A total of 352 participants were enrolled (179 and 173 participants in CoronaVac- and ChAdOx1-prime group), 285 (81%) were female, and the median age was 39 (interquartile range, IQR: 31-47) years. The demographic of the study participants receiving different booster vaccine was shown in Table 1.
Adverse events (AEs)
Among CoronaVac-prime group, the frequency of overall AEs was the highest after boosting with ChAdOx1 (98%), followed by 30 µg BNT162b2 (92%), 15 µgBNT162b2 (80%), and BBIBP-CorV (70%); whereas in ChAdOx1-prime group, the highest rate of overall AEs was reported after boosting with 30 µg BNT162b2 (98%), followed by 15 µgBNT162b2 (88%), ChAdOx1 (72%), and BBIBP-CorV (61%) (Fig. 1, Supplementary Table 1). Systemic AEs were in the same trend as local AEs. All AEs were mild (grade 1) or moderate (grade 2) in severity and recovered within 2-3 days. No serious AEs was found in this study.
Anti-SARS-CoV-2 RBD IgG responses
Prior to the booster dose (8-12 weeks post-primary series), 175/179 (97.8%) participants in CoronaVac-prime and all (172/173, 99.4%) in ChAdOx1-prime were seropositive; however, the anti-RBD IgG geometric mean concentrations (GMCs) were lower in CoronaVac-prime group than ChAdOx1-prime group (36.31 BAU/mL range: 33-38 BAU/mL vs 98.27 BAU/mL, range: 90-116 BAU/mL) (Fig. 2A). For the CoronaVac-prime group, at 2 weeks after the booster vaccination, the anti-RBD IgG GMC in the 30 µg BNT162b2 group (5,152 BAU/mL) was significantly higher than other booster vaccine groups: 15 µg BNT162b2 (3,981 BAU/mL), ChAdOx1 (1,358 BAU/mL), and BBIBP-CorV (154 BAU/mL) (Fig. 2A). A booster dose of BBIBP, ChAdOx1, 30µg BNT162b2 or 15µg BNT162b2 increased anti-RBD IgG GMC by 4.5, 35.6, 154.7, and 105.7 times from baseline, respectively. The geometric mean ratio (GMR, 95% CI) between post-boost and post-primary series of CoronaVac were 0.94 (0.53, 1.67) for BBIBP-CorV, 8.26 (6.29, 10.85) for ChAdOx1, 31.34 (24.37, 40.30) for 30µg BNT162b2, and 24.22 (18.60, 31.54) for 15µg BNT162b2, respectively (Supplementary Table 2).
For the ChAdOx1-prime group, at 2 weeks after the booster dose, the anti-RBG IgG GMC was highest in the participants receiving a booster dose of 30 µg BNT162b2 at the level of 2,364 BAU/mL, followed by 15 µgBNT162b2 (1,962 BAU/mL), ChAdOx1 (246.4 BAU/mL), and BBIBP-CorV (128.1 BAU/mL). However, there was no statistical difference in the anti-RBD IgG GMC between 30 µg and 15 µg BNT162b2 groups (P=0.1381) (Fig. 2B). A booster dose of BBIBP, ChAdOx1, 30µg BNT162b2, or 15µg BNT162b2 increased GMCs for 1.2, 2.3, 25.1, and 21.8 times from baseline, respectively. The GMR (95% CI) between post-boost and post-primary series of ChAdOx1 were 0.46 (0.28, 0.65) for BBIBP-CorV, 0.88 (0.58, 1.13) for ChAdOx1, 8.49 (5.71, 10.44) for 30µg BNT162b2, and 7.04 (4.69, 8.84) for 15µg BNT162b2, respectively (Supplementary Table 2). The post-boost GMC levels in ChAdOx1-prime group were generally lower than those in the CoronaVac-prime group for all booster vaccines. (Fig. 2C).
Neutralising antibody response against SARS-CoV-2 variants
The PRNT50 GMTs against the Delta variant among CoronaVac-prime participants were highest after a booster dose of 15 µg BNT162b2 (499.1), followed by 30 µg BNT162b2 (411.1), ChAdOx1 (271.2), and BBIBP-CorV (61.3); whereas among ChAdOx1-prime group, the highest PRNT50 GMT was after a booster dose of 30 µg BNT162b2 (470), followed by 15 µg BNT162b2 (358.4), ChAdOx1 (69.1), and BBIBP-CorV (49.1) (Fig. 3A). A similar trend was observed for PRNT50 GMT against the Beta variant (Fig. 3B). There was no statistical difference of PRNT50 GMT between boosting with 30 µg and 15 µg BNT162b2 regardless of the vaccine given as primary series and the type of variants. The PRNT50 GMT of the Beta variants were around 1.5-fold lower than the Delta variants for both CoronaVac-prime and ChAdOx1-prime groups (Fig. 3C). The GMRs (95% CI) of PRNT50 titres between post-boost and post-primary series were highest among the participants who received BNT162b2 boosting vaccination in both CoronaVac-prime and ChAdOx1-prime groups for both Delta and Beta variants (Table 2). The anti-SARS-CoV-2 RBD IgG levels and the PRNT50 titres against Delta variant (Fig. 4A and Fig. 4B) or Beta variant (Fig. 4C and Fig. 4D) were highly correlated (r = 0.49-0.89) particularly in the ChAdOx1-primed group.
Interferon gamma responses (Interferon gamma release assay, IGRA)
At 8-12 weeks following the primary series, 62/173 (35.8%) of participants in ChAdOx1-prime group had positive IGRA, while only 45/179 (25%) CoronaVac-prime group was positive (P=0.029). Among those who were seronegative in both CoronaVac-prime and ChAdOx1-prime group, IGRA conversion was the highest after a booster dose of 30 µg BNT162b2, followed by 15 µg BNT162b2, ChAdOx1, and BBIBP-CorV (Table 1). Interestingly, none of the study participants who were IGRA-negative at baseline in theChAdOx1-prime group had a positive IGRA response following boosting with BBIBP-CorV or ChAdOx1 (Table 1) (Supplementary Fig. S1).