Background: There are no studies providing head-to-head comparison across SARS-CoV-2 vaccines. Therefore, we compared the efficacy of candidate vaccines in inducing neutralizing antibodies against SARS-CoV-2. Methods: A network meta-analysis was performed to compare the peak levels of SARS-CoV-2 neutralizing antibodies across candidate vaccines. Data were reported as standardized mean difference (SMD) since the outcome was assessed via different metrics and methods across the studies. Results: Data obtained from 836 healthy adult vaccine recipients were extracted from 11 studies. BBIBP-CorV, AZD1222, BNT162b2, New Crown COVID-19, and Sputnik V induced a very large effect on the level of neutralizing antibodies (SMD > 1.3); CoVLP, CoronaVac, NVX-CoV2373, and Ad5-nCoV induced a large effect (SMD > 0.8 to ≤1.3); and Ad26.COV2.S induced a medium effect (SMD > 0.5 to ≤0.8). BBIBP-CorV and AZD122 were more effective (p < 0.05) than Ad26.COV2.S, Ad5-nCoV, mRNA-1237, CoronaVac, NVX-CoV2373, CoVLP, and New Crown COVID-19; New Crown COVID-19 was more effective (p < 0.05) than Ad26.COV2.S, Ad5-nCoV, and mRNA-1237; CoronaVac was more effective (p < 0.05) than Ad26.COV2.S and Ad5-nCoV; and Sputnik V and BNT162b2 were more effective (p < 0.05) than Ad26.COV2.S. In recipients aged ≤60 years, AZD1222, BBIBP-CorV, and mRNA-1237 were the most effective candidate vaccines. Conclusion: All the candidate vaccines induced significant levels of SARS-CoV-2 neutralizing antibodies, but only AZD1222 and mRNA-1237 were certainly tested in patients aged ≥70 years. Compared with AZD1222, BNT162b and mRNA-1237 have the advantage that they can be quickly re-engineered to mimic new mutations of SARS-CoV-2.
Keywords: COVID-19; SARS-CoV-2; network meta-analysis; neutralizing antibodies; vaccine.