All of these events were monitored by an independent, unblinded Data and Safety Monitoring Board (DSMB) that met approximately twice a year during the course of the study. In addition, Bangladesh required additional monitoring by a local DSMB. The common protocol surveillance system was designed to capture severe GE occurring among participants upon presentation to medical facilities in the GSK1210151A study areas. Infants who underwent randomization were visited at least monthly to remind parents to bring their child to a clinic or hospital if they developed symptoms
of gastroenteritis [4] and [5]. GE was defined as three or more watery or looser-than-normal stools within a 24-h period and/or forceful vomiting [7]. Upon presentation to a medical facility, stool samples CCI-779 purchase were collected; history of symptoms of the current illness was collected through interview with the parent/guardian; and physical signs were documented by medical staff caring for the subject via direct observation. Data on ongoing symptoms and signs were collected throughout the course of the episode. These data were used to define severity using the 20-point modified Vesikari Clinical Scoring System
(VCSS) (“severe” was defined as a score of ≥11) [8], [10] and [11]. For this analysis, we also looked at a score of ≥15 and ≥19, indicating “very Ketanserin severe” or “extremely severe” GE. Rotavirus antigens in stool specimens were detected by enzyme immunoassay (EIA) [12]. Wild-type rotavirus was confirmed by reverse-transcriptase-polymerase-chain-reaction (RT-PCR) for identification of the VP6 genotype. Identification of rotavirus P and G genotypes was performed by RT-PCR as previously described [13]. EIA assays were conducted in the laboratory of Dr. Richard Ward at Children’s Hospital Medical Center, Cincinnati, OH; RT-PCR assays were conducted at Merck Research Laboratories. Statistical analysis. Efficacy was defined as 1–(Rvaccine/Rplacebo) × 100%, where R represented the incidence for the respective groups. It was assumed that the
number of cases in each group followed a Poisson distribution; the statistical analysis then conditioned on the total number of subjects with severe gastroenteritis from both treatment groups, such that the number of subjects with severe gastroenteritis in the vaccine group followed a binomial distribution. For subjects with multiple episodes, only the most severe episode (identified by the VCSS) was used for analysis. For efficacy calculations, we counted cases starting 2 weeks after receipt of third dose of vaccine (per-protocol definition). We also calculated efficacy by specific serotype of rotavirus according to the same methods. Exact inference was used, and follow-up time was accounted for in the calculations.