BACKGROUND: Powders containing iron and other micronutrients are recommended as a strategy to prevent nutritional anaemia and other micronutrient deficiencies in children. We assessed the effects of provision of two micronutrient powder formulations, with or without zinc, to children in Pakistan. METHODS: We did a cluster randomised trial in urban and rural sites in Sindh, Pakistan. A baseline survey identified 256 clusters, which were randomly assigned (within urban and rural strata, by computer-generated random numbers) to one of three groups: non-supplemented control (group A), micronutrient powder without zinc (group B), or micronutrient powder with 10 mg zinc (group C). Children in the clusters aged 6 months were eligible for inclusion in the study. Powders were to be given daily between 6 and 18 months of age; follow-up was to age 2 years. Micronutrient powder sachets for groups B and C were identical except for colour; investigators and field and supervisory staff were masked to composition of the micronutrient powders until trial completion. Parents knew whether their child was receiving supplementation, but did not know whether the powder contained zinc. Primary outcomes were growth, episodes of diarrhoea, acute lower respiratory tract infection, fever, and incidence of admission to hospital. This trial is registered with ClinicalTrials.gov, number NCT00705445. RESULTS: The trial was done between Nov 1, 2008, and Dec 31, 2011. 947 children were enrolled in group A clusters, 910 in group B clusters, and 889 in group C clusters. Micronutrient powder administration was associated with lower risk of iron-deficiency anaemia at 18 months compared with the control group (odds ratio [OR] for micronutrient powder without zinc=0·20, 95% CI 0·11-0·36; OR for micronutrient powder with zinc=0·25, 95% CI 0·14-0·44). Compared with the control group, children in the group receiving micronutrient powder without zinc gained an extra 0·31 cm (95% CI 0·03-0·59) between 6 and 18 months of age and children receiving micronutrient powder with zinc an extra 0·56 cm (0·29-0·84). We recorded strong evidence of an increased proportion of days with diarrhoea (p=0·001) and increased incidence of bloody diarrhoea (p=0·003) between 6 and 18 months in the two micronutrient powder groups, and reported chest indrawing (p=0·03). Incidence of febrile episodes or admission to hospital for diarrhoea, respiratory problems, or febrile episodes did not differ between the three groups. INTERPRETATION: Use of micronutrient powders reduces iron-deficiency anaemia in young children. However, the excess burden of diarrhoea and respiratory morbidities associated with micronutrient powder use and the very small effect on growth recorded suggest that a careful assessment of risks and benefits must be done in populations with malnourished children and high diarrhoea burdens. FUNDING: Bill & Melinda Gates Foundation.