Cogswell 2003.

Methods	RCT, 2 arms with individual randomisation.
Participants	275 legally competent, non‐imprisoned, non‐anaemic, low‐income pregnant women at < 20 weeks of gestation with ferritin levels above 20 μg/L enrolled at the Cuyahoga County, MetroHealth Center, Supplemental Nutrition Program for Women, Infants and Children in Cleveland, Ohio, USA.
Interventions	Participants were randomly assigned to 1 of 2 groups: group 1 received 1 gelatin capsule containing 30 mg of elemental iron (as ferrous sulphate) daily; group 2 received 1 placebo soft gelatin capsule daily for 119 days. Supplementation started at an average of 11 weeks of gestation until delivery.
Outcomes	Maternal: prevalence of anaemia at 28 and 38 weeks, side effects, compliance to treatment, maternal weight gain, iron status (MCV, Hb concentration, serum ferritin, erythrocyte protoporphyrin concentrations at 28 and 38 weeks. Infant: birthweight, birth length, proportion of low birthweight, low birthweight and premature, small‐for‐gestational age. Setting and health worker cadre: the intervention was performed by a dietician at the Cuyahoga County, MetroHealth Medical Center, Supplemental Nutrition Program for Women, Infants and Children in Cleveland, Ohio, United States of America.
Notes	Unsupervised. Women were re‐evaluated at 28 weeks of gestation, and according to Hb concentrations at that time were prescribed treatment following the Institute of Medicine guidelines for iron supplementation during pregnancy. Compliance was 63.4% and 65.2% in groups 1 and 2 respectively. Gestational age at start of supplementation: early gestational age (supplementation started before 20 weeks' gestation). Anaemic status at start of supplementation: non‐anaemic. Daily iron dose: lower daily dose (30 mg of elemental iron). Iron release formulation: normal release preparation/unspecified. Iron compound: ferrous sulphate. Malaria setting: non‐malarial setting. As of 2011: Malaria: no risk.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	By computerised random numbers.
Allocation concealment (selection bias)	Low risk	Placebo‐controlled trial. Randomisation by study data manager. The placebo and active treatment were indistinguishable and all staff were blind to group allocation.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Placebo‐controlled trial, with placebo indistinguishable from the active treatment. Bottles were coded and treatment group was only known to a data manager. It was stated that women and staff were not aware of treatment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	It was stated that laboratory analysts and staff collecting information on side effects were blind to treatment allocation.
Incomplete outcome data (attrition bias) All outcomes	High risk	More than 20% lost to follow‐up.
Selective reporting (reporting bias)	Unclear risk	There is insufficient information to permit judgement.
Other bias	Unclear risk	No other bias apparent.