A double blind randomised controlled trial comparing standard dose of iron supplementation for pregnant women with two screen-and-treat approaches using hepcidin as a biomarker for ready and safe to receive iron.

WHO recommends daily iron supplementation (60mg) for all pregnant women where anaemia prevalence exceeds 40%. However, recent evidence suggests that iron supplementation may be harmful as it increases the risks of hypertension and of infection. Iron absorption is regulated by hepcidin, a key iron regulatory hormone with the potential to be a useful marker to determine if oral iron can be absorbed effectively and safely. We aimed to identify a hepcidin threshold to define ’safe and ready’ to receive iron and then test whether a hepcidin-guided screen-and-treat approach to iron supplementation is non-inferior to the WHO-recommended universal daily supplementation. Method: We established our screening threshold by measuring haemoglobin and serum hepcidin, ferritin, iron, soluble transferrin receptor (sTfR), and C-reactive protein (CRP) at 14, 20 and 30 week of gestation among 395 pregnant rural Gambian women using archived maternal blood samples (2010-2013), and analysed hepcidin’s diagnostic test accuracy [area under the receiver operating characteristic curve (AUCROC), sensitivity, specificity, cut offs] for iron deficiency at each time point. We established a threshold of 2.5μg/L. We then conducted a 3-arm randomised-controlled proof-of-concept trial in rural Gambia from June 2014 to March 2016. We recruited 498 pregnant women aged 18-45 years with 14-22 weeks gestation to receive either: (A) UNU/UNICEF/WHO international multiple micronutrient preparation (UNIMMAP) containing iron 60mg/d; (B) UNIMMAP containing iron 60mg/d but based on a weekly hepcidin screening <2.5μg/L indicating if iron can be given for the next 7 days or not; or (C): as in (B), but with iron 30mg/d. We report the per protocol analysis for primary and secondary outcomes at Day 84. We assessed non-inferiority with the primary endpoint being haemoglobin concentration at Day 84 with a non-inferiority margin of -5.0g/L. Results: The evidence for non-inferiority for screen-and-treat approaches using either 60mg iron (mean haemoglobin difference relative to Reference arm: -2.2g/L; 95% CI: -4.6, 0.1g/L) or 30mg iron (-2.7g/L; 95% CI: -5.0, -0.5g/L) was marginal. Anaemia (haemoglobin <110g/L) at the end of intervention was less prevalent in the Daily iron supplementation (Reference) arm than both Screen-and-treat arms. Among those without inflammation at the end of intervention, the prevalence of iron deficiency (ferritin <15ug/L) was less in the Reference arm compared to the two Screen-and-treat arms; corresponding prevalence values for transferrin saturation <16%; soluble transferrin receptor >4.4mg/L and hepcidin <2.5μg/L were lower in the Reference arm. The Screen-and-treat approaches had no added advantage than universal daily iron supplementation in terms of adherence, side effects or safety outcomes. Conclusion: The daily 60mg iron supplementation arm performed better than both screen-and-treat arms for anaemia and other iron markers (hepcidin, ferritin, soluble transferrin receptor, transferrin saturation). We therefore found no support for a screen-and-treat iron supplementation based on hepcidin concentration <2.5μg/L in pregnant Gambian women.