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

We read with interest the study by Ramachandran Thiruvengadam and colleagues. However, we feel there are some aspects of the study that require further input from the authors.

The mean age of cases is reported to be 35 years and of controls to be 32 years. In India, vaccination of the general population in the 18–45 years age group began in early May, 2021; in this age group, only frontline workers were vaccinated from mid-January to late April, 2021. Considering the quoted duration of the study (April 1 to May 31, 2021), most fully vaccinated individuals would have been frontline workers, and predominantly health workers. However, the unvaccinated group would be representative of the general population. Therefore, the two groups had different levels of exposure to SARS-CoV-2, making comparison and estimates of vaccine efficacy difficult.

The controls were selected on the basis of RT-PCR negativity in a defined time period. However, some of them might have been affected during the first wave of COVID-19 with mild or asymptomatic disease and might have been partially immune to reinfection during the study period, which would be an unknown variable in the study modifying reinfection rate or severity. Measurement of serum neutralising antibody titres against spike or nucleocapsid proteins at baseline could have been used to eliminate this group from the study.

Compared with the total study population, the number of people analysed for T-cell response was small (48 [1·1%] of 4360). Furthermore, the T-cell responses to spike peptide pools of wild-type SARS-CoV-2 and delta (B.1.617.2) SARS-CoV-2 were only studied in a healthy vaccinated group. An unvaccinated group could have been included to check whether cross-reactive T cells primed by endemic coronaviruses can also respond to wild-type SARS-CoV-2 or the delta variant. The concept of cross-immunity has been expanded in the context of COVID-19, both theoretically and experimentally.2, 3 Additionally, these data would have helped us to understand the intensity of the T-cell immune response in the vaccinated population compared with the unvaccinated population. Even those positive for antibodies against SARS-CoV-2 nucleocapsid at baseline could have been included in this testing to investigate how previous infection affects T-cell responses. Such an investigation becomes more relevant in a real-world study when breakthrough infections are known to be quite common.

Finally, an important prerequisite for test-negative case-control studies is matching of cases and controls for disease severity and confounders. Information on such potential confounders and symptom characterisation (especially disease severity) in the control group should be provided. Furthermore, the test-negative design can control for selection and information bias but is not effective in blocking bias due to health-seeking behaviour, which differs between vaccinated and unvaccinated individuals and is affected by the severity of COVID-19.

We declare no competing interests.