Patient blood management (PBM) embodies an evidence-based, multi-disciplinary approach to minimise inappropriate blood product transfusions, in turn improving patient outcomes and reducing overall healthcare costs (1, 2). It is an international initiative endorsed by the World Health Organisation (WHO), who in 2021 announced the global implementation of PBM as an urgent healthcare intervention (3). Initially developed with surgical procedures as its focus, current PBM policies apply to all patients in whom blood product transfusions maybe required (4). Kidney transplantation presents a particular challenge to PBM, as patients with a high prevalence of anaemia are subjected to a surgical procedure which is followed by the introduction of new factors e.g. immunosuppression, which could exacerbate anaemia further (5, 6). Consequently, peri-transplant anaemia is extremely common with reported rates as high as 90% (5). Despite its common occurrence, with transplant specific contributions, guidance on post-transplant anaemia management frequently defaults to replicate protocols in the non-dialysis chronic kidney disease (CKD) population (7, 8).

A fundamental aspect to PBM is robust evidence of the risk of transfusion, which is unfortunately lacking in the setting of kidney transplantation. Such evidence cannot be extrapolated from other surgical procedures, as a unique complexity of transplantation includes the potential risk of transfusion on de novo allosensitisation and rejection, contributing to premature allograft loss (9–13). Although transfusion is a well-recognised cause of allosensitisation pre-transplantation, in the presence of immunosuppression in the post-transplant setting, data are less consistent (9, 10, 14, 15). However, our group has previously demonstrated that transfused transplant recipients may develop HLA antibodies against their corresponding blood donors in the post-transplant period, providing evidence that de novo allosensitisation may occur (9, 16).

It is unknown whether the absence of transplant specific guidelines, lack of evidence of the risk of post-transplant blood transfusions or centre specific practices contribute to the large variation in the reported transfusion rates, which range from 18.1% to 74.6% in the early post-transplant period (17, 18). However, what is clear is that to incorporate the principles of PBM in kidney transplantation, there is an urgent need for data on who is being transfused, why they are being transfused and the impact, if any, that transfusion has on outcomes.

The aim of this study is to investigate the prevalence of post-transplant blood product transfusions across multiple centres in the UK, to analyse recipient factors associated with transfusion and to compare allograft and patient survival by transfusion status.

Seven hundred and twenty patients received a kidney transplant across the 4 centres between 2016 and 2017. Of these, 221/720 (30.7%) were transfused; 189 (26.3%) received a red blood cell (RBC) transfusion alone, 7 (1.0%) platelets alone and 25 (3.5%) received both blood and platelet transfusions. There was inter-centre variation in the proportion of patients transfused, ranging from 20% to 35%, p=0.005. The median time to transfusion was 4 (IQR:0-12) days post-transplant, and the median number of units transfused was 2 (IQR:2-5) units of blood and 1 (IQR:1-3) pool of platelets.

In this multicentre study we have demonstrated that post-transplant transfusions are common, with significant variation in rates across the different centres. Multivariable analysis of clinical characteristics associated with transfusions identified that recipient characteristics had a greater impact on transfusion risk than kidney donor characteristics. Furthermore, we have shown that transfusions, which mainly occur in the early post-transplant period are associated with inferior allograft function, allograft survival and patient survival at 12 months post-transplant, which is maintained even after adjusting for clinical characteristics recognised to be associated with poor allograft outcomes. Therefore, independent of the potential impact on allosensitisation, these data provide evidence that the transplant community needs to address these inequalities and explore the mechanisms behind the adverse outcomes to inform interventions which may improve patient health and transplant longevity.

The underlying principles or 3 pillars of PBM relate to the detection and management of anaemia, minimisation of blood loss including optimisation of coagulation, and optimising patient specific tolerance of anaemia (3). For patients undergoing kidney transplantation, each pillar requires the application of renal and transplant specific considerations. From optimisation of pre-transplant anaemia, to circumventing the dichotomy of increased bleeding and thrombotic risks, to balancing adverse effects of immunosuppression, prophylactic anti-microbials and infection (6, 20, 21). Despite this, collective evidence and therefore clinical guidelines are lacking, which is likely to be contributing to the variation of management of anaemia and transfusion rates (22, 23).

We found that the requirement for blood transfusions predominantly occurs in the early post-transplant period, which is supported by other reports (5, 9, 24). Although blood loss during surgery will contribute to the need for early transfusion, our data suggests that intraoperative transfusions occur less frequently than in the early post-transplant period. By measuring erythropoietin (EPO) levels post-transplant, it has been shown that EPO follows a bimodal distribution post-transplantation, with sustained levels from 28 days in the setting of primary graft function (25). Therefore, it is the first couple of weeks where efforts to optimise anaemia management needs to be prioritised in the first instance.

Recombinant erythropoietin (rEPO) and iron therapy have been and remain the cornerstones of the treatment of renal anaemia. However, rEPO use is not without risk, with data from clinical trials demonstrating an increased risk of cardiovascular events when normal haemoglobin or haematocrit thresholds are passed (26–29). Moreover, in one retrospective observational study, rEPO use to increase haemoglobin levels in renal transplant recipients was associated with increased mortality risk, suggesting careful dosing is required (30). In the past few years, iron therapy has had a renaissance, with clinical studies showing the benefit of more aggressive iron therapy in renal patients (CKD and dialysis), improving clinical outcomes whilst reducing rEPO use and blood transfusions, in the absence of increased infection rates (31, 32). In the setting of transplantation consideration must also be given to a recent meta-analysis of the administration of intravenous iron, which suggested its use to be associated with enhanced infection risk (33). Although this study was limited by standardised definitions of infection, the role of iron in the post-transplant period, when infection risk is high is not clear. Irrespective of its role in the post-transplant period, adequate iron therapy in the pre-transplant setting is of upmost importance, and proactive iron for those patients on the wait-list now seems a sensible approach to optimise patients prior to transplantation (32). In addition to administering adequate iron supplementation, it is also imperative to investigate, where indicated, occult causes of blood loss and iron deficiency; together with other vitamin deficiencies and causes of anaemia prior to transplantation.

Even in patients fully optimised pre-transplant, the need for blood transfusions will never be eradicated, and although not addressed in this study, the medium- and long-term impact of transfusion induced allosensitisation in kidney transplant risk is still an area of concern that warrants further investigation independently. The transplant community recognises that the prevention of allosensitisation is vital, given the lack of effectiveness of therapeutic interventions, both for desensitisation and treatment of antibody mediated rejection. We acknowledge that an issue remains of the conflicting evidence of the relevance of post-transplant transfusions on de novo HLA and donor specific antibody (DSA) development in the literature (10, 11, 13–15, 17, 34, 35). However, the major limitation of reported series include small sample size, short duration and the clinical characteristics of the patient cohorts studied. We have previously shown that transplant recipients can develop HLA antibodies against blood donors, and if the blood and transplant donor shared HLA mismatches with the transplant recipient, then clinical outcomes were worse (9). We hypothesise, that in the UK, where the majority of the blood and transplant donors are of white ethnicity, but there is a disproportionately high number of non-white kidney transplant recipients, this may contribute to the greater impact of allosensitisation (36, 37). Given we have reported that the risk factors for transfusion includes non-white ethnicity, and that it is recognised that non-white transplant recipients have inferior allograft survival compared with white transplant recipients in multi-ethnic communities, this again supports the call for research in this area.

This study would have been strengthened with additional data related to immunosuppression type, rejection and de novo DSA. Given the median time to transfusion was 4 days post-transplant, we did not employ a time-varying coefficient model, which is recognised as a further potential limitation. Key data on haemoglobin and iron levels are unavailable, and further information on causes of graft loss and death, may have helped to delineate cause from association. However, the multicentre nature of this study has highlighted the high frequency of transfusions is a common complication, and the impact of graft and patient survival at 12 months more consequential than previously considered. It needs to be acknowledged that this is not the first study to show an association of transfusion with graft and patient outcomes, however, it’s strength is the multicentre approach, incorporating analyses on both donor and recipient characteristics using granular data to report on multiple clinical outcomes including allograft function (9–13, 18, 38).

To summarise, PBM applied to general surgical procedures has been shown to reduce the need for transfusions and improve patient outcomes by lowering morbidity and mortality (1, 2). In the field of kidney transplantation, where patients at high risk of anaemia in whom the potential consequences of transfusion could be life long, evidence to aid PBM is relatively scarce. We have shown, in this multi-centre study, that blood transfusions are common and are independently associated with inferior allograft and patient outcomes at one-year post-transplant. We urge that further research is urgently needed to understand the mechanisms behind the outcomes, to help shape and develop transplant specific anaemia guidelines, responding to a call to action for PBM implementation by the WHO.

The data analysed in this study is subject to the following licenses/restrictions: The datasets generated and analysed during the current study are not publicly available due to analysis being performed as part of a service evaluation within NHSBT of routinely collected clinical data. Requests to access these datasets should be directed to m.willicombe08@imperial.ac.uk.

The data in this manuscript was acquired as part of routine clinic care, and analysed as part of a service evaluation, and obtained institutional approval from NHSBT under clinical governance arrangements, which does not need research ethics committee review under Health Research Authority guidance (39). As the data were obtained from routinely collected clinical information by the care teams, individual consent was not required.

SH, MW, FR, CB and DR conceptualised the study. All authors contributed to the research methodology. SR, DF, MM, DR, FR and LM participated in data acquisition. LM performed the data analysis. SH, MW, FR, DR wrote the initial draft. All authors reviewed and edited the manuscript.