Safety and efficacy of intravenous iron polymaltose, iron sucrose and ferric carboxymaltose in pregnancy: A systematic review

Transcription

1 Aust N Z J Obstet Gynaecol 2018; 58: DOI: /ajo REVIEW ARTICLE Safety and efficacy of intravenous iron polymaltose, iron sucrose and ferric carboxymaltose in pregnancy: A systematic review Alaa Qassim 1, Ben W. Mol 2, Rosalie M. Grivell 2,3,4 and Luke E. Grzeskowiak 2,5 1 School of Pharmacy and Medical Sciences, University of South Australia, Adelaide, SA, Australia 2 The Robinson Research Institute, Adelaide Medical School, The University of Adelaide, Adelaide, SA, Australia 3 Department of Obstetrics and Gynaecology, Flinders Medical Centre, Bedford Park, SA, Australia 4 Department of Obstetrics and Gynaecology, School of Medicine, Flinders University, Bedford Park, SA, Australia 5 SA Pharmacy, SA Health, Flinders Medical Centre, Bedford Park, SA, Australia Correspondence: Dr Luke Grzeskowiak, Centre for Perinatal Medicine, Flinders Medical Centre, Flinders Drive, Bedford Park, SA 5042, Australia. luke.grzeskowiak@adelaide.edu.au Conflict of Interest: The authors have no conflicts of interests to disclose. Received: 10 January 2017; Accepted: 2 August 2017 Background: Intravenous (IV) iron in pregnancy is useful where oral iron is not tolerated or a rapid replenishment of iron is required. Aims: To review the literature on the efficacy and safety of different IV iron preparations in the management of antenatal iron- deficiency anaemia (IDA). Materials and methods: We searched MEDLINE, Embase and Scopus from inception to June Eligible studies were randomised controlled trials (RCTs) and observational studies, involving administration of IV iron (ferric carboxymaltose (FCM), iron polymaltose (IPM) or iron sucrose (IS)), regardless of comparator, to manage antenatal IDA. Two independent reviewers selected studies, extracted data and assessed quality. Results: A total of 47 studies were eligible (21 RCTs and 26 observational studies), investigating IS (n = 2635; 41 studies), FCM (n = 276; four studies) and IPM (n = 164; three studies). All IV preparations resulted in significant improvements in haematological parameters, with a median increase of 21.8 g/l at 3 4 weeks and 30.1 g/l by delivery, but there was no evidence of any associated improvements in clinical outcomes. A greater median increase in Hb was observed with a high (25 g/l; range: g/l) compared with low (20 g/l; range: g/l). The median prevalence of adverse drug reactions for IPM (2.2%; range: 0 4.5%) was lower than FCM (5.0%; range: 0 20%) and IS (6.7%; range: %). Conclusion: While IV iron in pregnancy improves haematological parameters, there is an absence of evidence for improvements in important maternal or perinatal outcomes. No single preparation of IV iron appeared to be superior, with the current IV iron preparation of choice largely determined by cost and convenience around administration. KEYWORDS anemia, iron-deficiency/drug therapy, ferric compounds/administration and dosage, pregnancy, treatment outcome INTRODUCTION Iron deficiency represents the most widespread nutritional deficiency globally and is the leading cause of anaemia during pregnancy. 1 Anaemia is estimated to affect almost 40% of pregnancies worldwide and is associated with a significantly increased risk of perinatal morbidity and mortality, 1,2 including risk of low birth weight and preterm birth. 3 Further, the iron status of infants have been demonstrated to mirror that of the mother, with infants born to anaemic women at increased risk of developing iron The Royal Australian and New Zealand College of Obstetricians and Gynaecologists wileyonlinelibrary.com/journal/anzjog

2 A. Qassim et al. 23 deficiency anaemia (IDA) in the first year of life. 4 These findings are independent of birth weight or gestational age and occur despite infants born to anaemic women having normal haemoglobin status at birth. 4 Such findings are of particular interest given the previously noted association between IDA in the third trimester and impaired mental development of the offspring. 5 While haemoglobin status in pregnancy has been demonstrated to be of importance, the role of iron in ameliorating such adverse outcomes in the puerperal period is less clear, although it can be postulated that if the anaemia is due to iron deficiency, replenishing iron stores should restore normal haemoglobin and lead to improved perinatal outcomes. 6 Contrary to this belief, supplementation of anaemic or non- anaemic pregnant women with iron, folic acid or both, is not associated with improvements in birth weight or gestational age. 7 Despite such uncertainty, both oral and intravenous (IV) iron are commonly utilised in clinical practice. Use of oral iron is usually considered first- line due to convenience and low cost; however, its use is limited by significant gastrointestinal side effects affecting adherence, low absorption rate, and delay in replenishing iron stores. IV iron is considered to be a useful option when treatment with oral iron has failed or when rapid replenishment is required, albeit without strong evidence supporting its role in pregnancy. Increasing use of IV iron in clinical practice has raised some concerns regarding potential harms as iron itself can lead to oxidative stress and inflammatory changes and evidence that administration of oral iron is associated with an increased risk of preterm birth in observational studies but not RCTs. 2 Multiple formulations of IV iron are available, differing on aspects such as the maximum single that can be administered in one sitting, the total infusion time and associated administration and medication- related costs (Table 1). In light of the frequency of use in clinical practice and the multiple formulations of IV iron that are available, we undertook a systematic review to evaluate the efficacy and safety of IV iron polymaltose (IPM), iron sucrose (IS) and ferric carboxymaltose (FCM) in the management of antenatal IDA. MATERIALS AND METHODS Search strategy MEDLINE, Embase and Scopus were searched for studies investigating the use of IV iron in pregnancy. The search included an obstetric concept and an IV iron concept. Terms were searched as controlled vocabulary in applicable databases and as keywords in all databases. A complete list of search terms used is presented in Appendix S1. All searches were conducted on 8 June, Reference lists of all articles included in full- text review, as well as other review articles, were screened for additional studies. Eligibility criteria The study types for this review included randomised controlled trials, quasi randomised controlled trials, and observational studies, published in any language and any country. Eligible studies were those involving administration of IV iron (FCM, IPM or IS, regardless of comparator, to manage antenatal IDA. Studies only published in abstract form were not eligible for inclusion. Data abstraction Two independent, non- blinded authors (AQ and LEG) reviewed each title and abstract for inclusion eligibility. Full- text review was also conducted by two independent, non- blinded authors (AQ and LEG) and discrepancies were resolved through author consensus discussions. For non- English language studies included in the TABLE 1 Comparison of different intravenous iron preparations Preparations Iron sucrose Iron polymaltose Ferric carboxymaltose Trade (brand) name Venofer Ferrum H, Ferrosig Ferinject Concentration of elemental iron 20 mg/ml 50 mg/ml 50 mg/ml Maximum in single administration for patients 35 kg 100 mg Up to 2500 mg 20 mg/kg with max. of 1000 mg Total single infusion No Yes No (unless total body iron deficit is 1000 mg) Frequency of administration Not more than three times per week Not required Not less than seven days apart Infusion time for maximum 15 min 4 5 h 15 min Approximate cost $40/500 mg $20/500 mg $150/500 mg Eligible for PBS subsidy Yes Yes Yes PBS, Pharmaceutical Benefits Scheme. Data taken from Product Information (Available from accessed June 2017). Costs is based on dispensed price for maximum quantity which represents the base dispensed price for this medication as outlined by the Australian PBS as of June 2017 ( The cost to the patient is dependent on whether they are eligible for a subsidy. These prices are an approximation only as there is considerable variability depending on purchasing situation or country of origin.

3 24 Intravenous iron in pregnancy: a review full- text review, the primary author (AQ) translated the contents with computer translation software, which has previously been demonstrated to be effective for systematic reviews. 8 The study selection process was documented as per the Preferred Reporting Items for Systematic Review and Meta- Analysis (PRISMA). Data collection from eligible studies was undertaken using an electronic data extraction form. Data were collected in relation to study characteristics (including study type, eligibility and exclusion criteria), intervention type (ie type of IV iron), comparator treatment (if applicable), iron dosing regimen, and all outcome measures. Outcomes of interest included change in haematological parameters (ie haemoglobin, ferritin), treatment- related adverse drug reactions (ADRs), requirement for antenatal or postnatal blood transfusion, maternal wellbeing and quality of life (including postnatal depression and breastfeeding outcomes), and perinatal outcomes (eg gestational age, birth weight, postpartum haemorrhage, method of delivery). Study quality assessment Study quality assessment was conducted by two independent, non- blinded authors (AQ and LEG) and discrepancies were resolved through author consensus discussions. To assess the internal quality of eligible studies, the Cochrane Collaboration s tool for assessing risk of bias was used for RCTs, 9 which consists of the following domains: selection of participants, confounding variables, measurement of exposure, blinding of outcomes, incomplete outcome data and selective reporting. For non- randomised studies, the Risk of Bias Assessment Tool for Non Randomised Studies (RoBANS) was used. 10 The RoBANS tool is similar in construct to the Cochrane tool and comprises six domains: selection of participants, confounding variables, measurement of exposure, blinding of outcomes, incomplete outcome data, and selective reporting. For all studies, the risk of bias was evaluated according to the criteria outlined by each relevant tool and judgement was reported by assigning low risk, high risk and unclear risk to each domain. RESULTS Search results Our search identified 1263 abstracts, comprising 1035 unique abstracts. An additional 11 articles were identified from previous systematic reviews or from checking citations of included studies and retrieved for full- text review. A total of 76 articles were included in full- text review and of these, 47 were deemed eligible for inclusion in the systematic review (Fig. 1). The most common reason for exclusion in the full- text review was articles published in only abstract form (n = 22/29). A summary of all articles included in this review is presented in Tables 2 and 3, while a summary of exclusion criteria and outcomes investigated in each study is presented in Appendices S2 and S3. Overview of included studies Of the 47 eligible studies, 21 were RCTs and 26 were observational studies. Types of IV iron formulation studied included IS (n = 2635 women; 41 studies ), FCM (n = 276 women; four studies 13,52 54 ) and IPM (n = 164 women; three studies ). The majority of studies involved a comparison group of some kind including oral iron (n = 17; 36%), 11,14,16,19,24,26,27,29,31,32,34,35, 37,39,50,56,57 IV iron (n = 4; 9%), 13,15,36,52 intramuscular (IM) (n = 3; 6%) 25,28,38 or other (n = 5; 11%), 30,33,42,46,54 while 18 studies (38%) did not involve any comparator group. 12,17,18,20 23,40,41,43 45,47 49,51,53,55 More than half of the studies were published since 2010 (n = 29; 62%), with the countries most frequently represented being India (n = 19; 40%), USA (n = 7; 15%) and Pakistan (n = 5; 11%) (Appendix S4). Substantial variation was observed across studies with respect to the prescribed, related to differences in baseline anaemia severity and target Hb used in calculations, and administration regimen. The majority of studies used a variable of iron (n = 41; 87%), 12 18,21 35,37 42,44 46,48 57 taking into account factors such as the haemoglobin deficit and body weight, while the remainder used a fixed (n = 6; 13%). 11,19,20,36,43,47 Of the studies that administered a variable, 32 of 41 reported using the Ganzoni formula with a target haemoglobin of 100 (n = 3), 29,33, (n = 16), 14 16,18,23,24,26 28,30,38,39,41,44,49, (n = 5), 31,34,35,37, (n = 2), 22, (n = 2), 12,21 or 150 (n = 3) g/l. 45,55,57 Efficacy All preparations resulted in significant improvements in haematological parameters, with an overall median increase of 21 g/l (range: g/l) at 3 4 weeks 12,14 18,20,21,24 26,28,31 39,41,42,44,49 53,55 and 34.3 g/l (range: g/l) by delivery. 15,16,24,28,37,39,43,49,50 Ferritin increased by a median of 27 μg/l (range: μg/l) over the first 4 weeks following treatment initiation. 12,14,16,18,24,25,30,35,41,49,50,56 When comparing individual types of IV iron, administration of IPM was associated with the greatest median increase in haemoglobin by 3 4 weeks following initiation of treatment (34 g/l; range: g/l), 55 followed by IS (21 g/l; range: 9 51 g/l) 12,14 18,20,21,24 26,28,31 39,41,42,44,49 51 and FCM (15 g/l; range: 6 26 g/l). 52,53 When separated according to administered (based on target Hb used in dosing calculation of 130 or >130 g/l), a greater median increase in Hb was observed with a high (24.2 g/l; range: g/l) 12,21,32,55 compared with low (20 g/l; range: g/l) ,18,20,24,26,28,31,33 39,41,42,44,49,50,53 Only one study was identified that directly compared the administration of a full replacement of IV iron to a that was 300 mg less than the calculated full replacement. In comparing the two groups, administration of the full replacement was associated with a greater increase of 8 g/l 3 4 weeks after administration and 14 g/l at delivery. 28 Compared to oral iron, use of IV iron was associated with a greater increase in haemoglobin at 2 4 weeks following the

4 A. Qassim et al. 25 FIGURE 1 Flow diagram of included studies. infusion (median: 6 g/l; range: 1 20 g/l) 11,16,26,27,29,31,34,35,37,39,50,56 and at delivery (median: 6.8 g/l; range: 6 19 g/l). 14,16,32,56 When examining ferritin at 4 weeks, seven of eight studies demonstrated a greater median increase of 38 μg/l (range: μg/l) among the group receiving IV rather than oral therapy. 11,24,26,27,35,37,50,56 Only two studies reported on difference in ferritin at delivery, which was 43 and 108 μg/l, respectively. 32,56 More women reached target Hb by delivery who received IV compared with oral iron (76%; 62 88% vs 61%; range: 20 76%). 14,16,24,28,29,56 Data on bleeding and transfusion- related outcomes were scarce. Following the introduction of IS into clinical practice, Shrivastava et al. demonstrated a reduction in peripartum blood transfusions occurring among anaemic women from 29% to 20% (P < 0.001). 33 Perinatal outcomes Few studies incorporated an investigation of perinatal outcomes. Christoph et al. demonstrated a difference in gestational age at birth among two cohorts of women who received either IV FCM or IS (38.4 vs 39.6 weeks; P = 0.009), but this was not adjusted for important maternal characteristics that differed between the two groups. 13 Among other comparative studies, none demonstrated any statistically significant differences between groups across any perinatal outcomes investigated, including birth weight, gestational age, method of delivery, and postpartum haemorrhage. 11,14,16,24,30,31,37,38,54,56 Notably, the small study samples included in each of these studies limited the statistical power to detect clinically relevant differences between the two groups. Quality of life Quality of life measures were only reported in two studies. 53,58 Both studies included assessments following delivery. Froessler et al. reported outcomes for 29 (44.6%) women enrolled in their non- comparative study cohort who all received IV FCM, with 19 (65.5%) of them reporting improvement in their wellbeing and the remainder feeling no change. Khallafallah et al. reported outcomes for 126 (68.9%) women enrolled in their randomised controlled trial of IV IPM versus oral iron, demonstrating that, in

5 26 Intravenous iron in pregnancy: a review TABLE 2 Characteristics of randomised controlled trials of intravenous iron in pregnancy Study (country) Participant characteristics control IV dosing and infusion Efficacy outcomes IV side effects Abhilashini et al. 2014; India weeks gestation with IDA (Hb g/l). IV (n = 50): iron polymaltose variable Oral (n = 50): iron sulphate (80 mg elemental iron) three times daily Dose: (Weight (11 g/dl actual Hb) 2.4) mg Administration: 200 mg on alternate days (maximum of 600 mg/week) Test : yes Hb mean difference from baseline: Week 2: Hb increase: IV 1.27 ± 0.43 g/dl versus oral 1.07 ± 0.45 g/dl (P = 0.026) Week 4: IV 2.59 ± 0.72 g/dl versus oral 1.99 ± 0.68 g/dl (P < 0.001) Delivery: IV 3.95 ± 0.56 g/dl versus oral 2.93 ± 0.57 g/dl (P < 0.001) Urticaria (n = 1), myalgia (n = 1) Aggarwal et al. 2012; India >24 weeks gestation with Hb 7 g/dl, transferrin saturation 10% and ferritin 15 μg/l IV (n = 25): IS variable oral (n = 25): iron sulphate 60 mg elemental iron three times daily Dose: (pre- pregnancy weight (11 g/ dl actual Hb] 0.24) mg Administration: 200 mg on alternate days. Treatment continued until administration of calculated or reaching target Hb of 11 g/dl. Test : yes Hb treatment success: Week 4: IV 76% versus oral 32% reached Hb >11 g/dl by 4 weeks (P < 0.002) Hb mean difference from baseline: Week 4: IV 6.27 ± 0.48 g/dl to 11.3 ± 0.7 g/dl versus oral 5.94 ± 0.62 g/ dl to ± g/dl (P = ) Ferritin: Week 4: IV ± 45.1 μg/l versus oral ± 33.1 μg/l (P < ) Fever (n = 1; 4%), arthritis (n = 1; 4%), altered taste (n = 1; 16%) and thrombophlebitis (n = 1; 4%). Allergic reaction requiring treatment (n = 2; 8%) Al et al. 2005; Turkey weeks gestation with established IDA (Hb g/dl and ferritin <13 μg/l) IV (n = 45): IS variable Oral (n = 45): iron polymaltose (100 mg elemental iron) three times daily Dose: (pre- pregnancy weight (110 g/l actual Administration: 200 mg on alternate days : Week 2: IV 0.6 versus oral 0.2 g/dl; P = Week 4: IV 1.2 versus oral 0.6 g/dl; P < Delivery: IV 2.1 versus oral 1.5 g/dl; P = Postpartum: IV 0.8 versus oral 0.2 g/dl; P = 0.01 Hb treatment success: Delivery: IV 95.6% versus oral 62.2% reached Hb>11 g/dl by delivery (P < 0.001) Metallic taste (n = 11), hot flush (n = 12), arthralgia (n = 1), dizziness (n = 8), nausea (n = 5), vomiting (n = 1). Bayoumeu et al. 2002; France Hb between 8 and 10 g/dl at 6 months of pregnancy, MCV <100 fl and ferritin level <50 μg/l. IV (n = 24): IS variable Oral (n = 23): ferrous sulphate 80 mg elemental iron three times daily for 4 weeks Dose: (pre- pregnancy weight (120 g/l actual Administration: 200 mg on days 1, 4, 8, 12, 15 and 21. Treatment continued until administration of calculated or reaching target Hb of 12 g/dl. Test : yes Hb change from baseline: Week 4: IV 9.6 ± 0.79 g/dl to ± 1.3 g/dl versus oral 9.7 ± 0.5 g/ dl to 11 ± 1.25 g/dl (P > 0.05) Metallic taste (n = 1) (Continues)

6 A. Qassim et al. 27 TABLE 2 (Continued) Study (country) Participant characteristics Bhargava and Maheshwari 2013; India Unknown gestation with Hb 7 9 g/dl, MCV<78 fl and serum ferritin <20 μg/l Breymann et al. 2001; Switzerland >21 weeks gestation with iron- resistant IDA, ferritin level <15 μg/l, and Hb levels remained <10 g/dl despite receiving 160 mg elemental iron orally for previous 2 weeks Deeba et al. 2012; India weeks gestation with IDA, Hb 6 10 g/dl and ferritin <15 ng/ml Dhanani et al. 2012; India weeks gestation with IDA with Hb <8.5 g/dl Divakar 2011; India weeks gestation with Hb <11 g/dl control IV dosing and infusion Efficacy outcomes IV side effects IV (n = 50): IS variable replacement Oral (n = 50): ferrous ascorbate (100 mg elemental iron) variable daily not specified Dose: (weight (11 g/dl actual Hb) 2.21) mg Administration: 200 mg on alternate days, maximum 600 mg/week Test : yes Week 2: IV 3.2 ± 0.6 g/dl versus oral 2.0 ± 0.3 g/dl (P < 0.001) Ferritin Week 2: IV 46.5 ± 14.1 μg/l versus oral 13.3 ± 5.3 μg/l (P < 0.001) No side effects IV + rhepo (n = 20): IS and erythropoietin IV only (n = 20): IS as above Dose: calculation not specified Administration: 200 mg twice weekly (72 to 96 h apart) Test : yes Hb treatment success Week 4: IV + EPO 95% (n = 19) versus IV only 75% (n = 15) reached target Hb 11 g/dl Metallic taste (n = 3); feeling warm (n = 2); not reported individually for each treatment group IV (n = 100): IS variable replacement Oral (n = 100): ferrous ascorbate 200 mg elemental iron daily IV : (weight (12 g/dl actual Hb) 2.4) mg Administration: 200 mg on alternate days Week 2: IV 1.72 ± 0.48 g/dl versus oral 0.58 ± 0.47 g/dl Week 4: IV 2.18 ± 0.87 g/dl versus oral 1.39 ± 0.44 g/dl Week 6: IV 2.89 ± 0.60 g/dl versus oral 1.90 ± 0.30 g/dl Ferritin: Week 2: IV 40.0 ± 17.0 ng/ml versus oral 8.5 ± 4.5 ng/ml Week 4: IV 52.6 ± 19.9 ng/ml versus oral 15.2 ± 8.1 ng/ml Week 6: IV 78.5 ± 19.8 ng/ml versus oral 26.6 ± 8.6 ng/ml Metallic taste (n = 5); Hot flushes (n = 2); Arthralgia (n = 1); Dizziness (n = 1); nausea (n = 4) IV (n = 29): IS variable IM (n = 23): iron sorbitol 300 mg in four divided s IV : (weight (target Hb* actual Hb) 0.24) +500 mg. *Unspecified target Hb Administration: 200 mg in 100 ml IS infused over half an hour. Test : yes Hb difference between groups Week 4: IV 7.59 ± 1.38 g/dl to 9.25 ± 1.91 g/dl versus 8.32 ± 1.34 to 9.77 ± 1.64 g/dl; P > 0.05 Burning/pain at injection site (n = 1); swelling at site of injection (n = 1) IV (n = 75): IS as slow intravenous infusion IV (n = 77): IS as rapid bolus push Dose: 400 mg in two divided s given 2 4 days apart Administration: given either as slow infusion of 100 ml over 30 min or rapid bolus over 2 5 min. Week 4: IV slow 0.92 ± 0.75 g/dl versus IV rapid 1.10 ± 0.87 g/dl Side effects reported by five (6.7%) in slow infusion group and 15 (19.5%) in rapid bolus group, including local pain, itching/burning, nausea, swelling in extremities and chest discomfort. (Continues)

7 28 Intravenous iron in pregnancy: a review TABLE 2 (Continued) Study (country) Participant characteristics Froessler et al. 2013; Australia weeks gestation with IDA (Hb <110 g/l) (ferritin <12 mcg/l) Gupta et al. 2014; India weeks gestation with Hb 7 9 g/dl and ferritin levels <15 ng/ml Halimi et al. 2011; Pakistan weeks gestation with Hb <11 g/dl Khalafallah et al. 2010; Australia <28 weeks gestation with moderate IDA (Hb <115 g L but >85 g L with serum ferritin <30 g L) control IV dosing and infusion Efficacy outcomes IV side effects IV (n = 69): IS fixed Oral (n = 51): ferrous sulphate 80 mg elemental iron twice daily Dose: 400 mg IV IS given as two 200 mg infusions over 30 min at least 24 h apart Hb difference between groups Day 1: IV 101 g/l versus oral 107 g/l (P = 0.2) Day 14: IV 128 g/l versus oral 129 g/l (P = 0.4) Day 42: IV 127 g/l versus oral 127 g/l (P = 0.9). Ferritin difference between groups Day 1: IV 33 versus oral 21 (P = 0.06) Day 14: IV 39 versus oral 40 (P = 0.4) Day 42: IV 27 versus oral 41 (P = 0.4). No side effects IV (n = 50): IS Oral (n = 50): ferrous sulphate (80 mg elemental iron) three times daily for 4 weeks Dose: (weight (11 g/dl actual Hb) 2.4) mg Administration: 200 mg on alternate days with 600 mg max/week Day 14: IV 0.58 g/dl versus oral 0.23 g/ dl (P = 0.004) Day 28: IV 1.9 g/dl versus oral 1.3 g/dl (P < 0.001) Delivery: IV 3.53 g/dl versus oral 2.43 g/ dl (P < 0.001) Hb treatment success Delivery: IV 76% versus oral 54% reached Hb>11 g/dl by delivery (P = 0.003) Local pain (n = 1); thrombophlebitis (n = 2); fever (n = 1); urticaria (n = 1) IV (n = 50): IS variable Oral (n = 50): ferrous sulphate 80 mg elemental iron three times daily Dose: (weight (12 g/dl Administration: diven in two divided s on days 0 and 15. Test : yes Week 4: IV 3.45 ± 1.06 versus oral 1.85 ± 0.28 g/dl Nausea (n = 17; 34%), unpleasant taste (n = 9; 18%) IV + oral (n = 92): IV polymaltose variable s followed by oral iron as in control Oral alone (n = 91): iron sulphate (80 mg elemental iron) daily Dose: (pre- pregnancy weight (max 90 kg) (120 g/l Administration: single Test : yes Hb difference between groups Delivery: IV 6.6 g/l (95% CI: g/l) higher Postnatal: IV 5.8 g/l (95% CI: g/l) higher Hb treatment success Delivery: reduction in proportion of women with anaemia pre- delivery (IV 16% versus oral 29%: IRR 0.55: 95% CI: ) Ferritin delivery: reduction in proportion of women with low ferritin pre- delivery (IV 4.5% versus oral 79%; P < 0.001) Severe urticarial reaction (n = 2; 2.2%); requiring infusion cessation (Continues)

8 A. Qassim et al. 29 TABLE 2 (Continued) Study (country) Participant characteristics Kochhar et al. 2013; India weeks gestation with moderate IDA (Hb 7 9 g/dl; ferritin <15 ng/ml) Mehta and Shah 2014; India weeks gestation with IDA (Hb <8 g/dl) Neeru et al. 2012; India weeks gestation with IDA (Hb levels g/ dl and ferritin <27 ng/dl) Rudra et al. 2016; India weeks gestation with IDA (Hb levels g/dl) control IV dosing and infusion Efficacy outcomes IV side effects IV (n = 49) IS as variable Oral (n = 49): ferrous sulphate 60 mg elemental iron three times daily Dose: (weight (12 g/dl actual Administration: 200 mg on alternate days Test : yes Hb difference between groups Week 4: IV 12.8 ± 1.1 g/dl versus oral 10.7 ± 0.7 g/dl (P = 0.002) Delivery: IV 13.4 ± 0.9 g/dl versus oral 11.2 ± 0.9 g/dl (P = 0.002) Ferritin Week 4: IV ± 13.4 g/dl versus oral 77.6 ± 13.7 g/dl (P = 0.005) Delivery: IV ± 15.8 g/dl versus oral 94.6 ± 14.2 g/dl (P = 0.001) Heartburn (n = 1), constipation (n = 2). IV (n = 75): IS variable s Oral (n = 75): ferrous sulphate 60 mg elemental iron two tablets three times daily Dose: (weight (10 g/dl Actual Hb) 2.4) mg. Administration: 100 mg on alternate days. Week 6: IV 3.93 ± 0.6 g/dl versus oral 3.45 ± 0.68 g/dl Hb treatment success Week 6: IV 88% (n = 66) versus 76% (n = 57) Pain at injection side (n = 15; 20%), phlebitis (n = 6; 8%), fever (n = 5; 7%). IV (n = 45): IS Oral (n = 44): ferrous fumarate 100 mg elemental iron once daily Dose: ((11 g/dl actual Hb) 250) mg Administration: 200 mg on alternate days Hb change from baseline Timing not specified: IV 9.18 ± 0.94 g/dl to ± 0.70 g/dl versus oral 9.75 ± 0.83 g/dl to ± 0.63 g/dl Ferritin timing not specified: IV 8.60 ± 5.17 g/dl to ± g/ dl versus oral ± 7.55 g/dl to ± g/ dl Hb treatment success Delivery: IV 66% versus oral 61% reached target Hb 11 g/dl Side effects reported by six (13%) women, with four (9%) requiring treatment cessation; individual reactions not specified IV (n = 100): IS variable replacement. Oral (n = 100): ferrous ascorbate 100 mg twice daily. Dose: (weight (11 g/dl Actual Hb) 2.4) mg Administration: 200 mg on alternate days Week 2: IV ± 0.13 g/dl versus oral 0.22 ± 0.16 g/dl (P < ) Week 4: IV 1.99 ± 0.18 g/dl versus oral 1.27 ± 0.30 g/dl (P < ) Delivery: IV 3.67 ± 0.17 g/dl versus oral 3.02 ± 0.17 g/dl (P < ) ferritin Week 4: IV ± 1.46 ng/ml to ± 4.37 ng/ml versus oral ± 1.86 ng/ml to ± 1.86 ng/ ml (P < ) Fever (n = 2); itching (n = 1); injection site swelling, redness or pain (n = 4); arthralgia (n = 1); nausea (n = 4) (Continues)

9 30 Intravenous iron in pregnancy: a review TABLE 2 (Continued) Study (country) Participant characteristics control IV dosing and infusion Efficacy outcomes IV side effects Singh et al. 1998; Singapore weeks gestation with IDA (Hb <90 g/l and ferritin <20 μg/l) IV (n = 50): iron polymaltose variable replacement Oral (n = 50): ferrous fumarate (200 mg) three times daily Dose: (weight (150 g/l actual Administration: pre- treatment with promethazine. Hb difference between groups 36 weeks gestation: IV 110 ± 0.1 g/l versus oral 99 ± 0.2 g/l Ferritin 36 weeks gestation: IV ± 57.2 μg/l versus oral 20.2 ± 2.1 μg/l None reported Singh et al. 2013; India weeks gestation with Hb <8 g/dl and serum iron <60 μg/dl IV (n = 50): IS given as variable replacement IM (n = 50): iron sorbitol Dose: (weight (11 g/dl actual Hb) 0.3) + (weight 10 mg/kg) Administration: 150 mg every third day Hb difference between groups Week 2: IV: 2.3 g/dl versus IM 1.3 g/dl; P < 0.01 Week 4: IV: 3.5 g/dl versus IM: 2.3 g/dl; P < 0.01 Hb treatment success Week 8: IV 90% versus IM 34% reached target Hb (P < 0.001) IV group achieved target Hb significantly quicker (6.4 vs 9.0 weeks; P < 0.01). Side effects reported by four (8%) women; individual reactions not specified Tariq et al. 2015; Pakistan >12 weeks gestation with confirmed IDA (Hb <10.5 g/ dl and ferritin <12 μg/l), unable to tolerate or non- compliant to oral iron or suffering from IBS included IV (n = 93): IS variable IV (n = 105): ID variable Dose: (weight (11 g/dl actual Administration: given in divided s Test : yes 4 weeks: IS 9.09 ± 0.83 g/dl to ± 1.10 g/dl versus dextran 8.74 ± 0.96 g/dl to ± 1.22 g/dl Delivery: IS 9.09 ± 0.83 g/dl to ± 0.87 g/dl versus dextran 8.74 ± 0.96 g/dl to ± 1.11 g/dl Palpitation (n = 1), shivering (n = 2), hypotension (n = 1), flushing (n = 1), arthralgia (n = 1). Wali et al. 2002; Pakistan weeks gestation with IDA (Hb 5 10 g/dl) IV Group A (n = 15): IS as full replacement IV Group B (n = 20): IS as reduced replacement IM (n = 25): iron sorbitol fixed 750 mg Dose: IV A (full replacement ): (weight (11 g/dl actual IV B (reduced replacement ): (weight (11 g/dl actual Hb) 0.24) mg Administration: up to 500 mg per infusion; usually in divided s of mg Test : yes Week 3 4: IV A 2.7 ± 1.1 g/dl versus IV B 1.9 ± 0.8 g/dl Delivery: IV A 8.0 ± 1.1 g/dl to 11.8 ± 1.1 g/dl (P < 0.001), IV B 8.9 ± 0.7 g/dl to 11.3 ± 0.9 g/dl (P < 0.001), IM 8.8 ± 0.9 g/dl to 10.2 ± 1.2 g/dl (P > 0.05) Hb treatment success Delivery: IV A 80% versus IV B 70% vs IM 28% reached target Hb Abdominal pain (n = 1); shivering/weakness within few hours of infusion (n = 2); Phlebitis at injection site (n = 3) IBS, irritable bowel syndrome; IDA, iron deficiency anaemia; IM, intramuscular; IRR, incidence rate ratio; IS, iron sucrose; IV, intravenous; MCV, mean corpuscular volume; rhepo, recombinant human erythropoietin.

10 A. Qassim et al. 31 TABLE 3 Characteristics of observational studies of intravenous iron in pregnancy Study (country) Participant characteristic control IV and infusion Efficacy outcome IV side effects Akther and Afroj 2010; Bangladesh weeks gestation with moderate severe anaemia IV (n = 35): IS variable If Hb <9 g/dl after 4 weeks of oral therapy then 400 mg IS given (200 mg in 200 ml once and repeated after 3 days). Further two s of sucrose was given if Hb remained below 9 g/dl 2 4 weeks later. : significant improvement (P < 0.5) in Hb after treatment with IV IS: 8.6 g/dl to 9.66 g/dl Upper limb weakness (n = 2; 5.7%); Urticarial reaction (n = 1; 2.9%); both requiring cessation of treatment Al- Momen et al. 1996; Saudi Arabia <32 weeks gestation with severe IDA (Hb <90 g/l; ferritin <20 mcg/l) IV (n = 52): IS variable Oral (n = 59): Ferrous sulphate 60 mg elemental iron three times a day Dose: (130 g/dl actual Hb) weight mg/kg. Administration: 200 mg in every 1 3 days Test Dose: No : timing not specified: IV 75.8 ± 7.9 g/l to ± 6.6 g/l versus oral 76.6 ± 7.8 g/l to ± 12.4 g/l; P < Ferritin: timing not specified: IV 11.9 ± 5.0 μg/l to 95.5 ± 38.1 μg/l versu oral 12.0 ± 5.3 μg/l to 52.4 ± 3.1 μg/l; P < Fever after fifth in one patient (1.9%); treated with paracetamol Tightness and discomfort in the skin of whole body at end third in one patient (1.9%); infusion rate modified for subsequent s Bare and Sullivan 1960; USA 30 weeks gestation with Hb <10 g/dl IV (n = 54): IS variable Blood transfusion (n = 50): 500 ml Dose: (100% Hb actual Hb%) 25 mg. Administration: 100 mg for first infusion, then 200 mg alternate daily Test : yes : Timing not specified: Baseline Hb <8 4 g/dl versus transfusion 2.5 g/dl Baseline Hb = 8 9 g/dl IV 2.1 g/dl versus transfusion 2.1 g/dl Baseline Hb = 9 10 g/dl IV 2.1 g/dl versus transfusion 1.3 g/dl Side effects reported for 16 of 279 injections (5.7%) Local pain (n = 4), nausea (n = 1), flushing (n = 1), syncope (n = 5), dyspnoea (n = 2), forearm numbness (n = 1), urticaria (n = 1) and acute back pain (n = 1). Briscoe 1952; USA Unknown gestation with Hb 10 g/dl IV (n = 40): IS variable Placebo (n = 77) Dose: (100% Hb actual Hb%) 25 mg. Administration: initially three s were given 30 mg, 60 mg and 100 mg. Doses changed to 100 mg initially then 200 mg daily 3 days after. Dose was given over 4 7 days. : Week 4: IV 2.8 g/dl ( g/ dl) versus placebo 1.2 g/dl Not specifically reported for antenatal group. Christoph et al. 2012; Switzerland All pregnant women who received IV ferric carboxymaltose or IS over a defined period IV ferric carboxymaltose (n = 103) variable IV IS (n = 103) variable Dose: FCM calculation not specified (mean 13 ± 3.1 mg/kg/week) IS calculation not specified (mean 6 ± 0.9 mg/kg/week) Administration: not specified Timing not specified: FCM 15.4 ± 11.7 g/l IS 11.7 ± 10.0 g/l Side effects reported in eight (7.8%) women receiving FCM and 11 (10.7%) receiving IS (Continues)

11 32 Intravenous iron in pregnancy: a review TABLE 3 (Continued) Study (country) Participant characteristic Dewan et al. 2012; India Any gestation with anaemia Evans and Waltman 1953; USA Unknown gestation with anaemia Froessler et al. 2014; Australia weeks gestation with anaemia (Hb <115 g/l) Govan and Scott 1949; UK Unknown gestation with anaemia (Hb <96 g/l) Gupta et al. 2015; India weeks gestation with IDA (Hb 7 9 g/dl; ferritin <12 ng/ml) Holly 1951; USA weeks gestation with (n = 11) and without (n = 5) anaemia control IV and infusion Efficacy outcome IV side effects IV (n = 118): IS variable Dose: body weight (target Hb* actual Hb) 2.4 *Unspecified target Hb (additional 500 mg administered in case of chronic anaemia. Administration: Not Specified Week 4: 6.98 ± 1.35 g/dl to 9.28 ± 1.45 g/dl (P < ) Not evaluated in pregnancy specifically IV (n = 32): IS fixed Dose: 500 mg Administration: 100 mg every second day Test : yes Week 2 (n = 15): 1.04 g/dl (range: 1 5 g/dl) Delivery (n = 17): 2.47 g/dl (range: g/dl) Not specifically reported for antenatal group. IV (n = 65): ferric carboxymaltose variable Dose: up to 15 mg/kg Administration: single Hb Difference from baseline Week 3 6: 12 g/dl According to anaemia severity: Mild: ± 1 versus ± 2.9 g/l Moderate: 92.6 ± 0.4 versus ± 3.8 g/l Severe: 83.7 ± 0.9 versus 110 ± 8.1 g/l Minor side effects experienced by 13 women (20%); one patient required treatment with metoclopramide. IV (n = 25): IS variable Oral (n = 62): ferrous sulphate (unknown ) daily Dose: calculation not specified Administration: daily during first week 30 mg first day, 60 mg second day and 100 mg thereafter given on alternate days Week 1: IV 13.2 g/l versus oral 8.0 g/l Generalised infusion reaction (n = 3) Anaphylaxis (n = 1) IV (n = 45): IS variable Dose: ((11 g/dl actual Hb) weight 2.3) mg Administration: 200 mg on alternate days (max 600 mg/week) Week 4: 7.81 ± 0.43 to 9.81 ± 0.46 g/dl Delivery: 7.81 ± 0.43 to ± 0.78 g/dl ferritin Week 4: ± 3.8 to ± 5.7 ng/ml Shivering (n = 1) IV (n = 16): IS variable (ranged from 420 mg to 1430 mg). Dose: ((13.4 g/dl actual Hb) 0.2) mg. Administration: max single of 200 mg. Timing unclear: with anaemia 3.93 g/dl (range: 0.9 to 6 g/ dl); without anaemia 2.2 g/dl (range: g/dl) Generalised reaction (n = 1) sharp epigastric pain, dizziness, throbbing headache, shortness of breath (Continues)

12 A. Qassim et al. 33 TABLE 3 (Continued) Study (country) Participant characteristic control IV and infusion Efficacy outcome IV side effects Kartchner and Holmstrom 1950; USA Unknown gestation including anaemic women (A: Hb g/ dl; n = 9, B: Hb g/dl; n = 8) and non- anaemic women (C: Hb g/dl; n = 9) IV (n = 26): IS variable Dose: (15 g/dl actual Hb) g Administration: 100 mg, then 200 mg alternate daily Test : yes Delivery: Group A g/ dl Group B g/dl Group C g/dl Mild (27%): suffusion, weakness, light- headedness, minimal pain, cough and backache Moderate (6%): mild reactions lasting up to 5 min or accompanied by dizziness, pain along the vein or vomiting Severe (2%): symptoms lasting >5 min or shock- like symptoms Kriplani et al. 2013; India IDA (Hb 5 9 g/dl) IV (n = 97): IS given as variable s Dose: (pre- pregnancy weight (14 g/dl actual Hb) 2.4) mg. Administration: 200 mg twice weekly Week 4: 7.63 ± 0.61 g/dl at baseline to 9.90 ± 0.80 g/dl Week 8: 7.63 ± 0.61 g/dl to ± 0.73 g/dl ferritin Week 8: 11.2 ± ± 23.1 μg/l Nausea (n = 5), vomiting (n = 3), diarrhoea (n = 1), thrombophlebitis (n = 1), mild giddiness and restlessness (n = 1), fever (n = 2), haematemesis (n = 1) Myers et al. 2012; UK >12 weeks gestation with symptomatic IDA unresponsive to oral iron IV (n = 44): FCM variable IV (n = 48): ID 20 mg/kg up to 1000 mg Dose: calculation not specified, up to 1000 mg single Administration: not specified t specified Week 2: FCM 1.73 g/dl versus ID 1.34 g/dl Week 4: FCM 2.57 g/dl versus ID 2.34 g/dl Week 6: FCM 3.01 g/dl versus ID 3.2 g/dl Syncope (n = 1) Naz et al. 2010; Pakistan weeks gestation with suspected IDA (no Hb cut- off specified) IV (n = 130): IS variable Dose: (weight (11 g/dl actual Administration: 200 mg on alternate days Hb difference: Week 3: 8.7 ± 1.2 g/dl to 11.1 ± 1.9 g/dl (P < 0.001) Hb treatment success: delivery: 75.3%; n = 98 No side effects Nicholson and Assali 1952; USA Unknown gestation including anaemic women (Hb < 10 g/dl; n = 6) and non- anaemic controls (Hb > 10 g/dl; n = 4) IV A (n = 6): IS variable to anaemic women IV B (n = 4): IS 1000 mg to non-anaemic controls Dose: ((14.5 g/dl actual Hb) ) mg Administration: 200 mg on alternate days Week 3: Group A: 2.0 g/dl; Group B 0.7 g/dl Reactions not separated according to antenatal or postnatal administration (Continues)

13 34 Intravenous iron in pregnancy: a review TABLE 3 (Continued) Study (country) Participant characteristic Patel et al. 2013; India weeks gestation with IDA (Hb <5 8 g/dl) Pels and Ganzevoort 2015; Netherlands Received FCM infusion during pregnancy and matched non- anaemic controls Perewusnyk et al. 2002; Switzerland IDA (Hb <10 g/l; ferritin <15 μg/l) Raja et al. 2003, Pakistan weeks gestation, with IDA (Hb <8 g/dl) Scott and Govan 1951; UK > 24 weeks gestation with anaemia Shah et al. 2015; India weeks gestation with anaemia (Hb <9 g/ dl) control IV and infusion Efficacy outcome IV side effects IV (n = 61): IS variable Dose: (weight (11 g/dl actual Administration: 200 mg on alternate days Test : yes Week 2: 7.0 g/dl to 9.1 g/dl Week 4: 7.0 g/dl to 10.3 g/dl ferritin Week 4: 9.6 μg/l to 16.4 μg/l Epigastric pain (n = 1); restlessness (n = 1) IV (n = 64) FCM given at variable s up to 20 mg/kg Control (n = 64) no treatment group Dose: max single of 1000 mg given over at least 15 min Delivery: IV g/l No side effects IV (n = 100): IS variable s Dose: calculation not specified; s to reach target Hb of 110 g/l with maximum of 1600 mg Administration: 200 mg every second day Week 4: 92 ± 6 g/l to 109 ± 5 g/l ferritin Week 4:7 ± 4 μg/l to 342 ± 119 μg/l No data on side effects IV (n = 50): IS variable s Dose = (weight (11 g/dl actual Hb) 0.3) + 10 mg/kg Administration: 200 mg every 3 7 days. Test : yes Unknown timing: g/dl ferritin Unknown timing: μg/l Epigastric pain (n = 1; 2%); restlessness (n = 1; 2%); allergic reactions (n = 2; 4%); both required treatment IV (n = 91): IS variable s Dose: 790 mg (additional s optional for women with severe anaemia) Administration: over 2 weeks 30 mg day 1, 60 mg day 2, 100 mg every second day Week 1: moderate anaemia (n = 29): good responders (n = 19) 1.5 g/dl; poor responders (n = 10) 0.2 g/dl Severe anaemia (n = 49): good responders (n = 35) 1.3 g/dl; poor responders (n = 14) 0.3 g/ dl Anaemia resistant to oral iron (n = 13) 1.2 g/dl Not specific for antenatal group IV (n = 100): IS variable s Dose: (weight (target Hb* actual Hb) 2.4) mg *Target Hb unspecified Administration: 100 mg on alternate days Week 6: g/dl Skin rash (n = 4; 4%); headache (n = 8; 8%); vomiting (n = 6; 6%); injection site thrombophlebitis (n = 2; 2%) (Continues)

14 A. Qassim et al. 35 TABLE 3 (Continued) Study (country) Participant characteristic control IV and infusion Efficacy outcome IV side effects Shrivastava et al. 2012; India weeks gestation and Hb 6 8 g/dl [A] IV (n = 256): IS given as variable [B] 797 versus 880 anaemic women from 2007 and 2008 (n = 256 received IV IS in 2008) Dose: weight (10 g/dl actual Hb) 0.24 Administration: not stated Hb difference from baseline: Week 1: 1.1 ± 0.2 g/dl Week 2: 2.3 ± 0.8 g/dl Week 3: 3 ± 0.4 g/dl Blood transfusions: Introduction of IV IS associated with decrease in number of transfusions from 2007 (29.36%) to 2008 (20%) Muscle cramps (n = 39; 15.2%); arthralgia (n = 31; 12.1%); headache (n = 28; 10.9%); breathlessness/chest pain (n = 19; 7.4%); fever (n = 17; 6.64%); hypotension (n = 3; 1.1%) Sogbanmu 1976; Nigeria Known IDA IV (n = 22): iron polymaltose variable s Dose: (0.3 weight (150 g/l actual Hb)) mg Administration: pre- medicated with promethazine Unknown timing: Hb g/ dl increase; Hb < g/dl increase Chest pain (n = 1); required treatment cessation Sortor 1957; USA >12 weeks gestation and Hb <9.5 g/dl IV (n = 26): IS fixed Dose: 500 mg Administration: 100 mg daily for five consecutive days Week 1: average increase of 1 g/ dl Week 2: average increase of 1.82 g/dl (n = 20) Side effects reported in 10 of 130 injections: mild chill (n = 1), flushing and nausea (n = 3), perivascular extravasation (n = 6) Thakor et al. 2015; India weeks gestation and Hb between 5 and 9 g/dl IV (n = 75): IS variable Dose: (weight (10 g/dl actual Administration: 200 mg on alternate days Test : yes Week 6: 7.8 ± 0.61 g/dl to 10.1 ± 0.73 g/dl None reported FCM, ferric carboxymaltose; ID, iron dextran; IDA, iron deficiency anaemia; IS, iron sucrose; IV, intravenous.

15 36 Intravenous iron in pregnancy: a review addition to improved haematological outcomes, receipt of IV iron was associated with improvements in general health (P = 0.021), improved physical energy (P = 0.016), less psychological downheartedness (P = 0.005), less clinical depression (P = 0.003), and overall improved mental health status (P < 0.001). 58 In addition, duration of breastfeeding was greater (hazard ratio (HR) for cessation: 0.70, 95% CI: ; P = 0.046) among women who received IV IPM. Cost- effectiveness While a small number of studies reported costs of treatment, 24,25,36,55,56 none of them included a formal cost- effectiveness analysis. Tolerability While the way in which ADRs were monitored and reported was highly variable, the median prevalence of ADRs for IPM (2.2%; range: 0 4.5%) was lower than FCM (5.0%; range: 0 20%) 13,52 54 and IS (6.7%; range: %). 11,13,16,20,22 25,27,31,32,34,36,38,40,44,49 Across all studies, the median reported prevalence of ADRs was similar among prospective (4.5%) 11,16,19,20,22,23,27,29,31,32,34,36,38,40,44,49,53,55 57 and retrospective (5%) 13,52,54 studies. Eighteen studies administered a test, 15,23,25 31,34,37,39,40,45,46,50,56 while only two studies specifically mentioned administering pre- medication. 38,55 The median reported prevalence of ADRs was similar between studies that did (4%; range: 0 18%) and did not (6.7%; range: 0 20%) administer a test of iron prior to administration of the full. Among the nine studies comparing IV iron to oral iron and reporting ADRs to oral therapy, in all cases ADRs were more common with oral therapy (median 33.5%; range: 27 46%) than IV therapy (median 7.5%; range: %). 14,16,29,32,34,35,39,50,56 Two studies compared IV to IM administration, noting a higher rate of ADRs in the IM group (27% and 24%) compared with the IV group (3% and 8%). 25,38 The prevalence of moderate or severe adverse reactions requiring treatment or discontinuation of IV iron was slightly higher in observational studies (13/1352; 9.6 per 1000 women; n = 20 studies) 12,13,19,20,22,23,32,33,40,41,44 47,49,51 55 compared with RCTs (8/1273; 6.3 per 1000 women; n = 22 studies). 11,14 16,24 31,34 39,50,56,57 When separated according to IV iron type, moderate or severe adverse reactions were more commonly reported with IPM (3/214; 14.0 per 1000 women), compared with IS (17/2157; 7.9 per 1000) 11 16,19,20,22 41,44 47,49 51 or FCM (1/276; 3.6 per 1000). 13,52 54 Risk of bias Included studies were noted to be of variable methodological quality (Appendices S5 and S6). One of the key challenges consistent to RCTs and observational studies relates to blinding of subjects and clinicians to the intervention, but it is unclear what impact lack of blinding to IV iron administration may have on evaluated outcomes. One quarter (5/21) of the included RCTs were considered to be at high risk of bias in at least one domain. Most common sources of possible bias include concerns related to selective reporting (eg perinatal outcomes evaluated but not published) and incomplete outcome data (eg high loss to follow up). Observational studies were assessed as being at greater risk of bias, with 12 of 26 studies identified as being at high risk of bias in at least one domain. These most often related to concerns regarding selection of participants, inadequate adjustments for potential confounders, incomplete outcome data and selective outcome reporting. Overall, the incomplete manner in which many studies were described meant that many aspects were regarded as unclear risk of bias. Further, while a small number of studies were registered in an international clinical trials registry, none of them had a published protocol available, making it difficult to evaluate aspects such as incomplete outcome data and selective reporting which were a concern for a subset of studies. DISCUSSION Our systematic review finds that use of IV iron during pregnancy is associated with significant improvements in haematological parameters, but there is an absence of evidence for improvements in important maternal and perinatal clinical outcomes. IV iron was generally well tolerated, with serious adverse reactions requiring medical treatment or infusion cessation occurring at an observed rate of seven per 1000 women. Based on the findings of this review, no single preparation of IV iron appeared to be superior, but the vast majority of literature (>85% of included studies and exposed women) involves the use of IV IS. While further research is needed to determine the optimal dosing and administration regimens, the current IV iron preparation of choice in pregnancy will be largely determined by cost and convenience around administration. A comparison of each IV iron preparation included in this review is provided in Table 1. Favourable properties of FCM including the ability to administer a high (up to 1000 mg) in a single infusion over a short period of time (15 min) has seen its use become increasingly common across obstetric settings in Australia, despite an absence of RCT evidence identified in this review and lack of data relating to cost- effectiveness. Our findings remain consistent with that of a 2011 Cochrane review on interventions for managing IDA in pregnancy. 59 Notably, 60% of identified studies were published since 2010 and therefore were not included as part of this previous review. Compared to oral iron, IV iron was associated with an additional increase of 4.4 g/l (95% CI: g/l; n = 4) 4 weeks following treatment and 7.5 g/l (95% CI: g/l; n = 1) at delivery. With the additional studies included in our review the median observed increase in haemoglobin was similar at 7.5 g/l at 4 weeks and 6.8 g/l at delivery. This provides evidence that the additional benefit offered by IV iron in comparison to oral iron is modest, as oral iron itself is associated with an increase of 13.4 g/l compared to placebo in clinical trials. 59

16 A. Qassim et al. 37 Looking outside pregnancy, two recent systematic reviews have been published examining outcomes associated with the use of IV iron in adults, regardless of clinical setting. One review identified that compared with oral iron, use of IV iron was not associated with an increased risk of adverse events (relative risk (RR): 1.1; 95% CI: ); however, among 32 studies specifically reporting on anaphylaxis, this occurred in eight of 2186 participants (rate of 3.7 per 1000). This review also found a significant increase in the risk of infection (RR: 1.33; 95% CI: ). 60 A more recent review which included additional studies not included in the previous review, again found no overall increased risk of serious adverse events with the use of IV iron (RR: 1.04; 95% CI: ). 61 However, this finding was strongly influenced by clinical setting, with a significant increased risk of serious adverse events observed in the obstetric and gynaecology setting (RR: 2.0; 95% CI: ). Further, severe infusion reactions were more common with IV iron (RR: 2.47; 95% CI: ). In contrast to the previous review, IV iron use was associated with a much smaller and no longer statistically significant increased risk of infection (RR: 1.17; 95% CI: ). That said, considering infection was not a specified outcome in any of the studies undertaken in the antenatal setting, it remains unclear whether this is of specific concern and highlights an area for future research. Taken together, it is clear that the risk of adverse reactions to IV iron are going to be different from those of oral iron. Oral iron is associated with a high rate of gastrointestinal disturbance which can impact on adherence to treatment and subsequent treatment outcomes. While IV iron was generally well tolerated, significant adverse reactions to IV iron have been published in case reports, including serious anaphylactic reactions, 62 that in one case was fatal. 63 Based on studies included in this review, the prevalence of severe adverse reactions was seven per 1000 women receiving IV iron, highlighting the importance of women receiving IV iron in a setting where rapid treatment can be provided should adverse reactions occur. Despite noted improvements in haematological parameters, the existing literature does not adequately address the question of whether treatment of anaemia with iron leads to improvements in maternal and neonatal clinical outcomes. Therefore, there is an urgent need for large, adequately powered, RCTs addressing the efficacy of IV iron formulations during pregnancy in improving clinical outcomes. Such studies should address key aspects such as the diagnosis of iron deficiency anaemia and identification of women most likely to obtain benefit (ie ferritin and haemoglobin cut- points for treatment), as well as the cost- effectiveness of IV iron. Greater focus should also be placed on optimisation of existing dosing strategies. 64 While identifying key clinical outcomes should be determined with consumer involvement, important clinical outcomes such as improvements in maternal wellbeing and quality of life measures (eg depressive symptoms, maternal postnatal attachment), or breastfeeding outcomes (eg duration of breastfeeding) would likely require randomisation of a few hundred women, while investigation of rarer outcomes such as postpartum haemorrhage or requirement for a blood transfusion would require randomising a thousand or more women. Strengths of our review include the comprehensive literature search strategy and inclusion of a broad range of observational studies in additional to RCTs to explore the breadth of research previously undertaken on IV iron in pregnancy. This is of particular usefulness in exploring different treatment regimens and outcomes to guide future research priorities. Further, while such observational studies may be prone to more bias than RCTs, they provide useful information on the incidence/prevalence of adverse events related to IV iron, and potential usefulness of different IV infusion regimens. Despite the comprehensive nature of our systematic review, several limitations bear consideration. First, significantly variability was evident across all included studies, with differences in study population, trial methods including dosing regimens, timing of assessment and outcome evaluation. Such heterogeneity has been observed in previous systematic reviews of iron supplementation in pregnancy, 65 precluding meta- analytical techniques. Second, a large number of studies (50% overall, and 70% of RCTs) were undertaken in low resources settings, where study participants were more likely to have moderate or severe anaemia, and does not necessarily reflect practice in higher- resource settings where IV iron may be utilised in the management of mild anaemia. In low- resource settings, where moderate to severe anaemia is more common, the benefits of IV iron administration may prove to be much greater compared with that in higher- resource settings where IV iron is commonly utilised in the management of mild anaemia or non- anaemic iron deficiency. 66 A systematic review of routine oral iron supplementation in pregnancy identified that implementation of routine supplementation recommendations were likely to produce heterogenous results depending on the populations background risk for adverse pregnancy outcomes and the same could be expected for IV iron. 67 In conclusion, given the noted negative impacts of moderate and severe anaemia on adverse maternal and neonatal outcomes, high- quality data on the effect of IV iron repletion on ameliorating adverse outcomes are necessary for informing clinical practice. Yet, our systematic review has found that the evidence base from a combination of randomised and observational studies is limited, with a noted focus on evaluation of haematological response to treatment and maternal tolerability of infusions, with very limited data on clinically relevant outcomes. There is no clear evidence that administration of IV iron is associated with improved perinatal outcomes, and only limited evidence that administration is associated with improved quality of life measures. This review highlights a number of substantive gaps in the primary literature, including insufficient evidence on levels of anaemia requiring treatment with IV iron, optimal gestational timing of administration, and IV iron dosing regimens. Currently, considerable resources are utilised in the diagnosis and treatment of iron deficiency anaemia, despite a lack of evidence of clear benefit and potential harms. As such, widespread