In a double blind trial, 319 fully immunized children received two doses of either placebo or 10(6.7) focus-forming units of the attenuated RIX4414 human rotavirus (RV) vaccine ("all-in-one" formulation). Plasma RV-specific IgA (RV IgA), stool RV IgA, and circulating total and RV memory B cells (CD19+ IgD+/- CD27+) with an intestinal homing phenotype (alpha4beta7+ CCR9+/-) were measured, after the first and second doses, as potential correlates of protection. After the first and/or second dose, 54% of vaccinees and 13% of placebo recipients had plasma RV IgA. Before vaccination, most (95%) of the children (of both study groups) were breast-fed and had stool RV IgA (68.64%). Coproconversion (4-fold increase) after the first and/or second dose was observed in 32.7% of vaccinees and 17.4% of placebo recipients. No significant difference was seen when comparing the frequencies of any subset of memory B cells between vaccinees and placebo recipients. Statistically significant weak correlations were found between plasma RV IgA titers and coproconversion, and several subsets of memory B cells. The vaccine provided 74.8% protection (95% confidence interval, 30.93-92.62) against any RV gastroenteritis and 100% protection (95% confidence interval, 14.53-100) against severe RV gastroenteritis. When vaccinees and placebo recipients were considered together, a correlation was found between protection from disease and plasma RV IgA measured after dose 2 and RV memory (IgD- CD27+ alpha4beta7+ CCR9+) circulating B cells measured after dose 1. However, the correlation coefficients for both tests were low (<0.2), suggesting that other factors are important in explaining protection from disease.