CoronaVac induces lower neutralising activity against variants of concern than natural infection

The inactivated whole-virus CoronaVac vaccine (Sinovac Biotech, Beijing, China) has been approved for emergency use in mass vaccination programmes in Thailand and is widely available in many low-income countries. Results from a phase 1–2 clinical trial of CoronaVac were recently published in this journal,¹ and a large, observational study in Chile further estimated that two doses of CoronaVac had vaccine effectiveness of 65·9% against COVID-19, 87·5% against hospitalisation, 90·3% against intensive care unit admission, and 86·3% against death, with values adjusted for potential effects of age and sex.² Variants of concern (VOCs) circulating in Thailand as of writing include B.1.1.7 (alpha), B.1.351 (beta), and B.1.617.2 (delta). To assess the impact of SARS-CoV-2 variants on vaccine-induced and infection-induced antibodies, we evaluated titres of SARS-CoV-2 S1-receptor-binding domain (RBD)-binding IgG, as well as neutralising antibody (NAb) titres against the SARS-CoV-2 prototypic vaccine strain (wild-type [WT]) and VOCs in sera from health-care workers who had received two doses of CoronaVac; we compared these with sera from unvaccinated, naturally infected patients who had been hospitalised in March–May, 2020 (hereafter denoted the natural infection 2020 cohort), or April–May, 2021 (hereafter denoted the natural infection 2021 cohort). We used a live-virus microneutralisation assay for NAb titre quantification. Details regarding cohort demographics, methods, and statistical analyses can be found in the appendix (pp 4, 6–8).

We found that 100% of participants in all cohorts were seropositive for virus-specific IgG. We next assessed NAb-afforded protection against WT and VOCs in our cohorts. Overall, the percentage of participants with quantifiable NAb titres (≥20 units) was highest against the WT strain, followed by much lower titres against the alpha, beta, and delta variants (appendix p 5). This pattern was consistently observed in all cohorts, and notably, the percentages of individuals with detectable NAbs were lower in CoronaVac recipients than in the naturally infected cohorts (appendix p 5). In adjusted analyses, we observed that, in all cohorts, geometric mean NAb titres were significantly lower against all VOCs than against WT (appendix p 2). NAb titres against the alpha and beta variants were not significantly different from each other, and NAb titres against the delta variant were the lowest and significantly different from the rest (appendix p 2).

We further found that WT was best neutralised by natural infection 2020 sera and the alpha variant was best neutralised by natural infection 2021 sera (appendix p 2). These results are consistent with the predominant strains circulating in Thailand in early to mid-2020 and mid-2021 at the time of sample collection for each respective cohort. The beta variant was neutralised equally well by natural infection 2020 and 2021 sera, with geometric mean NAb titres that were higher than those elicited by CoronaVac (appendix p 2). Similarly, the delta variant was neutralised equally well by natural infection 2020 and 2021 sera, but with markedly lower NAb titres than those obtained with the beta variant. Titres against the delta variant in CoronaVac recipients were lower still, almost at the limit of detection (appendix p 2). Together, these results highlight the relatively low NAb titres elicited by CoronaVac compared with natural infection.

Although NAb titres are not an exclusive immune correlate of protection, they are highly predictive of immune protection from symptomatic SARS-CoV-2 infection.³ Based on our data, although there was robust production of S1-RBD-binding IgG and 100% seropositivity across the board, NAb-mediated protection was markedly reduced (and in many cases undetectable) against the three VOCs compared with WT in sera from all groups. Furthermore, NAb potency against alpha and beta VOCs was comparable in our CoronaVac vaccinee sera; this finding is inconsistent with a previous report showing that the beta variant is more resistant to neutralisation than the alpha variant with sera from CoronaVac recipients collected 14 days after the second dose when tested using a pseudovirus neutralisation assay.⁴ Worryingly, the delta variant, which is the most transmissible, possibly among the most virulent of all VOCs identified to date,⁵ and a dominant variant in many countries, appears to be most refractory to neutralisation. Lastly, our study highlights a low degree of neutralisation-afforded protection mounted by CoronaVac when compared with natural infection. Further booster doses, heterologous or otherwise, beyond the conventional two-dose regimen might be needed for recipients of CoronaVac to maintain a long-term anamnestic response. Amid steady NAb decay over time³ and the continued emergence of divergent SARS-CoV-2 variants, it is imperative to maintain effective mitigation strategies and to continue monitoring vaccine efficiency in areas with circulating VOCs.

We declare no competing interests. This work was supported by the National Research Council of Thailand, the Mahidol University for Integrated and Multidisciplinary Research Cluster grant (MRC-IM 02/2564), the Program Management Unit C, and the Ramathibodi Foundation.