Trial Design and Participants
From August 17, 2020, through November 25, 2020, we enrolled participants at 16 sites in South Africa. The trial was designed to provide a preliminary evaluation of vaccine safety and efficacy during ongoing pandemic transmission of SARS-CoV-2. Participants were healthy adults between the ages of 18 and 84 years without human immunodeficiency virus (HIV) infection or a subgroup of adults between the ages of 18 and 64 years with HIV infection whose condition was medically stable. Baseline IgG antibodies against the spike protein (anti-spike IgG antibodies) were measured at study entry to help determine baseline SARS-CoV-2 serostatus for the analysis of vaccine efficacy. As a safety measure, enrollment was staggered into stage 1 (defined by the first third of targeted enrollment) and stage 2 (the remainder of enrollment) for both HIV-negative and HIV-positive participants. Progression from stage 1 to stage 2 in each group required a favorable review of safety data through day 7 from the previous stage against prespecified rules that would trigger a pause in vaccine administration. (Details regarding the participants in each stage are provided in Table S1 in the Supplementary Appendix, available with the full text of this article at NEJM.org.)
Key exclusion criteria were pregnancy, long-term receipt of immunosuppressive therapy, autoimmune or immunodeficiency disease except for medically stable HIV infection, a history of confirmed or suspected Covid-19, and SARS-CoV-2 infection as confirmed on a nucleic acid amplification test (NAAT) performed as part of screening within 5 days before anticipated initial administration of the vaccine or placebo. All the participants provided written informed consent before enrollment. Additional details regarding the trial design, conduct, oversight, and analyses are provided in the Supplementary Appendix and the protocol (which includes the statistical analysis plan), available at NEJM.org.
The NVX-CoV2373 vaccine was developed by Novavax, which sponsored the trial and was responsible for the overall design (with input from the lead investigator), site selection, monitoring, and analysis. Trial investigators were responsible for data collection. The protocol was approved by the South African Health Products Regulatory Authority and by the institutional review board at each trial center. Oversight of safety, which included monitoring for specific vaccination-pause rules, was performed by an independent safety monitoring committee.
The first author wrote the first draft of the manuscript with assistance from a medical writer who is an author and an employee of Novavax. All the authors made the decision to submit the manuscript for publication and vouch for the accuracy and completeness of the data and for the fidelity of the trial to the protocol.
Participants were randomly assigned in a 1:1 ratio to receive two intramuscular injections, 21 days apart, of either NVX-CoV2373 (5 μg of recombinant spike protein with 50 μg of Matrix-M1 adjuvant) or saline placebo (injection volume, 0.5 ml), administered by staff members who were aware of trial-group assignments but were not otherwise involved with other trial procedures or data collection. All other staff members and trial participants remained unaware of trial-group assignments. Participants were scheduled for in-person follow-up visits on days 7, 21, and 35 and at 3 months and 6 months to collect vital signs, review any adverse events, discuss changes in concomitant medications, and obtain blood samples for immunogenicity analyses. A follow-up telephone visit was scheduled for 12 months after vaccination.
The primary safety end points were the occurrence of all unsolicited adverse events, including those that were medically attended, serious, or of special interest, through day 35 (Tables S2 and S3) and solicited local and systemic adverse events that were evaluated by means of a reactogenicity diary for 7 days after each vaccination (Tables S4 and S5). Safety follow-up was ongoing through month 12.
The primary efficacy end point was confirmed symptomatic Covid-19 that was categorized as mild, moderate, or severe (hereafter called symptomatic Covid-19) and that occurred within 7 days after receipt of the second injection (i.e., after day 28) (Table S6). Starting on day 8 and continuing through 12 months, we performed active surveillance (telephone calls every 2 weeks from trial sites to participants) and passive surveillance (telephone contact at any time from participants to trial sites) for symptoms of suspected Covid-19 (Table S7 and Fig. S1). A new onset of suspected symptoms of Covid-19 triggered initial in-person and follow-up surveillance visits to perform clinical assessments (vital signs, including pulse oximetry, and a lung examination) and for collection of nasal swabs (Fig. S2). In addition, suspected Covid-19 symptoms were also assessed and nasal swabs collected at all scheduled trial visits. Nasal-swab samples were tested for the presence of SARS-CoV-2 by NAAT with the use of the BD MAX system (Becton Dickinson). We used the InFLUenza Patient-Reported Outcome (FLU-PRO) questionnaire to comprehensively assess symptoms for the first 10 days of a suspected episode of Covid-19.
In a blinded fashion, we performed post hoc whole-genome sequencing of nasal samples obtained from all the participants who had symptomatic Covid-19. Details regarding the whole-genome sequencing methods and phylogenetic analysis are provided in Fig. S3.
The safety analysis population included all the participants who had received at least one injection of NVX-CoV2373 or placebo; regardless of group assignment, participants were evaluated according to the intervention they had actually received. Safety analyses were presented as numbers and percentages of participants who had solicited local and systemic adverse events through day 7 after each vaccination and who had unsolicited adverse events through day 35.
We performed a per-protocol efficacy analysis in the population of participants who had been seronegative for SARS-CoV-2 at baseline and who had received both injections of NVX-CoV2373 or placebo as assigned, had no evidence of SARS-CoV-2 infection (by NAAT or anti-spike IgG analysis) within 7 days after the second injection (i.e., before day 28), and had no major protocol deviations affecting the primary efficacy outcome. A second per-protocol efficacy analysis population was defined in a similar fashion except that participants who were seropositive for SARS-CoV-2 at baseline could be included.
Vaccine efficacy (calculated as a percentage) was defined as (1–RR)×100, where RR is the relative risk of Covid-19 illness in the vaccine group as compared with the placebo group. The official, event-driven efficacy analysis targeted a minimum number of 23 end points (range, 23 to 50) to provide approximately 90% power to detect vaccine efficacy of 80% on the basis of an incidence of symptomatic Covid-19 of 2 to 6% in the placebo group. This analysis was performed at an overall one-sided type I error rate of 0.025 for the single primary efficacy end point. The relative risk and its confidence interval were estimated with the use of Poisson regression with robust error variance. Hypothesis testing of the primary efficacy end point was performed against the null hypothesis of vaccine efficacy of 0%. The success criterion required rejection of the null hypothesis to show a statistically significant vaccine efficacy.