Effectiveness of ChAdOx1 nCoV-19 vaccine during the delta (B.1.617.2) variant surge in India - Authors' reply.

We thank Sasanka Chakrabarti and colleagues for their interest in our Article and would like to clarify some issues. Their concern is that younger vaccinated participants were probably heath-care workers with a higher exposure risk than the controls in our study, making estimation of vaccine effectiveness difficult. However, we did clarify in the appendix of our Article that we accounted for this potential higher risk of exposure by adjusting for the confounding factors of age, sex, and exposure.

The possibility of previous asymptomatic or mildly symptomatic infections remains similar in both the groups and is not confined to controls. Serology for nucleocapsid antibodies could have been useful for diagnosing asymptomatic infections but vaccine effectiveness studies have generally focused on symptomatic infections. Estimates of vaccine effectiveness that consider asymptomatic infections can be quite imprecise. Furthermore, considering the large sample size, collecting samples and performing serological testing would have been prohibitively time consuming and would have defeated the purpose of generating estimates of vaccine effectiveness in a timely manner during the SARS-CoV-2 delta (B.1.617.2) variant surge in India.

Regarding the number of participants tested for cellular responses, the chosen sample size for T-cell assays was one of the largest to date to show that the T-cell responses were conserved between the ancestral virus and the delta variant of SARS-CoV-2, as has been confirmed by recent reports as well. Studying a larger sample size would have been challenging because of sample availability, resources, and the time required. Most vaccine effectiveness studies have tested around 10% of the vaccinated population for cellular responses.

Studying the T-cell responses against endogenous coronaviruses in unvaccinated individuals was beyond our study objectives, and we would welcome such a study. Nonetheless, published literature indicates that the magnitude of T-cell responses in unexposed individuals is considerably lower than in COVID-19 convalescent and vaccinated individuals. It is possible, although speculative, that the cross-reactivity of T cells against endogenous coronaviruses might provide some protection against SARS-CoV-2 and its variants.

The test-negative case-control design is a WHO-recommended and well established design for assessing real-world vaccine effectiveness—eg, for influenza, rotavirus, and SARS-CoV-2 vaccines. This design balances risk profile, health-care seeking behaviour, and access to care among vaccinated and non-vaccinated people. We considered the severity of COVID-19 as a secondary outcome and not as a confounder, and therefore it was not included in the multivariable model.

We declare no competing interests.