Journal of Perinatology

https://doi.org/10.1038/s41372-019-0320-2

REVIEW ARTICLE

Intravenous compared with oral iron for the treatment of

iron-deficiency anemia in pregnancy: a systematic review
and meta-analysis
Adam K. Lewkowitz1 Anjlie Gupta2 Laura Simon
● ●
3 ●
Bethany A. Sabol1 Carrie Stoll2 Emily Cooke4
● ● ●

Roxanne A. Rampersad1 Methodius G. Tuuli5

●

Received: 10 September 2018 / Revised: 14 December 2018 / Accepted: 3 January 2019

© Springer Nature America, Inc. 2019

Abstract
Objective To assess the effect of intravenous versus oral iron on hematologic indices and clinical outcomes for iron-
deficiency anemia (IDA) in pregnancy.
Study design Searches in Ovid Medline, Embase, SCOPUS, Cochrane Database, and ClinicalTrials.gov identified randomized-
1234567890();,:
1234567890();,:

controlled trials comparing intravenous to oral iron for treating IDA in pregnancy. Primary outcomes were maternal hematologic
indices at delivery. Secondary outcomes were blood transfusion, cesarean delivery, neonatal outcomes, and medication reactions.
Results Of 15,637 studies, 20 randomized trials met inclusion criteria and were analyzed. Mean hemoglobin at delivery
(9 studies: WMD 0.66 g/dL (95% confidence Interval 0.31 –1.02 g/dL)) was significantly higher after intravenous iron
therapy. Intravenous iron was associated with higher birthweight (8 studies: WMD 58.25 g (95% CI: 5.57–110.94 g)) but no
significant differences in blood transfusion, cesarean delivery, or neonatal hemoglobin. There were fewer medication
reactions with intravenous iron (21 studies: RR 0.34% (95% CI: 0.20–0.57)).
Conclusion Intravenous iron therapy is associated with higher maternal hemoglobin at delivery with no difference in blood
transfusion and fewer mild medication reactions.

Introduction

Per the World Health Organization, more than 2 billion

people globally suffer from nutritional iron deficiency;
Supplementary information The online version of this article (https:// 38.2% of pregnant women worldwide affected by anemia in
doi.org/10.1038/s41372-019-0320-2) contains supplementary pregnancy is 38.2% [1]. In the United States, anemia affects
material, which is available to authorized users. nearly 20% of pregnancies, and iron-deficiency anemia is the
* Adam K. Lewkowitz
most common anemia of pregnancy [2]. If present at
lewkowitza@wustl.edu delivery, iron-deficiency anemia is associated with increased
risk of blood transfusion, preterm delivery, cesarean deliv-
1
Department of Obstetrics and Gynecology, Washington University ery, and neonatal intensive care admission [3]. As such, iron
in St. Louis, 4566 Scott Avenue, Campus Box 8064, St. Louis,
supplementation for the treatment of iron-deficiency anemia
MO 63110, USA
2
in pregnancy is recommended [2]. However, the preferred
Washington University School of Medicine in St. Louis, 660
route of iron administration is unclear. Specifically, the role
South Euclid Avenue, St. Louis, MO 63110, USA
3
of intravenous iron in treatment algorithms lacks consensus.
Becker Medical Library, Washington University School of
The American College of Obstetricians and Gynecologists
Medicine in St. Louis, 660 South Euclid Avenue, St. Louis, MO
63110, USA (ACOG) and the United Kingdom (UK) guidelines convey
4 that intravenous iron should be used in women who are
Department of Pharmacy, Barnes-Jewish Hospital, One Barnes-
Jewish Hospital Plaza, St. Louis, MO 63110, USA intolerant of [1] or cannot tolerate [4] oral iron supple-
5 mentation, but the gestational age or hemoglobin level at
Department of Obstetrics & Gynecology, Indiana University
School of Medicine, 550N. University Boulevard, UH 2440, which intravenous iron would be preferred is unclear, as is
Indianapolis, IN 46202, USA the anticipated hemoglobin increase per route of iron [2, 4].

The reason for the lack of consensus is twofold. First,
both available treatment routes have shortcomings, with a
low rate of systemic absorption and high rate of gastro-
intestinal side effects for oral iron [5] and more medical
resources and a risk of transfusion reaction for intravenous
iron [6]. Notably, the risk of transfusion reaction is thought
to be lower with the second- and third-generation intrave-
nous iron formulations currently used in transfusions [6].
Second, current data on the risks and benefits of intravenous
iron in pregnancy are limited. The most recent Cochrane
review on treatment of iron-deficiency anemia in pregnancy
—published in 2011—concluded that intravenous iron may
increase hematologic response compared to oral iron but
that more good quality trial assessing clinical outcomes as
well as adverse effects are needed [7]. More recent meta-
analyses either included a small sample size (six trials [8])
or both randomized controlled trials and observational stu-
dies [9]. Both of these previous meta-analyses also focused
on hematologic indices and adverse events with little on
obstetric and pediatric outcomes [8, 9].
Multiple randomized controlled trials have been pub-
lished comparing intravenous iron to oral iron for the
treatment of iron-deficiency anemia in pregnancy since
2011 [10–23]. Moreover, many of these reported clinical
outcomes in addition to hematologic parameters and
adverse medication reactions [10, 13–16, 20, 23, 24]. Thus,
an updated meta-analysis is warranted. The primary aim of
this study is to assess the effect of intravenous iron com-
pared to oral iron on hematologic indices, maternal and
neonatal outcomes, and medication reactions among women
with iron-deficiency anemia in pregnancy.
Materials/subjects and methods
This study was registered on PROSPERO (#CRD42018090276)
and exempt from institutional board review as only deidentified
data available in the public domain through prior publications
were included. All search strategies and written methodology for
this systematic review and meta-analysis were created using the
standards and guidelines for conducting and reporting systematic
reviews set forth by the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) [25], the Institute for
Medicine Standards for Systematic Reviews [26], the Cochrane
Handbook of Systematic Reviews [27], and the Peer Review of
Systematic Search Strategies (PRESS) [28].
Eligibility criteria, information sources, and search
strategy
A medical librarian (L.E.S.) searched published literature
for records discussing iron-deficiency anemia, intravenous
iron, or oral iron therapies in pregnant women. The librarian
A. K. Lewkowitz et al.
created search strategies using a combination of keywords
and controlled vocabulary in Ovid Medline 1946-, Embase
1947-, Scopus 1823-, Cochrane Database of Systematic
Reviews (CDSR), Cochrane Central Register of Controlled
Trials (CENTRAL), and Clinicaltrials.gov 1997. All search
strategies were completed in February 2018 and executed
again in October 2018. Reproducible search strategies for
each database can be found in Appendix 1.
Study selection
Titles and abstracts were screened independently by the co-
first authors (A.K.L. and A.G.), and the full-text articles
were retrieved if they appeared relevant. The same two
authors independently reviewed all full-text articles, and
disagreements were resolved with discussion with a third
author (M.G.T.). Case reports, case series, review articles,
studies without comparison groups, abstracts, and articles
published in languages other than English were excluded,
while prospective randomized controlled trials comparing
any formulation of intravenous iron to any formulation of
oral iron were included. We also searched the biblio-
graphies of included studies to identify additional eligible
studies that were not identified in our literature search.
Studies were limited to having been published after 1980 to
ensure treatment algorithms analyzed were more relevant to
current clinical practice patterns.
Data extraction
Two authors (A.K.L. and A.G.) independently reviewed
eligible articles for data extraction; our data extraction tool
is included as Appendix 2. The primary outcomes were
mean maternal hemoglobin and ferritin at delivery. Sec-
ondary outcomes were mean maternal hemoglobin and
ferritin 2–6 weeks after treatment, rate blood transfusion,
cesarean delivery, neonatal outcomes, and medication
reactions. Study design and study location, inclusion criteria
including hemoglobin and ferritin threshold for iron-
deficiency anemia diagnosis, exclusion criteria, number of
patients, and pre and post-treatment mean hemoglobin and
mean ferritin were abstracted. Extracted maternal outcomes
included the rate of blood transfusion and mode of delivery.
Extracted neonatal outcomes included gestational age at
delivery, birthweight, low-cord pH, total APGARS,
APGARS < 5 at 5 min, neonatal hemoglobin and ferritin,
and neonatal intensive care unit (NICU) admission. All
reported adverse medical reactions were also abstracted. In
studies in which both a composite of adverse medication
reactions and rates of individual adverse medication reac-
tions were reported, we used the rates of individual adverse
reactions for our outcome to allow for individual patients to
have more than one adverse reaction.
Intravenous compared with oral iron for the treatment of iron-deficiency anemia in pregnancy: a. . .
The same two reviewers assessed the quality of each
study based on predefined criteria adapted from the
Cochrane Handbook [27], in a classification system that has
been reported previously [29, 30]. Higher-quality studies
were defined as randomized trials with appropriate rando-
mization method, clear definition of primary outcomes,
intention-to-treat analysis, and less than 10% loss to follow
up; lower-quality studies were those missing one or more of
these attributes [29, 30]. Studies that were missing infor-
mation necessary to judge an attribute were classified as
missing that attribute. Disagreements on study
classification were resolved by discussion with a third
author (M.G.T.).
Data synthesis and assessment of risk of bias
All statistical analyses were performed with the METAN
add-on programs in STATA version 14.2 (StataCorp LP,
College Station, TX). Heterogeneity between studies was
assessed using the Cochran’s Q and Higgins I2 tests [27].
Heterogeneity was considered significant based on con-
servative thresholds of Q > df for the Q tests or I2 > 30%
[27]. Pooled weighted mean differences (WMD) and
pooled relative risks (RR) with 95% confidence intervals
(CI) were calculated using raw data from each study. To
account for clinical heterogeneity between the studies and
produce a more conservative estimate of effect size [31],
we pooled data using the DerSimonian–Laird random-
effects models regardless of whether there was
evidence of statistical heterogeneity. Since we allowed for
the same woman to report multiple adverse medication
events, the proportion of cured women was at 1 in some
studies; had we used traditional statistical methods, these
studies would have been excluded in our pooled analyses,
leading to a biased pooled estimate. To include all studies
in our pooled estimate of adverse medication reactions—
even those with an estimated proportion of adverse
medication events at 1 or greater—we enabled the Free-
man–Tukey double arcsine transformation [32].
To assess whether the effect of intravenous iron was
impacted by developed versus developing country setting,
pretreatment mean hemoglobin or study quality, we con-
ducted prespecified stratified analyses to determine whe-
ther our pooled estimates of the impact of intravenous iron
on obstetric, maternal, or neonatal outcomes would
change using adjusted odds ratios. These analyses were
based on study quality (higher- versus lower-quality),
location of study completion (developed versus develop-
ing country, as categorized by the United Nations [33]), or
pretreatment mean hemoglobin (≤9 g/dL versus >9 g/dL).
Publication bias was assessed visually via a funnel plot,
and the Egger’s test was used to statistically assess sym-
metry [34].
Results
Study selection
The initial literature search in February, 2018 returned
14,798 results including clinicaltrials.gov, and the updated
search in October, 2018 identified an additional 931 results:
the total number of results initially identified was 15,637.
Totally, 5787 duplicate records were removed using the
automatic duplicate finder in Endnote, with an additional
767 duplicate records immediately removed manually after
the initial Endnote de-duplication (Fig. 1). A total of
1263 studies were published prior to 1980 and were
excluded. The titles and abstracts for each of the remaining
8150 citations were reviewed for relevance and screened
against inclusion and exclusion criteria. Most citations
excluded did not include the population of interest
(4688 studies) or treatment of interest (1893 studies). Most
excluded abstracts were from conference proceedings and
had no associated full-text article published. Ultimately, 20
full-text articles were retrieved, and through review of their
bibliographies an additional five relevant articles were
identified. These articles had not been indexed and thus
were not identified by our comprehensive literature search
[12, 13, 23, 35, 36]. Of the 25 studies, five were excluded:
two lacked relevant data for extraction, one was a com-
mentary, one was a follow-up study of another included
study, and one was a randomized controlled trial comparing
two different intravenous formulations. All of the five
unindexed studies met inclusion criteria. Thus, a total of
20 studies were included in our analysis.
Study characteristics
Baseline characteristics of included studies are described in
Table 1. All 20 studies were randomized controlled trials
published as full-text manuscripts. One study was a multi-
national randomized controlled trial conducted in both
developed and developing countries [18] and was thus
excluded from our stratified analysis of location of study
completion. Another was a three-pronged randomized
controlled trial (oral iron, iron polymaltose, and ferric car-
boxymaltose) [14], and, as per methodological strategies
described in prior meta-analyses [30], data were stratified
into oral iron compared to each individual intravenous
formulation. Of the 21 analyzed studies, 14 met the quality
criteria and were determined to be higher quality [10–17,
21, 24, 37, 38] and 7 were lower-quality [18–20, 22, 23, 35,
36, 39, 40]. The reason for a study’s designation as lower-
quality is described in Table 1. Fourteen studies were
conducted in developing countries [10–13, 16, 17, 19–23,
35–37, 39]. and only five studies were conducted in
developed countries [14, 15, 24, 38, 40]. Eleven studies
 A. K. Lewkowitz et al.

Fig. 1 Flow diagram for study Citations identiied through search of

selection Ovid, Embase, Scopus, Cochrane Database
of Systematic Reviews, Cochrane Central
Register of Controlled Trials, and
Clinicaltrials.gov
(n=15,637)
Studies excluded (n=7,487)
Duplicates: 6,224
Unique citations reviewed using titles and Published prior to 1980: 1,263
abstracts
(n=8,150)
Studies excluded (n=8,130)
Duplicates: 190
Not in English language: 528
Animal studies: 98
No population of interest: 4,688
No treatment of interest: 1,893
Outcomes reported but not extractable: 51
Not randomized controlled trial: 638
Conference Abstract without full text: 44
Full texts identiied
(n=20)

Relevant Articles from Review of Bibliographies

(n=5)

Full texts reviewed

(n=25)

Full texts excluded (n=5)

Commentary: 1
Follow-up of an already included study: 1
No comparison of oral to IV iron: 1
No extractable data: 2

Studies included in meta-analysis

(n=20)

reported initial mean hemoglobin ≤9.0 g/dL [10–13, 16, 17,
19, 21, 23, 39, 40] and nine >9.0 g/dL [14, 18, 20, 22, 24,
36–38].
Synthesis of results and risk of bias of included
studies
A total of 1359 women received intravenous iron and 1357
women received oral iron across the 21 studies, with sample
sizes ranging from 23 to 126 women per arm. Inclusion and
exclusion criteria differed significantly across studies.
Whereas the majority required women to have iron-
deficiency anemia, the hemoglobin and ferritin thresholds
varied: the highest hemoglobin included was a range of 8.5–
11.5 g/dL while the lowest was ≤7 g/dL, and the highest fer-
ritin included was <100 µg/L and the lowest was <13 µg/L. In
addition, various formulations of iron were used, with fer-
rous sulfate (n = 13) the most common oral and iron
sucrose (n = 14) the most common intravenous formulation.
The treatment effects assessed varied and included
hemoglobin or ferritin 2–6 weeks after treatment compared
to baseline (n = 14), hemoglobin or ferritin at admission
compared to baseline (n = 9).
Reporting of maternal or neonatal outcomes varied
across studies (Table 1). Of the nine studies that reported
these outcomes, seven included blood transfusion, two
reported cesarean delivery, eight studies reported birth-
weight and neonatal hemoglobin, seven reported neonatal
ferritin, four reported gestational age at delivery, and two
included APGARs. No studies included extractable data on
cord blood pH, neonatal morbidity or NICU admission. All
studies reported rates of mild adverse medication reactions,
and ten reported rates of severe adverse medication
reactions.
Compared to oral iron, intravenous iron was associated
with a significantly higher hemoglobin (9 studies: WMD
0.66 g/dL (95% CI: 0.31–1.02 g/dL)) and ferritin (5 studies:
WMD 45.68 µg/L (95% CI: 26.21–65.16 µg/L)) on admis-
sion to labor and delivery (Fig. 2). A high amount of inter-
study heterogeneity was identified for both hemoglobin and
Table 1 Study Characteristics
Author Country (UN Inclusion criteria Exclusion criteria Iron formulation Quality BL mean Hgb and ferritin mean (SD) units: Hgb (g/ Maternal Neonatal outcomes: mean (SD) units:
(year) category) (sample size) Hgb dL), ferritin (mcg/L) outcomes GA (wks), BW (g), Hgb (g/dL),
ferritin (mcg/L)

Khalafallah Australia Pregnancy, Age 18+, IDA, Hgb Anemia from causes other than IDA, known Oral: Ferrous Higher >9.0 g/dL Oral: BL Hgb: 10.93 (0.48) Delivery Hgb: None Oral: GA at delivery: 39.1 (−) BW:
(2010) (developed) 8.5–11.5 g/dL, Ferritin < 30 mcg/L malabsorptive syndromes, iron overload, sulfate (N = 91) 12.18 (0.87)BL ferritin: 17.7 (17.7) IV: BL 3420 (−) Neonatal Hgb: 16.5 (0.96)
multiple gestation, first Hgb after 28 wks, IV: Iron Hgb: 10.74 (0.54) Delivery Hgb: 12.66 (0.97) Neonatal ferritin: 142 (86) IV: GA at
severe iron deficiency polymaltose BL ferritin: 18.1 (16.3) delivery: 38.9 (−) BW: 3440 (−)
(N = 92) Neonatal Hgb: 15.7 (1.4) Neonatal
ferritin: 185 (101)
Kocchar India Singleton pregnancy, Age 18+, GA Anemia from causes other than IDA, severe Oral: Ferrous Higher ≤9.0 g/dL Oral: BL Hgb: 7.6 (0.8); 2 wk Hgb: 8.9 (0.6); Oral: Blood Oral: GA at delivery: 37 (2) BW:
(2013) (developing) 24–34 wks, Moderate IDA, Hgb 7– anemia (Hgb < 7.0 g/dL), asthma, viral sulfate (N = 50) 3 wk Hgb: 9.6 (0.9); 4 wk Hgb: 10.7 (0.7) transfusion (n): 1 2695 (765) Neonatal Hgb: 15.9 (2.2)
9 g/dL, Ferritin < 15 mcg/L, hepatitis, cirrhosis, cardiovascular disease, IV: Iron sucrose Delivery Hgb: 11.2 (0.9) BL ferritin: 16.5 IV: Blood Neonatal ferritin: 138 (98) IV: GA at
Negative naked eye single tube red autoimmune disease, suspected acute (N = 50) (5.9) 4 wk ferritin: 77.6 (13.7) Delivery transfusion (n): 0 delivery: 38 (1) BW: 2870 (680)
cell osmotic fragility test, no infection, intolerance to iron derivatives, ferritin: 94.6 (14.2) IV: BL Hgb: 7.7 (0.5) 2 Neonatal Hgb: 16.3 (2.1) Neonatal
associated obstetric or medical receipt of earlier parenteral iron treatment, wk Hgb: 9.7 (0.8) 3 wk Hgb: 10.9 (0.8) 4 wk ferritin: 141 (101)
complications those unwilling to participate Hgb: 12.8 (1.1) Delivery Hgb: 13.4 (0.9) BL
ferritin:18.1 (4.6) 4 wk ferritin: 104 (13.4)
Delivery ferritin: 128.8 (15.8)
Abhilashini India GA 30–34 wks, IDA, Hgb 6–8 g/dL, Anemia from causes other than IDA, iron Oral: Ferrous Higher ≤ 9.0 g/dL Oral: BL Hgb: 7.16 (0.6); 2 wk Hgb: 8.22 Oral: Cesarean None
(2014) (developing) peripheral smear suggestive of IDA hypersensitivity, history of blood transfusion sulfate (N = 50) (0.7); 4 wk Hgb: 9.15 (0.7) IV: BL Hgb: 6.89 delivery (n): 4 IV:
in current pregnancy, liver disease, anemia in IV: Iron sucrose (0.86) 2 wk Hgb: 8.15 (0.6) 4 wk Hgb: 9.48 Cesarean delivery
failure (N = 50) (0.7) (n): 3
Abhilashini India GA 30–34 wks, IDA, Hgb 6–8 g/dL, Anemia from causes other than IDA, iron Oral: Ferrous Higher ≤ 9.0 g/dL Oral: BL Hgb: 7.16 (0.6); 2 wk Hgb: 8.22 Oral: Cesarean None
(2014) (developing) peripheral smear suggestive of IDA hypersensitivity, history of blood transfusion sulfate (N = 50) (0.7); 4 wk Hgb: 9.15 (0.7) IV: BL Hgb: 6.89 delivery (n): 4 IV:
in current pregnancy, liver disease, anemia in IV: Iron sucrose (0.86) 2 wk Hgb: 8.15 (0.6) 4 wk Hgb: 9.48 Cesarean delivery
failure (N = 50) (0.7) (n): 3
Gupta India Singleton pregnancy, GA 24–34 Known iron hypersensitivity, thalassemia, Oral: Ferrous Higher ≤ 9.0 g/dL Oral: BL Hgb: 7.88 (0.42); 2 wk Hgb: 8.11 Oral: Blood Oral: GA at delivery: 38.31 (1.47)
(2014) (developing) wks, Hgb 7–9 g/dL, ferritin < 15 clinical evidence of inflammation or sulfate (N = 50) (0.45); 4 wk Hgb 9.18 (0.55) Delivery Hgb: transfusion (n): 0 BW: 2568 (244.19) Neonatal Hgb:
mcg/L autoimmune disease IV: Iron sucrose 10.84 (1.12) BL ferritin: 10.4 (1.89) 4 wk IV: Blood 15.9 (2.2) Neonatal ferritin: 147.68
(N = 50) ferritin: 13.96 (1.88) IV: BL Hgb: 7.81 (0.43) transfusion (n): 0 (39.05) IV: GA at delivery: 38.48
2 wk Hgb: 8.39 (0.43) 4 wk Hgb: 9.8 (0.46) (1.36) BW: 2607 (253.28) Neonatal
Delivery Hgb: 11.5 (0.78) BL ferritin: 10.7 Hgb: 15.8 (0.7) Neonatal
(1.47) 4 wk ferritin: 37.45 (5.73) ferritin:155.77 (46.34)
Darwish Egypt GA 14–28 wks, Age 18+, IDA Anemia from causes other than IDA, iron Oral: Ferrous Higher ≤9.0 g/dL Oral: BL Hgb: 8.58 (0.73) 4 wk Hgb: 9.51 None None
(2017) (developing) (confirmed with clinical and overload or disturbances in utilization of iron fumarate (N = (0.77) BL ferritin: 2.76 (1.03) 4 wk ferritin:
laboratory evidence), Hgb 7–10 g/ (e.g., hemochromatosis, hemosiderosis), 33) IV: Iron 58.79 (18.5) IV: BL Hgb: 8.22 (0.87) 4 wk
dL, consideration of MCV, MCH, decompensated liver cirrhosis, active dextran (N = Hgb: 10.29 (0.86) BL ferritin: 8.22 (0.87) 4
MCHC, serum iron, serum ferritin, hepatitis, active acute or chronic infections, 33) wk ferritin: 78.59 (27.4)
and TIBC rheumatoid arthritis w sx or signs of active
inflammation, hx of multiple allergies, GI
tract disorders or known hypersensitivity to
parenteral iron, recipients of investigational
drug product, Hgb < 7 g/dL, EPO within 8
wks prior to screening visit, other iron tx or
Intravenous compared with oral iron for the treatment of iron-deficiency anemia in pregnancy: a. . .

blood transfusion within 4 wks prior to

screening visit, planned elective surgery
decided during study
Breymann South Korea, GA 16–33 wks, age 18+, Hgb 8– Anemia from causes other than IDA, Oral: Ferrous Lower (17% >9.0 g/dL Oral: BL Hgb: 9.9 (0.12) 3 wk Hgb: 10.68 None Oral: (N = 109) BW: 3400 (500)
(2017) Russia, 10.4 g/dL for GA 16–26 wks, Hgb ≤ significant bleed or surgery within 3 months sulfate (N = loss to follow- (1.38) 6 wk Hgb: 11.22 (1.54) IV: BL Hgb: Neonatal Hgb: 14.6 (1.7) Neonatal
Australia, 11.0 g/dL for GA 27–33 wks, prior to screening, receipt of blood 126) IV: Ferric up) 9.8 (0.08) 3 wk Hgb: 11.03 (0.95) 6 wk Hgb: ferritin: 236.7 (135.9) IV: (N = 112)
Singapore, Ferritin ≤ 20 mcg/L transfusion, erythropoietin treatment, oral carboxymaltose 11.55 (1.18) BW: 3400 (500) Neonatal Hgb:14.8
Sweden, iron (unless dose was ≤100 mg iron per day) (N = 126) (1.9) Neonatal ferritin: 255.6 (187.2)
Switzerland, or parenteral iron tx during month prior to
Turkey screening, or were anticipated to need blood
(mixed) transfusion during study period, multiple
pregnancy, fetal abnormalities evident on US,
any serious medical condition that may
prevent completion of the trial
Al-Momen Saudi Arabia GA < 32 wks, severe IDA defined as Anemia from causes other than IDA, bleeding Oral: Ferrous Lower (Rand- ≤9.0 g/dL Oral: BL Hgb: 7.66 (0.78) BL ferritin: 12 None None
(1996) (developing) Hgb < 9.0 g/dL, MCV < 78 fl, MCH tendency, hemolytic anemia, hypersplenism, sulfate (N = 59) omiza-tion, (5.3) IV: BL Hgb: 7.58 (0.79) BL ferritin:
< 30, serum ferritin < 20 mcg/L low inflammation, liver or renal disease IV: Iron sucrose prima-ry out- 11.9 (5)
iron and elevated TIBC (N = 52) come, and loss
to follow-up not
defined)
Al (2005) Turkey GA 26–34 wks, Hgb 8–10.5 g/dL, Anemia from causes other than IDA, serum Oral: Iron Higher >9.0 g/dL Oral: BL Hgb: 9.8 (0.6) 2 wk Hgb: 10 (0.7) 4 Oral: Blood
(developing) Ferritin < 13 mcg/L folate < 4 pg/mL, vitamin B12 < 1000 pg/mL, polymaltose wk Hgb: 10.4 (0.6) Delivery Hgb: 11.3 (1) transfusion (n): 1
Table 1 (continued)
Author Country (UN Inclusion criteria Exclusion criteria Iron formulation Quality BL mean Hgb and ferritin mean (SD) units: Hgb (g/ Maternal Neonatal outcomes: mean (SD) units:
(year) category) (sample size) Hgb dL), ferritin (mcg/L) outcomes GA (wks), BW (g), Hgb (g/dL),
ferritin (mcg/L)

multiple pregnancy, previous blood (N = 45) IV: BL ferritin: 5 (2.2) 4 wk ferritin: 11 (11) Cesarean delivery Oral: GA at delivery: 39.1 (1.2) BW:
transfusion, history of hematological disease, Iron sucrose (N Delivery ferritin: 18.1 (11) IV: BL Hgb: 9.9 (n): 27 IV: Blood 3439 (451) IV: GA at delivery: 39.2
risk of preterm labor, intolerance to iron = 45) (0.5) 2 wk Hgb: 10.5 (0.6) 4 wk Hgb: 11.1 transfusion (n): 0 (1.5) BW: 3498 (452)
derivatives, recent administration of iron for (0.6) Delivery Hgb: 12 (0.8) BL ferritin: 4.1 Cesarean delivery
treatment of IDA, current usage of iron (2.5) 4 wk ferritin: 28 (26) Delivery ferritin: (n): 20
supplement 23.7 (13.8)
Bayoumeu France GA 24 wks, Age 18+, Hgb 8–10 g/ Anemia from causes other than IDA, asthma, Oral: Ferrous Higher >9.0 g/dL Oral: BL Hgb: 9.7 (0.5) 4 wk Hgb: 11 (1.25) Oral: Blood Oral: GA at delivery: 37.1 (−) BW:
(2002) (developed) dL, MCV < 100 fl, ferritin < 50 mcg/ cirrhosis, viral hepatitis, multiple pregnancy, sulfate (N = 23) Delivery Hgb: 11.5 (1.2) IV: BL Hgb: 9.6 transfusion (n): 1 3220 (570) Neonatal Hgb: 15.3 (2.17)
L risk of premature birth, suspected acute IV: Iron sucrose (0.79) 4 wk Hgb: 11.11 (1.3) Delivery Hgb: IV: Blood Neonatal ferritin: 134 (107) IV: GA at
infection, parenteral iron treatment before (N = 24) 12 (1.1) transfusion (n): 0 delivery: 37.2 (−) BW: 3595 (785)
inclusion, intolerance to iron derivatives, Neonatal Hgb:15.15 (2.1) Neonatal
transport difficulties, difficulties in ferritin: 132 (104)
comprehension of study guidelines, no
consent, participation in clinical trial in
previous month
Abdelazim Kuwait GA 24–30 wks, Age 18+, Hgb 8– Anemia from causes other than IDA, receipt Oral: Proferrin- Lower (Not >9.0 g/dL Oral: BL Hgb: 8.5 (3.5) BL ferritin: 19.4 (4.9) None None
(2017) (developing) 10 g/dL of blood transfusion in current pregnancy ES (N = 124) inten-tion to IV: BL Hgb: 8.7 (2.5) BL ferritin: 15.3 (5.6)
IV: Iron treat)
saccharate (N =
126)
Neeru India GA 14–36 wks, Hgb 6.5–10.9 g/dL, Severe anemia requiring blood transfusion, Oral: Ferrous Lower (11% >9.0 g/dL Oral: BL Hgb: 9.75 (0.415) 4 wk Hgb: 11.06 Oral: Blood None
(2012) (developing) Ferritin < 27 ng/dL bronchial asthma, suspected acute infection fumarate (N = loss to follow- (0.315) Delivery Hgb: 11.65 (0.455) BL transfusion (n): 0
44) IV: Iron up) ferritin: 14.74 (3.775) 4 wk ferritin: 27.33 IV: Blood
sucrose (N = (7.48) IV: BL Hgb: 9.18 (0.47) 4 wk Hgb: transfusion (n): 4
45) 11.24 (0.35) Delivery Hgb: 11.76 (0.4) BL
ferritin: 8.6 (2.585) 4 wk ferritin: 139.93
(61.065)
Deeba India GA 28–37 wks, established IDA, Anemia from causes other than IDA, multiple Oral: Ferrous Higher >9.0 g/dL Oral: BL Hgb: 7.925 (0.862) 2 wk Hgb: 8.5 None None
(2012) (developing) Hgb 6–10 g/dL, ferritin < 15 ng/mL pregnancy, previous blood transfusion, ascorbate (N = (0.862) 4 wk Hgb: 9.32 (0.8707) 6 wk Hgb:
history of hematologic disease, risk of 100) IV: Iron 9.903 (0.8848) BL ferritin: 8.13 (1.45) 2 wk
preterm labor, iron derivative intolerance, sucrose (N = ferritin: 16.65 (4.87) 4 wk ferritin: 23.36
recent administration of iron, current use of 100) (8.57) 6 wk ferritin: 34.78 (8.793) IV: BL
iron supplements Hgb: 7.9 (0.8741) 2 wk Hgb: 9.63 (0.885) 4
wk Hgb: 10.09 (0.8072) 6 wk Hgb: 10.79
(0.8432) BL ferritin: 8.44 (1.35) 2 wk ferritin:
48.46 (16.66) 4 wk ferritin: 61.05 (19.662) 6
wk ferritin: 86.98 (19.939)
Singh (1998) Singapore GA 20–28 wks, Age 16+, Hgb < 9 Anemia from causes other than IDA, acute Oral: Ferrous Lower (Rand- ≤9.0 g/dL Oral: BL Hgb: 8.6 (0.01) BL ferritin: 8.1 (0.5) None None
(developed) g/dL, Ferritin < 20 mcg/L, serum infections, parasitosis, severe liver, kidney, or fumarate (N = omiz-ation not IV: BL Hgb: 8.1 (0.01) BL ferritin: 7.1 (0.4)
iron < 10 mmol/L, MCH < 27 pg, cardiovascular disease, severe psychiatric 50) IV: Iron described)
MCV < 80 fL disorders, any form of parenteral iron therapy dextran (N =
for anemia within 20 days prior to intended 50)
recruitment, history of allergy or abnormal
reaction to iron therapy, history of alcohol
and drug abuse, inability to comply with
therapy
Bhavi India Singleton pregnancy, Age 18–45, Hemoglobinopathy or other red blood cell Oral: Ferrous Lower >9.0 g/dL Oral: BL Hgb: 9.14 (1.1) 4 wk Hgb: 10.65 None None
(2017) (developing) Hgb 7–11 g/dL, ferritin < 15 mcg/L disorders, history of bleeding tendency, fumarate (N = (Randomization (1.03) BL ferritin: 9.1 (3.42) 4 wk ferritin:
history of blood transfusion within prior 56) IV: Iron not described) 30.62 (9.88) IV: BL Hgb: 8.9 (1.07) 4 wk
120 days, allergy, asthma, acute inflammatory sucrose (N = Hgb: 10.64 (1.3) BL ferritin: 8.84 (3.47) 4 wk
state 56) ferritin: 120.85 (87.91)
Aggarwal India GA > 24 wks, Age 18+, Hgb ≤ 7 g/ Placenta previa, placental abruption, Oral: Ferrous Lower ≤9.0 g/dL Oral: BL Hgb: 5.94 (0.62) 2 wk Hgb: 7.64 Oral: Blood None
(2012) (developing) dL, transferrin saturation ≤ 10%, preeclampsia, clotting disorder, patients who sulfate (N = 25) (Randomization (1.21) 3 wk Hgb: 8.6 (0.79) 4 wk Hgb: 10.26 transfusion (n): 0
ferritin ≤ 15 mcg/L participated in another clinical study within IV: Iron sucrose not described) (1.077) BL ferritin: 10 (1.9) 4 wk ferritin: IV: Blood
the previous 3 months, iron intolerance or (N = 25) 160.8 (33.1) IV: BL Hgb: 6.27 (0.48) 2 wk transfusion (n): 0
hypersensitivity, hemolytic anemia, Hgb: 8.62 (0.6) 3 wk Hgb: 9.9 (0.66) 4 wk
hemoglobinopathies (thalassemia, sickle Hgb: 11.3 (0.7) BL ferritin: 9.44 (3.01) 4 wk
cell), bleeding tendency, hypersplenism, ferritin: 295.5 (45.1)
chronic heart failure, class II–IV heart
disease, uncontrolled arterial hypertension
(DBP ≥ 115 mmHg), DVT, thrombocytosis,
chronic renal disease, severe renal failure
(2.5× or higher plasma creatinine level than
high limit of normal state), severe liver
A. K. Lewkowitz et al.
Table 1 (continued)
Author Country (UN Inclusion criteria Exclusion criteria Iron formulation Quality BL mean Hgb and ferritin mean (SD) units: Hgb (g/ Maternal Neonatal outcomes: mean (SD) units:
(year) category) (sample size) Hgb dL), ferritin (mcg/L) outcomes GA (wks), BW (g), Hgb (g/dL),
ferritin (mcg/L)

dysfunction (2.5× or higher AST or ALT than

ULN), cirrhosis, viral hepatitis, CK level
greater than double or more of ULN, asthma,
seizure, hemochromatosis, hemosiderosis
Halimi Pakistan GA 26–30 wks, Hgb < 11 g/dL, Hct Anemia from causes other than IDA, history Oral: Ferrous Lower (Loss-to- >9.0 g/dL Oral: BL Hgb: 9.35 (1.62) 4 wk Hgb: 11.2 None None
(2011) (developing) < 33% of IV iron therapy sulfate (N = 50) follow up and (0.28) IV: BL Hgb: 9.2 (1.69 (4 wk)) Hgb:
IV: Iron sucrose inten-tion to 12.65 (1.06)
(N = 50) treat not
described)
Mehta India GA < 34 wks, Hgb < 8 g/dL, IDA Anemia from causes other than IDA, GA < Oral: Ferrous Higher ≤9.0 g/dL Oral: BL Hgb: 6.72 (0.67) 6 wk Hgb: 10.17 None None
(2014) (developing) evidenced by hypochromic 12 wks, GA > 34 wks, any other medical or sulfate (N = 75) (0.54) IV: BL Hgb: 6.71 (0.65) 6 wk Hgb:
microcytic anemia, low MCV, obstetric complicating factors including IV: Iron sucrose 10.64 (0.71)
MCH, and MCHCC values, hypertension, diabetes, malaria, and infective (N = 75)
increased RDW, low serum iron hepatitis
Rudra India GA 24–34 wks, Hgb 7–9 g/dL, Chronic infection, chronic lung, liver, renal, Oral: Ferrous Higher ≤9.0 g/dL Oral: BL Hgb: 7.88 (0.45) 2 wk Hgb: 8.11 None Oral: GA at delivery: 38.22 (1.59)
(2016) (developing) peripheral smear features suggestive or cardiac disease ascorbate (N = (0.45) 4 wk Hgb: 9.17 (0.47) Delivery Hgb: BW: 2684 (361.6) Neonatal Hgb: 15.8
of IDA 100) IV: Iron 10.9 (0.62) BL ferritin: 10.43 (1.86) 4 wk (0.9) Neonatal ferritin: 155.72 (41.94)
sucrose (N = ferritin: 14.04 (1.86) Delivery ferritin: 46.89 IV: GA at delivery: 38.24 (1.67) BW:
100) (11.86) IV: BL Hgb: 7.81 (0.44) 2 wk Hgb: 2784 (428.4) Neonatal Hgb: 15.6 (0.9)
8.36 (0.43) 4 wk Hgb: 9.8 (0.36) Delivery: Neonatal ferritin: 158.4 (40.2)
11.48 (0.61) BL ferritin: 10.48 (1.46) 4 wk
ferritin: 35.47 (4.37) Delivery ferritin: 70.92
(13.79)
Shim (2018) Korea Age 18+, GA 16–33 wks, ferritin ≤ Anemia from causes other than IDA, Oral: Ferrous High Not Not extractable (primary outcome was change None Oral: BW: 3200 (400) Apgars: 9.0
(Developed) 20 mcg/L, IDA (defined as Hgb 8.0– significant bleed or surgery within 3 months sulfate (N = 44) extractable after treatment). (1.4) Neonatal Hgb: 15.0 (1.3) IV:
10.4 g/dL for GA 16–26 wk or ≤ prior to screening, receipt of blood IV: ferric BW: 3200 (400) Apgars: 9.5 (0.8)
11.0 g/dL for GA 27–33 wk), transfusion, erythropoietin treatment, oral carboxymaltose Neonatal Hgb: 14.9 (1.2)
participant in FER-ASAP study in iron (unless dose was ≤100 mg iron per day) (N = 45)
Korea or parenteral iron tx during month prior to
screening, or were anticipated to need blood
transfusion during study period, multiple
pregnancy, fetal abnormalities evident on US,
any serious medical condition that may
prevent completion of the trial
Khallafallah Australia Second or third trimester Known hypersensitivity to any of trail meds, Oral: Ferrous High >9.0 g/dL Oral: BL Hgb: 11.49 (0.452) Change at 4 wk Oral: 2 women Oral: BW 3518 (602), Apgars 9 (9,10)
PM (2018) (Developed) pregnancy, Hgb ≥ 85 g/L but ≤ 120 known malabsorptive syndromes, anemia sulfate (N = 81) Hgb: 0.461 (0.611) Change at Delivery Hgb: transfused IV: no Neonatal hgb: 16.2 (1.91) Neonatal
g/L, ferritin ≤ 100 mcg/L, from non-ID, severe anemia (Hgb < 85 g/L), IV: iron 0.983 (0.928) BL ferritin: 12.5 (9.6) Change transfusion ferritin 151.3 (88.5) IV: BW 3581
nonanemic patients with Hgb pts first trimester polymaltose (N at 4 wk ferritin: 4.03 (12.0) Change at (469) Apgars 9 (9,10) Neonatal hgb:
110–120 g/L with ID assessed to see = 83) Delivery ferritin: 4.03 (12.0) IV: BL Hgb: 16.26 (1.62) Neonatal ferritin: 214.1
if benefit from therapy indicated 11.4 (0.541) Change at 4 wk Hgb: 0.869 (116.7)
(0.702) Change at Delivery Hgb: 1.46 (1.1)
Intravenous compared with oral iron for the treatment of iron-deficiency anemia in pregnancy: a. . .

BL ferritin: 13.0 (9.08) Change at 4 wk

Khallafallah Australia Second or third trimester pregnancy,
PM (2018) (Developed) Hgb ≥ 85 g/L but ≤ 120 g/L,
ferritin ≤ 100 mcg/L, nonanemic
patients with Hgb pts 110–120 g/L
with ID assessed to see if benefit
from therapy indicated
Known hypersensitivity to any of trail meds,
known malabsorptive syndromes, anemia
from non-ID, severe anemia (Hgb < 85 g/L),
first trimester
ferritin: 148 (130.8) Change at Delivery
ferritin: 90.4 (88.2)
Oral: Ferrous High >9.0 g/dL Oral: BL Hgb: 11.49 (0.452) Change at 4 wk Oral: 2 women Oral: BW 3518 (602), Apgars 9 (9,10)
sulfate (N = 81) Hgb: 0.461 (0.611) Change at Delivery Hgb: transfused IV: no Neonatal hgb: 16.2 (1.91) Neonatal
IV: ferric 0.983 (0.928) BL ferritin: 12.5 (9.6) Change transfusion ferritin 151.3 (88.5) IV: BW 3525
carboxymaltose at 4 wk ferritin: 4.03 (12.0) Change at (444) Apgars 9 (9,10) Neonatal hgb:
(N = 82) Delivery ferritin: 4.03 (12.0) IV: BL Hgb: 15.7 (2.08) Neonatal ferritin: 187.4
11.33 (0.647) Change at 4 wk Hgb: 0.895 (101.6)
(0.822) Change at Delivery Hgb: 1.55 (1.02)
BL ferritin: 12.0 (10.0) Change at 4 wk
ferritin: 170 (106.8) Change at Delivery
ferritin: 99.6 (69.2)

IDA iron deficiency anemia, Hgb hemoglobin, Hct hematocrit, GA gestational age, wks weeks, ferritin serum ferritin, ULN upper limit of normal, BL baseline, BW birthweight, Sx symptoms, Tx
therapy, Hx history, mcg micrograms, EPO erythropoietin

Fig. 2 Weighted mean
difference of maternal
hemoglobin (a) and ferritin
(b) at delivery after oral versus
intravenous iron for the
treatment of iron deficiency
anemia in pregnancy
ferritin (I2 was 92.8% and 97.7%, respectively). In stratified
analyses the treatment effect of intravenous iron on
admission was higher among higher-quality studies, studies
from developing countries, and studies with mean pre-
treatment hemoglobin of ≤9.0 g/dL, with high amount of
heterogeneity identified in each analysis (Table 2). The
funnel plot for the primary outcome of hemoglobin at
delivery did not show asymmetry, and the Egger test con-
firmed there was no evidence of publication bias (p =
0.294) (Fig. 3).
A. K. Lewkowitz et al.
Forest plots for stratified analyses for primary outcomes
and for all secondary outcomes and pooled or stratified
WMD for maternal and neonatal outcomes are included in
Appendix 3 and Appendix 4, respectively. Of note, all
stratified analyses and analyses for secondary outcomes
demonstrated high amount of inter-study heterogeneity
(Table 2). In terms of hematologic indices, intravenous iron
was also associated with increased hemoglobin and ferritin
2–6 weeks after treatment (14 studies: WMD 0.67 g/dL
(95% CI: 0.44–0.90 g/dL) and 11 studies: WMD 58.95 µg/L
Intravenous compared with oral iron for the treatment of iron-deficiency anemia in pregnancy: a. . .

Table 2 Pooled and stratified analyses of the effect of intravenous iron versus oral iron for the treatment of iron deficiency anemia in pregnancy
Outcome No. of Intravenous iron Oral Measure of effect Effect size (95% confidence I2 P
studies (n) iron (n) intervaI)

Hematologic indices
Hemoglobin (Hgb) at 9 571 565 Weighted mean 0.66 (0.31–1.02) 92.8 <0.001
admission for delivery (g/dL) difference (WMD)
Higher quality 8 526 521 WMD 0.74 (0.37–1.11) 91.4 <0.001
Lower quality 1 45 44 WMD 0.11 (−0.07 to 0.29) – –
Developing countries 5 290 289 WMD 0.84 (0.24–1.44) 96.3 <0.001
Developed countries 4 281 276 WMD 0.43 (0.27–0.60) 0 0.956
Mean pretreatment 3 200 200 WMD 1.14 (0.16–2.12) 97 <0.001
hemoglobin ≤ 9.0 g/dL
Mean pretreatment 6 371 365 WMD 0.39 (0.20–0.58) 56.4 0.04
hemoglobin > 9.0 g/dL
Ferritin at admission for 5 360 357 WMD 45.68 (26.21–65.16) 97.7 <0.001
deliverya (µ/L)
Mean Pretreatment 2 150 150 WMD 28.83 (18.88–38.79) 88.1 0.004
hemoglobin ≤ 9.0 g/dL
Mean Pretreatment 3 210 207 WMD 59.95 (−5.29 to 125.19) 98.7 <0.001
hemoglobin > 9.0 g/dL
Hgb at 2–6 weeks after 14 829 827 WMD 0.67 (0.44–0.90) 92.1 <0.001
treatment (g/dL)
Feritin at 2–6 weeks after 11 669 665 WMD 58.95 (50.08–67.82) 98.0 <0.001
treatment (µ/L)
Maternal outcomes
Blood transfusion 8 404 399 Relative Risk 1.02 (0.99–1.04) 7.5 0.37
Cesarean delivery rate 2 95 95 Relative Risk 1.03 (0.94–1.13) 0 0.39
Neonatal outcomes
Gestational age at delivery 4 245 245 WMD 0.29 (−0.12 to 0.71) 57 0.073
(wks)
Neonatal birthweight (g) 8 534 521 WMD 58.25 (5.57–110.94) 0 0.663
Neonatal Hgb (g/dL) 8 582 567 WMD −0.12 (−0.43 to 0.18) 72.2 0.001
Neonatal ferritin (µ/L) 8 582 567 WMD 21.38 (5.50–37.25) 63.8 0.007
Adverse medication reactions
Mild side effects: overall 21 1359 1357 Relative Risk 0.34 (0.20–0.57) 88.0 <0.001
Mild gastrointestinal side 13 960 957 Relative Risk 0.06 (0.02–0.16) 89 <0.001
effects
Mild nongastrointestinal side 21 1359 1357 Relative Risk 2.65 (1.42–4.92) 57.6 0.001
effects
a
All included studies were high-quality in developing countries

(95% CI: 50.08–67.82 µg/L), respectively) (Table 2). Stra-
tified analyses suggested this impact persisted regardless of
study quality, setting of the study, and initial hemoglobin
(Appendix 4).
Few studies reported maternal or neonatal outcomes
(Table 1). There were no significant differences in rates of
blood transfusion at delivery and cesarean delivery among
women receiving intravenous and oral iron (Table 2).
Intravenous iron was associated with a higher neonatal
birthweight overall (8 studies: WMD 58.25 g (95% CI:
5.57–110.94 g)), and, in particular, among higher-quality
studies (n = 7: WMD 69.32 g (11.85–126.79 g))
(Appendix 4). Intravenous iron was also associated with
increased neonatal ferritin (8 studies: WMD 21.38 µg/L
(95% CI: 5.50–37.25 µg/L)). There were no significant
differences in gestational age at delivery or neonatal
hemoglobin (Table 2).
The majority of studies reported severe (10 studies) or
mild (21 studies) medication reactions (Table 2). Within the
studies reporting severe medication reactions, 624 women
received intravenous iron and 622 women received oral
iron, and no severe medication reactions or anaphylaxis
occurred. Table 3 describes the specific medication reac-
tions reported in each study. When pooled, the rate of all

Fig. 3 Funnel plot for hemoglobin at delivery and Egger test for all
studies utilized in our meta-analysis
mild-medication reactions was 13.0% among the intrave-
nous iron group compared to 40.0% among the oral iron
group (18 studies: pooled RR 0.34 (95% CI: 0.20–0.57))
(Tables 2 and 3). When stratified by type of side effects,
intravenous iron was associated with a higher risk of non-
gastrointestinal side effects (9.0% versus 2.0%); pooled RR
2.65 (95% CI: 1.42–4.92), but lower risk of gastrointestinal
side effects (1.0% versus 38.0%); pooled RR 0.06 (95% CI:
0.02–0.16) compared to oral iron (Tables 2 and 3). Overall,
0.02% of those on intravenous iron and 2% of those on oral
iron reported completely discontinuing therapy due to
adverse medication reactions.
Discussion
In this systematic review and meta-analysis, we found intra-
venous iron therapy is associated with a statistically sig-
nificant increase in maternal hemoglobin and ferritin at
delivery and birthweight and neonatal ferritin when compared
to oral iron supplementation to treat iron-deficiency anemia in
pregnancy. However, the magnitude of the hemoglobin
increase was modest (0.66 g/dL), and there were no differ-
ences in the rate of maternal blood transfusion. No severe
allergic reactions were reported in either group, and overall
mild-adverse medication reactions were significantly lower
among women who received intravenous iron. Studies were
heterogeneous, and stratified analyses suggested the effect of
intravenous iron on hemoglobin and ferritin was higher
among studies conducted in developing country settings,
higher-quality studies, and studies with pretreatment hemo-
globin of ≤9.0 g/dL.
Our findings are consistent with those reported from prior
meta-analyses [8, 9] and the most recent Cochrane review
[7] showing improved hematologic indices with the use of
intravenous iron for the treatment of iron-deficiency anemia
A. K. Lewkowitz et al.
in pregnancy. However, we also found a statistically sig-
nificant improvement in mean hemoglobin with intravenous
iron that extended from 2–6 weeks after treatment until
delivery, which differs from the prior meta-analysis which
found no differences in hemoglobin levels at delivery [9].
This difference could be due to our more stringent inclusion
criteria: we limited our analyses to randomized controlled
trials while the prior meta-analysis included observational
trials as well [9].
Unlike the prior meta-analyses that reported only pooled
outcomes [7–9], we conducted stratified analyses based on
clinical factors that could plausibly impact the effectiveness
of iron therapy. Our findings that the higher impact of
intravenous iron on maternal hemoglobin at admission and
2–6 weeks post-treatment within studies that were con-
ducted in developing countries, higher quality, and had
mean pretreatment hemoglobin ≤9.0 g/dL are biologically
plausible. Patients in developing countries are more likely
to have larger iron deficits and therefore more likely to
benefit from intravenous iron. Patients with lower pre-
treatment hemoglobin are also more likely to similarly
benefit from intravenous iron. The higher impact seen in
higher quality may be attributable to a reduction in random
error.
Similar to the Cochrane Review [7], we found that
maternal and neonatal outcomes remain underreported in
the medical literature. While we found no differences in
maternal and most neonatal outcomes, similar to prior meta-
analyses [8, 9], only eight studies reported data on rates of
blood transfusion, neonatal hemoglobin and ferritin, and
birthweight. Even fewer published studies included rates of
cesarean delivery (two studies), gestational age at delivery
(four studies) or APGAR scores (two studies), and no stu-
dies reported cord blood pH, neonatal morbidity, or rates of
admission to the NICU. Though the lack of difference in
maternal or most neonatal outcomes from the two iron
preparations could be due to the limited amount of pub-
lished data on these clinical outcomes, it is also possible that
these outcomes are similar because of the modest hemo-
globin difference between intravenous and oral iron. More
research is needed to clarify whether there is a difference in
maternal or neonatal outcomes using intravenous versus
oral iron therapy for iron-deficiency anemia in pregnancy.
Our findings support those from prior meta-analyses that
intravenous iron is a safe therapy for the treatment of iron-
deficiency anemia in pregnancy with a similar risk overall
of mild-adverse medication reactions [6, 8, 9]. One recent
meta-analysis by Qassim et al. [9] concluded that intrave-
nous iron was associated with a range of moderate or severe
adverse medication reactions ranging from 3.6 per 1000
women to 14.0 per 1000 women, depending on the iron
formulation. However, our meta-analysis included 1359
women who received intravenous iron with no severe
Table 3 Comparison of nonsevere medication adverse events (AE) or reactions due to intravenous versus oral iron
Author (N) Khalafallah Kocchar Abhilashini Gupta Darwish Breymann Al- Al Bayoumeu Abdelazim Neeru Deeba Singh Bhavi Aggarwal Halimi Mehta Rudra Khalafallah Khalafallah Shim Total Total Pooled
(91) (50) (50) (50) (33) (126) Momen (45) (23) (124) (44) (100) (50) (56) (25) (50) (75) (100) PM (81) CM (81) (44) Oral Oral Rate of
(59) AE (N) (N) AEa (%)

Oral iron
Nonsevere 0 12 21 23 57 26 18 17 1 2 11 27 0 14 15 23 32 28 153 153 19 652 1357 40.0%
medication
reaction
GI AE 25 12 16 21 57 23 18 17 1 2 10 27 0 14 12 23 29 24 142 142 19 624 1163 38.0%
composite
Nausea 3 4 0 11 6 0 0 0 2 0 0 18 18 4 66 791
Vomiting 3 1 13 2 0 0 0 8 16 4 12 12 3 74 891
Heartburn 2 6 10 13 8 0 4 3 16 41 41 6 150 796
Constipation 4 1 9 20 3 2 23 13 0 45 45 3 168 889
Diarrhea 2 2 1 4 4 1 5 2 5 4 26 26 82 787
Non GI AE 0 1 5 2 0 3 0 0 0 0 1 0 0 0 3 0 6 4 11 11 0 47 1357 2.0%
composite
Headache 1 1 0 0 0 0 2 398
Chest 0 2 0 0 2 323
tightness
Faintness/ 0 0 0 0 0 1 0 1 512
lethargy
Fever 0 0 0 0 0 0 0 0 0 432
Itching 0 0 0 0 0 0 0 0 0 365
Urticaria 0 0 0 0 0 0 8 8 16 528
Flushing 0 0 0 145
Arthralgia/ 0 0 0 0 0 0 0 0 0 517
myalgia
Tachycardia 0 0 0 0 0 197
Hypertension 0 0 0 25
Hypotension 0 0 0 0 0 0 0 0 497
Metallic taste 5 2 0 0 0 0 0 0 3 0 6 4 3 3 26 901
IV site 0 0 0 0 0 0 0 0 0 488
irritation
Discontinuation 4 0 4 0 0 8 371 2.0%
from AE
Severe 0 0 0 0 0 0 0 0 0 0 0 0 759
Intravenous compared with oral iron for the treatment of iron-deficiency anemia in pregnancy: a. . .

medication
reaction
Intravenous iron
Nonsevere 2 3 2 5 2 13 4 27 11 1 0 8 5 6 5 17 35 12 5 10 6 179 1359 13.0%
medication
reaction
GI AE 0 3 0 0 0 0 0 6 0 0 0 4 0 0 0 17 0 4 0 0 1 35 1170 1.0%
composite
Nausea 0 0 0 0 0 5 4 0 0 17 0 4 1 31 805
Vomiting 0 0 0 0 1 0 0 0 0 0 1 685
Heartburn 1 0 0 0 0 0 0 0 0 0 1 640
Constipation 2 0 0 0 0 0 0 0 0 0 2 685
Diarrhea 0 0 0 0 0 0 0 0 0 0 0 0 626
Non GI AE 2 0 2 5 2 13 4 21 11 1 0 4 5 6 5 0 35 8 5 10 5 144 1359 9.0%
composite
Headache 0 4 0 0 0 0 3 7 446
Chest 0 0 0 0 0 326
tightness
 A. K. Lewkowitz et al.

AEa (%)
Rate of
Pooled

0.02%
allergic reactions reported. This difference may be

0%
Total
Oral
because we included only randomized controlled trials

561

471
326
350
145
685

200

500
735
534

368

966
(N) while Qassim et al. [9] included observational studies.

25
Indeed, Qassim et al. [9] noted that serious medication
AE (N)
Total
Oral

13

11

14
20

28
42
5
4

0
0
0

0
reactions were higher in observation trials compared to
randomized controlled trials.
Shim
(44)

0
Our study offers several strengths. We used a prede-
Khalafallah

signed protocol with a comprehensive search strategy con-

CM (81)

10

0
ducted by an expert research librarian, and two researchers
independently screened all articles for eligibility before
Khalafallah
PM (81)

independently extracting data, which reduced bias. Unlike

5

0
recent meta-analyses [9], all studies included were rando-
mized controlled trials published after 1980, further redu-
Rudra
(100)

2
1

0
0
4

0
cing bias. We also updated our search during the review
process, which identified another three eligible randomized
Mehta
(75)

21
5

0
9

controlled trials, allowing us to base our conclusions on up-

Halimi

to-date analyses. A unique feature of our study is the stra-

(50)

tified analyses to assess the effect of study quality, study

Aggarwal

setting, and pretreatment hemoglobin on the impact of

(25)

0
0
0
1

intravenous iron. Finally, we pooled data from studies using

the more conservative random-effects model to account for
Bhavi
(56)

the clinical heterogeneity between studies.

Singh
(50)

There are also limitations that should be considered.

0
0

First, the findings of this meta-analysis carry forward the

Deeba
(100)

2
1

limitations of the primary studies, which include lower-

quality studies. However, to assess the impact that lower-
Neeru
(44)

quality studies may have on our results, we conducted a

Abdelazim

prespecified stratified analysis by study quality. Second,

(124)

0
1

side effects were not consistently reported in the studies,

which may have impacted our findings. Third, included
Bayoumeu

studies were heterogeneous with regard to type of intrave-

(23)

11
0

0
0
0

nous iron utilized in the primary studies. However, we do

not anticipate that this heterogeneity would impact our
Momen (45)

12
Al

findings, as intravenous iron formulations have been shown

(59)
Al-

to equally efficacious [7, 9]. Lastly, while cost considera-

1
1

Calculated using Freeman–Tukey double arcsine transformation

tions are critical in the decision to use oral or intravenous

Breymann
(126)

iron, we did not conduct a cost-effectiveness analysis.

3

In conclusion, the results of this meta-analysis suggest

intravenous iron therapy in pregnant women with iron-
Darwish
(33)

1
0

0
1

deficiency anemia is associated with a statistically sig-

nificant, but modest, increase in maternal hemoglobin and
Gupta
(50)

1
1

0
0
3

ferritin at delivery, with no differences in maternal blood

Abhilashini

transfusion, as well as a statistically significant but poten-

(50)

0
1

0
0

tially less clinically relevant difference in birthweight and

neonatal ferritin. No severe allergic reactions were reported
Kocchar

in either group, and overall mild adverse medication reac-

(50)

0
0

tions were higher among women receiving oral iron. How-

Khalafallah

ever, because few studies were conducted in developed

Table 3 (continued)

(91)

countries and data remain limited on clinical outcomes,

additional high-quality randomized trials reporting maternal
Metallic taste
Hypertension

Discontinuation
Hypotension
Tachycardia

and pediatric outcomes as well as adverse medication effects

Arthralgia/
Faintness/

Medication
Urticaria
Flushing
Author (N)

irritation
lethargy

myalgia
Itching

IV site

are needed to clarify the role of intravenous iron for the

from AE

Reaction
Fever

Severe

treatment of iron-deficiency anemia in pregnancy.

a
Intravenous compared with oral iron for the treatment of iron-deficiency anemia in pregnancy: a. . .
Acknowledgments Dr. Lewkowitz is supported in part by a National
Institutes of Health training grant T32-HD-55172-9. Dr. Tuuli is
supported by National Institutes of Health U01 (U01HD077384-03)
and R01 (1Ro1HD0867001-01) grants. The contents of this publica-
tion are solely the responsibility of the authors and do not necessarily
represent the official view of the NIH.
Author contributions AKL, AG, BAS, EC, RR, and MGT designed
the work that led to the submission. AKL, AG, LS, CS, and MGT
acquired data and conducted the analyses. AKL and AG drafted the
manuscript. All authors revised the manuscript and approved its final
submission.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of
interest.
Publisher’s note: Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
References
1. WHO. The Global Prevalence of Anaemia in 2011. Geneva:
World Health Organization; 2015.
2. Anemia in Pregnancy. ACOG Practice Bulletin No. 95. Obstet
Gynecol. 2008;112:201–7.
3. Drukker L, Hants Y, Farkash R, Ruchlemer R, Samueloff A,
Grisaru-Granovsky S. Iron deficiency anemia at admission for
labor and delivery is associated with an increased risk for
Cesarean section and adverse maternal and neonatal outcomes.
Transfusion. 2015;55:2799–806.
4. Pavord S, Myers B, Robinson S, Allard S, Strong J, Oppenheimer C.
UK guidelines on the management of iron deficiency in preg-
nancy. Br J Haematol. 2012;156:588–600.
5. Achebe MM, Gafter-Gvili A. How I treat anemia in pregnancy:
Iron, cobalamin, and folate. Blood. 2017;129:940–9.
6. Tomer A, Amir B, Alon G, Hefziba G, Leonard L, Anat GG. The
safety of intravenous iron preparations: Systematic review and
meta-analysis. Mayo Clin Proc. 2015;90:12–23.
7. Reveiz L, Gyte GML, Cuervo LG. Treatments for iron-deficiency
anaemia in pregnancy. Cochrane Database Syst Rev. 2011: Art
No.: CD003094.
8. Shi Q, Leng W, Wazir R, Li J, Yao Q, Mi C, et al. Intravenous
iron sucrose versus oral iron in the treatment of pregnancy with
iron deficiency anaemia: a systematic review. Gynecol Obstet
Invest. 2015;80:170–8.
9. Qassim A, Mol BW, Grivell RM, Grzeskowiak LE. Safety and
efficacy of intravenous iron polymaltose, iron sucrose and ferric
carboxymaltose in pregnancy: a systematic review. Aust New Zeal
J Obstet Gynaecol. 2017;58:1–18.
10. Kochhar PK, Kaundal A, Ghosh P. Intravenous iron sucrose
versus oral iron in treatment of iron deficiency anemia in preg-
nancy: a randomized clinical trial. J Obstet Gynaecol Res.
2013;39:504–10.
11. Abhilashini GD, Sagili H, Reddi R. Intravenous iron sucrose and
oral iron for the treatment of iron deficiency anaemia in preg-
nancy. J Clin Diagn Res. 2014;8:Oc04–7.
12. MN M, Shah J. Iron deficiency anemia in pregnancy: Intravenous
versus oral route. Natl J Community Med. 2014;5:10–12.
13. Rudra S, Chandna A, Nath J. Comparison of intravenous iron
sucrose with oral iron in pregnant women with iron deficiency
anaemia. Int J Reprod Contracept Obstet Gynecol. 2016;5:747–51.
14. Khalafallah AA, Hyppa A, Chuang A, Hanna F, Wilson E, Kwok C,
et al. A prospective randomised controlled trial of a single intra-
venous infusion of ferric carboxymaltose vs. single intravenous
iron polymaltose or daily oral ferrous sulphate in the treatment of
iron deficiency anaemia in pregnancy. Semin Hematol.
2018;55:223–34.
15. Shim JY, Kim MY, Kim YJ, Lee Y, Lee JJ, Jun JK, et al. Efficacy
and safety of ferric carboxymaltose versus ferrous sulfate for iron
deficiency anemia during pregnancy: subgroup analysis of Korean
women. BMC Pregnancy Childbirth. 2018;18:1–8.
16. Gupta A, Manaktala U, Rathore AM. A randomised controlled
trial to compare intravenous iron sucrose and oral iron in treatment
of iron deficiency anemia in pregnancy. Indian J Hematol Blood
Transfus. 2014;30:120–5.
17. Darwish AM, Khalifa EE, Rashad E, Farghally E. Total dose iron
dextran infusion versus oral iron for treating iron deficiency
anemia in pregnant women: a randomized controlled trial. J
Matern Neonatal Med. 2017;7058:1–6.
18. Breymann C, Milman N, Mezzacasa A, Bernard R, Dudenhausen J.
Ferric carboxymaltose vs. oral iron in the treatment of pregnant
women with iron deficiency anemia: an international, open-label,
randomized controlled trial (FER-ASAP). J Perinat Med.
2017;45:443–53.
19. Abdelazim IA, Abu-Faza M, Elbiaa AAM, Othman HS, Alsharif DA,
Elsawah WF. Heme iron polypeptide (Proferrin®—ES) versus
iron saccharate complex (ferrosac) for treatment of iron deficiency
anemia during pregnancy. Acta Med Int. 2017;4:55–60.
20. Neeru S, Nair NS, Rai L. Iron sucrose versus oral iron therapy in
pregnancy anemia. Indian J Community Med. 2012;37:214–8.
21. Shafi D, Purandare SV, Sathe AV. Iron deficiency anemia in
pregnancy: intravenous versus oral route. J Obstet Gynaecol India.
2012;62:317–21.
22. Bhavi SB, Jaju PB. Intravenous iron sucrose v/s oral ferrous
fumarate for treatment of anemia in pregnancy. A randomized
controlled trial. BMC Pregnancy Childbirth. 2017;17:137.
23. Aggarwal R, Mishra V, Panchal N, Patel N, Deshchougule V,
Anil J. Evaluation of iron sucrose and oral iron in management of
iron deficiency anaemia in pregnancy. 2012;3:55–60.
24. Khalafallah A, Dennis A, Bates J, Bates G, Robertson IK, Smith L,
et al. A prospective randomized, controlled trial of intravenous
versus oral iron for moderate iron deficiency anaemia of preg-
nancy. J Intern Med. 2010;268:286–95.
25. Moher D, Liberati ATJ, AD. ThePRISMA Group. Preferred
reporting items for systematic reviews and meta-analyses: the
PRISMA statement. Ann Intern Med. 2009;151:264–9.
26. Institute of Medicine. Finding what works in health care: stan-
dards for systematic reviews. Medicine. 2011;26:2–4.
27. Higgins JPT, Green S (editors). Cochrane Handbook for Sys-
tematic Reviews of Interventions Version 5.1.0 [updated March
2011]. The Cochrane Collaboration, 2011. www.handbook.cochra
ne.org.
28. McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V,
Lefebvre C. PRESS peer review of electronic search strategies:
2015 guideline statement. J Clin Epidemiol. 2016;75:40–46.
29. Yu L, Kronen RJ, Simon LE, Stoll CRT, Colditz GA, Tuuli
MG. Prophylactic negative-pressure wound therapy after cesarean
is associated with reduced risk of surgical site infection: a
systematic review and meta-analysis. Am J Obstet Gynecol.
2017;218:200–210.e1.
30. Tuuli MG, Frey HA, Odibo AO, MacOnes GA, Cahill AG.
Immediate compared with delayed pushing in the second stage of
labor: a systematic review and meta-analysis. Obstet Gynecol.
2012;120:660–8.
31. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-
analysis. Stat Med. 2002;21:1539–58.

32. Nyaga VN, Arbyn M, Aerts M. Metaprop: a Stata command to
perform meta-analysis of binomial data. Arch Public Heal.
2014;72:1–10.
33. UN. Country Classification System. 2014. http://web.worldbank.
org/WBSITE/EXTERNAL/DATASTATISTICS/0,contentMDK:
20421402~pagePK:64133150~piPK:64133175~theSitePK:
239419,00.html#Low_income. Accessed February 5, 2007.
34. Stuck AE, Rubenstein LZ, Wieland D, Vandenbroucke JP, Irwig L,
Macaskill P, et al. Bias in meta-analysis detected by a simple,
graphical. Br Med J. 1998;316:469–469.
35. Bhargava R, Maheshwari M. Evaluation of intravenous iron versus
oral iron in management of iron deficiency anemia in pregnancy
with specific reference to body iron store. J Evol Med Dent Sci.
2013;2:2750–5.
36. Halimi S, Halimi SMA, Shoaib M. Oral versus parenteral iron
therapy for correction of iron deficiency anaemia in pregnancy.
Gomal J Med Sci. 2011;9:3–5.
A. K. Lewkowitz et al.
37. Al R, Unlubilgin E, Kandemir O, Yalvac S, Cakir L, Haberal A.
Intravenous versus oral iron for treatment of anemia in
pregnancy: a randomized trial. Obstet Gynecol. 2005;106:1335–
40.
38. Bayoumeu F, Subiran-Buisset C, Baka N-E, Legagneur H,
Monnier-Barbarino P, Laxenaire MC. Iron therapy in iron defi-
ciency anemia in pregnancy: intravenous route versus oral route.
Am J Obstet Gynecol. 2002;186:518–22.
39. al-Momen AK, Al-Meshari A, Al-Nuaim L, Saddique A, Abotalib
Z, Khashogji T, et al. Intravenous iron sucrose complex in the
treatment of iron deficiency anemia during pregnancy. Eur J
Obstet Gynecol Reprod Biol. 1996;69:121–4.
40. Singh K, Fong YF, Kuperan P. A comparison between intrave-
nous iron polymaltose complex (Ferrum Hausmann) and oral
ferrous fumarate in the treatment of iron deficiency anaemia in
pregnancy. Eur J Haematol. 1998;60:119–24.