The effect of BCG vaccination on infection and antibody levels against SARS-CoV-2-The results of ProBCG: a multicenter randomized clinical trial in Brazil

Abstract

Objectives: Evatuate if Bacillus Calmette-Guérin (BCG) vaccine could be used as a tool against SARS-CoV-2 based on the concept of trained immunity.

Methods: A multicenter, double-blinded, randomized clinical trial recruited health care workers (HCWs) in Brazil. The incidence rates of COVID-19, clinical manifestations, absenteeism, and adverse events among HCWs receiving BCG vaccine (Moreau or Moscow strains) or placebo were compared. BCG vaccine-mediated immune response before and after implementing specific vaccines for COVID-19 (CoronaVac or COVISHIELD) was analyzed. Cox proportional hazard and linear mixed effect modeling were used.

Results: A total of 264 volunteers were included for analysis (BCG = 134 and placebo = 130). The placebo group presented a COVID-19 cumulative incidence of 0.75% vs 0.52% of BCG. The Moreau strain also presented a higher incidence rate (1.60% × 0.22%). BCG did not show a protective hazard ratio against COVID-19. In addition, the log (immunoglobulin G) level against SARS-CoV-2 presented a higher increase in the BCG group, whether or not participants had COVID-19, but also without statistical significance.

Conclusion: Our results suggest that BCG has a tendency of protection against SARS-CoV-2 and higher immunoglobulin G levels than placebo. The clinical trial was registered at https://clinicaltrials.gov/ (NCT04659941).

Keywords: BCG; COVID-19; Vaccine.

Figure 1

Flowchart of study participants. The flow diagram was constructed according to the CONSORT guidelines and illustrates the progress of the participants through the trial (recruient, enrollment, randomization, withdrawal and analysis). BCG, Bacillus Calmete-Guérin; HMB, Hospital Muninical de Barueri; UERJ, Universidade do Estado do Rio de Janeiro; UFRJ, Universidade Federal do Rio de Janeiro.

Figure 2

Multivariate Cox regression results of variables associated with COVID-19 according to the BCG strain used for intervention. Forest plots illustrate the hazard ratios of each variable (BCG vaccine, BCG strains, COVID-19 vaccine, age, sex, comorbidities and IGRA results) on the risk of COVID-19 and the sensitivity analysis using Moreau and Moscow BCG strains separately. AIC, Akaike information criterion; BCG, Bacillus Calmete-Guérin; IGRA, interferon-γ-release assay.

Figure 3

Estimated log (IgG) values of intervention groups for (a) non-COVID-19 and (b) COVID-19 cases. Graphic plots of the log(IgG) slopes representing the linear mixed effect models according to COVID-19 development or not during follow-up. Solid lines represent log(IgG) means of BCG group with its respective 95% confidence interval in red, while dashed lines represent log(IgG) means of placebo group with its respective 95% confidence interval in gray. (a) Log(IgG) levels according to the moment of follow-up (M0 = intervention, M1 = 10 days after intervention, M3 = 90 days after intervention, M6 = 180 days after intervention) in non-COVID-19 cases. (b) Log(IgG) levels according to the moment of COVID-19 diagnosis (time 0) and the visits before and after it (each interval 30 days apart). BCG, Bacillus Calmete-Guérin; Ig, immunoglobulin.