CoronaVac efficacy data from Turkey – Authors' reply

We thank Martina McMenamin and Benjamin Cowling for raising important issues on vaccine trials in the context of our Article.¹ The work they refer to by Palacios and colleagues² has not been published in a peer reviewed journal; thus we cannot comment on the accuracy or the comparability of its methods. The Indonesian trial data have been published,³ and although the main method of this study was similar to ours, the case definition of COVID-19 and the methods used for active surveillance were different. Only 1620 volunteers were included, but over a longer period of follow-up (approximately 2·5 months) precluding a direct comparison of the primary outcome.³ Nevertheless, the efficacy of CoronaVac against severe disease in this study was 100% because there were no critical cases or deaths attributable to COVID-19, which is undoubtedly similar to our results.

Regarding post-vaccination neutralising antibody titres in the immunogenicity subset of our trial, the seroconversion rate was 89·7% in the vaccine group, of whom 92% had neutralising antibodies. This result might translate into 82·5% neutralising antibody positivity in these volunteers. The efficacy against symptomatic disease reported as 83·5% is compatible with this immunogenicity result. We performed active surveillance to detect COVID-19 in patients; however, because the primary outcome was symptomatic COVID-19, it is indeed possible that we missed asymptomatic patients. In fact, most of the COVID-19 vaccine trials target a similar outcome, focusing on efficacy to prevent symptomatic and severe disease rather than preventing infection. We were aware of the short follow-up period in our interim analysis and hence discussed this as a major limitation in the Article,¹ stating that the study would not allow for commenting on the long-term protection.

Real-world effectiveness data from pragmatic study designs will add value to phase 3 trials to see the performance of the vaccines in non-selected populations, as complementary rather than competing studies. For instance, Jara and colleagues⁴ reported the analysis of real-life data from Chile, including approximately 10·2 million people vaccinated with CoronaVac. The adjusted vaccine effectiveness among the fully immunised people was 65·9% (95% CI 65·2–66·6) for the prevention of COVID-19 and 87·5% (86·7–88·2) for the prevention of hospital admission.

During a pandemic where only 2·3% of people in low-income countries had received at least one dose of a COVID-19 vaccine as of October, 2021, every single effort to make safe COVID-19 vaccines available is valuable. We believe that our data are an important contribution to the scientific literature in a world where we can no longer establish placebo-controlled randomised trials for COVID-19 vaccines for ethical reasons. The way forward to build confidence in vaccines is by reporting real-world data transparently.

We declare no competing interests.