aDuring the study, women entered the cohort as they immigrated into the study area or attained the minimum eligible age (15 years) and left the cohort as they emigrated or attained the maximum eligible age (45 years).
bSample sizes <215 are due to missing data, which varied by outcome. In the intention-to-treat (ITT) analysis, missing values were replaced by multiple imputation.
cMaternal samples were collected at delivery so that primary outcome could be established.
Analyses were by intention-to-treat, with multiple imputations to replace missing values. Where pooling of results from multiple imputations led to noninteger counts, we rounded those values. P values indicate the 2-sided probability that group effects are as different as observed or more extreme when assuming that they are identical. Dashed lines indicate overall effect; solid lines, no effect.
aAnalysis restricted to those who were iron-deficient or iron-replete.
Analyses were by intention-to-treat, with multiple imputations to replace missing values. Where pooling of results from multiple imputations led to noninteger counts, we rounded those values. P values indicate the 2-sided probability that group effects are as different as observed or more extreme when assuming that they are identical. Dashed lines indicate overall effect; solid lines, no effect; IPT, intermittent preventive treatment.
Statistical analysis plan
eTable 1. List of variables included in the multiple imputation model
eTable 2. Frequency of serious adverse events in mothers and children
eTable 3. Baseline characteristics of study participants, by intervention group
eTable 4. Effect of iron on maternal and neonatal outcomes, per protocol analysis
eTable 5. Effect of iron supplementation on Plasmodium infection risk, various case definitions, per protocol analysis
eFigure 1. Effect of iron supplementation on selected outcomes, by subgroups (per protocol analysis)
eFigure 2. Effect of iron supplementation on selected outcomes, by IPT use (per protocol analysis)
Mwangi MN, Roth JM, Smit MR, Trijsburg L, Mwangi AM, Demir AY, Wielders JPM, Mens PF, Verweij JJ, Cox SE, Prentice AM, Brouwer ID, Savelkoul HFJ, Andang’o PEA, Verhoef H. Effect of Daily Antenatal Iron Supplementation on Plasmodium Infection in Kenyan WomenA Randomized Clinical Trial. JAMA. 2015;314(10):1009-1020. doi:10.1001/jama.2015.9496
Copyright 2015 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Anemia affects most pregnant African women and is predominantly due to iron deficiency, but antenatal iron supplementation has uncertain health benefits and can increase the malaria burden.
To measure the effect of antenatal iron supplementation on maternal Plasmodium infection risk, maternal iron status, and neonatal outcomes.
Design, Setting, and Participants
Randomized placebo-controlled trial conducted October 2011 through April 2013 in a malaria endemic area among 470 rural Kenyan women aged 15 to 45 years with singleton pregnancies, gestational age of 13 to 23 weeks, and hemoglobin concentration of 9 g/dL or greater. All women received 5.7 mg iron/day through flour fortification during intervention, and usual intermittent preventive treatment against malaria was given.
Supervised daily supplementation with 60 mg of elemental iron (as ferrous fumarate, n = 237 women) or placebo (n = 233) from randomization until 1 month postpartum.
Main Outcomes and Measures
Primary outcome was maternal Plasmodium infection at birth. Predefined secondary outcomes were birth weight and gestational age at delivery, intrauterine growth, and maternal and infant iron status at 1 month after birth.
Among the 470 participating women, 40 women (22 iron, 18 placebo) were lost to follow-up or excluded at birth; 12 mothers were lost to follow-up postpartum (5 iron, 7 placebo). At baseline, 190 of 318 women (59.7%) were iron-deficient. In intention-to-treat analysis, comparison of women who received iron vs placebo, respectively, yielded the following results at birth: Plasmodium infection risk: 50.9% vs 52.1% (crude difference, −1.2%, 95% CI, −11.8% to 9.5%; P = .83); birth weight: 3202 g vs 3053 g (crude difference, 150 g, 95% CI, 56 to 244; P = .002); birth-weight-for-gestational-age z score: 0.52 vs 0.31 (crude difference, 0.21, 95% CI, −0.11 to 0.52; P = .20); and at 1 month after birth: maternal hemoglobin concentration: 12.89 g/dL vs 11.99 g/dL (crude difference, 0.90 g/dL, 95% CI, 0.61 to 1.19; P < .001); geometric mean maternal plasma ferritin concentration: 32.1 µg/L vs 14.4 µg/L (crude difference, 123.4%, 95% CI, 85.5% to 169.1%; P < .001); geometric mean neonatal plasma ferritin concentration: 163.0 µg/L vs 138.7 µg/L (crude difference, 17.5%, 95% CI, 2.4% to 34.8%; P = .02). Serious adverse events were reported for 9 and 12 women who received iron and placebo, respectively. There was no evidence that intervention effects on Plasmodium infection risk were modified by intermittent preventive treatment use.
Conclusions and Relevance
Among rural Kenyan women with singleton pregnancies, administration of daily iron supplementation, compared with administration of placebo, resulted in no significant differences in overall maternal Plasmodium infection risk. Iron supplementation led to increased birth weight.
clinicaltrials.gov Identifier: NCT01308112
Anemia in pregnancy is a moderate or severe health problem in more than 80% of countries worldwide, but particularly in Africa, where it affects 57% of pregnant women.1 Iron deficiency is the most common cause and may be even more widespread, but global or regional estimates are lacking.
Although universal antenatal iron supplementation has been recommended since 1959,2 delivery and adherence have been notably poor in low- and middle-income countries.3,4 It is well established that antenatal iron supplementation leads to increased maternal hemoglobin concentrations and a reduced risk of anemia at term,5 but the functional significance of this hematologic response is uncertain.6 Iron deficiency in pregnancy has been associated with severe anemia and maternal death, but causal evidence from randomized trials is inconclusive.5 A meta-analysis of randomized trials of iron supplementation during pregnancy found no evidence of benefits for maternal and neonatal health outcomes.5
Added to these uncertainties, a trial among children7 reinforced earlier concerns that iron supplementation can increase rates of infectious diseases, including malaria.8 Antenatal supplementation nonetheless continues to be recommended,9 despite reports that iron deficiency is associated with protection against Plasmodium infection in placental blood.10 In a previous randomized trial, antenatal iron supplementation did not increase susceptibility to Plasmodium infection,11 but its design had limitations in allocation concealment, blinding, and data completeness.5 In highly endemic areas, P falciparum infections in pregnancy are usually asymptomatic, but they increase the risk of adverse birth outcomes (reduced birth weight, intrauterine growth retardation, preterm delivery, increased neonatal mortality) and adverse maternal outcomes (severe anemia and death).12
We aimed to measure the effect of daily iron supplementation during pregnancy on maternal Plasmodium infection risk. We conducted preplanned analysis of effect modification by baseline iron status, gravidity, age, and HIV infection. We also assessed effects of iron supplementation on gestational age at delivery, newborn size, and maternal and neonatal iron status at 1 month postpartum.
The study was a double-blind, randomized trial comparing daily supplementation with iron vs placebo, with 2 parallel groups of pregnant women receiving daily supplementation with and without iron. (Additional details are provided in the trial protocol in Supplement 1, the statistical analysis plan in Supplement 2, and the eMethods in Supplement 3.) The study received ethical clearance in Kenya (Kenyatta National Hospital/University of Nairobi) and England (London School of Hygiene and Tropical Medicine). All women in the trial provided written informed consent.
The study was conducted (October 2011–April 2013) in Nyanza Province, Kenya, where malaria is highly endemic and transmission is perennial. As per national and international guidelines,13- 15 pregnant women should receive daily iron supplementation and intermittent preventive treatment (IPT) for malaria with sulfadoxine-pyrimethamine. In 2008 to 2009, 31% of women in Nyanza Province reported not having taken any iron supplements during their last pregnancy, while 54% took them for less than 60 days.16
Local women aged 15 to 45 years who were married or cohabitating were included in a community surveillance program and invited for pregnancy screening when having missed their menstrual period for 10 weeks. Those who were not in stable relationships were invited for pregnancy testing every 12 weeks. At screening, we collected stool, and we administered a urine pregnancy test, a medical examination including obstetric ultrasonography, and preventive antihelminth chemotherapy with praziquantel and albendazole (Figure 1).
At 14 to 21 days after the initial visit but before randomization, we collected venous blood and measured hemoglobin concentration (HemoCue301) and zinc protoporphyrin (ZPP; Aviv206d). Erythrocytes and plasma were stored in liquid nitrogen and dry ice until analysis.
Recruitment continued until the target number of women was reached. We calculated that we would need 225 pregnant women per intervention group (450 women total) to give 92% probability of excluding no effect, assuming a Plasmodium infection risk of 50% in the placebo group, a relative risk increase due to iron of 35%, 5% drop-out of the iron group, and no “drop-in” of women crossing over from the placebo group to the iron group. To enroll this number of women, we anticipated an 18-month recruitment period.
Women were included when they were aged 15 to 45 years, written informed consent had been obtained, they were likely to be available for study until 1 month postpartum, and they were planning to deliver in the predesignated health facility. Women were excluded if they had obvious intellectual disabilities or a metabolic disorder (eg, diabetes); they had a medical history of sickle cell anemia or epilepsy or an obstetric history suggestive of eclampsia or preeclampsia; they were carrying multiples; gestational age at the second visit was less than 13 weeks or more than 23 weeks; no venous blood was collected; or hemoglobin concentration was less than 9 g/dL (which has previously been associated with adverse events such as prematurity, low birth weight, and fetal death).17,18
Supplements were given as capsules that were identical in appearance except for shell color. The code linking each color to supplement type was kept in sealed envelopes. One of us (H.V.) not involved in the field work used tables with random numbers to produce sequentially numbered, sealed, opaque envelopes containing the code, using simple randomization with a 1:1 allocation ratio. Eligible women were allocated in order of enrollment to the color indicated in the next available envelope. Participants and field staff were blinded to intervention until data analysis.
Supplements contained iron (60 mg of elemental iron as ferrous fumarate, which has similar iron bioavailability as ferrous sulfate) or placebo; they contained no other micronutrients. Research assistants administered supplements and daily observed that women swallowed their supplements.
From screening until end of intervention, local mill operators added fortificant iron (target dose: 20 mg/kg flour) to grain routinely brought for milling by homestead members of participating women. Based on weighed intake studies, we estimate that fortification supplied on average 5.7 mg of elemental iron as NaFeEDTA (ferric sodium ethylenediaminetetraacetate) daily in pregnant women.
Women were referred to regular health services for routine antenatal and medical care during intervention, including IPT for malaria and (as required) antiretroviral therapy, but they were instructed not to take micronutrient supplements other than those supplied by our field staff. They were asked to ensure that all diagnoses, treatments, and drugs administered were recorded in antenatal health booklets and to alert the field team as soon as they went into labor. An obstetric nurse assisted in delivery; complicated cases were brought to a tertiary facility. Birth weight was measured (within 10 g) immediately after delivery or, for those born at home, on presentation in the research clinic.
We collected maternal venous blood, maternal placental blood,19 cord blood, and placental biopsies within 1 hour postpartum. For home deliveries, samples were collected within 2 hours postpartum. All women received therapy with artemether-lumefantrine, praziquantel, and albendazole immediately postpartum.
Supplementation with iron or placebo continued for 1 month postpartum. We then collected maternal venous blood and neonatal capillary blood. We also extracted information from antenatal health booklets about IPT use, possession of mosquito nets, and micronutrient supplements supplied during antenatal visits.
We used dipstick tests (Access Bio) to detect histidine-rich protein-2 (HRP2) and lactate dehydrogenase (pLDH) specific to either P falciparum or to nonfalciparum human Plasmodium species. Whereas HRP2-based tests detect current or recent P falciparum infection, pLDH-based tests only indicate current infection.20 Placental biopsies were examined histologically21 to detect Plasmodium infection. We used real-time polymerase chain reaction (PCR) to detect P falciparum–specific DNA in erythrocytes and DNA in stool of intestinal helminths associated with blood loss (Schistosoma spp, Trichuris trichiura, hookworm).
We assessed plasma iron markers (concentrations of ferritin, soluble transferrin receptor, transferrin) and plasma inflammation markers (concentrations of C-reactive protein [CRP] and α1-acid glycoprotein) on a Beckman Coulter UniCel DxC 880i analyzer and HIV infection by point-of-care antibody tests.
The primary outcome was defined as past or present maternal Plasmodium infection assessed at parturition, regardless of species, as indicated by 1 or more positive results for the presence of pLDH or HRP2 in plasma or by placental histopathology or P falciparum DNA in maternal erythrocytes from venous or placental blood by PCR test. Secondary (exploratory) outcomes were patent Plasmodium infection (similarly defined as primary outcome but restricted to ≥1 positive result from dipstick tests or placental histopathology); presence of P falciparum DNA (as primary outcome, but restricted to ≥1 positive result for PCR tests); current or recent Plasmodium infection (as primary outcome, but restricted to ≥1 positive result for dipstick tests or PCR tests); birth weight; gestational age at delivery; intrauterine growth as indicated by birth-weight-for-gestational-age z score; and maternal and neonatal iron status at 1 month postpartum.
We used SPSS version 21 (IBM) and a predefined plan (see the statistical analysis plan in Supplement 2). We estimated effects when possible; P values, where reported, are 2-sided. Birth-weight-for-gestational-age z scores were derived with Kenyan children as a reference22 and a coefficient of variation of 12.8% (estimated from the present study).
Anemia was defined as a hemoglobin concentration less than 11 g/dL. A participant had an iron deficiency if ferritin concentration was less than 15 μg/L and was “iron-replete” if ferritin concentration was 15 μg/L or greater without inflammation. Iron status was considered uncertain if ferritin concentration was 15 μg/L or greater with inflammation. Inflammation was defined as concentrations of CRP of greater than 10 mg/L or α1-acid glycoprotein greater than 1 g/L.
The preplanned primary analysis was per protocol. It included all individuals with singleton pregnancies who complied with intervention (ie, consumed >90% of scheduled supplements) and for whom outcomes were available. We also conducted intention-to-treat analysis with multiple imputations to replace missing values. Results reported are by intention-to-treat analysis unless indicated otherwise. In the analysis of birth weight, infants born at referral facilities and those born at home whose weight was measured more than 24 hours postpartum were included in intention-to-treat analysis but excluded from per-protocol analysis.
For binary outcomes, effects are reported as the absolute difference in proportions except for the risk of low birth weight, which we also report as a relative effect to facilitate extrapolation of results to different settings. Continuous outcomes were log-transformed as needed to normalize distributions; effects and corresponding confidence interval limits were exponentiated and expressed as proportional differences in geometric means. We used multiple linear regression and multiple logistic regression to compare effect estimates with and without adjustment for baseline factors (gravidity, maternal age, HIV infection, Plasmodium infection status, hemoglobin concentration, iron deficiency, and gestational age). For complete accounting, we report the adjusted effect for the primary outcome, with the odds ratio re-expressed as a risk difference; for secondary outcomes (birth weight, gestational age at delivery, and maternal and infant hemoglobin concentration at 30 days after birth), adjusted results are not reported because adjustment did not markedly affect the magnitude of the intervention effects.
We anticipated that iron absorption and thus effects of administered iron would depend on baseline iron status and that baseline proxy markers of immunity against malaria (gravidity, age, HIV infection) might determine the ability to suppress a possible increase in parasitemia resulting from administered iron. We used stratified and direct multivariate analyses to assess effect modification by these baseline factors, with Plasmodium infection, birth weight, and maternal hemoglobin concentration at 1 month after birth as outcomes. We also assessed to what extent the use of IPT influenced the magnitude of intervention effects.
Of 2015 women invited for screening, 470 (23%) were randomized (237 to the iron group and 233 to the placebo group) and were included in the intention-to-treat analysis (Figure 1). Per-protocol analysis included 430 or fewer cases, depending on the volume of missing data for each outcome. For the primary outcome, 363 women (77%) were available for analysis with missing data from 55 in the iron group and 52 in the placebo group. Adherence in the iron and placebo groups was 100% and 99.1%, respectively. The groups were similar regarding the percentage who received iron supplements supplied by external sources (9.3% vs 9.9%) and who possessed insecticide-treated bed nets (15.2% vs 15.9%). Only on 1 occasion (at baseline) did we find a participant to be infected by a Plasmodium species other than P falciparum.
We found no evidence that iron supplementation caused serious adverse events (Figure 1 and eTable 2 in Supplement 3). Serious adverse events were reported for 9 and 12 women who received iron and placebo, respectively. There were 2 maternal deaths, which both occurred in the placebo group; one woman was reported to have died due to postpartum hemorrhage and another at 2 weeks postpartum due to pneumonia and cardiac arrest. There were 7 fetal or neonatal deaths in each group; in addition, 1 child died in each group after 28 days of life.
Four hundred fifteen women (88.3%) delivered newborns at the research clinic, 38 (8.1%) in referral hospitals, and 17 (3.6%) at home. Placenta was missing or poorly preserved for 42 women (8.9%) due to delivery at home or in referral hospitals; in addition, placental biopsies were unavailable for 85 women first enrolled (18.1%) due to incorrect preservation of samples. Two mothers refused consent for neonatal blood collection. Plasma sample volumes from 2 infants in the control group were insufficient for all biochemical tests.
Intervention groups were similar except for mild imbalances in marital status (married or living together: 86.1% vs 79.0% in groups receiving iron or placebo) and gravidity (secundigravida: 24.1% vs 15.0% in groups receiving iron or placebo) (Table 1). Of women without inflammation, 59.7% (190/318) were iron-deficient. Infection by Schistosoma spp and T trichiura was relatively common, whereas Necator americanus was rarely found. Ancylostoma DNA was absent in all stool samples.
Forty women (22 in the iron group and 18 in the placebo group) were lost to follow-up or excluded at birth. There was no evident effect of iron supplementation on Plasmodium infection risk (50.9% vs 52.1% for iron vs placebo groups, respectively; difference, −1.2%, 95% CI, −11.8% to 9.5%; P = .83) (Table 2). This effect was not influenced by baseline imbalances (crude and adjusted risk difference, −1.2% vs −0.5%). Subgroup analysis showed no evidence that the effect of iron on Plasmodium infection risk depended on any of the baseline factors investigated (Figure 2) or on IPT use (Figure 3).
Iron supplementation led to improved maternal iron status as indicated by hemoglobin concentrations, prevalence of anemia, and ZPP–heme ratios (Table 2).
Iron supplementation resulted in increased birth weight by 150 g (95% CI, 56 to 244) (Table 2) and reduced the risk of low birth weight by 58% (95% CI, 22% to 87%). The absolute risk reduction was 6.0% (95% CI, 0.8% to 11.1%) (Table 2). Thus, on average, 16.8 women (95% CI, 9.0 to 61.3) needed to receive supplementation to prevent 1 case of low birth weight. The effect of iron on birth weight was larger in women with iron deficiency at baseline than in those who were initially iron-replete (234 g vs 39 g; difference, 195 g; 95% CI, −3 to 393; P = .05) (Figure 2). There was no evidence that the effect of iron on birth weight was influenced by IPT use (Figure 3).
Iron supplementation led to an increased gestational age at delivery by 3.4 days and neonatal length by 0.9 cm; it reduced the risk of premature birth by 7% (Table 2). We found no effect of iron on neonate iron markers (hemoglobin concentration, ZPP-heme ratios).
Twelve mothers were lost to follow-up postpartum (5 in the iron group and 7 in the placebo group). Iron supplementation improved maternal iron status. Hemoglobin concentration and geometric mean plasma ferritin concentration increased by 0.90 g/dL and 123%, respectively; geometric mean plasma transferrin receptor concentration was reduced by 29% (Table 3). Gains in iron status were greater in women with poor iron status at baseline. Iron supplementation increased hemoglobin concentration by 1.52 g/dL and 0.44 g/dL in women with and without initial anemia, respectively. Corresponding values were 1.29 g/dL and 0.60 g/dL for women who were initially iron-deficient and -replete, respectively (Figure 2).
Twenty-two infants were lost to follow-up (13 in the iron group and 9 in the placebo group). Iron supplementation increased geometric mean plasma ferritin concentration by 18%. There was no evidence that it affected hemoglobin and plasma transferrin receptor concentrations.
Results of the per-protocol analysis were similar to those obtained by intention-to-treat analysis (eTables 3 and 4, eFigures 1 and 2 in Supplement 3). There was no evident effect of iron on Plasmodium infection, however defined (eTables 4 and 5 in Supplement 3). Iron supplementation led to a decrease in the risk of low birth weight by 64% (95% CI, 12%-86%). Among women who were initially iron deficient, iron supplementation increased birth weight by 263 g (95% CI, 136-362). There was some evidence that IPT use modified the effect of iron on maternal hemoglobin concentration (P = .04 for interaction), but there was no monotonic trend in effect by IPT dose.
Overall, we found no effect of daily iron supplementation during pregnancy on risk of maternal Plasmodium infection. Iron supplementation resulted in an increased birth weight (by 150 g), gestational duration, and neonatal length; enhanced maternal and infant iron stores at 1 month after birth; and a decreased risk of low birth weight (by 58%) and prematurity. The effect on birth weight was influenced by initial maternal iron status. Correction of maternal iron deficiency led to an increase in birth weight by 239 g.
Supplementation with iron-containing micronutrients can increase malaria rates in young children with iron deficiency,23 perhaps through transiently increased counts of young erythrocytes that are more susceptible to invasion and propagation by P falciparum merozoites than mature erythrocytes.24 Although we found no evidence in the present study that iron increased Plasmodium infection risk in women who were iron-deficient or anemic at baseline, we could not exclude such an effect (see upper limit of 95% CIs in Figure 2). Should this raise concerns? P falciparum infections in pregnancy notably cause increased risks of low birth weight and severe anemia.12 In our study, however, women who were initially iron-deficient benefited most in terms of birth weight, indicating that any potential decrease in birth weight due to the effect of iron supplements on Plasmodium infection was outweighed by beneficial effects of iron on birth weight through other mechanisms.
Depending on local conditions, approximately one-quarter of severe anemia cases among pregnant women are attributable to malaria,12 suggesting that most cases are due to iron deficiency. Our data show that women who were initially iron-deficient or anemic benefited particularly in terms of hemoglobin concentrations (Figure 2), thus decreasing their risk of severe anemia. Women with increased hemoglobin concentrations due to iron supplementation may also be less prone to develop severe anemia during or following episodes of malarial illness.
Per-protocol analysis suggested that IPT use may have modified the effect of iron on hemoglobin concentration (eFigure 2 in Supplement 3), but the absence of a clear trend with dose indicates that this finding may be spurious. Interpretation of these data is further complicated because IPT use was not a baseline factor and may have acted as a mediating factor.
One limitation of our study concerns the missing data. Multiple imputation yields valid effect estimates when data are missing at random,25 which seems a reasonable assumption given the reasons for missingness of most primary outcome data (exclusion of twin pregnancies, women leaving the study area, absent placental samples because of delivery at home or in tertiary facilities, or incorrect preservation of samples).
Another limitation of our study concerns the lack of detailed, reliable morbidity data during follow-up. To avoid ethical dilemmas, we did not collect blood samples during intervention, and we referred women for antenatal care after enrollment to regular health services, where they received standard service. Plasmodium infection was determined in samples collected at delivery, which may not have captured all infections or symptomatic episodes during intervention. Plasmodium pigment persists in the placenta, however, and placental histopathology probably captures most infections in the second half of pregnancy, even though it may have low sensitivity in capturing earlier infections.21,26
The use of placebo in a population that included women with anemia and iron deficiency was justified by our concern that risks inherent to iron supplementation outweighed possible benefits; however, it also provided a unique opportunity to assess benefits of antenatal iron supplementation. Recent meta-analyses, published after we started our trial, found no27 or only small effects (31 g28 and 41 g29) of antenatal iron supplementation on birth weight. In these meta-analyses, however, subgroup analysis by initial iron status could not be investigated, because women with anemia or iron deficiency were excluded from the studies reviewed, women (also in the placebo group) received iron as rescue therapy during intervention, or initial anemia status and iron status had not been specified.27 Consistent with the findings reported in these meta-analyses,27- 29 our data showed no evident effect of iron on birth weight among iron-replete women.
The beneficial effect of maternal iron supplementation on birth weight found in our study may also be explained in part by a superior adherence compared with previous studies.27- 29 Whereas we observed daily whether women swallowed supplemental capsules, supervision was unclear in other studies, or contact with participants was limited to repeated visits or telephone calls. Although several studies used capsules or coated tablets,30,31 iron was mostly given as tablets with ferrous salts, which produce a strong, unpleasant taste that may have discouraged adherence. Attrition was high in many studies, although it is unclear whether this occurred selectively in the iron groups. Adherence in most studies was not reported or poor, or supplement use was assessed by tablet counts or self-reports, which tend to overestimate adherence.32,33
The gain in birth weight was due at least in part to a longer gestational duration, with perhaps some contribution by improved fetal growth. In magnitude, this effect (150 g) is comparable or exceeds effects of interventions to prevent malaria in pregnancy, namely, IPT (79-135 g, depending on frequency of administration) and insecticide-treated nets (55 g),34 and is particularly important because low birth weight is associated with neonatal and postneonatal mortality, morbidity, inhibited growth, cognitive development, and potentially chronic diseases later in life.35 This comparison should not detract from the benefits of these malaria control interventions, but is rather made to indicate the importance that should be given to increased coverage of iron supplementation.
In accordance with our finding on birth weight, the hemoglobin response to iron was larger in women who were initially iron-deficient as compared with those who were iron-replete. A differential benefit of iron by anemia status was apparent for hemoglobin concentration but not birth weight (Figure 2), consistent with initial anemia being due only partly to iron deficiency and partly to other causes.
Our finding of increased plasma ferritin concentrations at 1 month postpartum adds to growing evidence,36 so far confirmed only in a single trial,37 that antenatal iron supplementation improves neonatal iron stores, thus delaying the age at which iron deficiency is likely to develop during infancy.
The baseline characteristics of our study population were typical for pregnant women in many rural settings in low- and middle-income countries. There was no evidence that gains in birth weight depended on gravidity, maternal age, HIV infection, anemia, and IPT use, suggesting benefits from iron for all subgroups thus defined, including primigravidae and those who did not receive IPT. Thus, our results may apply to pregnant women in other low- and middle-income countries, although the effect on birth weight can vary depending on the prevalence of iron deficiency. Our results cannot be extrapolated to daily antenatal supplementation with 120 mg, for which safety data are lacking, or to children, for whom there is substantial evidence that iron supplementation can increase malaria rates.38
In low- and middle-income countries, it is generally impractical to screen for iron status, and most countries have policies for universal iron supplementation for pregnant women. Based on our results, we believe that the benefits of universal supplementation outweigh possible risks.
Among rural Kenyan women with singleton pregnancies, administration of daily iron supplementation, compared with administration of placebo, resulted in no significant differences in overall maternal Plasmodium infection risk. Iron supplementation led to increased birth weight.
Corresponding Author: Martin N. Mwangi, PhD, Cell Biology and Immunology Group, Wageningen University, PO Box 338, 6700 AH Wageningen, the Netherlands (firstname.lastname@example.org).
Author Contributions: Drs M. N. Mwangi and Verhoef had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: M. Mwangi, Cox, Verhoef.
Acquisition, analysis, or interpretation of data: M. Mwangi, Roth, Smit, Trijsburg, Demir, Wielders, Mens, Verweij, Prentice, Savelkoul, Andang’o, Verhoef.
Drafting of the manuscript: M. Mwangi, Verhoef.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: M. Mwangi, Verhoef.
Obtained funding: Brouwer, Verhoef.
Administrative, technical, or material support: M. Mwangi, Roth, Smit, Trijsburg, A. Mwangi, Demir, Wielders, Mens, Verweij, Brouwer, Savelkoul, Andang’o, Verhoef.
Study supervision: Andang’o, Verhoef.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Verhoef reported having received grants from the European Union, having received nonfinancial support from Luchtvaart Zonder Grenzen, and having intellectual property rights to his invention relating to an iron supplement for use in the treatment and/or prevention of infant low birth weight. No other disclosures were reported.
Funding/Support: This work was supported by the INSTAPA project, which received funding from the European Union’s Seventh Framework Programme (FP7/2007–2013) under grant agreement 211484. Fortitech and Swiss Precision Diagnostics donated fortification premix and urine pregnancy tests, respectively. Laboratory and Allied, Nairobi, prepared supplemental capsules (without financial compensation).
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank local authorities, field staff, community workers, research assistants, and students; Stephen Rogerson, MBBS, MRCP, DTM&H, FRACP, PhD (University of Melbourne/Royal Melbourne Hospital, Australia), Paul Milligan, PhD, and Tim Clayton, PhD (London School of Hygiene and Tropical Medicine), and Meghna Desai, MPH, PhD (US Centers for Disease Control and Prevention, Kisumu, Kenya), for data and safety monitoring board oversight; Mark Londema, MD, and Stephen Rulisa, BCh MB, PhD, for assistance in staff training and trial implementation; Kephas Otieno, MSc (KEMRI/CDC, Kenya), for help in placental examinations; Paul Hulshof (Wageningen University, the Netherlands) for measuring iron content in flour; Jenny Howard (London School of Hygiene and Tropical Medicine), Chris van Kreij, and Lucy Elburg (Wageningen University, the Netherlands) for administrative assistance; Karlijn van Rijzingen and Rini Geurts (St Elisabeth Hospital, the Netherlands) for laboratory assistance; and Luchtvaart Zonder Grenzen for free logistics support. Data and safety monitoring board members, Stephen Rulisa, and Paul Hulshof did not receive financial compensation for their contributions; the other individuals did.
Correction: This article was corrected online October 1, 2015, to add 2 individuals to the Additional Contributions section.