Christian 2003.

Methods	This was a double‐blind cluster‐randomised trial, carried out in rural Nepal. Dates of study: December 1998‐April 2001
Participants	A total of 4926 pregnant women were enrolled in the study. The women were randomised into 5 groups as follows: group 1 (n = 941), group 2 (n = 957), group 3 (n = 999), group 4 (n = 1050) and group 5 (n = 1051) Women who were currently pregnant or those who were breastfeeding an infant < 9 months old were excluded from the study. Also excluded were menopausal, sterilised or widowed women
Interventions	Group 1 received folic acid 400 mcg and vitamin A Group 2 received folic acid 400 mcg, iron 60 mg as ferrous fumerate and vitamin A Group 3 contained the same minerals as group 2 in addition to 30 mg of zinc as zinc sulphate Group 4 received similar micronutrients as group 3 in addition to vitamin D 10 mcg, vitamin E 10 mg, vitamin B1 1.6 mg, vitamin B2 1.8 mg, niacin 20 mg, vitamin B6 2.2 mg, vitamin B12 2.6 mcg, vitamin C 100 mg, vitamin K 65 mcg, copper 2 mg and magnesium 100 mg Group 5 (control group) received 1000 mcg of vitamin A only All supplements were given orally from the time of pregnancy detection until 12 weeks after a live birth or 5 weeks after a still birth or a miscarriage.
Outcomes	Preterm births, SGA (weight < 10 percentile of gestational age), LBW (< 2500 g), side‐effects, fetal loss, perinatal mortality, neonatal mortality, 3‐month infant mortality It should be noted that the data for SGA were obtained from a separate report (Food and Nutrition Bulletin 2009) and not from the individual trial report.
Notes	All women were offered 2 x 400 mg single‐dose albendazole in the second and third trimester of pregnancy because of the high prevalence of hookworm infestation in this population. Hookworm infestation and vitamin A deficiency are one of the major causes of anaemia in this population. Due to this reason, vitamin A was given to all the participants including the control group. Baseline characteristics did not differ significantly among the various randomisation groups except for ethnicity and land holding. In this review, we used the group 4 data for the MMN group and group 2 data for the control group. All the estimates were adjusted for the cluster design and provided by the trial authors. Declarations of interest: the trial authors declared no conflict of interest. Funding sources: US Agency for International Development (USAID) and additional support from the Unicef Country Office, Kathmandu, Nepal, and the Bill and Melinda Gates Foundation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomisation was done in blocks of five within each village development community by the senior study investigators, who drew numbered chips from a hat" Comment: probably done
Allocation concealment (selection bias)	Unclear risk	Quote: "Randomisation was done in blocks of five within each village development community by the senior study investigators, who drew numbered chips from a hat" Comment: insufficient information to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "participants, investigators, field staff and statisticians did not know supplement codes", "supplements, which were of identical shape, size, and color" and "code allocation was kept locked at the Johns Hopkins University, Baltimore". Comment: participants and caregivers were blinded to the treatment assignment.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "participants, investigators, field staff and statisticians did not know supplement codes" Comment: outcome assessors were blinded to the treatment assignment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Exclusion (1.43%) and attrition (6.9%) were reported along with their reasons
Selective reporting (reporting bias)	Low risk	Comment: results of all outcomes mentioned in methods were presented in the various publications of this trial.
Other bias	Low risk	Comment: no other bias was identified, including cluster‐design specific biases (recruitment bias, baseline imbalance, loss of clusters, incorrect analysis, and comparability with individually randomised trials).