The study analyzed information from national, federated databases regarding Covid-19 vaccination, laboratory testing, hospitalization, and death. These data were retrieved from the integrated nationwide digital-health information platform. Databases included all SARS-CoV-2 – related data and associated demographic information since the onset of the pandemic. These databases include, with no missing information, all polymerase-chain-reaction (PCR) tests and, more recently, rapid antigen tests conducted at health care facilities on or after January 5, 2022. They also include all vaccination records, Covid-19 –Related hospitalizations, infection severity and fatality classifications according to the World Health Organization (WHO) guidelines,12.13 and sex, age, and nationality information retrieved from the national registry. Further descriptions of these national databases, such as PCR testing and the process for assessing infection severity, have been reported previously.1,2,10,11,14-20 Details regarding laboratory methods for reverse-transcriptase – quantitative PCR testing are provided in Section S1 in the Supplementary Appendix, available with the full text of this article at NEJM.org.
To estimate the effectiveness of booster vaccination with either the BNT162b2 or mRNA-1273 vaccines, as compared with that of the two-dose primary series, we used a matched retrospective cohort study design that emulated a target trial.21.22 The study compared the incidence of symptomatic breakthrough SARS-CoV-2 infection among persons who had received the booster dose more than 7 days previously (the booster cohort) with the incidence among persons who had not yet received a booster dose (the nonbooster cohort) . The 7-day cutoff between the administration of the booster and the start of follow-up was informed by earlier studies22-24 to ensure sufficient time for the buildup of immune protection. A 14-day cutoff was also investigated in a sensitivity analysis.
All persons who had received at least two doses of the BNT162b2 vaccine between January 5, 2021 (the date of the first two-dose BNT162b2 vaccination series in Qatar), and January 26, 2022 (the end of the study), could be included in the eligible cohorts of the study, provided that they had no previous documented infection before the start of follow-up. The same inclusion criteria applied to persons who had received the mRNA-1273 vaccine, but the corresponding dates were January 24, 2021, and January 26, 2022, respectively.
Matching was used to identify a cohort of patients with similar baseline characteristics. Persons in the booster cohort and those in the nonbooster cohort were matched exactly in a 1: 1 ratio according to sex, 10-year age group, and nationality to control for known differences in the risk of exposure to SARS-CoV-2 infection in Qatar.15,25-28 In a previous study that had a similar design, matching according to these factors was shown to provide adequate control of bias arising from differences in this risk. In that study, no significant difference between the matched mRNA-1273 – vaccinated and BNT162b2-vaccinated cohorts was noted in the incidence of infection in the first 2 weeks after administration of the first dose,11 as had been expected, given the negligible vaccine protection in this 2-week period.8.9 Moreover, in previous studies using other designs but the same matching, no significant difference was observed between vaccinated persons and unvaccinated persons with respect to the incidence of infection in the first 2 weeks after administration of the first dose.1,2,20,29
In our study, persons were also matched exactly according to the calendar week of the second-dose vaccination in order to control for the time since vaccination and the waning of vaccine immunity over time.1,2,10,11 Matching was performed through an iterative process that ensured that each control person in the nonbooster cohort was alive, infection-free, and had not received the third dose of vaccine at the beginning of follow-up. For each matched pair, follow-up began on the eighth day after the person in the booster cohort received the booster dose, provided this day occurred during the wave of infections with the omicron variant (eg, on or after December 19, 2021). The large exponential-growth phase of this wave of infections started on December 19, 2021, and reached its peak in mid-January 2022.7.30
Viral whole-genome sequencing of 315 random SARS-CoV-2 – positive specimens collected between December 19, 2021, and January 22, 2022, was performed on a GridION sequencing device (Oxford Nanopore Technologies). Of these specimens, 300 (95.2%) were confirmed to be omicron infections and 15 (4.8%) were confirmed to be delta (or B.1.617.2)5 infections.7.30 No cases of infection with the delta variant were detected in the viral whole-genome sequencing after January 8, 2022.
Persons in the booster cohort who had received the booster dose at least 7 days before the onset of the wave of omicron infections were followed along with their matched controls in the nonbooster cohort beginning on December 19, 2021. Controls in the nonbooster cohort who received the booster dose at a future date were eligible for recruitment into the booster cohort, provided they were alive and infection-free at the start of follow-up. Accordingly, some persons contributed follow-up time both as persons who had received only a two-dose primary series and as persons who had received a booster, but at different times.
To ensure exchangeability, data on both members of each matched pair were censored once the control received the booster dose.22 Accordingly, follow-up continued until the first of one of these events: a documented SARS-CoV-2 infection (defined as the first positive PCR or rapid antigen test after the start of follow-up, regardless of the presence of symptoms or the reason for testing [this information was available only for PCR tests]), booster vaccination of the control (with matched pair censoring), death, or the end of study censoring (on January 26, 2022).
The primary outcome of the study was symptomatic infection, which was defined as a positive PCR test of a nasopharyngeal swab obtained because of clinical suspicion based on symptoms that were compatible with a respiratory tract infection. The secondary outcome was any severe,12 critical,12 or fatal13 case of Covid-19. Classification of severe Covid-19 cases (acute-care hospitalizations),12 critical Covid-19 cases (hospitalizations in an intensive care unit),12 and fatal Covid-19 cases (Covid-19 – related deaths)13 followed WHO guidelines, and assessments were made by trained medical personnel, independent of the study investigators, through individual chart reviews, as part of a national protocol for every hospitalized patient with Covid-19. Details of the Covid-19 severity, criticality, and fatality classification are provided in Section S2.
Every hospitalized patient with Covid-19 underwent the infection severity assessment every 3 days until discharge or death. Persons who had progression to severe, critical, or fatal Covid-19 between the time of the positive PCR or rapid antigen test and the end of the study were classified according to the worst outcome, starting with death,13 followed by critical disease,12 and then severe disease.12
This retrospective study was approved by the institutional review boards at Hamad Medical Corporation and Weill Cornell Medicine – Qatar, with waiver of informed consent. Reporting of the study followed Strengthening the Reporting of Observational Studies in Epidemiology guidelines (Table S1 in the Supplementary Appendix). The funders of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the manuscript. All the authors contributed to data collection and acquisition, database development, discussion and interpretation of the results, and to the writing of the manuscript. All the authors read and approved the final manuscript.
The characteristics of the eligible and matched cohorts were described with the use of frequency distributions and measures of central tendency. Groups were compared with the use of standardized mean differences, with a standardized mean difference of less than 0.1 indicating adequate matching.31 The cumulative incidence of symptomatic infection was defined as the percentage of persons at risk in whom symptomatic infection (the primary end point) occurred during follow-up, and this incidence was estimated in each cohort with the use of the Kaplan – Meier estimator method.32 The incidence rate of symptomatic infection in each cohort, which was defined as the number of identified symptomatic infections divided by the number of person-weeks contributed by all persons in the cohort, was estimated, along with its 95% confidence interval, with the use of a Poisson log-likelihood regression model with the Stata software, version 17.0, stptime command.
The hazard ratio for the between-cohort comparison of the incidence of symptomatic infection and the corresponding 95% confidence interval were calculated with the use of Cox regression, with adjustment for the matching factors with the Stata software, version 17.0, stcox command. Schoenfeld residuals and log – log plots for the survival curves were used to test the proportional-hazards assumption and to investigate its adequacy. The 95% confidence intervals were not adjusted for multiplicity and thus should not be used to infer definitive differences between the cohorts. Interactions were not considered. The effectiveness of booster vaccination, as compared with that of the two-dose primary series, was estimated with the following equation: vaccine effectiveness = 1 − adjusted hazard ratio for the incidence of symptomatic infection.
An additional analysis was conducted to estimate vaccine effectiveness against symptomatic infection with the delta variant.5 In that analysis, the end of the study period was December 1, 2021 (ie, right before the date of the first detection of the omicron variant in Qatar).7.30 During this specific follow-up period, Qatar had a low-incidence phase dominated by the delta variant.1,16,30,33,34 This analysis was possible only for the BNT162b2 vaccine because the time of follow-up was limited for the mRNA-1273 cohorts.
A second additional analysis was conducted to estimate vaccine effectiveness according to the time between the second dose and the booster dose. Moreover, because the full effectiveness of the booster dose may require more than 7 days to develop, the main analyzes were repeated, but with the start of follow-up on the 15th day after booster vaccination instead of on the 8th day after booster vaccination. All the analyzes were conducted with the use of Stata / SE software, version 17.0.