TRIAL OBJECTIVES, PARTICIPANTS, AND OVERSIGHT
All the trial participants were a least 16 years of age and had previously received two doses of the BNT162b2 vaccine administered 19 to 42 days apart in the ongoing pivotal trial. Participants had received the second vaccine dose at least 175 days (approximately 6 months) earlier and had received no previous clinical or microbiologic diagnosis of Covid-19. Additional inclusion and exclusion criteria are provided in the Supplementary Appendix, available with the full text of this article at NEJM.org, along with the protocol. All the participants provided written informed consent; in adolescents younger than 18 years of age, written informed consent was provided by a parent or guardian.
Pfizer was responsible for the design and conduct of the trial; for the collection, analysis, and interpretation of the data; and for the writing of the manuscript. Both Pfizer and BioNTech manufactured the vaccine and placebo that were used in the trial. BioNTech was the regulatory sponsor and contributed to interpretation of the data and the writing of the manuscript. All aggregated analyses of the data were available to all the authors, who vouch for the accuracy and completeness of the data and for the adherence of the trial to the protocol.
Using an interactive Web-based response system, we randomly assigned the participants in a 1:1 ratio to receive an intramuscular injection of a third dose of the BNT162b2 vaccine or saline placebo. All the participants and site personnel were unaware of the trial-group assignments, except for staff members who prepared, dispensed, or administered the injections. The trial design originally called for all the participants to remain unaware of their assigned group until after a prespecified interim analysis. This planned analysis was to occur 2 months after the third dose, with subsequent review by the data and safety monitoring committee. In anticipation of recommendations by regulatory authorities that booster doses be made available, we revised the protocol to allow for the unblinding of the trial-group assignments before the 2-month cutoff. During the interim analysis, the data monitoring committee also recommended that participants be made aware of their trial-group assignments to allow those in the placebo group to receive a third vaccine dose. This recommendation was made in consideration of the observation of waning effectiveness after two doses, the observed benefit of the third dose, and the recommendation for boosters in some countries. On September 24, 2021, on the basis of FDA issuance of the EUA for a BNT162b2 booster and at the discretion of the sponsor, participants in the placebo group were offered the opportunity to receive the BNT162b2 vaccine.
The incidence of adverse events and serious adverse events was the primary safety end point. Data regarding unsolicited adverse events, including reactogenicity, were collected from the time that informed consent was provided through 1 month after the administration of dose 3. We collected assessments of reactogenicity and other adverse events by telephone follow-up. The collection of reports of serious adverse events was ongoing from the time that informed consent was provided through 6 months after dose 3.
The primary efficacy end point was the effectiveness of the BNT162b2 vaccine against laboratory-confirmed Covid-19 beginning at least 7 days after the administration of dose 3. This end point was analyzed both in participants who had no evidence of previous SARS-CoV-2 infection and in all participants. The efficacy of the BNT162b2 vaccine against severe Covid-19 was determined according to the definitions used by the FDA and the Centers for Disease Control and Prevention (CDC).18,19 The methods that were used for identifying SARS-CoV-2 infection and Covid-19 are summarized in the Supplementary Appendix.
The sample size was chosen to provide a relatively large database to assess safety, with the intent of accruing sufficient Covid-19 cases to assess vaccine efficacy. The number of participants was not based on a projected between-group difference. We determined that the enrollment of 10,000 participants would provide 4250 participants with data that could be evaluated in each group, assuming a loss of follow-up of 15%. Included in the primary analyses were all the participants who had undergone randomization and received either a third dose of the BNT162b2 vaccine or placebo with no important deviations from the protocol.
We calculated safety end points as counts, percentages, and associated Clopper–Pearson two-sided 95% confidence intervals in the safety population (Table S1 in the Supplementary Appendix). Adverse events and serious adverse events are presented according to the terms used in the Medical Dictionary for Regulatory Activities, version 24.0. Between-group differences in percentages of participants who reported having adverse events are presented with associated two-sided Miettinen–Nurminen 95% confidence intervals for adverse events with an incidence of at least 1% in either group.
Interim efficacy analyses were performed after all the participants had reached 2 months of blinded follow-up and through the data-cutoff date; these data were reviewed by the data monitoring committee. Data from participants who were made aware of trial-group assignments before the first interim analysis because of the regulatory decision were censored at the time of the unblinding of the trial group. Details regarding the procedures that were used to determine the relative efficacy of the third dose of vaccine as compared with placebo are provided in the Supplementary Appendix.
We calculated two-sided Clopper–Pearson 95% confidence intervals for relative vaccine efficacy after adjustment for surveillance time. Confidence intervals were not adjusted for multiplicity and thus cannot be used to infer effects. Instead, the trial was designed to provide information to assist policymakers in deciding whether to recommend a third dose of vaccine, as well as the timing of this dose if recommended. Missing efficacy data (e.g., the report of a symptom without laboratory testing data) were not imputed.