Kidney transplantation (KT) is currently the best treatment option, considering quality of life, life expectancy and cost-effectiveness in end-stage renal disease (1–3). However, there is an increased risk of death during the first months post-transplantation compared to dialysis, owing to surgical and infectious complications (2). This sensitive early post-transplant period brings along increased risks of bleedings due to surgery and anemia of multifactorial origins (infections, inflammation, medications, … ) (4). Early blood transfusion is therefore often necessary, in an era where the age of recipients is constantly increasing, as well as the use of anticoagulant drugs (5).

Blood transfusions are a well-known cause of allogenic sensitization, especially before transplantation. Even though red blood cells are reputed to carry only low levels of Human Leukocyte Antigens (HLA) antigens, blood transfusions also bring few lymphocytes or platelets which may carry class I or class II HLA molecules (6). This antigen exposure to the immune system causes the generation of long-lived alloantibody-producing memory B cells (7) and anti-HLA antibodies. This process is dose-dependent, as the level of pre-transplant sensitization is correlated with the number of pre-transplant transfusions (8). Early blood transfusion is frequent after kidney transplantation, and concerns up to 40–60% among recipients (9–12). Considering its impact, the interrelationship between early blood transfusion allogenic exposure, de novo Donor Specific Antibodies (DSA) formation and allograft outcomes is less clear as previous reported cohorts provide contradictory results (9–11). Furthermore, the detection of DSA has greatly evolved over time thanks to Luminex-based methods, and there is a lack of large-scaled studies which examined the link between transfusion and de novo DSA occurrence using Luminex.

Our objective was then to examine the impact of post-KT early blood transfusions on de novo DSA formation, using Luminex-based methods, in a large cohort of renal transplant recipients. We also evaluated the impact of post-KT blood transfusions on the risk of biopsy-proven acute rejection (BPAR) and graft failure.

In this single-center study comprising a large number of KT and a median time of follow-up of 4.52 years, we did not show any association between post-KT early blood transfusions and the occurrence of de novo DSA, BPAR or graft failure.

Allorecognition leads to the generation of alloantibodies targeting non-self antigens, i.e. de novo DSA (7), which are associated with an increased risk of antibody-mediated rejection and allograft failure (13). As far as blood transfusions are concerned, the risk of induced-alloimmunization seems to have decreased over the last decades (14). A blood product is composed of three distinct parts (1): the desired product, such as red blood cells (RBCs) or platelets (2); excipients (e.g. anticoagulant or residual plasma) (3); residual leukocytes that carry HLA antigens, and in rare cases unexpected residual cells, such as platelets in red blood cells products (15). Leukocytes carry most of the antigenic load in a blood unit, yet the systematic use of leukoreduction process implemented in the late 90’s to fight against Creutzfeldt-Jakob disease transmission dramatically reduced the amount of WBC into blood units (16, 17). As a consequence, the rate of post-transfusion HLA sensitization decreased from nearly 30% of transfused patients to 10–20% depending on studies (18–21). However, even with leukoreduction, the risk of sensitization still persists as erythrocytes constitutively express HLA class I molecules at low levels (22). The risk of transfusion-related sensitization also depends on the immunological history of the recipient. Indeed, transfused kidney transplant candidates with a history of pregnancy or previous transplantation have a higher risk of sensitization after transfusion compared to kidney transplant candidates with a sole history of blood transfusion (23). Moreover, there seems to be a dose-effect, as the level of immunogenicity correlates with the number of administered units (6, 24).

Even though blood transfusion seems to be clearly associated with the risk of HLA sensitization, its impact on allograft outcomes remains unclear. Paradoxically, throughout the early beginnings of solid organ transplantation and before the implementation of cyclosporine, pre-transplant blood transfusion was supposed to be associated with immunomodulatory properties and benefits on renal allograft outcomes (25, 26). Donor-specific transfusion in KT has long been part of routine practices for its supposed ability to prevent post-transplant rejection (27, 28). Animal models also provided evidence of transfusion-related immunomodulation properties owing to the generation of alloreactive CD25⁺CD4⁺ regulatory T cells that prevent graft rejection (29), from both related and unrelated donor blood. Nowadays, even if donor-specific transfusion is no longer used, the potential immunomodulatory properties of blood transfusion question the impact of early blood transfusion after KT.

In our study, we did not find any association between early blood transfusion post-KT and the further risk of de novo DSA development, on a large-scaled cohort using Luminex-based methods to identify DSA. On the contrary, HLA mismatches and HLA sensitization were significantly associated with de novo DSA and are a well-known risk factors of post-transplantation allosensitization (30, 31). Aging was also associated with the risk of de novo DSA occurrence. On the one hand, aging in recipients was associated with a lower risk of allosensitization, which may reflect the aging-related immunosenescence in recipients. On the other hand, aging in donors was associated with an increased risk of de novo DSA occurrence, which exhibits the aging-related immunogenicity of kidney donors (32). The transfusion status was also not associated with secondary outcomes such as the long-term risks of rejection or graft failure. Considering the literature, Scornik et al. (9) reported 746 patients transplanted followed for 6 years, including 45% transfused-recipients with 79% of blood transfusions performed during the first month post-KT. There was no significant difference regarding the incidence of rejection episodes or graft loss according to the transfusion status. There was also no difference regarding the frequency of de novo DSA between transfused and non-transfused recipients (17% vs. 15%, p = 0.67). Verghese et al. (12) reported then a pediatric study of 482, including 44% transfused patients. Among these, 134 recipients could be tested for HLA antibodies using solid-phase based methods, including 82 transfused recipients. In their study, blood transfusion was also not associated with the risks of de novo DSA after KT (HR 0.9; 95% CI 0.6–1.4; p = 0.65), rejection or graft failure. In the same way, Daloul et al. recently reported their experience of 273 recipients, including 127 transfused recipients before 1-month post-KT. They did not find any difference at 1-year post-KT regarding the incidence of de novo DSA using solid-phase based methods (12.8% in transfused recipients and 10.9% in non-transfused recipients, p = 0.48) (33), as well as with the risk of rejection or graft loss.

Conversely, Ferrandiz et al. showed opposite results regarding the association between blood transfusion and de novo DSA, with one of the largest cohorts studying HLA antibodies using Luminex. Three hundred and ninety non-sensitized kidney transplant recipients were included, of which 250 were transfused during the first year post-KT. 94.8% of them were transfused during the first month post-KT. During the first-year post-transplantation, 18 recipients (7.2%) in the transfusion group developed de novo DSA, compared to only one (0.7%) in the nontransfusion group (p < 0.0001). This higher prevalence of de novo DSA was also associated with a higher incidence of ABMR (15 transfused-recipients (6%) vs two non-transfused recipients (1.4%), p = 0.04). However, baseline characteristics significantly differed regarding notably the immunosuppressive regimen, with a higher proportion of transfused recipients treated with cyclosporine. Furthermore, they examined early outcomes as logistic regression at 1-year post-KT revealed that both the use of cyclosporin and blood transfusions were associated with the risk of DSA formation. In our study, both the evaluation of early and long-term outcomes did not find any association with the transfusion status. Yet, we acknowledge that it may be difficult to compare these two monocentric studies, as far as local practices, e.g. regarding the immunosuppressive regimen management, may influence the results. Recently, Hassan et al. reported a cohort of 1,104 recipients including 667 transfused recipients. 88.9% of blood transfusions were performed before 1-month post-KT. Blood transfusion was significantly associated with the development of de novo DSA (transfusion received: HR = 1.49 [1.10–2.04], p = 0.01) and graft failure (transfusion received: HR = 1.85 [1.19–2.77], p = 0.005). However, the prevalence of blood transfusion was surprisingly high, which could be linked to the baseline characteristics of the overall cohort (not provided). Nevertheless, they provided novel data dealing with the analysis of shared transfusion and kidney donors’ alloantibodies in transfused recipients. They analyzed a subgroup of 86 transplant recipients who received transfusion from 244 blood donors. Overall, 61.5% of transfused recipients developed de novo transfusion specific antibodies (TSA), of which 46.7% shared HLA antibody specificity with a DSA response in the recipient (DSA+/TSA+). DSA+/TSA + recipients had an increased risk of allograft loss or rejection compared to recipients with only TSA or DSA. This may suggest a need of blood donor HLA matching in kidney transplant recipients.

Compared to the existing literature, our study has two main strengths. First, we provide one of the largest cohort of recipients screened for HLA antibodies during their whole follow-up, combined with their transfusion status. Second, this cohort benefitted from a long-term follow-up with the evaluation of reliable time-dependent outcomes. Still, our findings need to be interpreted in the context of some caveats. Indeed, the retrospective nature of the study could be associated with information bias. Then, it is also limited by the lack of information regarding the hemoglobin levels. Post-KT anemia is indeed known to be associated with mortality and graft loss (34, 35). Yet, our primary outcome is focused on the emergence of de novo DSA and no association between anemia and de novo DSA has been reported to date. Thus, it does not seem to constitute a confounding factor. Finally, there are significant baseline differences between transfused and non-transfused recipients that should be considered to interpret our results. Donors from transfused recipients were significantly older, with longer cold ischemia times. Transfused recipients were significantly older, sensitized against HLA antibodies, more frequently treated with thymoglobulin induction and had a significant worse graft function at 1-month post-transplantation. These baseline differences may explain why univariate and short-term analyses revealed differences concerning rejection and graft failure. However, after adjusting on confounding factors on long-term analyses, transfusion was no longer associated with any of those outcomes. To be noted, as far as our primary criteria of judgement is concerned, none of our analyses, i.e. short-term or long-term, univariate and multivariate, revealed an association between transfusion and de novo DSA occurrence.

Ultimately, even if on a global scale we did not find any association between transfusion and the development of de novo DSA, it does not mean that this correlation does not exist at an individual level, as suggested by Hassan et al. The risk of allosensitization should be kept in mind, and strategies of HLA matching between blood and kidney donors may be of interest in the next few years. However, our data provide evidence that transfusion should not be limited in the early period post-KT when required.

In our large-scaled cohort of kidney transplant recipients, we did not find any association between post-KT early blood transfusions and the development of de novo DSA, nor the risks of rejection and graft failure.