Zeng 2008 (C).

Methods	Cluster‐randomised trial (3 arms). Villages were assigned to interventions. Villages were stratified and there was block randomisation to ensure geographical balance in 2 participating counties.
Participants	5828 eligible pregnant women with less than 28 weeks and resident in 2 poor rural counties in Shaanxi Province of north west China participated in the study. Village doctors recruited women by active surveillance. In the study areas there were no specific policies for the distribution of multiple micronutrients or iron‐folic acid supplements even in disadvantaged areas although folic acid supplements were promoted to prevent NTDs.Their villages were randomly assigned for women to receive 1 of 3 groups.
Interventions	Their villages were randomly assigned for participants to receive 1 of 3 groups: group 1, daily antenatal multiple micronutrients containing 30 mg elemental iron, 400 µg (0.4 mg) folic acid and 15 mg zinc, 2 mg copper, 65 µg selenium, 150 µg iodine, 800 µg vitamin A, 1.4 mg vitamin B1 (thiamine), 1.4 mg vitamin B2 (riboflavin), 1.9 mg vitamin B6, 2.6 µg vitamin B12, 5 µg vitamin D, 70 mg vitamin C, 10 mg vitamin E, and 18 mg niacin; group 2 who received a tablet containing 60 mg elemental iron and 400 μg (0.4 mg) of folic acid; group 3 received a tablet containing 400 μg (0.4 mg) folic acid alone (control). Setting and health worker cadre: the intervention was performed by local maternal and child health workers in rural, antenatal clinics and local health facilities in Shaanxi Province, China.
Outcomes	Birthweight within 1 hour of delivery, low birthweight, birth length, gestational age at birth, preterm delivery, small‐for‐gestational age babies, maternal Hb concentration in the third trimester (gestation 28‐32 weeks), anaemia in the third trimester, fetal losses during pregnancy, birth outcome, delivery information, neonatal and maternal deaths; neonatal survival at the 6 weeks, perinatal deaths, neonatal deaths, stillbirths. A follow‐up publication describes a follow‐up to 850 children born to women who participated in the study and focused on mental development outcomes using the Bayley scales of infant development at 3,6,12, and 24 months of age.
Notes	We have only included groups 2 (iron + folic acid) and 3 (folic acid alone) in the analyses. In the data tables we have adjusted the raw data presented in the paper to take account of the cluster‐design effect. We have calculated an effective sample size by dividing figures by the design effect calculated using the ICC for the trial’s primary outcome: birthweight ICC = 0.03. We have used the same sample adjustment for all outcomes. 65.9 of women in group 2 (iron + folic acid) and 65.2% of women in group 3 (folic acid) started supplementation before 16 weeks of gestational age. Gestational age at start of supplementation: mixed/unspecified gestational age. Anaemic status at start of supplementation: mixed anaemia status. Daily iron dose: higher dose group (60 mg elemental iron daily) in 1 group and 30 mg in another group (multiple micronutrients). Iron release formulation: normal release preparation/unspecified. Iron compound: unspecified. Malaria setting: yes. As of 2011: Malaria risk, including P. falciparum malaria, exists in Yunnan and to a lesser extent in Hainan. P. falciparum resistance to chloroquine and sulphadoxine–pyrimethamine reported. Limited risk of P. vivax malaria exists in southern and some central provinces, including Anhui, Ghuizhou, Henan, Hubei, Jiangsu. There is no malaria risk in urban areas.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“The randomisation schedule was generated off site with a pseudo‐random number generator.” Recruitment bias was unlikely.
Allocation concealment (selection bias)	Low risk	Off‐site randomisation.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Cluster trial all women in village received the same intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Low risk for laboratory outcomes but unclear for other outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Total clusters  (531). Total women  5828 (in 3 groups, 2 groups included in the analyses, total randomised 3929). Overall 133 women lost to follow‐up and 279 stopped taking supplements and were excluded (7% lost to follow‐up). 3270 women in groups 1 and 2 had live births (3306 babies). Approximately 6% further missing data for primary outcome (infant birthweight). Further missing data for other outcomes. Available case analysis for primary outcome (LBW).
Selective reporting (reporting bias)	Unclear risk	There is insufficient information to permit judgement.
Other bias	Unclear risk	The trial was stopped early because of funding constraints. The treatment groups appeared similar at baseline. Results were adjusted for cluster design effect.