All solicited injection site and systemic reactions were considered to be related to vaccination by definition. The following were denoted as AESIs (adverse events of specific interest) for the JE-CV vaccine and collected up to 28 days after vaccination: hypersensitivity/allergic reactions, neurological events including febrile convulsions, and vaccine failure. Guidelines were also

provided to the investigator as assistance in the assessment of AEs that may be indicative of viscerotropic/neurotropic disease. All serious adverse events (SAEs) were collected from Day 0 until 6 months after the last vaccination and only related SAEs (as per investigator) were collected from this time until 12 months after the first vaccination. All deaths were collected during the study. AEs were coded using the Medical Dictionary for Regulatory activities Antidiabetic Compound Library cell assay (MedDRA version 12.0) preferred term. Statistical analysis was performed using SAS® 9.2 software. The null hypothesis (to be rejected AZD5363 clinical trial to demonstrate the primary objective) was that at least one

of the antibody responses to the concomitant administration of JE-CV and MMR was inferior to that of JE-CV or MMR vaccination alone by more than a maximum clinically acceptable limit for non-inferiority. This limit was set at 10% based on available data and recommendations for the development of JE vaccines from a group of experts assembled by WHO [8] and [9]. The four non-inferiority tests were performed using two-sided, 95% confidence intervals (CI) of pairwise differences between groups, using Wilson score method without continuity correction [10]. Non-inferiority was demonstrated if the lower bounds of all four 95% CIs were above −10%. Non-inferiority was tested on the per-protocol (PP) population and confirmed in the full analysis set (FAS) of all children who to were randomized and received at least one dose of vaccine. In addition to protocol deviations, children

were excluded from the non-inferiority analysis of JE antibody response if they were JE-seropositive at baseline. The sample size was calculated using the Farrington and Manning method, and an alpha level of 2.5% (one-sided hypothesis) for each comparison, to provide an overall power of >90% [11]. Assuming a 10% protocol deviation rate, and that 3%, 20%, 10% and 15% of children would be seropositive at baseline for JE, measles, mumps, and rubella, respectively, the planned sample size was 110, 220, and 220 for the three groups, respectively. The sample size of the first group is smaller because this group is included in only one comparison (JE antibody response), compared to at least three in the other groups. No alpha adjustment for multiple comparisons was necessary in these calculations, but a power adjustment was performed.