University of Oxford – Iron Bioavailability Study (December 2024)

The study will also determine whether the relative bioavailability of iron from the three supplements remains consistent when two doses are given two days in a row vs. separated by a day. This consecutive-day testing helps estimate "steady state" effects that better reflect iron absorption patterns in pregnant women taking supplements daily. Research suggests the body reduces iron absorption from subsequent doses after receiving large initial doses. Researchers believe these regulatory effects are short-lived, making the comparison between the two dose spacings sufficient to measure this effect.10

All participants will undergo four sets of absorption studies separated by 14-day periods, ensuring thorough assessment across all supplementation conditions.11 The study will recruit healthy pregnant women from the community and antenatal clinics, in the 2nd trimester, without severe anemia, and without severe inflammation.12

• It will be pre-registered.13
• It is powered to detect a 20% difference in absorption, which researchers consider physiologically relevant.14
• It will include multiple linear regression analyses to identify predictors of iron absorption, including variables such as pregnancy stage, iron status, and inflammation markers.15
• The design aims to capture absorption at "steady state" conditions.16
• It will measure absorption both on an empty stomach and after meals containing iron-inhibiting foods, reflecting real-world conditions.17

The research team

The team will include:

• Dr. Nicole Stoffel, PhD, Department of Health Sciences and Technology, ETH Zurich, Switzerland
• Prof. Simon Karanja, PhD, Medical Epidemiology, Jomo Kenyatta University of Agriculture and Technology, Nairobi, Kenya
• Prof. Michael Zimmermann, MD, Radcliffe Department of Medicine, University of Oxford, Oxford, UK

This study will be conducted at the ETH Zurich/Oxford/JKUAT research facility in Msambweni Hospital in southern Kenya.18

Budget for grant activities

A summary of the study budget is below:19

Budget
Administration	$4,500
International Travel	$4,600
Rental fees	$1,400
Isotopes and analysis	$65,700
Laboratory analyses	$17,500
Miscellaneous	$8,900
Salaries, reimbursement	$67,700
Overhead (10%)	$17,030
Total	$187,330

The case for the grant

We are recommending this grant because:

• We think this study will impact our nutrition grantmaking. We think this study will inform our best guess of how much weight to put on the Smith et al. 2017 subgroup analysis reporting possible increases in early-life mortality from replacing IFA containing 60 mg of iron with MMS containing 30 mg of iron, by informing our view on the biological plausibility of the results. The findings could directly influence our funding decisions in several ways:
  • If results show that absolute iron absorption is similar from MMS with 30 mg of iron and IFA with 60 mg of iron, this would imply that mortality risks from replacing IFA with MMS are unlikely. In turn, it would make us more likely to support the shift to MMS due to the other benefits provided by MMS beyond iron.
  • If results show that absolute iron absorption is substantially lower from MMS with 30 mg of iron than IFA with 60 mg of iron, this would increase our concerns about possible risks of replacing IFA with MMS. In turn, this could lead us to focus on ways to mitigate these risks, such as improving anemia screening and treatment in areas where the shift from IFA to MMS is already happening.
  • Data on MMS with 60 mg iron could inform our view on whether this alternative formulation would provide similar bioavailability to IFA with 60 mg iron, informing how promising it would be to advocate for this specific formulation.
• This research could influence global health policy beyond our own grantmaking.
  • Based on conversations we had with the WHO, we think the results of this study could affect its antenatal care recommendations.20 This guidance will be important for countries looking to transition from IFA to MMS, especially for high-anemia settings where optimal iron dosage remains unclear.21
• We believe Dr. Zimmermann is well-positioned to conduct this research.
  • We have a positive impression from our conversations with him. He also has access to laboratory facilities capable of testing pregnant women in Kenya, which enhances the external validity of the study by conducting research in a relevant population with high anemia prevalence.
• We have confidence in the study design based on thorough consultation.
  • We've discussed the methodology with Dr. Zimmermann, WHO experts, and the MMS Technical Advisory Group (MMS TAG) and think it is sound. (More)

Risks and reservations

Our main reservations about this grant are:

• We’re unsure about global room for more funding for MMS. We haven't modeled how much funding this research might influence in detail. Our cost effectiveness estimates of MMS are very uncertain, which makes us uncertain about the size of funding opportunities. We also have not looked in detail into opportunities to fund anemia screening and treatment, or advocate for MMS 60mg. This makes it difficult to estimate the potential impact this study could have on redirecting MMS-related funding.
• The study might not have much influence on global MMS policy. It’s possible that part of the health community advocating for the shift from IFA to MMS might not find this study informative about the relative effects of IFA and MMS. For example, other researchers might not put much weight on iron absorption as an indicator of biological plausibility (and would instead prefer to focus on maternal iron status).22
• There might be some limitations to the study’s external validity, though we don’t think these are major. These include:
  • The study only includes women in the 2nd trimester and excludes those with high inflammation levels. This limits external validity, since in a program setting women would continue taking the supplements through the third trimester, and supplements would also be provided to women with very high inflammation.23 However, we think including these populations could increase study complexity and data variability. The latter may make it more difficult to reach statistically significant conclusions.
  • Researchers will use a regression analysis rather than separate groups to determine how gestational age, anemia levels and inflammation affect iron absorption from IFA and MMS. This method will provide less confidence in how these factors affect absorption than a trial with separate groups would. However, we think a regression analysis is reasonable given that additional groups would increase study complexity, and we also lack strong evidence that these factors affect iron absorption.
  • The "steady state" measurement approach depends on assumptions about iron absorption dynamics that we haven't looked into. Whether or not this is appropriate depends on the duration of body “reaction” after high doses of iron (i.e., how long the body continues to absorb less iron from subsequent doses). Researchers have argued this timeline is appropriate, and we found a paper supporting this thesis, but have not looked at this in depth.24
• There are some ethical concerns about fasting (skipping breakfast) among pregnant women and not treating women identified as having mild and moderate anemia. Researchers have addressed these by noting that vomiting is infrequent in the study area, women will remain under observation, and they will discuss treatment protocols with the ethics committee.25

Plans for follow up

• We expect to share results of the study with WHO and the MMS TAG as we receive them, and then publish results.

Internal forecasts

For this grant, we are recording the following forecasts:

Confidence	Prediction	By time	Resolution
35%	We decide to put less weight on the subgroup analysis as a result of this study	End of Dec 2025
30%	We do not change the weight on the subgroup analysis as a result of this study	End of Dec 2025
35%	We decide to put more weight on the subgroup analysis as a result of this study	End of Dec 2025

Our process

• We spoke to the WHO about existing evidence on MMS, and they suggested this study might help make progress on uncertainties about iron absorption from different supplement formulations.
• We held discussions with Dr. Zimmermann about conducting research to compare iron absorption from different MMS and IFA formulations, and then received and reviewed an initial proposal from Dr. Zimmermann's team.
• We shared specific questions with Dr. Zimmermann regarding study design, power calculations, and program relevance, and incorporated his responses.
• We solicited feedback on the study design from the MMS Technical Advisory Group and WHO technical experts, which was incorporated into the final trial design.

Sources

Document	Source
Conversation between GiveWell and the WHO, September 4, 2024	Unpublished
Dr. Michael Zimmermann, Responses to GiveWell's questions, October 21, 2024	Unpublished
Haider and Bhutta 2015	Source
Healthy Mothers Healthy Babies Consortium, "Multiple Micronutrient Supplementation in Pregnancy Technical Advisory Group"	Source (archive)
Smith et al. 2017	Source
Stoffel, Karanja, and Zimmermann, Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024	Source
Summary of MMS TAG feedback to Michael Zimmermann, November 2024	Unpublished
University of Oxford team, full budget, GiveWell IFA MMS iron absorption study, 2024	Source

• 1

  “Multiple‐micronutrient (MMN) deficiencies often coexist among women of reproductive age in low‐ to middle‐income countries. They are exacerbated in pregnancy due to the increased demands, leading to potentially adverse effects on the mother and developing fetus.” Haider and Bhutta 2015

• 2

  “This UNIMMAP tablet provides one recommended daily allowance of vitamin A, vitamin B1, vitamin B2, niacin, vitamin B6, vitamin B12, folic acid, vitamin C, vitamin D, vitamin E, copper, selenium and iodine with 30 mg of iron and 15 mg of zinc for pregnant women.” Haider and Bhutta 2015

• 3

  "In response to the 2016 WHO Guidelines for Antenatal Care, the New York Academy of Sciences (NYAS) assembled a scientific task force comprised of international experts in micronutrient deficiencies, public health, nutrition, pediatrics, and health economics to: 

  • Compile the evidence on the prevalence of micronutrient deficiencies in pregnant women or women of reproductive age. 
  • Review the evidence on the benefits and risks of multiple micronutrient supplements on maternal and perinatal outcomes.
  • Create a roadmap to guide decisions in countries considering the implementation of such programs. The first phase of this initiative revealed that shifting from iron and folic acid supplementation (IFA) to MMS in antenatal care (ANC) programs could yield substantial benefits in reducing maternal mortality and improving birth outcomes"

  Healthy Mothers Healthy Babies Consortium, "Multiple Micronutrient Supplementation in Pregnancy Technical Advisory Group"

• 4

  Smith et al. 2017. Supplemental Table 7. Effect of MMS on stillbirth, mortality, and birth outcomes stratified by iron dose provided in the MMS arm as compared to the IFA comparison arm

• 5

  For statistical significance, see 95% confidence intervals in Smith et al. 2017. Supplemental Table 7. Effect of MMS on stillbirth, mortality, and birth outcomes stratified by iron dose provided in the MMS arm as compared to the IFA comparison arm

• 6
  • Stoffel, Karanja, and Zimmermann, Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024
    • "In a follow-up analysis of data from 11 trials, Gomes et al. (2022) compared IFA to MMS with either 30 mg or 60 mg of iron. This analysis focused on Hb, third-trimester anemia, and iron deficiency anemia (IDA), considering daily iron dosage, total supplemental iron intake, and the baseline prevalence of anemia. The results showed no significant differences between MMS and IFA regarding third-trimester Hb or the risks of anemia and IDA, regardless of the iron dosage or total supplemental iron consumed. Notably, MMS providing 30 mg of iron was found to be comparable to IFA with 60 mg of iron…" (p. 1)
    • "Other micronutrients in the MMS, such as vitamins A, C and riboflavin, may improve the absorption and/or utilization of iron, particularly in inhibitory food matrices. Vitamin C enhances iron absorption from ferrous fumarate in nonpregnant women (Fidler et al 2003; von Siebenthal et al 2023). Fidler et al. (2003) compared iron absorption from ferrous fumarate-fortified meals with and without ascorbic acid added at a 4:1 molar ratio (relative to 5 mg iron): ascorbic acid increased Fe absorption from 6.3 to 10.4%, (P=0.02). Similarly, von Siebenthal (2023) reported that iron absorption from a dose of 100 mg iron as ferrous fumarate was increased 30% by an 80 mg dose of ascorbic acid (p < .001)" (p. 2)
  • "To overcome other possible maternal micronutrient deficiencies, the United Nations Children's Fund (UNICEF), United Nations University (UNU) and the WHO, in 1999, agreed on the composition of a proposed multiple‐micronutrient (MMN) tablet. This UNIMMAP tablet provides one recommended daily allowance of vitamin A, vitamin B1, vitamin B2, niacin, vitamin B6, vitamin B12, folic acid, vitamin C, vitamin D, vitamin E, copper, selenium and iodine with 30 mg of iron and 15 mg of zinc for pregnant womenIn contrast to the WHO recommendation, a lower dose of iron was recommended as the absorption of iron was expected to be enhanced due to vitamin C, vitamin A, and riboflavin, and given that the majority of pregnant women suffer from mild anaemia and the potential side‐effects associated with higher doses of iron." Haider and Bhutta 2015

• 7

  "The primary outcome will be TIA [total iron absorption] for each condition of iron administration, calculated as described above. The specified comparisons will be:
  In Study A
  (i) TIA from MMS with 30 mg iron versus IFA given with maize porridge
  (ii) TIA from MMS with 60 mg iron versus IFA given with maize porridge
  In Study B
  (i) TIA from MMS with 30 mg iron versus IFA given fasted with water
  (ii) TIA from MMS with 60 mg iron versus IFA given fasted with water
  (iii) TIA from MMS with 30 mg iron from the consecutive day versus alternate day dosing
  (iv) TIA from MMS with 60 mg iron from the consecutive day versus alternate day dosing
  (v) TIA from IFA with 60 mg iron from the consecutive day versus alternate day dosing"
  Stoffel, Karanja, and Zimmermann, Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024, p. 6

• 8

  “The study will be a controlled, open-label, randomized cross-over trial. During screening, 1 week before study start, we will assess women for eligibility. We will collect a venous blood sample for the determination of Hb concentration, Hb type and pregnancy status, and measure weight to the nearest 0.1 kg and height to the nearest 0.5 cm. Eligible and consented subjects will be enrolled in the iron absorption study.
  All women will undergo four sets of absorption studies separated by 14-day isotope incorporation periods. The first set (Study A) will assess iron absorption from MMS with 30 mg or 60 mg iron versus IFA with 60 mg iron when given with an inhibitory meal. The second, third and fourth sets (Study B) will assess iron absorption from the three formulations when given fasted and when given daily and on alternate days.” Stoffel, Karanja, and Zimmermann, Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024, p. 3

• 9
  • "Our study aims are, in pregnant women in the 2nd trimester:

  1. Quantify and compare iron absorption (bioavailability) from MMS with 30 mg or 60 mg iron
  versus IFA with 60 mg iron, in the fasted state and when given with an inhibitory meal."

  • "(iii) the relative effect of an inhibitory meal matrix, characteristic of diets in LMIC, on iron absorption from the MMS and IFA formulations has not been tested."

  Stoffel, Karanja, and Zimmermann, Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024, p. 2

• 10

  “In study 1, using a crossover design, we administered two iron challenges either as a single dose or as two doses given on consecutive days. Subjects were randomly assigned to start the study with one of the two treatments. Iron was administered at 8.00 a.m. in 4 different iron concentrations (40, 80, 160, and 240 mg as elemental Fe). In study 2, we administered two single doses of 60 mg elemental iron at 8:00 a.m. on two consecutive days and similarly assessed hepcidin response until 48 h post-administration. In studies 1 and 2, 24 h after doses ≥60 mg, hepcidin was increased (p < 0.01) and fractional iron absorption was decreased by 35%–45% (p < 0.01)... In summary, these short-term data suggested that, to maximize fractional absorption: (A) oral iron at doses ≥60 mg should be spaced by 48 h; and (B) twice-daily dosing should be avoided”
  Stoffel et al. 2020, p. 4.

• 11

  "All women will undergo four sets of absorption studies separated by 14-day isotope incorporation periods." Stoffel, Karanja, and Zimmermann, Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024, p. 3

• 12

  "We will recruit healthy pregnant women from the community and antenatal clinics. Inclusion criteria will be: female; pregnant at gestational age 12 (±1) weeks (dated by ultrasound); age 18 to 35 y; Hb concentration ≥80 g/L; CRP < 10 mg/L (absence of significant inflammation); body weight <80 kg; no major chronic diseases; no intake of outside vitamin and mineral supplements in the 2 weeks before study start and during the study; no blood transfusion, blood donation, or significant blood loss over the past 4 mo. Exclusion criteria will be: severe anemia (defined as Hb <80 g/L), malaria, sickle cell disease (SS and SC) or hemoglobin C disease (CC)." Stoffel, Karanja, and Zimmermann, Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024, p. 2

• 13

  The trial investigators confirmed to GiveWell that the trial will be registered on clinicaltrials.gov prior to starting the study. Dr. Michael Zimmermann, Responses to GiveWell's questions, October 21, 2024 (unpublished)

• 14

  “We feel a 20% difference would be physiologically relevant. We’ve changed the power calculation in the revised protocol to 38+20%attrition= 46 + 4 for uncertainty, so n=50 to detect a 20% difference” Dr. Michael Zimmermann, Responses to GiveWell's questions, October 21, 2024 (unpublished)

• 15

  “Separately, we will perform multiple linear regression analyses to determine predictors of TFIA, including variables such as week of pregnancy, hepcidin, iron status (serum ferritin, Hb, sTfR) and inflammation markers (CRP, AGP).” Stoffel, Karanja, and Zimmermann, Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024, p. 7

• 16

  See above.

• 17

  "Quantify and compare iron absorption (bioavailability) from MMS with 30 mg or 60 mg iron
  versus IFA with 60 mg iron, in the fasted state and when given with an inhibitory meal." Stoffel et al., Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024, p. 2

• 18

  "This study will be conducted in Q1-2 of 2025 at the ETH Zurich/Oxford/JKUAT research facility in Msambwemi Hospital in southern Kenya." Stoffel et al., Study protocol: Comparing iron absorption from multiple micronutrient supplements and iron-folate supplements: a stable isotope study in Kenyan pregnant women, 2024, p. 2

• 19

  See University of Oxford team, full budget, GiveWell IFA MMS iron absorption study, 2024 for budget details.

• 20

  Conversation between GiveWell and the WHO, September 4, 2024 (unpublished)

• 21

  Conversation between GiveWell and the WHO, September 4, 2024 (unpublished)

• 22
  • MMS TAG: “The study design looks good. They have indicated that fractional iron absorption is the primary outcome, which is fine, but they don’t discuss how they would interpret a finding that shows no difference in fractional iron absorption but higher total iron absorption in the comparison of 60 vs 30 mg iron. The ultimate outcome of greatest programmatic relevance is iron deficiency or iron-deficiency anemia, which may or may not track with differences in fractional or total iron absorption. However, because they are using a cross-over study design, using iron deficiency or iron-deficiency anemia in late pregnancy as key outcomes would not be a logical choice in this case. For those reasons, I think the study will provide useful information but I’m not sure whether it will yield definitive programmatic conclusions.” Summary of MMS TAG feedback to Michael Zimmermann, November 2024 (unpublished)
  • While we believe this is a valid concern, we think the trial will be quite valuable regardless. For example, it is plausible that the trial will find that the amount of iron absorbed from MMS with 30 mg of iron vs. IFA with 60 mg of iron is approximately the same. If so, that would substantially undermine the rationale for viewing MMS as less beneficial than IFA, and substantially update GiveWell’s perspective on MMS.

• 23

  Because supplements in a program setting would be provided to all women universally.

• 24
  • "The acute hepcidin response induced by 60 or 120 mg daily oral iron does not gradually increase when supplements are given over a week vs the increase seen after two successive days, in nonpregnant women. So, yes, we think the current design will ‘max’ hepcidin and represent daily dosing” Dr. Michael Zimmermann, Responses to GiveWell's questions, October 21, 2024 (unpublished)
  • “In study 1, using a crossover design, we administered two iron challenges either as a single dose or as two doses given on consecutive days. Subjects were randomly assigned to start the study with one of the two treatments. Iron was administered at 8.00 a.m. in 4 different iron concentrations (40, 80, 160, and 240 mg as elemental Fe). In study 2, we administered two single doses of 60 mg elemental iron at 8:00 a.m. on two consecutive days and similarly assessed hepcidin response until 48 h post-administration. In studies 1 and 2, 24 h after doses ≥60 mg, hepcidin was increased (p < 0.01) and fractional iron absorption was decreased by 35%–45% (p < 0.01)... In summary, these short-term data suggested that, to maximize fractional absorption: (A) oral iron at doses ≥60 mg should be spaced by 48 h; and (B) twice-daily dosing should be avoided" Stoffel et al. 2020, p. 4.4

• 25

  “MMS TAG: My first question relates to the study groups where pregnant women are asked to fast overnight, take a supplement on an empty stomach and continue to fast for a few hours. I feel this will be challenging for many women during a period when they are often not feeling well and are prone to vomiting, esp. After consuming an iron containing supplement without food. How do the investigators plan to manage this challenge, not only for women’s well being but also potential impact to the study? i.e. women may not retain/absorb the supplement and may not report that they vomited."
  Dr. Zimmermann: This is a valid consideration. Our sample size includes a 20% drop out rate; it is possible that a woman may vomit the study dose, but this is very infrequent based on our previous experience with women in this area, even when given higher doses. The women will remain at the study site under the care of the investigators for 2 hrs. After consuming the iron, participants will remain at the study site for 2 hrs under the supervision of the study team and will be instructed not to eat or drink for 3 h, except for 500 mL of bottled water. The water will be provided to the women by the study team, and they can begin to drink it 1 h after the test condition."
  MMS TAG: Are we comfortable to give women a preventive dose of iron when she is identified as anaemic (moderate and mild) knowing the potential risks of anaemia during pregnancy?"
  Dr. Zimmermann: This is a valid point. All women will be receiving iron, and we expect the iron doses will be well absorbed, although they are lower than might be given to a woman with moderate anemia. We
  will discuss this point with the ethical committee. Certainly, if we find during the study that a woman ‘s Hb has dropped below 80 g/L, she will be excluded from the study and referred to the antenatal care clinic for standard-of-care treatment with iron and other measures.”
  Summary of MMS TAG feedback to Michael Zimmermann, November 2024 (unpublished)