Design of a phase III multicenter trial to evaluate the efficacy of the RTS,S/AS01 malaria vaccine in children across diverse transmission settings in Africa

This is a phase III, randomized, controlled, multicentre, participant- and observer-blind study. Enrolment occurred between May 2009 and February 2011. Follow up is currently on-going at 11 centres covering a wide range of transmission settings in seven countries in sub-Saharan Africa (Figure 1). The study is conducted in accordance with the current Declaration of Helsinki, International Committee on Harmonization Good Clinical Practice guidelines and local rules and regulations of each country. The study is overseen by an Independent Data Monitoring Committee (IDMC), assisted by a Local Safety Monitor at each centre. Approval was obtained from 56 institutional review boards and national Regulatory Authorities. Prior to study inclusion, parents or guardians of all participants provided signed/finger-printed and witnessed informed consent.

An overview of the study design is shown in Figure 2. Children were enrolled in two age categories: 6-12 weeks old and 5-17 months old. A minimum of 6,000 children in both age categories (to a total maximum of 16,000 children) have been enrolled and randomized 1:1:1 to one of three study groups for primary and booster vaccination (Table 1). The control vaccine given depends upon the age of the child at enrolment (Table 1). Children in the younger age category receive their primary vaccination course at 6, 10 and 14 weeks of age, in co-administration with the other vaccines usually administered at these EPI visits (Table 1). The specified age range of 6-12 weeks at first vaccination allows for some flexibility in this schedule to align with local guidelines and practice. Bacillus-Calmette-Guérin (BCG) vaccine, neonatal dose of oral polio vaccine (OPV), measles vaccine and yellow fever vaccine are given according to local policy.

Primary immunizations are administered by intramuscular injection into the antero-lateral thigh (children aged 6-12 weeks) or the left deltoid (children aged 5-17 months); all children receive the booster injection in the left deltoid. Neither the study subjects and their parents/guardians nor the study personnel involved in evaluation of the study endpoints are aware of the group allocation of the subjects. Because the study vaccines differ in appearance, the study staff responsible for their preparation and administration is aware of treatment allocation and therefore perform no other role in the trial.

There is no routine testing for HIV infection in this study, HIV tests are performed only if clinically indicated. Voluntary counselling and testing, Highly Active Antiretroviral Therapy (HAART) and Prevention of Mother to Child Transmission (PMCT) are available at all study centres according to national policies.

In accordance with national policies, all centres use artemisinin-based combination therapy (ACT) as first line treatment for malaria cases. The use of insecticide-treated bed nets is optimized at all centres by the close collaboration between research staff and malaria control program managers or through distribution at screening. Other control interventions such as intermittent preventive treatment in infants (IPTi) and indoor residual spraying (IRS) are not currently part of policy in the study areas, but if this changes during the trial, their use will be recorded.

The overall sample size has taken into account the recent description of falling rates of malaria disease in several parts of Africa. To be assured of meeting the primary endpoints, conservative rates of disease were applied in the calculation of sample size and a case driven approach selected [3]. To control for the co-primary endpoint in each of the two age categories, evaluations will be performed at a 2.5% alpha level (Bonferroni correction). Assuming an attack rate in controls of 10/100 children years at risk (cyr), a 12 months follow up period, a true vaccine efficacy of 30% and a drop-out rate of 10% then the sample size of 6000 children in each age category has 90% power to detect a lower limit of the 97.5% CI around estimated VE above 0%. In the event that the attack rate is lower than anticipated, the analysis will be postponed until 450 cases have accumulated. Due to the uncertainty around the rate of severe malaria disease according to the case definitions used in the trial, the total sample size is up to 16000 children and the analysis will be conducted for both age categories pooled when 250 episodes have accumulated. This gives 80% power to detect 30% VE with a lower limit of the 95% CI above 0% or assuming 50% VE 90% power to detect a lower limit of the 95% CI above 25%.

Inclusion and exclusion criteria are shown in Table 2. The aim was to enrol a broad sample of children representative of the general population. Exclusion criteria have been kept to a minimum to mirror the general population as far as possible whilst minimizing participant safety risk exposure. Children with a history of simple febrile seizure, malnourished children not requiring hospitalization and HIV-infected children (other than those with HIV disease stage 3 or 4 severity as defined by the WHO 2005) were not excluded from study participation.

The choice of comparator vaccines was guided by the need to offer potential benefit to the control group without compromising the evaluation of study endpoints. The pros and cons of a number of options were debated and consensus reached by the CTPC after taking into account regional epidemiology and EPI programs. Rabies vaccine was chosen as the control for the 5-17 month age category because of the high burden of rabies across all of sub-Saharan Africa, its high fatality rate and the particular risk to children [17‐19]. Rabies vaccine has been evaluated according to several different vaccination schedules, and the 0, 1, 2-month schedule used in this trial is expected to produce acceptable antibody titers and provide protection. Rabies vaccine was not appropriate for children in the 6-12 week age category because co-administration with EPI antigens has not been evaluated. For the 6-12 week age group, consideration was given to vaccines against S. pneumoniae, which is a common cause of pneumonia in children in Africa. The reasons for not selecting pneumococcal vaccines were that they were expected to be implemented as policy in some countries prior to the enrolment and so would not provide additional benefit. In addition, there is a poorly understood interaction between malaria and pneumococcal infections. In paediatric hospital admissions, pneumonia and malaria co-occur more often than expected by chance [20]. This may be due to the overlapping clinical symptoms and signs of pneumonia and malaria, or the immunosuppressive effect of malaria infection on pneumococcal pneumonia [21]. The careful characterization of both malaria and pneumonia in this trial will also allow the study of the effect of malaria control on the incidence of pneumococcal disease [22]. Meningococcal disease in Africa is most commonly due to serogroup A, however there currently is no meningococcal A vaccine licensed for use in infants. Meningococcal C conjugate vaccine was chosen because it is acceptably safe and immunogenic when administered according to a 0, 1, 2-month schedule, can be safely co-administered with the other vaccines and will not compromise the analysis of the study endpoints. Although meningitis C is not common in sub-Saharan Africa outbreaks have been reported [23‐25] and, therefore, the vaccine may provide some benefit to study subjects.

Clinical malaria cases are detected through passive surveillance at local health facilities. A blood sample for evaluation of malaria parasites is taken from all children with axillary temperature of ≥37.5°C or those reported to have had a fever within 24 hours of presentation. All subjects attending hospital emergency departments in the study areas are evaluated as potential cases of severe malaria following an algorithm, and case assessment is standardized across centers [22]. The algorithm also allows identification of cases of anaemia, sepsis and pneumonia.

Two cross-sectional surveys will be conducted at study months 20 and 32 to assess vaccine efficacy against prevalent parasitaemia and anaemia. Data will be collected on potential covariates which may be included in the analysis of efficacy. These are bed net usage by direct observation, application of IRS, administered doses of IPTi, distance from nearest inpatient health facility, distance from nearest outpatient health facility, pneumococcal/Hib vaccination status, ethnicity, anthropometric measurements and feeding history.

Full quality systems are in place for all laboratory tests in the trial and these are described in a companion paper [26]. Anti-CS and anti-HB antibody titers are measured at all blood sampling time points in a subset of children from both age categories at all sites (See Figure 2 for blood sampling time points). In addition, a nested case control study will evaluate the association between CS-antibody response and protection against malarial disease. In a safety and immunogenicity trial of RTS,S/AS01 co-administered with EPI vaccines, pre-defined non-inferiority criteria compared to control were met for all the DTPwHepB/Hib+OPV, measles and yellow fever antigens, with the exception of polio 3 viruses when RTS,S/AS01 was administered at 0, 1, 2-months [27]. Although a post-hoc analysis showed that differences were explained by pre-vaccination titres, additional data will be collected in this phase III study. Titres will be assessed in a subset of infants in the 6-12 week age category at each site at three time points: study start, one month post primary vaccination, and one month post OPV booster vaccination.

Serious adverse events (SAE) are collected for all subjects for the entire study period. Completeness of SAE reporting is strengthened by monthly visits of field workers to the children's homes. SAEs are defined as AEs resulting in death, which are life-threatening or require hospitalization or prolongation of existing hospitalization or those that result in disability or incapacity. Unsolicited AEs occurring during the 30 days after each vaccine dose and solicited AEs occurring during the seven days after each vaccine dose are collected for the first 200 children enrolled in each age category at each centre. In the remainder of children, only AEs that are considered to be related to vaccination or those resulting in study withdrawal are recorded. Investigators will grade all AEs and SAEs as mild, moderate or severe based on a scale of interference with normal daily activities, and assess the relationship to vaccination.

Seizures occurring within 30 days of vaccination are also required to be reported as SAEs. For seizures occurring within seven days of vaccination, an analysis will be performed based on the Brighton Collaborations guidelines, which captures the features of the seizure and classifies the level of diagnostic certainty [28]. In the first 200 subjects enrolled at each site in the six to 12 weeks age category an analysis of rashes and mucocutaneous diseases within 30 days of vaccination will be performed based on the Brighton Collaboration Guidelines [29]. Due to a theoretical concern that the use of new adjuvanted vaccines may interfere with immunological self-tolerance, regulatory authorities have requested data collection on immune-mediated diseases (IMD). Therefore, all IMDs are reported as SAEs for all subjects over the entire study period. Diagnostic support at a referral laboratory is provided.

Solicited local (injection site) AEs recorded are pain, swelling and redness; grading of symptoms is on a scale of 0-3. Solicited general AEs recorded are drowsiness, fever, irritability/fussiness and loss of appetite; intensity of symptoms (except fever) are graded on a scale of 0-3 based on interference with normal daily activities; fever is defined as axillary temperature ≥37.5°C. Methods have been fully described previously [30].

Verbal autopsies are carried out for all children who die outside a health facility to ascribe the cause of death. The questionnaire used is based on the INDEPTH standard and adapted to be locally appropriate [31]. At study end, all forms will be read by a central panel to attribute cause of death. As a general health indicator, growth is monitored throughout the study according to standardized methods. The length (<2 years of age) and height (≥2 years of age), weight and mid-upper arm circumference are measured at first vaccination and at study months 3, 20 and 32.

Safety and immunogenicity will be described in the special sub-populations of malnourished and HIV-infected children. Weight at enrolment will be used to determine a subset of children who are low weight for age (weight for age z-score ≤-2) and very low weight for age (weight for age z-score ≤-3). HIV infections known at enrolment or diagnosed during the trial are recorded.