Abstract Objectives To investigate the effect of either a booster vaccine (one dose) or revaccination (two doses 3 weeks apart) on the antibody response to the COVID-19 mRNA vaccines in patients with rheumatic disease (RD) treated with rituximab (RTX) who had not produced vaccine-reactive antibodies after the initial two vaccine doses. Further, to examine if B cell levels in peripheral blood predicted seroconversion. Methods We included 91 RTX-treated RD patients previously vaccinated against COVID-19. Patients were offered revaccination or a single booster vaccination with an mRNA vaccine. Serum total antibodies against SARS-CoV-2 spike protein were measured before and 6 weeks after the last vaccine dose. B cells (CD19+CD45+) were measured by flow cytometry at inclusion. Results Of RD patients with undetectable SARS-CoV-2 antibody levels before inclusion, seroconversion was seen in 38% 6 weeks after the booster dose and 32% after revaccination. Patients receiving revaccination had significantly higher antibody levels than patients receiving a booster dose (P < 0.001). In both univariate and multivariate logistic regression analysis, only B cells higher than 10/µl before boost or revaccination were associated with seroconversion (P = 0.009 and P = 0.01, respectively). Seroconversion was independent of age, gender, diagnosis, cumulative RTX dose, RTX treatment time and time since last RTX treatment. Conclusion Continuously impaired humoral response to mRNA vaccines was found in most RTX-treated patients after a booster dose or revaccination. Seroconversion was observed in approximately one-third of the patients. Measurable B cells before boosting or revaccination was the strongest predictor of antibody response after boost or revaccination.