.23 [1.01, 1.48]
2 Mean maternal haemoglobin at birth	1	200	Mean Difference (IV, Fixed, 95% CI)	0.54 [0.30, 0.78]
3 Mean maternal hematocrit level at birth	1	200	Mean Difference (IV, Fixed, 95% CI)	1.40 [0.67, 2.13]
4 Caesarean section	1	200	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.66, 1.81]
5 Haematocrit (%) at 4 weeks of treatment	1	56	Mean Difference (IV, Fixed, 95% CI)	1.25 [‐0.03, 2.53]
6 Haematocrit (%) at 8 weeks of treatment	1	59	Mean Difference (IV, Fixed, 95% CI)	2.62 [1.26, 3.98]
7 Haematocrit (%) at 4 weeks of treatment	1	56	Mean Difference (IV, Fixed, 95% CI)	1.25 [‐0.03, 2.53]
8 Haematocrit (%) at 8 weeks of treatment	1	59	Mean Difference (IV, Fixed, 95% CI)	2.60 [1.02, 4.18]

12.4. Analysis.

Comparison 12 Intramuscular iron sorbitol citric acid versus oral iron, Outcome 4 Caesarean section.

12.5. Analysis.

Comparison 12 Intramuscular iron sorbitol citric acid versus oral iron, Outcome 5 Haematocrit (%) at 4 weeks of treatment.

12.7. Analysis.

Comparison 12 Intramuscular iron sorbitol citric acid versus oral iron, Outcome 7 Haematocrit (%) at 4 weeks of treatment.

Comparison 13. Intramuscular iron dextran versus oral iron + vitamin C + folic acid.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haematocrit	1	60	Mean Difference (IV, Fixed, 95% CI)	4.47 [3.67, 5.27]
2 Not anaemic at 6 weeks (packed cell volume > 33%)	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	11.0 [1.51, 79.96]

Comparison 14. Intramuscular iron sorbitol citric acid versus oral iron + folic acid.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mean haemoglobin at 36 weeks	1	150	Mean Difference (IV, Fixed, 95% CI)	‐0.26 [‐0.48, ‐0.04]
2 Haemoglobin > 11 g/dL at 36 weeks	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.64, 1.06]
3 Caesarean section	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	1.67 [0.41, 6.73]
4 Mean birthweight (kg)	1	150	Mean Difference (IV, Fixed, 95% CI)	‐20.0 [‐164.35, 124.35]
5 Diarrhoea	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	0.09 [0.01, 1.62]
6 Constipation	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	0.06 [0.00, 1.00]
7 Dyspepsia	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	0.05 [0.00, 0.89]
8 Local site mainly pain	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	125.0 [7.87, 1984.19]
9 Staining	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	113.0 [7.11, 1795.82]
10 Arthralgia	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	13.0 [0.75, 226.73]
11 Itching and rash	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	29.0 [1.76, 477.47]
12 Fever	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	17.0 [1.00, 289.34]
13 Malaise	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	15.0 [0.87, 258.02]
14 Vaso‐vagal due to apprehension	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	9.0 [0.49, 164.29]
15 Systemic ache	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	23.0 [1.38, 383.37]
16 Haemoglobin > 12 g/dL at 36 weeks	1	150	Risk Ratio (M‐H, Fixed, 95% CI)	0.52 [0.27, 1.01]

14.2. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 2 Haemoglobin > 11 g/dL at 36 weeks.

14.4. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 4 Mean birthweight (kg).

14.5. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 5 Diarrhoea.

14.6. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 6 Constipation.

14.7. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 7 Dyspepsia.

14.8. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 8 Local site mainly pain.

14.9. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 9 Staining.

14.10. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 10 Arthralgia.

14.11. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 11 Itching and rash.

14.12. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 12 Fever.

14.13. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 13 Malaise.

14.14. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 14 Vaso‐vagal due to apprehension.

14.16. Analysis.

Comparison 14 Intramuscular iron sorbitol citric acid versus oral iron + folic acid, Outcome 16 Haemoglobin > 12 g/dL at 36 weeks.

Comparison 15. Oral iron daily versus oral iron twice weekly.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at 4 weeks	1	160	Mean Difference (IV, Fixed, 95% CI)	0.54 [0.14, 0.94]
2 Haemoglobin level at 8 weeks	1	129	Mean Difference (IV, Fixed, 95% CI)	1.17 [0.67, 1.67]
3 Haemoglobin level at 12 weeks	1	105	Mean Difference (IV, Fixed, 95% CI)	1.27 [0.68, 1.86]
4 Haemoglobin level at 16 weeks	1	102	Mean Difference (IV, Fixed, 95% CI)	0.30 [‐0.01, 0.61]
5 Haemoglobin level > 11 g/dL at 16 weeks of treatment	1	102	Risk Ratio (M‐H, Fixed, 95% CI)	1.38 [0.86, 2.23]
6 Treatment failure (haemoglobin < 10 g/dL) at 16 weeks	1	102	Risk Ratio (M‐H, Fixed, 95% CI)	0.15 [0.02, 1.21]

15.1. Analysis.

Comparison 15 Oral iron daily versus oral iron twice weekly, Outcome 1 Haemoglobin level at 4 weeks.

15.2. Analysis.

Comparison 15 Oral iron daily versus oral iron twice weekly, Outcome 2 Haemoglobin level at 8 weeks.

15.6. Analysis.

Comparison 15 Oral iron daily versus oral iron twice weekly, Outcome 6 Treatment failure (haemoglobin < 10 g/dL) at 16 weeks.

Comparison 16. Oral iron daily versus oral iron once week.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at 16 weeks	1	97	Mean Difference (IV, Fixed, 95% CI)	0.70 [0.36, 1.04]
2 Haemoglobin level > 11 g/dL at 16 weeks of treatment	1	97	Risk Ratio (M‐H, Fixed, 95% CI)	1.73 [1.00, 3.01]
3 Treatment failure (haemoglobin < 10 g/dL) at 16 weeks	1	97	Risk Ratio (M‐H, Fixed, 95% CI)	0.05 [0.01, 0.35]

Comparison 17. Oral iron twice week versus oral iron once week.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at 16 weeks	1	101	Mean Difference (IV, Fixed, 95% CI)	0.40 [0.03, 0.77]
2 Haemoglobin level > 11 g/dL at 16 weeks of treatment	1	101	Risk Ratio (M‐H, Fixed, 95% CI)	1.25 [0.69, 2.28]
3 Treatment Failure (haemoglobin < 10 g/dL) at 16 weeks	1	101	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.15, 0.68]

Comparison 18. Intravenous iron sucrose 500 mg versus intravenous iron sucrose 200 mg.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at delivery	1	35	Mean Difference (IV, Fixed, 95% CI)	0.5 [‐0.18, 1.18]
2 Haemoglobin level > 11g/dL at delivery	1	35	Risk Ratio (M‐H, Fixed, 95% CI)	1.14 [0.78, 1.68]
3 Moderate abdominal pain	1	35	Risk Ratio (M‐H, Fixed, 95% CI)	1.33 [0.09, 19.64]

18.1. Analysis.

Comparison 18 Intravenous iron sucrose 500 mg versus intravenous iron sucrose 200 mg, Outcome 1 Haemoglobin level at delivery.

18.2. Analysis.

Comparison 18 Intravenous iron sucrose 500 mg versus intravenous iron sucrose 200 mg, Outcome 2 Haemoglobin level > 11g/dL at delivery.

18.3. Analysis.

Comparison 18 Intravenous iron sucrose 500 mg versus intravenous iron sucrose 200 mg, Outcome 3 Moderate abdominal pain.

Comparison 19. Intravenous iron sucrose 500 mg versus intramuscular iron sorbitol.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Maternal haemoglobin level at birth	1	40	Mean Difference (IV, Fixed, 95% CI)	1.60 [0.87, 2.33]
2 Haemoglobin level > 11g/dL at delivery	1	40	Risk Ratio (M‐H, Fixed, 95% CI)	2.86 [1.45, 5.63]

19.2. Analysis.

Comparison 19 Intravenous iron sucrose 500 mg versus intramuscular iron sorbitol, Outcome 2 Haemoglobin level > 11g/dL at delivery.

Comparison 20. Intravenous iron sucrose 200 mg versus intramuscular iron sorbitol.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at delivery	1	45	Mean Difference (IV, Fixed, 95% CI)	1.10 [0.49, 1.71]
2 Haemoglobin level > 11 g/dL at delivery	1	45	Risk Ratio (M‐H, Fixed, 95% CI)	2.5 [1.25, 4.99]

Comparison 21. Oral ferrous sulphate iron 1200 mg/day versus 600 mg/day.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haematocrit (%) at 4 weeks of treatment	1	56	Mean Difference (IV, Fixed, 95% CI)	0.37 [‐0.77, 1.51]
2 Haematocrit (%) at 8 weeks of treatment	1	56	Mean Difference (IV, Fixed, 95% CI)	0.02 [‐1.03, 1.07]

Comparison 23. Intramuscular iron sorbitol‐glu acid versus intravenous iron dextran.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haematocrit (%) at 4 weeks of treatment	1	59	Mean Difference (IV, Fixed, 95% CI)	2.18 [0.77, 3.59]
2 Haematocrit (%) at 8 weeks of treatment	1	43	Mean Difference (IV, Fixed, 95% CI)	1.48 [0.15, 2.81]
3 Neonatal jaundice	1	62	Risk Ratio (M‐H, Fixed, 95% CI)	0.94 [0.06, 14.33]
4 Viral hepatitis	1	62	Risk Ratio (M‐H, Fixed, 95% CI)	2.82 [0.12, 66.62]
5 Severe allergic reaction	1	62	Risk Ratio (M‐H, Fixed, 95% CI)	0.31 [0.01, 7.40]

23.3. Analysis.

Comparison 23 Intramuscular iron sorbitol‐glu acid versus intravenous iron dextran, Outcome 3 Neonatal jaundice.

23.4. Analysis.

Comparison 23 Intramuscular iron sorbitol‐glu acid versus intravenous iron dextran, Outcome 4 Viral hepatitis.

23.5. Analysis.

Comparison 23 Intramuscular iron sorbitol‐glu acid versus intravenous iron dextran, Outcome 5 Severe allergic reaction.

Comparison 24. Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at 8 weeks	1	100	Mean Difference (IV, Fixed, 95% CI)	‐0.05 [‐0.30, 0.20]
2 Constipation at 8 weeks	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.30 [0.14, 0.64]
3 Haemoglobin > 11 g/dL at 8 weeks of treatment	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	1.08 [0.82, 1.43]
4 Gastrointestinal intolerance	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.41 [0.25, 0.69]
5 Metallic taste	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.14 [0.02, 1.04]
6 Diarrhea	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.22 [0.01, 4.39]
7 Rashes	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.36 [0.02, 8.64]
8 Compliance	1	100	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.92, 1.21]
9 Cost due to hospital visits	1	100	Mean Difference (IV, Fixed, 95% CI)	‐7.05 [‐116.16, 102.06]
10 Cost due to loss of work	1	100	Mean Difference (IV, Fixed, 95% CI)	10.28 [‐2.77, 23.33]

24.1. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 1 Haemoglobin level at 8 weeks.

24.4. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 4 Gastrointestinal intolerance.

24.5. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 5 Metallic taste.

24.6. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 6 Diarrhea.

24.7. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 7 Rashes.

24.8. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 8 Compliance.

24.9. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 9 Cost due to hospital visits.

24.10. Analysis.

Comparison 24 Oral iron polymaltose complex (100mg) versus ferrous sulphate (120mg), Outcome 10 Cost due to loss of work.

Comparison 25. Oral bovine lactoferrin versus ferrous sulphate.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mean haemoglobin levels at 1 month	1	97	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐0.52, ‐0.08]

Comparison 26. Ferrous sulphate (elementary iron) 20 mg vs 40 mg.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at 8 weeks (or until birth)	1	114	Mean Difference (IV, Fixed, 95% CI)	‐0.30 [‐0.74, 0.14]
2 Rate of anaemia at 8 weeks	1	114	Risk Ratio (M‐H, Fixed, 95% CI)	1.45 [0.84, 2.52]
3 Rate of moderate anaemia at 8 weeks	1	114	Risk Ratio (M‐H, Fixed, 95% CI)	1.66 [0.58, 4.76]
4 Nausea at 8 weeks	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.67, 1.12]
5 Heartburn at 8 weeks	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.72, 1.13]
6 Stomach pain at 8 weeks	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	0.56 [0.36, 0.88]
7 Vomiting at 8 weeks	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	0.71 [0.39, 1.30]
8 Bowel habit: < 3 per week	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.42, 2.57]
9 Hard stool at 8 weeks	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	1.09 [0.64, 1.87]
10 Feeling unwell at 8 weeks	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	0.70 [0.50, 0.97]
11 Need transfusion	1	120	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.07, 16.15]
12 Caesarean section	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.55, 2.01]
13 Birthweight	1	119	Mean Difference (IV, Fixed, 95% CI)	‐32.0 [‐206.02, 142.02]
14 Preterm birth	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.70 [0.12, 4.05]
15 Small‐for‐gestational age	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.35 [0.07, 1.67]
16 Fetal distress	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.75 [0.25, 2.23]

26.1. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 1 Haemoglobin level at 8 weeks (or until birth).

26.2. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 2 Rate of anaemia at 8 weeks.

26.3. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 3 Rate of moderate anaemia at 8 weeks.

26.4. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 4 Nausea at 8 weeks.

26.5. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 5 Heartburn at 8 weeks.

26.6. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 6 Stomach pain at 8 weeks.

26.7. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 7 Vomiting at 8 weeks.

26.8. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 8 Bowel habit: < 3 per week.

26.9. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 9 Hard stool at 8 weeks.

26.10. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 10 Feeling unwell at 8 weeks.

26.11. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 11 Need transfusion.

26.12. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 12 Caesarean section.

26.13. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 13 Birthweight.

26.14. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 14 Preterm birth.

26.15. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 15 Small‐for‐gestational age.

26.16. Analysis.

Comparison 26 Ferrous sulphate (elementary iron) 20 mg vs 40 mg, Outcome 16 Fetal distress.

Comparison 27. Ferrous sulphate (elementary iron) 20 mg vs 80 mg.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at 8 weeks (or until birth)	1	110	Mean Difference (IV, Fixed, 95% CI)	‐0.80 [‐1.27, ‐0.33]
2 Rate of patients with anaemia at 8 weeks	1	110	Risk Ratio (M‐H, Fixed, 95% CI)	1.56 [0.87, 2.79]
3 Rate of moderate anaemia at 8 weeks	1	111	Risk Ratio (M‐H, Fixed, 95% CI)	1.76 [0.44, 7.01]
4 Nausea at 8 weeks	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.75 [0.59, 0.95]
5 Heartburn at 8 weeks	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.75, 1.20]
6 Stomach pain at 8 weeks	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.54 [0.35, 0.84]
7 Vomiting at 8 weeks	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.53 [0.30, 0.93]
8 Bowel habit: < 3 per week	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.37, 2.18]
9 Hard stool at 8 weeks	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.84 [0.51, 1.37]
10 Feeling unwell at 8 weeks	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.74 [0.53, 1.04]
11 Need transfusion	1	119	Risk Ratio (M‐H, Fixed, 95% CI)	0.51 [0.05, 5.46]
12 Caesarean section	1	118	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.46, 1.57]
13 Birthweight	1	119	Mean Difference (IV, Fixed, 95% CI)	‐39.0 [‐212.83, 134.83]
14 Preterm birth	1	118	Risk Ratio (M‐H, Fixed, 95% CI)	0.41 [0.08, 2.05]
15 Small‐for‐gestational age	1	118	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.15, 7.10]
16 Fetal distress	1	118	Risk Ratio (M‐H, Fixed, 95% CI)	0.57 [0.20, 1.61]

27.1. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 1 Haemoglobin level at 8 weeks (or until birth).

27.2. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 2 Rate of patients with anaemia at 8 weeks.

27.3. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 3 Rate of moderate anaemia at 8 weeks.

27.4. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 4 Nausea at 8 weeks.

27.5. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 5 Heartburn at 8 weeks.

27.6. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 6 Stomach pain at 8 weeks.

27.7. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 7 Vomiting at 8 weeks.

27.8. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 8 Bowel habit: < 3 per week.

27.9. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 9 Hard stool at 8 weeks.

27.10. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 10 Feeling unwell at 8 weeks.

27.11. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 11 Need transfusion.

27.12. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 12 Caesarean section.

27.13. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 13 Birthweight.

27.14. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 14 Preterm birth.

27.15. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 15 Small‐for‐gestational age.

27.16. Analysis.

Comparison 27 Ferrous sulphate (elementary iron) 20 mg vs 80 mg, Outcome 16 Fetal distress.

Comparison 28. Ferrous sulphate (elementary iron) 40 mg vs 80 mg.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Haemoglobin level at 8 weeks (or until birth)	1	112	Mean Difference (IV, Fixed, 95% CI)	‐0.5 [‐0.93, ‐0.07]
2 Rate of anaemia at 8 weeks	1	110	Risk Ratio (M‐H, Fixed, 95% CI)	1.56 [0.87, 2.79]
3 Rate of moderate anaemia at 8 weeks	1	107	Risk Ratio (M‐H, Fixed, 95% CI)	1.64 [0.41, 6.50]
4 Nausea at 8 weeks	1	121	Odds Ratio (M‐H, Fixed, 95% CI)	0.54 [0.23, 1.26]
5 Heartburn at 8 weeks	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.85, 1.30]
6 Stomach pain at 8 weeks	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.69, 1.31]
7 Vomiting at 8 weeks	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	0.75 [0.46, 1.21]
8 Bowel habit: < 3 per week	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.36, 2.11]
9 Hard stool at 8 weeks	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.47, 1.27]
10 Feeling unwell at 8 weeks	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	1.06 [0.81, 1.39]
11 Need transfusion	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	0.49 [0.05, 5.28]
12 Caesarean section	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	0.81 [0.44, 1.49]
13 Birthweight	1	119	Mean Difference (IV, Fixed, 95% CI)	‐7.0 [‐194.82, 180.82]
14 Preterm birth	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	0.59 [0.15, 2.36]
15 Small‐for‐gestational age	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	2.95 [0.62, 14.04]
16 Fetal distress	1	121	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.30, 1.92]

28.1. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 1 Haemoglobin level at 8 weeks (or until birth).

28.2. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 2 Rate of anaemia at 8 weeks.

28.3. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 3 Rate of moderate anaemia at 8 weeks.

28.4. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 4 Nausea at 8 weeks.

28.5. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 5 Heartburn at 8 weeks.

28.6. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 6 Stomach pain at 8 weeks.

28.7. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 7 Vomiting at 8 weeks.

28.8. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 8 Bowel habit: < 3 per week.

28.9. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 9 Hard stool at 8 weeks.

28.10. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 10 Feeling unwell at 8 weeks.

28.11. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 11 Need transfusion.

28.12. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 12 Caesarean section.

28.13. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 13 Birthweight.

28.14. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 14 Preterm birth.

28.15. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 15 Small‐for‐gestational age.

28.16. Analysis.

Comparison 28 Ferrous sulphate (elementary iron) 40 mg vs 80 mg, Outcome 16 Fetal distress.

Comparison 29. Intravenous iron + oral iron versus oral iron.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Mean predelivery maternal haemoglobin	1	183	Mean Difference (IV, Fixed, 95% CI)	0.48 [0.21, 0.75]
2 Mean maternal haemoglobin after delivery	1	183	Mean Difference (IV, Fixed, 95% CI)	0.39 [0.02, 0.76]
3 Tolerate oral iron	1	183	Risk Ratio (M‐H, Fixed, 95% CI)	0.28 [0.06, 1.32]

29.3. Analysis.

Comparison 29 Intravenous iron + oral iron versus oral iron, Outcome 3 Tolerate oral iron.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Al 2005.

Methods	Group allocation was predetermined by one of the authors who was not involved with women's care. Authors used opaque envelopes that were consecutively‐numbered by means of a computer‐generated randomisation table. As each woman gave consent for the study, the next envelope was opened to assign the participant to either of the 2 groups. A sample‐size analysis was performed before initiation of the study. The analysis was based on the ITT principle. No participants were lost to follow up, and there were no dropouts.
Participants	90 pregnant women, between the 26th and 34th weeks of gestation, with established iron‐deficiency anaemia who had Hb levels between 8 and 10.5 g/dL and ferritin levels less than 13 g/ L. Women were excluded when serum folate and vitamin B12 levels were found to be less than 4 pg/mL and 100 pg/mL respectively. Anaemia from causes other than iron‐deficiency, multiple pregnancy, previous blood transfusion, history of haematological disease, risk of preterm labour, intolerance to iron derivatives, recent administration of iron for the treatment of iron‐deficiency anaemia, or current usage of iron supplement were the reasons for other exclusions.
Interventions	Experimental group: the dose for total iron sucrose was calculated from the following formula: weight x (target Hb – actual Hb) x 0.24 + 500 mg. In each infusion, the maximum total dose administered was 200 mg elemental iron in 100 mL 0.9% NaCl, infused in 20–30 minutes. Total dose was administered over 5 days and maximum daily dose administered was 400 mg elemental iron. Most of the women received iron sucrose at the rate of 200 mg every other day. Control group: 3 x 100 mg iron polymaltose complex (300 mg elemental iron per day) tablets per day orally administered iron were given throughout pregnancy. Both groups were supplemented by 0.5 mg folic acid treatment per day.
Outcomes	The primary outcome measure was Hb concentration on day 28 and at birth. Secondary outcome measures included ferritin levels, the recorded adverse effects, and fetal birthweight.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Group allocation was predetermined by one of the authors who was not involved with patient care. Opaque envelopes were consecutively numbered by means of a computer‐generated randomisation table. As each patient gave consent for the study, the next envelope was opened to assign the patient to either of the 2 groups.
Allocation concealment (selection bias)	Low risk	Consecutive numbered opaque envelopes.
Blinding (performance bias and detection bias) All outcomes	High risk	Not done.
Incomplete outcome data (attrition bias) All outcomes	Low risk	The analysis was based on the ITT principle. No participants were lost to follow up, and there were no dropouts. Relevant neonatal outcomes were not reported.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Maternal and infant baseline characteristics were provided. Competing interests and funding were not declared.

Bayoumeu 2002.

Methods	Women were assigned to the group treatment by a randomisation table. Sample size and power calculation were described. Neither the women nor caregivers were blinded to the interventions. It is unclear if the outcome assessor was blinded to the interventions. The trialists reported that 3 (6%) women were excluded from the study and that 2 others where lost to follow up.
Participants	50 pregnant women. Inclusion criteria: pregnant women at 6 months of pregnancy > 18 years, with Hb 8‐10 g/dL at 6 months; MCV < 100 fl; ferritin < 50 ug/l (corresponds to iron store of < 500 mg). Exclusion criteria: anaemia not linked to iron deficiency; asthma; cirrhosis; viral hepatitis; multiple pregnancy; risk of premature birth; suspected acute infection; parenteral iron treatment before inclusion; intolerance to iron. Also transport problems, etc.
Interventions	Experiment group: IV iron sucrose. Total dose calculated from weight before pregnancy in kg x (target Hb ‐ actual Hb) x 0.24 + 500 mg rounded up to nearest 100 mg. Target Hb set at 120 g/L because of physiological haemodilution during pregnancy. Given in 6 slow IV injections (days 1, 4, 8, 12, 15 and 21). Control group: oral iron. 3 x 80 mg iron sulphate tablets (Tardyferon) per day for 4 weeks (i.e. 240 mg elemental iron a day for 4 weeks). Women asked to note compliance in calendar. "Women were also given 15 mg folic acid per day to prevent an eventual folic‐acid deficiency and to eliminate the influence of such a deficiency on the results." "After 4 weeks, physician or midwife decided duration and dose of any continuing iron treatment."
Outcomes	The trial measured haematological response, transferrin level and saturation coefficient, erythrocytic folates, ferritin level, baby's ferritin level and full blood cell count and adverse reactions.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Women were assigned to the group treatment by a randomisation table.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Neither the women nor caregivers were blinded to the interventions. It is unclear if the outcome assessor was blinded to the interventions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	The trialists reported that 3 (6%) women were excluded from the study and that 2 others where lost to follow up. No clinical outcomes were measured.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Baseline characteristics were provided. Supported by Maternity Hospital Nano. Competing interests not declared.

Breymann 2001.

Methods	Women randomly assigned to 2 treatment groups by means of a computer‐generated list. It is unclear whether participants, clinicians and outcome assessor were blinded to the interventions. Trial had no withdrawals. No description of the sample size or power calculation was recorded.
Participants	40 pregnant women. Inclusion criteria: pregnant women with Hb < 10 g/dL in 2nd trimester and < 11 g/dL in 3rd trimester; ferritin < 15 ug/l. Exclusion criteria: women with anaemia not caused by iron deficiency, e.g. B12 or folate deficiency; chronic bleeding; renal failure.
Interventions	Women were randomised to receive IV administered iron sucrose (200 mg IV administered twice weekly 72 to 96 hours apart) with vs. without adjuvant recombinant human erythropoietin (300 U/kg body weight). All women received orally administered iron sulphate (80 mg twice daily) for = 2 weeks before starting. Random assignment was initiated when the Hb dropped to < 10.0 g/dL despite orally administered iron supplementation. Median durations of therapy were 18 days in group 1 and 25 days in group 2.
Outcomes	Blood index values, maternal outcomes (which cannot be obtained separately for both groups of treatment) and safety were reported.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Women randomly assigned to 2 treatment groups by means of a computer‐generated list.
Allocation concealment (selection bias)	Low risk	Not described in the article but the author provided additional information upon request.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	It is unclear whether participants, clinicians and outcome assessor were blinded to the interventions. Probably not done.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Trial had no withdrawals.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Baseline characteristics were provided. Competing interests and funding were not declared.

Dawson 1965.

Methods	Method of allocation was a random‐number table. It is not clearly stated how allocation was concealed. Loss to follow up: at 2 weeks: loss: A = 24%; B = 9%; C = 27%. At 4 weeks: loss: A = 48%; B = 39%; C = 38%. At 8 weeks: loss: A = 88%; B = 70%; C = 85%.
Participants	74 pregnant women in the 3rd trimester with Hb less than 10 g/dL, MCHC under 30% and a marrow aspiration indicating iron deficiency. Women with toxemia, infection or antepartum haemorrhage were excluded. All women received prophylactic folate and oral iron was stopped prior to randomisation. Women were followed for 8 weeks and outcomes assessed at admission, 2, 4 and 8 weeks. Side effects were assessed during the treatment period.
Interventions	Iron sorbitol‐citric acid complex (Jectofer) IM 25 women. Iron dextran (Imferon) IM 23 women. Iron dextran (Imferon) IV 26 women. Dosages of all preparations were calculated to replace iron stores.
Outcomes	Side effects: 1. pain at injection site; 2. skin discolouration; 3. venous thrombosis; 4. nausea or vomiting; 5. headaches; 6. shivering; 7. itching; 8. metallic taste in mouth.
Notes	Other outcomes were not considered due to a high dropout rate. Hb level data are not included due to 81% dropouts at predelivery. The authors made an additional trial of iron dextran IV vs. iron dextran + 50 mg of hydrocortisone in the infusion to assess if this had any effect on the rate of side effects.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Method of allocation was a random‐number table.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Not done.
Incomplete outcome data (attrition bias) All outcomes	High risk	Loss to follow up: at 2 weeks: loss: A = 24%; B = 9%; C = 27%. At 4 weeks: loss: A = 48%; B = 39%; C = 38%. At 8 weeks: loss: A = 88%; B = 70%; C = 85%.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Competing interests and funding were not declared.

De Souza 2004.

Methods	Method of allocation generation and concealment are unclear. Neither the women nor treating physicians were blinded to the interventions. Trialists reported that the laboratory was blind to the interventions. 41 (21.5%) women were reported to drop the trial or were lost to follow up and the reasons described. ITT analysis was not used. Sample size and power calculation were described.
Participants	150 pregnant women at 16‐20 weeks of gestation, with an initial Hb of < 11 g/dL and > 8 g/dL.
Interventions	Women were randomly distributed into 3 groups, 1 receiving daily (300 mg of ferrous sulphate), the 2nd twice‐weekly and the 3rd once‐weekly iron supplementation for 16 weeks.
Outcomes	The trial measured Hb concentration, MCV and ferritin before and after the treatment.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Neither the women nor treating physicians were blinded to the interventions. Trialists reported that the laboratory was blind to the interventions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	21.5% of women were reported to drop the trial or were lost to follow up. The reasons were described.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Baseline characteristics were provided. Competing interests and funding were not declared.

Digumarthi 2008.

Methods	Open randomised controlled trial.
Participants	30 pregnant women with moderate anaemia. Singleton pregnancy, 20 to 32 weeks with Hb from 7.1 to 9 g/dL.
Interventions	IV iron sucrose (Weight in kg x (target Hb ‐ actual Hb) x 0.25 + 500) vs. oral ferrous fumerate (300 mg of iron + 1.5 mg folic acid, twice day).
Outcomes	Hb rise, serum ferritin, serum iron, transferrin saturation.
Notes	Abstract publication.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Randomized". Not described.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Not done.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Losses to follow up not reported. No clinical outcomes were reported.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	High risk	No baseline characteristics were reported. Competing interests and funding were not declared.

Kaisi 1988.

Methods	A randomisation list was used to generate the randomisation sequence.
Participants	630 pregnant women. The study was done in a population of indigenous women. Inclusion criteria were: 1. diagnosis of iron‐deficiency anaemia defined as Hb under 10 g/dL; MCHC under 32%; hypochromia; poikilocytosis and anisocytosis; 2. age 16 years and above; 3. gestational age under 36 weeks. Exclusion criteria were: 1. history of reaction to parenteral iron; 2. hypersensitivity to iron dextran; 3. asthma history; 4. allergic conditions; 5. hepatic impairment; 6. renal impairment; 7. rheumatoid arthritis; 8. fever. 314 women received the full dose while 309 women received the 2/3 dose. Age, gravidity, parity, duration of pregnancy and basal Hb levels were similar for the groups. Nearly one fourth of the women had Hb levels under 7.0 g/dL in both groups.
Interventions	The dose of the 2 studied treatments was determined according to the recommendations of providers. In the intervention group, women received 2/3 of the total dose calculated of iron dextran 'Imferon' while in the control group they received the total dose of iron dextran plus 10 additional ml as suggested for pregnant women. The iron dextran was diluted in 500 ml of 5% dextrose and infused at a steady rate of 40 drops per minute. A test dose was given at the start of each infusion. This test dose was administered at a rate of 5 drops per minute over 10 minutes.
Outcomes	Women were followed up regularly throughout the remaining part of their pregnancy, during delivery and for 16 weeks postpartum. Infants were examined at the time of birth. Maternal Hb levels were assessed at each visit to the antenatal clinic and 6 and 16 weeks after delivery. Cord Hb was measured as well.
Notes	Women were analysed by ITT. Loss to follow up for Hb result was 47% so these results were not included in this review. For other clinical outcomes, loss to follow up was 18% and 20% so they were included.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation list.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Not done.
Incomplete outcome data (attrition bias) All outcomes	High risk	High rate of loss to follow up.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available. . Loss to follow up for Hb result was 47% so these results were not included in this review. For other clinical outcomes, loss to follow up was 18% and 20% so they were included.
Other bias	Unclear risk	Baseline characteristics were reported.Competing interests and funding were not declared.

Khalafallah 2010.

Methods	Single site, prospective, non blinded randomised‐controlled trial.
Participants	Two hundred women matched for Hb concentration and serum ferritin level were recruited. Women aged 18 years or above were identified with moderate iron deficiency anaemia defined as Hb <115 g ⁄L and low iron stores based on a serum ferritin level < 30 lg ⁄L. Most of the women (75%) had ferritin levels below 20 lg ⁄L and were equally distributed in both groups.
Interventions	Patients were randomised to daily oral ferrous sulphate 250 mg (elemental iron 80 mg) with or without a single intravenous iron polymaltose infusion. The IV arm required a single intravenous infusion of iron polymaltose (Ferrosig, Sigma Pharmaceuticals, Sydney, Australia) within 1 week after booking followed by oral iron identical to the other arm.
Outcomes	Primary outcomes were Hb level and mean serum ferritin level. Secondary outcomes were tolerability, safety and compliance and the effect of iron therapy on birth weight and gestational age at delivery. Cord blood Hb and iron studies data were collected at 37 consecutive deliveries to assess the effects of IV plus oral iron.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Central randomisation by computer will follow thereafter" (data from the Australian New Zealand clinical trials Registry. The trial was retrospectively registered: ACTRN12609000177257).
Allocation concealment (selection bias)	Low risk	"If the patients agree to participate in the trial, randomisation will follow in different department by the Pharmacy clinical trial personnel where a special computer programme will assign the treatment in random fashion without influence of treating team. This is usually conducted in blocks until full recruitment occurs".
Blinding (performance bias and detection bias) All outcomes	Low risk	Not done.
Incomplete outcome data (attrition bias) All outcomes	Low risk	8.5% of participants did not complete the trial (8 and 9 participants in the oral and IV iron group respectively).
Selective reporting (reporting bias)	Unclear risk	The protocol was not available.
Other bias	Low risk	All authors declare no financial conflicts of interest.

Komolafe 2003.

Methods	The women were assigned randomly by offering them a choice from sealed envelopes containing computer‐generated random numbers. No description of the sample size or power calculation was described and neither the participants nor treating physicians were blinded to the interventions. This trial had no withdrawals.
Participants	60 women at 14‐32 weeks of pregnancy were included. Inclusion criteria: PCV 22% to 26% due to iron‐deficiency anaemia. Iron‐deficiency anaemia defined as: Hb genotype AA; MCV < 75 pg; MCHC < 32 g/dL; blood film ‐ picture of iron‐deficiency anaemia. Exclusion criteria: symptomatic anaemia; acute malaria; acute urinary tract infection; history of allergy to parenteral iron; multiple pregnancy atopic individual and haemoglobinopathy.
Interventions	Experiment group: 30 women. IM iron dextran. 50 mg iron dextran into buttock preceded by 25 mg promethazine tablet 30 minutes before. If no untoward reaction after 48 hours, 250 mg (5 ml) iron dextran thrice weekly I = until total dose given. Total dose = iron (mg) = weight (kg) x Hb deficit (g/dL) x 4.4 + 500. Control group: 30 women. Oral Fe: 200 mg ferrous sulphate 3 times daily between meals, with vitamin C 100 mg 3 times daily and 5 mg folic acid. Both groups were treated for 6 weeks.
Outcomes	The trial measured the mean PCV, corrected anaemia at the end of the follow‐up period, the cost of the treatment and the side effects.
Notes	Results given in % and not specific numbers.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated random numbers.
Allocation concealment (selection bias)	Unclear risk	"Sealed envelopes".
Blinding (performance bias and detection bias) All outcomes	High risk	Neither the participants nor treating physicians were blinded to the interventions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	This trial had no withdrawals.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Unclear risk	Baseline characteristics were reported.

Kumar 2005.

Methods	Participants were randomly allocated to 2 groups. Method of allocation generation and allocation concealment were unclear. Both the participants and treating physicians were not blinded to the interventions. Sample size considerations were not provided. 47 women in parenteral iron group and 23 women in oral iron group were lost to follow up.
Participants	220 pregnant women were including according to the following criteria. Inclusion criteria: women with gestation period of 16‐24 weeks, were selected according to the following inclusion criteria: singleton pregnancy, moderate anaemia (Hb 8‐11 g%) by cyanmethaemoglobin method, microcytic hypochromic blood smear and willingness for enrolment to the study. Exclusion criteria: the women with anaemia due to haemoglobinopathies, chronic bleeding, parasitosis, diseases of liver, cardiovascular system and kidney; medical disorders like tuberculosis, diabetes mellitus; women who had any form of parenteral iron therapy for anaemia during pregnancy; women with antepartum haemorrhage and intolerance to test dose (0.5 ml) of IM administration of iron were excluded from the study.
Interventions	Experiment group: IM iron: parenteral iron group were given 2 IM injections of 250 mg elemental iron as iron sorbitol citric acid in a injection volume of 5 ml at an interval of 4‐6 weeks in the antenatal clinic. An initial test dose of 0.5 ml was given. If there was no adverse reaction to the test dose, then a full 250 mg dose was given deeply in the outer quadrant of the buttock using Z‐tract technique. Control group: oral iron: tablets of 100 mg elemental iron (ferrous sulphate) and 500 Ag of folic acid daily.
Outcomes	The trial measured values of blood indices at 36 weeks as well as mean birthweight, mode of delivery and side effects.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported.
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding (performance bias and detection bias) All outcomes	High risk	Both the participants and treating physicians were not blinded to the interventions.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	47 women in parenteral iron group and 23 women in oral iron group were lost to follow up.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Unclear risk	Baseline characteristics were reported. Competing interests and funding were not declared.

Mumtaz 2000.

Methods	Randomisation was performed using a random‐number generator, and each women was assigned a unique identifier. The women and the investigator were blinded to the allocation of treatment group (daily vs. twice weekly) at initial recruitment and the 3 follow‐up visits. The appearance of the capsules and the blister packs of the 2 groups were identical. The randomisation code was opened only after the follow‐up for all participants had been completed. Sample size considerations were provided. This trial had 86 participants, (45%) that did not complete the entire duration of follow up (i.e. 4 follow‐up visits). However, data on 83.8% of the participants were available for 4 weeks of follow‐up. Analysis by ITT.
Participants	191 pregnant women between the age of 17‐35 years, with an initial Hb of < 110 g/L were included. Uneventful obstetric history. All given health education materials on importance of diet in pregnancy.
Interventions	Experiment group: daily iron ‐ plus daily folate. 200 mg iron sulphate (60 mg elemental iron) each day and 1 mg folate. Control group: twice weekly iron ‐ plus daily folate. 200 mg iron sulphate (60 mg elemental) twice weekly and 1 mg folate. Placebo was given for the rest of the days.
Outcomes	Venous blood samples were taken for complete blood count at each visit and for serum ferritin at the 1st, 3rd and 4th visits. A peripheral film was made to rule out congenital disorders such as thalassaemia minor. No information about maternal and fetal outcomes and side effects was provided.
Notes	There was no difference between the women who dropped out compared with those who continued in the trial in terms of age, initial Hb, parity and in the 2 treatment groups (41 versus 45).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random‐number generator.
Allocation concealment (selection bias)	Low risk	Randomisation was performed using a random‐number generator, and each women was assigned a unique identifier. The randomisation code was opened only after the follow up for all participants had been completed.
Blinding (performance bias and detection bias) All outcomes	Low risk	The women and the investigator were blinded to the allocation of treatment group. The appearance of the capsules and the blister packs of the 2 groups were identical.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	This trial had 86 participants, (45%) that did not complete the entire duration of follow up (i.e. 4 follow‐up visits). However data on 83.8% of the participants were available for 4 weeks of follow‐up.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Characteristics of the women at the beginning of the study were provided. Competing interests and funding were not declared.

Nappi 2009.

Methods	Prospective, randomised, controlled, double blind trial.
Participants	100 pregnant women with iron deficiency anaemia. Inclusion criteria were: physiological course of pregnancy, singleton pregnancy, gestational age 12 weeks and 36 weeks, Hb values11 mg/dl, serum iron 30 µg/dl, serum ferritin12 µg/dl and TIBC 450 µg/dl. Exclusion criteria were: gestational (such as hypertension, gestational diabetes) or pre‐existent maternal diseases and fetal abnormalities.
Interventions	Group A received bovine lactoferrin at the oral dose of one capsule of 100 mg twice a day for 4 weeks. Group B received ferrous sulphate at the daily oral dose of one tablet of 520 mg (containing 100 mg elementary iron) for 4 weeks. Both groups received an additional supplementation with calcium mefolinate at the daily oral dose of 1 x 15 mg tablet.
Outcomes	The primary outcome measure was the Hb level before and after 1 month of treatment. Secondary outcomes were serum ferritin, serum iron and TIBC levels and the difference in symptom scores between groups.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation list.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	Low risk	Both researchers and patients were blinded to group assignment. The tablets were similar and neither the researchers nor the patients could differentiate the treatment assigned.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	2 patients (4%) from the group assigned to ferrous sulphate were lost to follow up. No clinical outcomes were collected. The total number of adverse events were not reported for each group; only the median gastrointestinal side effects score were presented.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Baseline characteristics were provided. The authors report no conflicts of interest. Funding was not reported.

Ogunbode 1980.

Methods	There was no information on methods of randomisation. No description of the sample size or power calculation was described and both the participants and treating physicians were blinded to the interventions. No participants were reported to have dropped out or were lost to follow up.
Participants	91 women in the first or second trimester of pregnancy with a PCV of 33% or less were randomly allocated to 3 treatment groups.
Interventions	In group A, 32 participants received 200 mg of oral ferrous sulphate thrice daily; in group B 28 participants received 400 mg of oral ferrous sulphate 3 times daily; in group C, 31 women received IM iron poly (sorbitol gluconic acid) complex ferastral (500 mg iron) on alternate days until completion of the required dose (between 1250 to 2500 mg of iron). No formal formula to calculate the dose was described. All participants received 5 mg of folic acid and 25 mg of pyrimethamine once weekly.
Outcomes	The trial measured the reticulocyte response and haematocrit level.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	Low risk	Both the participants and treating physicians were blinded to the interventions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No participants were reported to have dropped out or were lost to follow up.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Unclear risk	Baseline characteristics were reported.

Oluboyede 1980.

Methods	Participants were allocated by restricted random allocation. There was no further information on methods of randomisation. No description of the sample size or power calculation was described and neither the participants nor treating physicians were blinded to the interventions. 1 participant from the imferon group was reported to have dropped out due to a severe reaction.
Participants	63 pregnant women with established iron‐deficiency anaemia defined as a PCV of 30% or less were included. Hb AA genotype only. Exclusion criteria: if previously had iron therapy.
Interventions	Experiment group: 32 women. IM ferastral (sorbitol gluconic acid). 500 mg ferastral (5 ml in each buttock) IM on alternate days until completion of required dose. Control group: 31 women. IV imferon (iron dextran). Calculated dose given in 540 ml normal saline and 50 ml promethazine hydrochloride (phenegran) was given IM before infusion. Drip ran slowly for first 30 minutes then 60 drops a minute until completion. Discharged home though 10 kept in. All women received anti‐malarial of 25 mg pyrimethamine and 5 mg of folic acid throughout pregnancy.
Outcomes	Weekly or 2‐weekly PCV estimations were done on all participants, after 4 weeks of treatment liver function tests were repeated and after 6 weeks bone marrow aspirations were repeated on 21 women. The trial also measured the reticulocyte response in 10 women randomly selected from each group and babies birth weights and any complication within the first week of life.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Restricted random allocation."
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Neither the participants nor treating physicians were blinded to the interventions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	1 participant from the imferon group was reported to have dropped out due to a severe reaction.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Unclear risk	Competing interests and funding were not declared.

Saha 2007.

Methods	A double blinded randomised control trial.
Participants	100 pregnant women between 20 and 40 years; gestation 14‐27 weeks; Hb < 9 and > 7 g/dL; and serum ferritin < mcg/L. Patients with a history of anaemia due to any other causes were excluded (i.e. chronic blood loss, haemolytic anaemia, thalassaemia). Women with other chronic or severe diseases were also excluded (hepatic, renal, cardiovascular etc.).
Interventions	Group 1 (48 participants): iron polymaltose, 100 mg elemental iron + folic acid 500 mcg for 8 weeks (1 tablet orally once day) + placebo (1 tablet orally once day). Gourp 2 (52 participants): 60 mg elemental iron + folic acid 500 mcg for 8 weeks (1 tablet orally twice day).
Outcomes	The primary efficacy parameter was the proportion of women achieving normal Hb level (> 11 g/dL). Adverse events and costs were also collected. No clinical outcomes were collected.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Randomized."
Allocation concealment (selection bias)	Unclear risk	Not reported.
Blinding (performance bias and detection bias) All outcomes	Low risk	Participants and outcome assessors were blind. Both of the drugs were dispensed in identical capsules to ensure blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No losses to follow up or discontinuation were reported. No clinical outcomes were reported.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Baseline characteristics were provided. Competing interests were declared.

Singh 1998.

Methods	Women were allocated using sealed envelopes with consecutive numbers.
Participants	First 100 women with diagnosed iron‐deficiency anaemia while attending for antenatal care at the National University Hospital, Singapore. Data provided by 1 of the authors reveals that compared groups had similar age distribution, parity, mean total income, weight, height, history of anaemia in previous pregnancies, history of intrauterine growth retardation in previous pregnancies and similar time‐gap between pregnancies. Races were distributed as follows: Chinese (10% parenteral and 6% of oral iron therapy), Malayan (46% and 78% in the same order), Indian (16 and 8%). History of preterm delivery was seen in 16% of the parenteral treatment group and 12% of oral iron group.
Interventions	Total dose iron polymaltose complex ‐ iron dextrin (ferrum Hausmann) infusion vs. oral therapy with 3 x 200 mg/day iron fumarate. The dose was determined according to the body weight and estimated iron deficiency.
Outcomes	This paper provided data at 36 weeks, delivery and 6 weeks' postpartum. The paper provided haematological outcomes. In addition, the publication mentioned some side effects and similar clinical outcomes in both groups. The authors were contacted and provided precise data on clinical outcomes and side effects that were included too.
Notes	Dr Kuldip Singh at the National University Hospital ‐ University of Singapore was the contacted author. Contact was established through a search on Internet. Further contact is being undertaken to check the units used for serum iron and ferritin estimations, then further comparison tables can be added to this review.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Randomised.
Allocation concealment (selection bias)	Low risk	Sealed envelopes with consecutive numbers.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	No sufficient detail to decide.
Incomplete outcome data (attrition bias) All outcomes	Low risk	The paper provided haematological outcomes. In addition, the publication mentioned some side effects and similar clinical outcomes in both groups. The authors were contacted and provided precise data on clinical outcomes and side effects that were included too.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Data provided by 1 of the authors reveals that compared groups had similar age distribution, parity, mean total income, weight, height, history of anaemia in previous pregnancies, history of intrauterine growth retardation in previous pregnancies and similar time‐gap between pregnancies.

Sood 1979.

Methods	There was no information on methods of randomisation. No description of the sample size or power calculation was described and researchers kept unmasked the 5 groups. 2 women were reported to have dropped in the group receiving IV iron due to severe delayed adverse allergic reaction.
Participants	151 pregnant women. Inclusion criteria: pregnant women, 26 + 2 weeks' gestation. Divided into 3 strata according to their Hb concentration: 50‐79; 80‐109; 110 or above. Exclusion criteria: women with chronic illness; with Hb < 50 g/L; who had received haematinics during the last months.
Interventions	Within each stratum they were randomised to 1 of the following groups. 1. Oral ferrous sulphate providing 120 mg of 120 mg elemental iron, given once per day 6 weeks. 2. Iron dextran complex providing 100 mg iron given IM twice per week. 3. Iron as in group 1 + pteroylmonoglutamic acid (5 mg, 6 day/week) + cyanocobalamin (100 µg IM once per 14 day). 4. Iron IM as in group 2 + pteroylmonoglutamic acid + cyanocobalamin as in group 3. 5. Iron dextran complex given IV as a single total dose infusion according to the formula (15 ‐ women's Hb (g/dL)) x (body‐weight (kg) x 3) + pteroylmonoglutamic acid + cyanocobalamin as in group 3. Treatment was continued for 10‐12 weeks.
Outcomes	Hb, PCV was estimated 48 before the treatment was started and at least 48 hours after the last injection or tablet was given.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	High risk	Not done.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	The paper provided haematological outcomes.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Unclear risk	Competing interests and funding were not declared.

Suharno 1993.

Methods	Allocation was done using a random‐number list from 1 to 305 and allocating women sequentially in the list. The manufacturers of the active treatments provided placebos. An independent researcher randomly labelled the active and placebo preparations. Coding colours were given to the preparations and these codes were opened once the data for all analyses had been entered in the computer and cleaned.
Participants	305 pregnant women. The study was conducted from April to September 1992 in 20 rural villages in 3 sub districts of Bogor, West Java. Participants came from middle and low socio‐economic groups. They were aged between 17 and 35 years, with parity in the ranges of 0 to 4 and gestational age of 16 to 25 weeks. 572 women met inclusion criteria. 305 participated in the study. Hb levels of participants were in the range of 8.0 to 10.9 g/dL. Women receiving iron or vitamin A treatments or supplements in the 6 months prior to study were excluded. Outcomes were assessed 2 and 7 days after the last dose of treatment was given (24 to 33 weeks). Participants had similar age, height, weight, pregnancies, parity and gestational age at admission.
Interventions	4 different groups received 2 active treatments at most (factorial design). All preparations were given daily for 8 weeks. 1. 60 mg elemental oral iron (as ferrous sulphate) + vitamin A (2.4 mg of retinol as retinyl palmitate). 2. 60 mg elemental oral iron (as ferrous sulphate ) + placebo of vitamin A. 3. Placebo of oral iron + vitamin A (2.4 mg of retinol as retinyl palmitate). 4. Placebo or oral iron + placebo or vitamin A.
Outcomes	Hb, ferritin and serum iron mean values and standard deviations were extracted for this review from the published paper. Percentage of women that became non‐anaemic (Hb > 10.9 g/dL) was counted as a dichotomised variable. Numbers needed to treat were calculated by the authors of this review.
Notes	Results for vitamin A + placebo were not considered since this is part of a different review. Vitamin A combined with iron was included in this review. Serum iron levels were given in umol/l and converted to mg/l by the authors (umol/l x 0.056 = mg/l). Loss to follow up accounted for 17% of the women. Analysis was done by ITT.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random‐number list.
Allocation concealment (selection bias)	Low risk	An independent researcher randomly labelled the active and placebo preparations. Coding colours were given to the preparations and these codes were opened once the data for all analyses had been entered in the computer and cleaned.
Blinding (performance bias and detection bias) All outcomes	Low risk	An independent researcher randomly labelled the active and placebo preparations. Coding colours were given to the preparations and these codes were opened once the data for all analyses had been entered in the computer and cleaned.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Some outcomes not fully reported.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Baseline characteristics were provided.

Sun 2010.

Methods	Double‐blind randomised trial
Participants	Pregnant women, 12 to 24‐week gestation, age between 20‐30 years, were examined for eligibility. Of these women, 186 anaemic pregnant women with Hb concentration ≥ 80 and < 110 g/L were included.
Interventions	Group I (n=47) was supplemented daily with 60 mg iron as ferrous sulfate. Group II (n=46) with 60 mg iron and 0.4 mg folic acid. Group III (n=46) with 60 mg iron, 2.0 mg retinol and 0.4 mg folic acid. Group C (n=47) was the placebo control group. Only group 1 and 4 were included for analysis in the review.
Outcomes	‐ Hb concentration at 2 months of treatment ‐ Plasma ferritin ‐ Plasma iron concentrations ‐ Peripheral blood lymphocytes No clinical outcomes or adverse events were reported
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Randomized". Not described.
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	“The capsules were coloured red, yellow, green and blue during manufacture by Hurun (a Chinese food‐additive company, Beijing). Both trial participants and the research team were unaware of the group assignment. The trial was decoded after analysis of the primary outcomes”.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Few participants were lost to follow up. Standard deviations of placebo group were not provided and had to be calculated.
Selective reporting (reporting bias)	High risk	No clinical outcomes and adverse events were informed.
Other bias	Low risk	Baseline characteristics were provided.

Symonds 1969.

Methods	Women were assigned randomly although the method is not clearly specified. Baseline data for the 4 groups compared are very similar.
Participants	103 women attending the Queen Elizabeth Hospital in Woodville, Australia. Inclusion criteria were a gestational age of 32 weeks or less and a Hb level of 10.8 g/dL or less.
Interventions	4 treatment groups were assembled. 1. Ferrous gluconate 108 mg of elemental iron daily divided in 3 doses given orally throughout pregnancy. 2. Ferrogradumet tablets (controlled‐release) iron tablets with 105 mg elemental iron given once daily throughout pregnancy. 3. Placebo for the controlled‐release iron tablets provided by the same pharmaceutical laboratory. 4. IV iron‐dextran 2% solution. Initial test dose of 2 ml IV followed by 5 injections of 5 ml (100 mg). Participants received controlled‐release iron or placebo for the first month. After that time side effects were evaluated, and then all participants were given a daily dose of active controlled‐release oral iron. The trial was masked only the first 2 months and only for these 2 groups.
Outcomes	For this review, side effects were considered. The other data were incomplete, without reported standard deviations of mean values and irrelevant due to important flaws in the design of the study.
Notes	Hb results were presented as increases in Hb. Since standard deviations for the values cannot be added to baseline data, the data were not included.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Randomly".
Allocation concealment (selection bias)	Unclear risk	Not described.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	The trial was masked only the first 2 months and only for these 2 groups.
Incomplete outcome data (attrition bias) All outcomes	High risk	For this review, side effects were considered. The other data were incomplete, without reported standard deviations of mean values.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Low risk	Baseline characteristics of participants were provided.

Wali 2002.

Methods	There was no information on methods of randomisation described and neither the participants nor treating physicians were blinded to the interventions. No description of the sample size or power calculation was provided.
Participants	60 pregnant women with anaemia were included in the study. Inclusion criteria: pregnant women; 12‐34 weeks; iron‐deficiency anaemia; Hb 5‐10 g/dL; PCV < 30%; MCV < 80 fl; MCH < 28 pg. Women with Hb < 7 g/dL (n = 2) were given IV iron sucrose (Venofer) as an alternative to blood transfusion; those with Hb 7‐10 g/dL were randomised for IV iron or IM iron. Exclusion criteria: chronic diseases; anaemic failure.
Interventions	Experiment group A: 15 women. IV iron sucrose 500 mg for iron storage. Experiment group B: 20 women. IV iron sucrose 200 mg iron. Experiment group C: 25 women IM iron sorbitol with varying doses depending on Hb level; 5 g/dL ‐ 24 injections; 6 g/dL ‐ 22 injections; 7 g/dL ‐ 20 injections; 8 g/dL ‐ 17 injections; 9 g/dL ‐ 14 injections; 10 g/dL ‐ 12 injections; 11 g/dL ‐ 10 injections. After parenteral iron, oral iron given till birth of baby.
Outcomes	The trial outcomes focused on laboratory values (Hb) and side effects.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description other than "randomised".
Allocation concealment (selection bias)	Unclear risk	No description.
Blinding (performance bias and detection bias) All outcomes	High risk	Neither the participants nor treating physicians were blinded to the interventions.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	The trial outcomes focused on laboratory values (Hb) and side effects.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Unclear risk	Competing interests and funding were not declared.

Zhou 2007.

Methods	Double blind randomised controlled trial.
Participants	Pregnant women (N = 180) diagnosed with anaemia (Hb < 110 g/L) at the routine mid‐pregnancy blood test were approached to take part in our study. They were eligible if between 24 and 32 weeks' gestation and had a singleton pregnancy. Women were excluded if they were taking iron or vitamin and mineral supplements containing iron, other causes of anaemia, history of thalassaemia or drug and alcohol abuse and had diabetes requiring insulin or a known fetal abnormality.
Interventions	Ferrous sulphate containing 20, 40 or 80 mg of elemental iron per tablet, which did not contain other micronutrients (1 tablet daily between meals from enrolment for 8 weeks or until birth).
Outcomes	The primary outcomes of the study were Hb levels and incidence of anaemia at the end of the intervention, and gastro‐intestinal side effects during treatment. Secondary outcomes included incidence of iron deficiency, iron‐deficiency anaemia and haemoconcentration at the end of treatment as well as pregnancy outcomes.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Computer‐generated randomizations schedule was generated by an independent consultant with stratification for Hb (Hb < 100 or > 100 g/L) to ensure that there were equal proportions of women with more severe anaemia in all groups."
Allocation concealment (selection bias)	Low risk	"The iron tablets were packed in bottles with childproof lids, which were labelled with a unique study number according to the randomizations schedule by the Pharmacy Department at the CYWHS."
Blinding (performance bias and detection bias) All outcomes	Low risk	"Trial participants and the research team were unaware of the treatment group assignment". The trial was unblinded after the analysis of the primary outcomes." "The tablets were identical in size, colour and shape."
Incomplete outcome data (attrition bias) All outcomes	Low risk	Only 1 patient was lost to follow up. Pregnancy outcomes and neonatal complications at birth were reported.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available. The Australian Clinical Trials Registry (ACTR) (Reference no. 12606000357550).
Other bias	Low risk	Baseline characteristics were provided. "Funding from the Adelaide Women’s & Children’s Hospital Research Foundation. The iron and placebo tablets used in the trial were manufactured and donated by Soul Pattinson Manufacturing, Kingsgrove, NSW, Australia. The funding organizations had no role in the design and conduct of the study, the analysis and interpretation of the data, or the preparation, review and approval of the manuscript. MM and RAG occasionally provide advice to manufacturers of nutritional supplements. SJZ and CAC have no known conflict of interest.".

Zutschi 2004.

Methods	There was no information on methods of randomisation. No women were reported to have dropped out or were lost to follow up. Neither the women nor treating physicians were blinded to the interventions. No description of the sample size or power calculation was described.
Participants	200 women with uncomplicated pregnancy enrolled at 24‐26 weeks of gestation with a Hb of > 8 gm% but < 11 gm% were included. Women dropped out of study if Hb fell below 8 g/l or if severe problems arose.
Interventions	Group A (100 women) received injectable iron‐sorbitol‐citrate in 3 IM doses of 150 mg each at 4‐weekly intervals and group B (100 women) were given oral iron having 100 mg elemental iron daily for at least 100 days.
Outcomes	The trial measured Hb levels at the time of inclusion into the study, 4‐weekly thereafter and at delivery as well as the proportion of caesarean delivery.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No description other than "randomized".
Allocation concealment (selection bias)	Unclear risk	No description.
Blinding (performance bias and detection bias) All outcomes	High risk	Neither the women nor treating physicians were blinded to the interventions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropout or lost to follow up were reported. Few clinical outcomes were provided.
Selective reporting (reporting bias)	Unclear risk	Protocol of the study is not available.
Other bias	Unclear risk	Baseline characteristics were partially presented. Competing interests and funding were not declared.

Hb: haemoglobin 
 IM: intramuscular 
 ITT: intention‐to‐treat 
 IV: intravenous 
 MCH: mean corpuscular haemoglobin 
 MCHC: mean corpuscular haemoglobin concentration 
 MCV: mean corpuscular volume 
 PCV: packed cell volume 
 TIBC: total iron binding capacity 
 vs: versus

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Al Momen 1996	Not a randomised controlled trial. Sequential allocation.
Allaire 1961	Loss to follow up of 56%. Used quasi‐random allocation.
Amir 1983	Excluded: inclusion criteria Hb < 12 g/dL. No baseline characteristics that lead to infer levels of Hb at the beginning of the study.
Ayub 2008	Non‐randomised study.
Bare 1960	They allocated participants using alternate order. This is not considered random.
Barrada 1991	Insufficient information for critical appraisal was provided. There are no explicit inclusion or exclusion criteria.
Basu 1973	Randomisation method not provided. No information is provided regarding blinding or number of women that completed the trial or that were accounted for each result.
Bencaiova 2009	Non‐anaemic pregnant women.
Bhatla 2009	Non‐anaemic women.
Bhutta 2006	Mebendazol and multivitamin trial. Ongoing.
Breymann 1998	Randomised open‐label trial.
Breymann 2002	Not a randomised controlled trial.
Buglanov 1984	No mention of randomisation.
Chan 2009	Iron supplementation from early pregnancy in anaemic and non‐anaemic women. No subgroup data.
Chanarin 1965	This study addressed megaloblastic anaemia.
Christiansen 1961	Alternate participants; not Hb < 11.0 g/d/L.
Coelho 2000	No mention that women were anaemic.
De Souza 2009	A convenience sample was used and allocation was consecutively and casually unblinded by the main researcher to each of the three experimental groups.
Dede 2005	Postpartum iron‐deficiency anaemia.
Dochi 1988	Non‐iron trial.
Dommisse 1982	Folic acid trial.
Ekstrom 2002	Centers were randomly assign (not women) and only some women had Hb < 11.5 g/dL.
Finzi 1972	No random allocation.
Fochi 1985	Randomisation is not well balanced. The paper does not explain unbalanced groups.
Govan 2005	Case series.
Graham 2007	A vitamin A‐fortified rice curry dish with iron + vitamins vs. placebo.
Halksworth 2003	No randomisation method. Evaluated the absorption of iron.
Hamilton 1973	Allocation was done using a haphazard strategy, not a random one.
Hampel 1974	Used a haphazard allocation method (day of diagnosis).
Hancock 1968	Non randomised trial.
Harrison 1971	Non‐iron RCT.
Hawkins 1970	The study evaluates the use of medication in women with Hb levels over 10.5 g/dL and uses non‐random strategy for allocation.
Holly 1955	The study is in women with Hb levels over 10 g/dL. Treatment allocation was not random.
Hosokawa 1989	Vitamin C trial.
Izak 1973	Although the authors state that allocation was random, the groups are very different (184, 76 and 22). The authors used an inappropriate control group of healthy women. They do not explain the randomisation method.
Jackson 1982	More than 50% of their women were lost to follow up.
Jaud 1979	Used open‐randomisation list. We tried to contact to verify data but it was impossible. Open‐randomisation lists are considered inadequate since there is no concealment and it is prone to bias.
Juarez‐Vazquez 2002	The trial evaluated folic acid + iron versus iron.
Kore 2006	Non randomised study. Only the abstract was available.
Krafft 2009	Non randomised study. No comparison.
Ma 2008	Retinol and Riboflavin supplementation trial.
Mahale 1993	This study has dropouts that surpasses the limit established for this review.
Mahmoudian 2005	Zinc trial.
Mathan 1979	The trial evaluated group 1 iron + pteroylmonoglutamic acid and cyanocobalamin versus group 1 + ascorbic acid versus group 1 + calcium caseinate.
Milman 2006	Non‐anaemic pregnant women.
Minganti 1995	Sample size: 15 participants. Data not available.
Mukhopadhyay 2004	The trial excludes women with Hb level < 10 g/dL. The mean baseline Hb was 11.3 and 11.6 g/dL in both groups of treatment.
Mukhopadhyay 2004a	Double publication.
Ogunbode 1984	Folic acid trial.
Preziosi 1997	No fundamental data to assess validity.
Reddy 2000	No comparison group.
Ridwan 1996	Health centres were randomised, rather than individuals.
Sagaonkar 2009	Assessed two different iron drugs and varied types of vitamins at the same time.
Schoyerer 1975	Folic acid trial.
Scott 2005	Series of cases description.
Sharma 2004	No randomisation.
Siega‐Riz 2001	Trialists reported that a subgroup of anaemic patients (women with Hb > 9 g/dL and ferritin < 40 g/l) was randomised  to receive prenatal supplements with 30 or 60 mg of iron. A subsequent publication only focus on the non‐anaemic group (Siega‐Riz 2006). The principal author was contacted to get data on the subgroup of anaemic patients. The author said "that it will not be possible".
Siega‐Riz 2006	Non‐anaemic pregnant women.
Sigridov 2005	Not randomised.
Sood 1975	This study has a high proportion of women lost to follow up exceeding the cut‐off point established for this review.
Stein 1991	No fundamental data to assess validity.
Steiner 1977	No information regarding random allocation or allocation concealment. Insufficient baseline data provided.
Szarfarc 2001	Non‐anaemic women.
Tange 1993	Non‐randomised trial. The number of anaemic participant was not provided for each group of treatment.
Thirunavukkarasu 2009	Not a controlled trial.
Tomar 2006	Non‐randomised study. Only the abstract was available.
Valli Rani 1995	They used sequential strategy for allocation and not a random one.
Van den Broek 2006	Non‐iron trial.
Visca 1996	No fundamental data to assess validity.
Von Peiker 1986	The differences between the groups was not iron, but vitamins.
Wu 1998	The numbers of participants in the 3 groups are not similar (93 for maternal, 50 for ferrous sulphate and 35 for ferroids) and is higher at follow‐up than at the beginning of the trial. It is unclear how they were allocated to the groups and whether therefore they are similar at baseline.
Young 2000	A weekly iron/folate supplement was compared with a standard daily iron/folate supplement in pregnant women living in rural Malawi. Acid folic trial.
Zhou 2006	Non‐anaemic women.

Hb: haemoglobin 
 vs: versus

Characteristics of studies awaiting assessment [ordered by study ID]

Paleru 2011.

Methods	Randomised open label controlled trial. No further details on the methods to generate randomisation and allocation concealment were provided.
Participants	80 antenatal women with anaemia (Hb levels between 8 and 10.5 g/dL) and serum ferritin < 13 mcg/L.
Interventions	300 mg oral iron sulphate per day vs. IV iron sucrose
Outcomes	Levels of Hb and ferritin on the day 28th and at delivery. No data on findings were provided.
Notes	Poster presentation. No numerical data were provided for analysis.

Pandit 2009.

Methods	Randomised controlled trial. No further details on the methods to generate randomisation and allocation concealment were provided.
Participants	100 anaemic pregnant women were recruited. All women were selected between 16 to 24 weeks of gestational age, irrespective of their parity, socio‐economic status and dietary habits.
Interventions	Women were randomly divided into study and control groups and were given study group (N = 50) – IV iron sucrose injections. Control group (N = 50) – oral ferrous sulphate tablets.
Outcomes	The outcome of pregnancy, birthweight of babies and side effects of both methods were also studied.
Notes	Poster presentation. No numerical data were provided for analysis. We attempted to find the authors on the Web but this strategy was unsuccessful.

Sarkate 2007.

Methods	Double blind randomised controlled trial.
Participants	Pregnant women with gestation period between 12 and 26 weeks having serum Hb < 10 g/dL, serum ferritin < 12 mcg/L.
Interventions	Group 1: sodium feredetate (33 mg elemental iron) + vitamin B12 (15 mcg) + folic acid (1.5 mcg), twice day. Group 2: sodium feredetate (66 mg elemental iron) + vitamin B12 (15 mcg) + folic acid (1.5 mcg), twice day. Group 3: Ferrous fumarate (100 mg elemental iron) + vitamin B12 (15 mcg) + folic acid (1.5 mcg), twice day.
Outcomes	Haematological levels; adverse events. 77% of patients completed all the visits. 11 patients withdrawn due to treatment failure at day 45. The data for their subsequent follow‐up visits were computed using the principle of last observation. More patients were recruited due to drop‐outs. Trialists did not report the total number of patients that were finally enrolled. Data of 48 patients were available for analysis.
Notes	Trialist did not report the total number of participants that were analysed in each group and no further analysis was possible. We attempted to contact some of the authors by electronic email but no response was obtained.

Hb: haemoglobin 
 IV: intravenous 
 vs: versus

Characteristics of ongoing studies [ordered by study ID]

Ruangvutilet 2009.

Trial name or title
Methods	Open randomised controlled trial. Phase IV study. No further details on the methods to generate randomisation and allocation concealment were provided.
Participants	Women aged between 18 and 45 years old, singleton pregnancy at 32 weeks' gestation, having anaemia (less than 11 g/dL). No underlying disease.
Interventions	Intravenous versus oral iron (Venofer vs Ferli‐6). No further detail was provided.
Outcomes	Serum ferritin, Hb level.
Starting date
Contact information
Notes	Ongoing study. We attempted to contact some of the authors by electronic email but no response was obtained.

Hb: haemoglobin

Differences between protocol and review

We have modified the 'Outcome measures' as a result of suggestions from a referee.

Contributions of authors

For the 2011 update, Ludovic Reveiz and Alexandra Casasbuenas appraised the papers independently. Ludovic Reveiz took the lead on writing the review, with Gill Gyte, Alexandra Casasbuenas and Luis Gabriel Cuervo providing comments on the various drafts, and revised the review in response to the editorial feedback.

The first published review was prepared by Luis Gabriel Cuervo and Kassam Mahomed (Cuervo 2001). The second published review was prepared by Ludovic Reveiz, Gill Gyte and Luis Gabriel Cuervo (Reveiz 2007).

Sources of support

Internal sources

• None, Not specified.

External sources

• None, Not specified.

Declarations of interest

Ludovic Reveiz and Luis Gabriel Cuervo have contributed to this systematic review in a personal capacity and during their spare time. Most of their contributions were made before joining the Pan American Health Organization. The Pan American Health Organization does not assume responsibility for the statements contained therein.

Edited (no change to conclusions)