Abstract INTRODUCTION Immunocompromised patients have been excluded from initial trials evaluating SARS-CoV-2 mRNA vaccines and there is a critical need to warrant vaccine efficacy in hematopoietic stem cell transplant (HSCT) recipients. In this study, we evaluated antibody responses to 2 doses mRNA SARS-CoV-2 vaccine in allogeneic HSCT recipients. METHODS We retrospectively enrolled successive hematopoietic cell transplant recipients across France who completed the 2-dose SARS-CoV-2 mRNA vaccine (BNT162b2 or mRNA-1273) between January 1 st and July 15 th 2021. All included patients had an available semi-quantitative antispike serologic testing after the second dose (from Roche, DiaSorin, Abbott or Siemens). We excluded patients with a prior COVID-19 confirmed by serology or PCR. For detectable antibody, we calculated the binding antibody units per milliliter (BAU/mL) according to the WHO International Standard by applying conversion factors given by the manufacturers (Kristiansen et al. , The Lancet 2021). Antibody response was categorized as "weak" or "good" with a threshold of 264 BAU/mL which has been associated to an estimate of 80% of mRNA vaccine-induced protection against symptomatic COVID-19 in immunocompetent patients (Feng S. et al., medRxiv 2021). We built a multivariate logistic regression model to assess factors independently associated with the absence of antibody response after the second dose of mRNA vaccination. RESULTS Overall, 620 allogeneic HSCT recipients from 12 hospitals across France were included in the analysis (60% male with a median age of 59 years old [IQR 47-66]), most with a myeloid (69%) or lymphoid (26%) malignancies. Donors were matched unrelated for 51%, HLA-identical sibling for 31% and haplo-identical for 18%. Thirty-one percent of HSCT recipients underwent a myeloablative conditioning, while 69% received a reduced intensity conditioning. The two doses of vaccines were given one month apart and the median time between transplantation and the initiation of vaccination was 29 months [IQR 14-58]. At a median of 33 [IQR 27-50] days after dose 2, an antibody response was detectable in 496 patients (80% [95CI: 77 to 83%]). Median [IQR] antibody levels was 243 BAU/mL [29.4-1391]. We classified detectable antibody responses as "weak" in 189 patients (30% [95CI 27 to 34%]) and as "good" in 306 (49% [95CI: 45 to 53%]). In the multivariate analysis including 533 patients (420 with detectable antibodies), factors associated with the absence of humoral responses were a time-interval from HSCT < 12 months (ajusted Odds-Ratio (aOR) 2.8 [95CI 1.6 to 4.8]), absolute lymphocyte count <1G/L (aOR 3.0 [95CI 1.7 to 5.0]), systemic immunosuppressive treatments within 3 months of vaccination (aOR 4.5 [95CI 2.7 to 7.5]), together with the use of rituximab within 6 months (aOR 15.1 [95CI 4.3 to 52.7]). In a subsequent multivariate analysis conducted a subset of 227 patients (170 with detectable antibodies) with available gammaglobulinemia as well as B and T lymphocytes counts, factors remaining associated with the absence of antibody response were only low B-lymphocytes count (aOR 5.5 [95CI 2.4 to 12.3]) and time-interval from HSCT < 12 months (aOR 3.3 [95CI 1.5 to 7.2]). CONCLUSION After 2 dose mRNA vaccination, the majority of allogeneic HSCT recipients developed an antibody response although a significant proportion of these responses may be insufficient. Studies are still needed to investigate the effect of a third vaccine dose in patients with a null or weak humoral response. Disclosures Loschi: Servier: Ended employment in the past 24 months, Honoraria; Novartis: Ended employment in the past 24 months, Honoraria; Gilead: Ended employment in the past 24 months, Honoraria; AbbVie: Ended employment in the past 24 months, Honoraria; CELGENE/BMS: Honoraria; MSD: Honoraria.