Vaccine efficacy against severe COVID-19 in relation to delta variant (B.1.617.2) and time since second dose in patients in Scotland (REACT-SCOT): a case-control study

Abstract

Background: Reports have suggested that the efficacy of vaccines against COVID-19 might have fallen since the delta (B.1.617.2) SARS-CoV-2 variant replaced the alpha (B.1.1.7) variant as the predominant variant. We aimed to investigate, for the two main classes of vaccine, whether efficacy against severe COVID-19 has decreased since delta became the predominant variant and whether the efficacy of two doses of vaccine against severe COVID-19 wanes with time since second dose.

Methods: In the REACT-SCOT case-control study, vaccine efficacy was estimated using a matched case-control design that includes all diagnosed cases of COVID-19 in Scotland up to Sept 8, 2021. For every incident case of COVID-19 in the Scottish population, ten controls matched for age rounded to the nearest year, sex, and primary care practice, and alive on the day of presentation of the case that they were matched to were selected using the Community Health Index database. To minimise ascertainment bias we prespecified the primary outcome measure to assess vaccine efficacy as severe COVID-19, defined as diagnosed patients with entry to critical care within 21 days of first positive test, death within 28 days of first positive test, or any death for which COVID-19 was coded as underlying cause. Although the data extracted for this study included cases presenting up to Sept 22, 2021, the analyses reported here are restricted to cases and controls presenting from Dec 1, 2020, to Sept 8, 2021, ensuring follow-up for at least 14 days after presentation date to allow classification of hospitalisation and (for most cases) severity based on entry to critical care or fatal outcome.

Findings: During the study period, a total of 5645 severe cases of COVID-19 were recorded; these were matched to 50 096 controls. Of the severe cases, 4495 (80%) were not vaccinated, and of the controls, 36 879 (74%) were not vaccinated. Of the severe cases of COVID-19 who had been vaccinated, 389 had received an mRNA vaccine and 759 had received the ChAdOx1 vaccine. The efficacy of vaccination against severe COVID-19 decreased in May, 2021, coinciding with the replacement of the alpha SARS-CoV-2 variant by the delta variant in Scotland, but this decrease was reversed over the following month. In the most recent time window centred on July 29, 2021, the efficacy of two doses was 91% (95% CI 87-94) for the ChAdOx1 vaccine and 92% (88-95) for mRNA (Pfizer or Moderna) vaccines. The efficacy of the ChAdOx1 vaccine against severe COVID-19 declined with time since second dose to 69% (95% CI 52-80) at 20 weeks from second dose. The efficacy of mRNA vaccines declined in the first ten weeks from second dose but more slowly thereafter to 93% (88-96) at 20 weeks from second dose.

Interpretation: Our results are reassuring with respect to concerns that vaccine efficacy against severe COVID-19 might have fallen since the delta variant became predominant, or that efficacy of mRNA vaccines wanes within the first 5-6 months after second dose. However, the efficacy of the ChAdOx1 vaccine against severe COVID-19 wanes substantially by 20 weeks from second dose. Efficacy of mRNA vaccines after 20 weeks and against newer variants remains to be established. Our findings support the case for additional protective measures for those at risk of severe disease, including, but not limited to, booster doses, at times when transmission rates are high or expected to rise.

Funding: None.

Figure 1

Association between vaccine efficacy and calendar time (A) Severe COVID-19 (single-dose categories for ChAdOx1 and mRNA vaccines have been combined as the numbers of events in those with one dose of mRNA vaccine were small). (B) Hospitalised or fatal COVID-19 cases. RRs in conditional logistic regression model, adjusted for covariates. The efficacy of vaccination was calculated as 1 minus the RR. For each effect, line thickness is proportional to precision (inverse variance) of estimate, scaled to the same maximum thickness for each effect. RR=rate ratio.

Figure 2

Association between vaccine efficacy and time since last vaccine dose (A) Severe COVID-19. RRs in conditional logistic regression model, adjusted for covariates. (B) Hospitalised or fatal COVID-19 cases. RRs in the 42-day time window centred on 20 weeks from the most recent vaccine dose are presented. The efficacy of vaccination is 1 minus the RR. For each effect, line thickness is proportional to precision (inverse variance) of estimate, scaled to the same maximum thickness for each effect. RR=rate ratio.