Natural infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) elicits strong protection against reinfection with the B.1.1.7 (alpha),1,2 B.1.351 (beta),1 and B.1.617.2 (delta)3 variants. However, the B.1.1.529 (omicron) variant harbors multiple mutations that can mediate immune evasion. We estimated the effectiveness of previous infection in preventing symptomatic new cases caused by omicron and other SARS-CoV-2 variants in Qatar. In this study, we extracted data regarding coronavirus disease 2019 (Covid-19) laboratory testing, vaccination, clinical infection data, and related demographic details from the national SARS-CoV-2 databases, which include all results of polymerase-chain-reaction (PCR) testing, vaccinations, and hospitalizations and deaths for Covid-19 in Qatar since the start of the pandemic.
The effectiveness of previous SARS-CoV-2 infection in preventing reinfection was defined as the proportional reduction in susceptibility to infection among persons who had recovered from infection as compared with those who had not been infected.4 Previous SARS-CoV-2 infection was defined as a positive result on PCR assay at least 90 days before a new positive PCR finding.4 We used a test-negative, case–control study design to assess the effectiveness of previous infection in preventing reinfection on the basis of a method that had recently been investigated and validated for derivation of robust estimates for such comparisons4 (Section S1 of the Supplementary Appendix, available with the full text of this letter at NEJM.org). In addition, we performed sensitivity analyses that included adjustment for vaccination status and that excluded vaccinated persons from the analysis. Case patients (defined as persons with positive PCR results) and controls (defined as persons with negative PCR results) were matched according to sex, 10-year age group, nationality, and calendar time of PCR testing to control for known differences in the risk of exposure to SARS-CoV-2 infection in Qatar.4
To ensure that epidemiologically relevant reinfections were considered in the analysis, only documented infections with a PCR cycle threshold (Ct) value of 30 or less were included as cases in our study. (Reinfection often occurs with negligible symptoms and high Ct values, indicating reduced epidemiologic significance.)5 We also estimated the effectiveness of previous infection in preventing hospitalization or death caused by reinfection.
The selection of the study population for various analyses is shown in Figures S1 through S4 and the population characteristics in Tables S1 and S2. The overall study population was broadly representative of the total population of Qatar (Table S3), with a median age of 31 to 35 years across the study samples. The median interval between previous infection and PCR testing among cases and controls was 279 days (interquartile range [IQR], 194 to 313) for analysis of the alpha variant, 285 days (IQR, 213 to 314) for analysis of the beta variant, 254 days (IQR, 159 to 376) for analysis of the delta variant, and 314 days (IQR, 268 to 487) for analysis of the omicron variant.
The effectiveness of previous infection in preventing reinfection was estimated to be 90.2% (95% confidence interval [CI], 60.2 to 97.6) against the alpha variant, 85.7% (95% CI, 75.8 to 91.7) against the beta variant, 92.0% (95% CI, 87.9 to 94.7) against the delta variant, and 56.0% (95% CI, 50.6 to 60.9) against the omicron variant (Table 1). Sensitivity analyses confirmed the study results, as expected for this study design, which is robust regardless of the approach that is used to control for vaccine-induced immunity.4 An additional analysis that was adjusted for the interval since previous infection also confirmed the study results (Table S4).
Among the patients with reinfection, progression to severe Covid-19 occurred in one patient with the alpha variant, in two patients with the beta variant, in no patients with the delta variant, and in two patients with the omicron variant. None of the reinfections progressed to critical or fatal Covid-19. The effectiveness with respect to severe, critical, or fatal Covid-19 was estimated to be 69.4% (95% CI, −143.6 to 96.2) against the alpha variant, 88.0% (95% CI, 50.7 to 97.1) against the beta variant, 100% (95% CI, 43.3 to 100) against the delta variant, and 87.8% (95% CI, 47.5 to 97.1) against the omicron variant. (For the delta variant, the calculation of the 95% confidence interval is clarified in a footnote in Table 1.) Limitations of the estimations (e.g., the relatively young population of Qatar) are discussed in Section S1.
Overall, in a national database study in Qatar, we found that the effectiveness of previous infection in preventing reinfection with the alpha, beta, and delta variants of SARS-CoV-2 was robust (at approximately 90%), findings that confirmed earlier estimates.1-3 Such protection against reinfection with the omicron variant was lower (approximately 60%) but still considerable. In addition, the protection of previous infection against hospitalization or death caused by reinfection appeared to be robust, regardless of variant.