Children were weighed to the nearest 0.1 kg and height was measured to the nearest 0.1 cm. Mean values for weight and height were calculated for each group, vaccine or placebo. The data were used to calculate weight-for-age Z scores (WAZ), height-for-age Z scores (HAZ), and find more weight-for-height Z scores (WHZ). Z scores were calculated using the WHO Child Growth Standards “igrowup” package for Stata, which uses the standard formula of the observed measure (weight or height) minus the reference measure taken from standard growth charts, divided by the standard deviation of the reference measure [1]. Malnutrition was defined as two or more Z scores below the reference [1]. Following anthropometry

study completion and data verification, data were linked with Phase 3 trial treatment arm assignment, birth weight, and age and weight from the other four study visits using the study allocation number and HDSS number assigned

to each child. For the primary analysis we assumed a 10% loss to follow-up from the original study enrollment of 1136 children, for a sample size of 1022 at the March–April 2010 visit. Given this sample size, we expected to have greater than 90% power to detect a difference in mean WAZ of 0.25, a difference in mean HAZ of 0.25, and a difference in mean WHZ of 0.23 between trial treatment groups at the March–April 2010 visit. The differences needed for statistical significance were assumed to be equivalent to a 15% or greater change in WAZ, HAZ, or WHZ. The t-test was used for the difference in mean WAZ, HAZ, or WHZ between vaccine and placebo groups at the March–April 2010 visit, as well as mean birth weight and mean weight at each of the four selleck Phase 3 trial visits. Chi-square and Fisher’s exact test were used to check for imbalances in the follow-up between males and females and trial treatment groups. Logistic regression was used to calculate odds ratios for the odds of

being malnourished between treatment groups. To check for a difference in growth patterns between treatment groups, longitudinal analyses were conducted using GEE with robust variance estimation. All analyses were conducted using Stata 11 (StataCorp Resminostat LP, College Station, TX). A total of 1136 infants were enrolled in the PRV study in Bangladesh beginning in March 2007 [21]. Three doses of vaccine or placebo were administered with the standard EPI vaccines at a mean age of 7.6, 11.8, and 16.0 weeks. Infants were evenly randomized to vaccine or placebo, and 54% of vaccine recipients and 49% of placebo recipients were male. Birth weight was available for 391 (34.4%) enrollees, of whom 18% were considered low birth weight. Weight was recorded at the three trial vaccination visits, at the trial closeout visit in March of 2009 (median age 20.1 months, IQR 18.0–22.5), and at a follow-up visit in March–April of 2010 (median age 32.3 months, IQR 30.1–34.8) for 1136 (100%), 887 (78.1%), 860 (75.7%), 1125 (99.0%), and 1033 (90.