Trial Design and Oversight
This phase 2b, randomized, double-blind, placebo-controlled trial involving 1649 participants was conducted from January 2013 through July 2016 in primary health centers and hospitals in Central Java and Yogyakarta, Indonesia. Indonesia is a low-middle–income country; the per capita gross regional product in Yogyakarta and Central Java is $2,164 to $2,326 (in U.S. dollars), and the mortality rate is 30 to 38 deaths per 1000 live births among children younger than 5 years of age.15,16
The protocol, which is available with the full text of this article at NEJM.org, was approved by the ethics committees at Universitas Gadjah Mada, the Royal Children’s Hospital Melbourne, and the National Agency of Drug and Food Control, Republic of Indonesia. The use of a placebo was deemed to be acceptable because vaccination against rotavirus disease is not currently being implemented under the Indonesian National Immunization Program and the cost of the vaccines limits private purchase.17
The trial was conducted in accordance with International Conference on Harmonisation Good Clinical Practice guidelines and was monitored by an independent contract research organization (Quintiles). The conduct of the trial was overseen by Murdoch Children’s Research Institute, with local input from PT Bio Farma. An independent data and safety monitoring board regularly reviewed the safety data. Data management was performed by Biophics Thailand. Statistical analysis was conducted by INC Research, Australia, and by an independent statistical consultant. The National Health and Medical Research Council, the Bill and Melinda Gates Foundation, and PT Bio Farma funded the trial but had no role in the trial design, data collection, or data interpretation, or in the decision to submit the manuscript for publication. All the authors reviewed the manuscript and vouch for the accuracy and completeness of the data and analysis and for the fidelity of the trial to the protocol.
Participants, Randomization, and Blinding
Pregnant women provided preliminary written informed consent before a sample of cord blood was obtained. Final written informed consent was obtained from the parent or guardian after birth, before eligibility for the infant’s participation in the trial was confirmed. Infants were eligible if they were healthy, full-term babies 0 to 5 days of age who had a birth weight of 2.5 to 4.0 kg. Eligible infants were randomly assigned, in a 1:1:1 ratio, to one of three groups: a neonatal-schedule vaccine group, an infant-schedule vaccine group, or a placebo group. Randomization was performed according to a computer-generated code with a block size of 6, with stratification according to province. The doses of vaccine (RV3-BB) or placebo were drawn into syringes for dispensing by a pharmacist who was located at the central pharmacy in each province and was aware of the trial-group assignments. Investigators, trial monitors, data managers, statisticians, and other trial staff, as well as the families of the participants, remained unaware of the trial-group assignments for the duration of the trial.
Participants received four 1-ml oral doses of vaccine or placebo according to their trial-group assignment, with doses administered at 0 to 5 days of age (dose 1), 8 to 10 weeks of age (dose 2), 14 to 16 weeks of age (dose 3), and 18 to 20 weeks of age (dose 4) (see Fig. S1 in the Supplementary Appendix, available at NEJM.org). Each of the two vaccine groups received three doses of RV3-BB and one dose of placebo. In the neonatal-schedule vaccine group, doses 1, 2, and 3 were RV3-BB and dose 4 was placebo, and in the infant-schedule vaccine group, dose 1 was placebo and doses 2, 3, and 4 were RV3-BB. Doses 2, 3, and 4 were preceded by a 2-ml dose of an antacid solution (Mylanta Original). Feeding was withheld for 30 minutes before and after each dose. The vaccine or placebo was administered at the same time as vaccines that were provided as part of the Indonesian National Immunization Program. Participants were followed by means of weekly telephone contact and monthly visits until the age of 18 months. All the participants received oral polio vaccine, with the exception of a subgroup of 282 participants (the first cohort of participants recruited) who received an inactivated polio vaccine.
Clinical trial lots of RV3-BB were prepared at Meridian Life Science to a titer of 8.3×106 to 8.7×106 focus-forming units per milliliter in a serum-free medium that was supplemented with 10% sucrose. Placebo consisted of the same medium with 10% sucrose and was visually indistinguishable from RV3-BB. Vials of vaccine or placebo were stored at −70°C until they were thawed within 6 hours before administration.
Gastroenteritis of any severity was defined as having three or more stools that were looser than normal for a given child within a 24-hour period. The severity of gastroenteritis was defined on the basis of the Vesikari clinical severity scoring system (scores range from 0 to 20, with higher scores indicating more severe disease) that takes into account clinical symptoms (diarrhea and vomiting), clinical signs (elevated body temperature and dehydration), and type of treatment, if any.18 A modified Vesikari score was applied in cases in which intravenous, nasogastric rehydration or 6 hours of supervised oral rehydration was scored as hospitalization, regardless of whether the rehydration was administered at a primary health center or at a hospital. Rotavirus gastroenteritis was defined as gastroenteritis coincident with the presence of rotavirus antigen in the stool that was detected with the use of an enzyme-linked absorbent assay (ProSpecT Rotavirus Microplate Assay, Oxoid). Severe rotavirus gastroenteritis was defined as rotavirus gastroenteritis with a modified Vesikari score of at least 11.
Vaccine Response and Immunogenicity
Vaccine response (often called “vaccine take”), as evidenced by serum immune response or shedding of RV3-BB in the stool, was assessed in the first cohort recruited (282 participants). A blood sample was obtained from the cord (which represented baseline for the neonatal schedule), immediately before dose 2 of vaccine or placebo (which represented baseline for the infant schedule), 28 days after dose 3, and 28 days after dose 4. Serum rotavirus IgA antibody titers and serum neutralizing antibody titers were measured with the use of previously described methods.14,19 The methods used to detect serologic responses to RV3-BB also detect responses to wild-type rotavirus strains. To determine the background exposure to wild-type rotavirus strains, serum immune response and shedding of RV3-BB in the stool were also assessed in participants in the placebo group. The shedding of RV3-BB in the stool was detected with the use of a rotavirus VP6–specific reverse-transcriptase–polymerase-chain-reaction assay and confirmed by sequence analysis.14 Vaccine response was defined quantitatively as a serum immune response (a serum rotavirus IgA antibody titer or a serum neutralizing antibody titer three times as high as the titer at baseline) 28 days after administration of the vaccine or shedding of RV3-BB between days 3 and 7 after administration of the vaccine. Cumulative vaccine response was defined as
evidence of vaccine response after dose 1, 2, or 3 in the neonatal-schedule vaccine group and after dose 2, 3, or 4 in the infant-schedule vaccine group.
Vital signs were evaluated before, and in the 30 minutes after, administration of the vaccine or placebo. Parents reported the participant’s temperature and solicited gastrointestinal and systemic symptoms on diary cards for 7 days after each dose. Parents were instructed to contact the trial staff immediately if blood was present in the stool so that further investigations to exclude intussusception, including ultrasonography, could be performed if clinically indicated. All unsolicited adverse events that were reported up to 28 days after administration of a dose of the vaccine or placebo were assessed according to the Division of AIDS grading table, version 1.0 (updated August 2009)20 for the grading of laboratory abnormalities reported as adverse events and according to standard criteria defined in the protocol for the grading of clinical adverse events. A serious adverse event was defined as an adverse event that resulted in death or in new or prolonged hospitalization or was considered to be medically significant or life threatening and occurred within 28 days after a dose of vaccine or placebo. Causality and severity grading of adverse events were determined by the local Indonesian investigators.
In the primary analysis of efficacy, we compared the percentage of participants in the neonatal-schedule vaccine group and infant-schedule vaccine group combined (combined vaccine group) who had an episode of severe rotavirus gastroenteritis during the period from 2 weeks after the administration of dose 4 through 18 months of age with the percentage in the placebo group who had such an episode during the same time period, using Pearson’s chi-square test. The primary analysis was conducted in the per-protocol population, which included only the participants who received all four doses of vaccine or placebo within the visit windows. In a secondary analysis, which was conducted in the intention-to-treat population (all participants who underwent randomization), we compared events of severe rotavirus gastroenteritis that occurred from randomization through 18 months of age. Vaccine efficacy is presented as 1 minus the relative risk of an event in the vaccine group as compared with that in the placebo group and multiplied by 100, and its exact 95% confidence interval was calculated with the use of the Clopper–Pearson method.21
Efficacy was assessed in the neonatal-schedule vaccine group from 2 weeks after dose 3 of the vaccine to 12 months and to 18 months and in the infant-schedule vaccine group from 2 weeks after dose 4 of the vaccine to 12 months and to 18 months. The assessment schedule resulted in two different presentations of data in the placebo group (denoted as neonatal-schedule placebo group and infant-schedule placebo group). In the analysis of vaccine response, data were considered to be missing for a given participant only if data on all the components of the outcome were missing for that participant. The Kaplan–Meier method was used to estimate the cumulative risk of a first episode of severe rotavirus gastroenteritis from the time of randomization, and the trial groups were compared with the use of a log-rank test. All statistical tests were two-sided.
On the basis of local surveillance data, we assumed that 3% of the participants in the placebo group would have an episode of severe rotavirus gastroenteritis during the trial,22,23 and we calculated that an enrollment target of 549 participants in each of the three trial groups would provide the trial with 80% power to reject the null hypothesis of no difference between the combined vaccine group and the placebo group if the true efficacy of the vaccine was 60%, at a one-sided alpha level of 0.1. The estimated sample size would allow for a rate of nonadherence to the trial regimen of 10%. We calculated that a minimum of 282 participants would be required to reject the null hypothesis of no difference in the percentage of participants with vaccine response, at a two-sided alpha level of 0.05, assuming that 25% of participants in the placebo group would be exposed to rotavirus14 and that 50% of the participants in each of the two vaccine groups would have vaccine response, and allowing for a rate of nonadherence of 10%.