LAUSR

B-depleting therapies increase the risk of severe COVID-19 [1] and have been shown to induce an impaired immune response to the SARS-COV2 vaccine [2]. There is a lack of data exploring the T-cell response to SARS COV2 vaccine [3,4], especially since it may be preserved in patients treated with rituximab [5]. Here we analyzed the T-cell immune response to SARS Cov2 vaccination measured by interferon-γ release assay (see details in sup file) after three Pfizer BNT162B2 mRNA vaccine doses (see details in sup file) in 28 long term rituximab-treated patients and compared the characteristics of patients according to the good or poor T-cell response to the vaccine. Patients received rituximab for ANCA-associated vasculitis (n = 13), Ig-G4 related disease (n = 4), immune thrombocytopenia (n = 3), antiMAG neuropathy (n = 3), dermatomyositis (n = 1), systemic sclerosis (n = 1), rheumatoid arthritis (n = 1), Waldenström macroglobulinemia (n = 1) and chronic polyradiculoneuropathy (n = 1). Median disease duration at inclusion was 6 years [2.0–10.75] without difference between vaccine responders and non-responders. Patients had received a median of 6 infusions [IQR: 4.0–13.0] of rituximab at inclusion with a median duration between the last infusion and inclusion less than 6 months (125.5 days [86.25–182.0], without any difference between the 2 groups (P = 0.9454). Sixteen patients (57%) had a preserved T-cell response after the third dose of vaccine. These patients were younger (median 56.5 years IQR [33.75–63.5] vs. 71.0 [66.5–76.25]; P = 0.0003). There was no difference depending on the underlying disease, the number of infusions of rituximab, previous therapies or low dose concomitant steroid therapy (n = 8, median daily dose 5 mg [5.0–8.75]). Total lymphocytes count (x109/L) was higher in responder’s patients (1.53 [1.170–1.890] vs. 1.010 [0.6525–1.170]; P = 0.0003). Number of patients with B-cell reconstitution was higher in responders patients (7 vs 0; P = 0.003). The percentages of CD3+ lymphocytes, CD3+CD4+ lymphocytes, CD3+CD8+ lymphocytes and NK cells were similar between groups (P = 0.4345, P = 0.7692, P = 0.6548 and P = 0.0550, respectively). A total of 10 patients had serological response toward SARS Cov2 vaccine, as defined by anti-SARS Cov2 Ig-G titers >50 UA/ ml. This titer tends to be higher in patients with a T-cell response (11.10 [2.125–1893] vs 1.30[0.25–107.7]; P = 0.0833). All data are reported in Table 1. Multivariate analysis (age≥65 yes/no, T lymphocytes<1 × 109/L yes/no, B cell reconstitution yes/no) using backward binary logistic regression showed that age ≥65 years (odds ratio[OR]=0.034 [95% CI=0.003–0.428]; P = 0.009) and T lymphocytes<1 × 109/L (OR=0.047 [0.003–0.887]; P = 0.047) were significantly associated with a poor T-cell response. Our data show that most rituximab-treated patients have a poor humoral response to SARS-Cov2 after complete vaccination with 3 doses of Sars-Cov2 BNT162b2 Pfizer vaccine and surprisingly, some have an additional T-cell immune response defect, suggesting that the risk of a severe form is further increased in these patients. Among factors specifically associated with poorer T-cell immune response, we identified the number of T lymphocytes and age ≥65 at the time of vaccination. By contrast and as previously reported [2], there was no correlation between disease duration, previous drug exposure or time since last rituximab infusion and T-cell response. It is of concern that some rituximab-treated patients do not develop a humoral or T cell response after vaccination and should therefore be prioritized to receive prophylactic cocktails of neutralizing antibodies, especially during epidemic periods [6]. In case a study of the anti-SARS Cov2 cellular immune response is not possible, consideration of age and circulating T cell count may predict a poor T-cell immune response after SARS Cov2 vaccination. To conclude, T-cell response to the BNT162B2 mRNA vaccine in rituximab-treated patients is primarily determined by age and T-cell count.