The trial protocol11 (available with the full text of this article at NEJM.org) was approved by the ethics committees of the London School of Hygiene and Tropical Medicine; the Ministry of Health of Burkina Faso; the University of Sciences, Techniques, and Technologies of Bamako; and the national regulatory authorities of Burkina Faso and Mali. A data and safety monitoring board reviewed serious adverse events, approved the statistical analysis plan, and archived the locked databases before unblinding. A steering committee provided scientific advice and monitored the progress of the trial. The trial was conducted in accordance with the International Council for Harmonisation Good Clinical Practice guidelines and all applicable local regulations. The authors vouch for the accuracy and completeness of the data and for the adherence of the trial to the protocol. GlaxoSmithKline (GSK) Biologicals donated the RTS,S/AS01E and Havrix vaccines. Dispersible sulfadoxine–pyrimethamine and amodiaquine and matching placebos were donated by Guilin Pharmaceutical.

The trial was conducted in Bougouni district and neighboring areas in Mali and in Houndé district in Burkina Faso.12 Information regarding the trial sites is provided in the Supplementary Methods section and Figure S1 in the Supplementary Appendix, available at NEJM.org.

All households with children who would be 5 to 17 months of age on April 1, 2017, within the trial areas were enumerated from February through March 2017. Inclusion and exclusion criteria are listed in the Supplementary Appendix. After written informed consent had been obtained from parents or guardians, an independent statistician randomly assigned eligible children to receive chemoprevention (chemoprevention-alone group), the RTS,S/AS01E vaccine (vaccine-alone group), or chemoprevention plus RTS,S/AS01E (combination group). The randomization list used permuted blocks after sorting according to age, sex, area of residence, and previous receipt of chemoprevention. Tablet computers with the randomization list were accessible only to the chief pharmacists. All other investigators and trial staff were unaware of treatment assignments until the locked database for analysis had been archived with the data and safety monitoring board in June 2020. All participating children were given an identity card containing their photograph and a quick response (QR) code that included the child’s trial identification number, name, and date of birth. At the time of vaccination or administration of chemoprevention, these cards were scanned to ensure that the correct intervention was administered.

All the participating children were given a long-lasting insecticide-treated bed net at the time of enrollment. Children in the vaccine-alone group and the combination group received three doses of RTS,S/AS01E in April, May, and June 2017, followed by a fourth and fifth dose in June 2018 and June 2019 (Fig. S2). Syringes containing vaccines were prepared by a chief pharmacist and masked with tape to conceal the contents from the administrator, caretakers, and children. The pharmacist and the vaccine administrators had no further role in the trial.

Children in the chemoprevention-alone group and the combination group received four courses of sulfadoxine–pyrimethamine and amodiaquine at monthly intervals each year; children in the vaccine-alone group received four courses of sulfadoxine–pyrimethamine and amodiaquine placebos on that same schedule. Children 12 months of age or older in the chemoprevention-alone group and the combination group received 500 mg of sulfadoxine, 25 mg of pyrimethamine, and 150 mg of amodiaquine on day 1, and an additional 150-mg dose of amodiaquine on days 2 and 3; infants received 250 mg of sulfadoxine, 12.5 mg of pyrimethamine, and 75 mg of amodiaquine on day 1 and 75 mg of amodiaquine on days 2 and 3. The trial drugs were prepared by a pharmacist, who had no further role in the trial, and were placed in resealable envelopes labeled with the QR code. Administration of each dose of sulfadoxine–pyrimethamine and amodiaquine or placebo was directly observed by trial staff at distribution points in trial villages. Children in the chemoprevention-alone group also received three doses of inactivated rabies vaccine (Rabipur)13 in 2017 and a dose of hepatitis A vaccine (Havrix)14 in 2018 and 2019.

The primary outcome was uncomplicated clinical malaria, defined as a measured temperature of at least 37.5°C or a history of fever within the previous 48 hours and P. falciparum parasitemia (parasite density ≥5000 per cubic millimeter) in children who presented to a trial health facility. Prespecified secondary outcomes were hospital admission with malaria, death from malaria, and malaria parasitemia or anemia at the end of the malaria transmission season (see the Supplementary Methods section of the Supplementary Appendix).

Trial staff based at trial health facilities tested children with suspected malaria with the use of a rapid diagnostic test. Children who were positive were treated with artemether–lumefantrine, and a blood film was obtained for subsequent microscopic examination. Blood films were read by two independent microscopists according to a standardized algorithm.15 Discrepant readings were resolved by a third reader. The quality of the blood film readings in each country was confirmed by an external reference laboratory (see the Supplementary Methods section in the Supplementary Appendix and Table S1 and Fig. S3).

Each week, 24 randomly selected children in each country were visited at home (8 children per trial group), and a blood film was obtained. Children were also evaluated during a cross-sectional survey conducted 1 month after the last course of chemoprevention at the end of each malaria transmission season to measure hemoglobin level and to obtain a blood film. At the end of the 2018 and the 2019 transmission seasons, 200 randomly selected school-age children who were 6 to 12 years of age (and therefore too old to receive chemoprevention), resided in the trial areas, and were in good health were tested for malaria by means of microscopic examination. If a child was identified as having clinical malaria at a home visit or in a cross-sectional survey, the child was treated with artemether–lumefantrine.

To determine the curative efficacy of the chemoprevention regimen, further informed consent was obtained, and children with asymptomatic malaria parasitemia at the time of the final cross-sectional survey were treated with the same doses of sulfadoxine–pyrimethamine and amodiaquine as those used for the chemoprevention intervention. Blood films were obtained for microscopic analysis on days 1, 2, 4, 7, 14, and 28 after treatment.

Serious adverse events were reported within 72 hours after identification. Deaths that occurred outside a health care facility were assessed by means of verbal autopsy.16 Assignment of the causes of hospital admissions or deaths that occurred inside or outside the hospital was performed by two physicians who were unaware of the trial-group assignments. A third independent physician reviewed cases for which there was a disagreement, and a consensus was reached.

The rationale for the trial’s sample size is described in the statistical analysis plan, available with the protocol. For the noninferiority comparison, we determined that 2000 children per group would provide 80% power to exclude, at the 2.5% significance level, a difference in the hazard ratio for clinical malaria between the vaccine-alone group and the chemoprevention-alone group of 20% (favoring vaccine alone) over the 3-year trial period. For the superiority comparisons, assuming that the difference in the hazard ratio between the combination group and the vaccine-alone group or the chemoprevention-alone group would be 30% (favoring the combination), we calculated that this sample size would provide close to 100% power to exclude a minimum difference in the hazard ratios of 0% and would give the trial 90% power to exclude a minimum difference in the hazard ratios of 15%.

The primary analysis was performed in the modified intention-to-treat population, which included all eligible children whose parents or guardians provided consent and who received a first dose of trial vaccine or placebo in April 2017. The per-protocol population for each trial year included all children who received all doses of the vaccine and attended all four chemoprevention visits in that year. Secondary outcomes were assessed only in the modified intention-to-treat population. Person-time at risk was calculated from the date of first vaccination until the date of death, the date of permanent emigration, the date consent was withdrawn, the date last seen for children lost to follow-up or who temporarily traveled out of the trial area, or the end of the trial (March 31, 2020).

The hazard ratio for the primary outcome was estimated with the use of Cox regression models, adjusted for trial center, with a robust standard error to account for potential clustering of recurrent episodes of malaria. Protective efficacy (the percent difference in the total number of events over the trial period) was estimated as (1−hazard ratio)×100. Effect modification according to trial center and year, prespecified in the statistical analysis plan, was assessed with the use of the Wald test for the interaction term without adjustment for multiple comparisons. Two-sided 90%, 95%, and 99% confidence intervals for the hazard ratio for the comparison of RTS,S/AS01E alone with chemoprevention alone were calculated and compared with the prespecified noninferiority margin of 1.20. To preserve the type I error rate at 5%, a closed testing procedure was used: the Wald test of the null hypothesis of equal hazard ratios comparing all three groups was performed. If the null hypothesis was rejected at the 5% significance level, pairwise comparisons were performed, also with a 5% significance level. Incidence rate differences and prevalence ratios were calculated with the use of published methods.17,18 An analysis was conducted to explore patterns of missingness in the outcome data and to assess sensitivity to missing outcome data (Table S8). Full details of the conduct of the trial are provided in the protocol.