Estimating the risks of prehospital transfusion of D‐positive whole blood to trauma patients who are bleeding in England

Abstract Background and Objectives D‐negative red cells are transfused to D‐negative females of childbearing potential (CBP) to prevent haemolytic disease of the foetus and newborn (HDFN). Transfusion of low‐titre group O whole blood (LTOWB) prehospital is gaining interest, to potentially improve clinical outcomes and for logistical benefits compared to standard of care. Enhanced donor selection requirements and reduced shelf‐life of LTOWB compared to red cells makes the provision of this product challenging. Materials and Methods A universal policy change to the use of D‐positive LTOWB across England was modelled in terms of risk of three specific harms occurring: risk of haemolytic transfusion reaction now or in the future, and the risk of HDFN in future pregnancies for all recipients or D‐negative females of CBP. Results The risk of any of the three harms occurring for all recipients was 1:14 × 103 transfusions (credibility interval [CI] 56 × 102–42 × 103) while for females of CBP it was 1:520 transfusions (CI 250–1700). The latter was dominated by HDFN risk, which would be expected to occur once every 5.7 years (CI 2.6–22.5). We estimated that a survival benefit of ≥1% using LTOWB would result in more life‐years gained than lost if D‐positive units were transfused exclusively. These risks would be lower, if D‐positive blood were only transfused when D‐negative units are unavailable. Conclusion These data suggest that the risk of transfusing RhD‐positive blood is low in the prehospital setting and must be balanced against its potential benefits.


INTRODUCTION
UK guidelines recommend that for females of childbearing age, group O D-negative red blood cells (RBC) should be administered if blood group is unknown, to prevent D alloimmunization, which can lead to haemolytic disease of the foetus and newborn (HDFN) in future pregnancies [1]. Group O D-negative RBC are frequently transfused in the prehospital phase of resuscitation, as it is not possible to predict in advance if the patient is going to be a male or a woman of childbearing age or determine their D type. In the last decade our improved understanding of the biology of acute traumatic coagulopathy has resulted in the development of damage control resuscitation [2][3][4][5], which advocates for the rapid and balanced administration of RBC, platelet and plasma as early as possible in the patient's resuscitation. This has improved outcomes for patients compared to the standard of care alone, especially in the prehospital phase of the resuscitation [6][7][8][9]. The need to provide early haemostatic resuscitation with plasma has awakened significant interest in the reintroduction of low-titre group O whole blood (LTOWB), due in part to logistical advantages for the clinical team, and the ability to transfuse RBC, plasma and platelets in a 1:1:1 ratio [10][11][12]. The demand for blood products to be used in the prehospital phase of the resuscitation is likely to grow and blood collectors will need to optimize their procedures to ensure sufficient blood is available for patients both in and out of hospital [13].
Currently, most air ambulances in the United Kingdom carry D-negative RBCs and either dried or thawed plasma. Supply of O D-negative RBCs is a challenge for most blood operators due to the limited numbers of donors (around 8% blood donors in the United Kingdom and United States are Group O D-negative). While in England demand for RBCs has fallen by 24% since 2014, this is not mirrored by a similar decrease in demand for O D-negative RBCs, as noted by other blood collectors [14]. Currently the demand for group O D-negative RBCs is 14% of total RBCs in England, many issued as the next best alternative when those of RhcDe phenotype are unavailable for the treatment of sickle cell disease [15]. The disparity between demand for group O-negative RBCs and the proportion of such donors in the population necessitates significant recruitment activity to enrich the donor pool to meet demand [16]. There is concern regarding the long-term sustainability of supplying group O D-negative RBCs, especially if the demand for prehospital and treatment of sickle cell disease transfusion increases. Due to limited availability, group O D-positive RBCs are sometimes used to resuscitate bleeding trauma patients especially in the prehospital phase of the resuscitation in some jurisdictions [17]. This balance of the risk of D-alloimmunization against providing early balanced resuscitation to injured patients is a particular consideration in relation to the use of LTOWB, which has a shorter shelf life than additive solution containing RBCs, may result in more wastage of the product, but may be more efficacious than current prehospital standard of care in England, that is, RBCs and plasma.
The risk of significant harm due to transfusion of D-positive blood to bleeding trauma patients of unknown D group is likely smaller than historically thought since (a) 75% to 80% of trauma patients are male [18,19], (b) the proportion of trauma patients that become D-alloimmunized is lower than values for healthy individuals [20,21] and (c) improvements in the diagnosis and treatment of HDFN have markedly reduced the foetal mortality rate where access to modern foetal-maternal care is available [22]. The aim of this paper was to model the risks of harm from transfusing group O D-positive RBC components to trauma recipients in England who are bleeding in the prehospital setting.

Model inputs and key assumptions used
We modelled the chain of upstream events that are necessary for the various harms to occur as a probabilistic graphical model ( Figure 1). For each of the events (X1-14) that must happen for one of the three harms in Figure 1 to occur, we can estimate the expected number of events seen per recipient assuming the upstream events have occurred (model inputs N1-14, Table S1). The model inputs are specified in the form of probability distributions, to reflect the level of uncertainty regarding the true value of the input. Where possible, model inputs used published data from the United Kingdom, or if unavailable then published information from the literature was used as described in Table S1.  Table 1. The main risk in this population was that of HDFN, where for every 570 transfusions, one foetal death or disability would occur (CI 260-2300), which is higher than the risk of the other two specific harms (contemporary and future HTR).

Risk estimate for D-negative females <50 years of age
The data from Table 1   at that centre [17]. During this time, 500 females of CBP would die from haemorrhage if prehospital transfusions were not available. We predicted that a case of HDFN would be expected to occur in England more frequently than at these American centres, likely due to the total higher number of females of CBP, approximately 20-fold that would be transfused throughout the country of England compared to the number transfused at a few regional trauma centres.
The strength of our study is that we have, for the first time,  The current philosophy of resuscitating trauma patients early with blood transfusion is shifting towards providing a balanced resuscitation that includes platelets in addition to RBCs and plasma, similar to whole blood [10][11][12]. Multiple donor qualifications are required in order to be eligible to donate LTOWB, which makes it challenging for blood providers to supply RhD-negative LTOWB to all trauma patients. As a result of these supply issues, the use of D-positive LTOWB and RBCs for injured patients early in their resuscitation is becoming a common practice at least in the United States; in an international survey of LTOWB practice (mainly from the United States), 27% of the respondents reported using D-positive LTOWB for females of any age [27], while a survey of American Level 1 trauma centres revealed that 51% would administer a D-positive LTOWB unit to a female of CBP whose D-type was unknown in a bleeding emergency [28].  [29]. However, in a survey of staff at a large American university, 90% of the respondents who were females of CBP indicated that they would accept a lifesaving transfusion even with the knowledge that it could harm future pregnancies [30]. Further work is needed to understand the views of patients and the general public as to whether these risk are acceptable, so that their views can be taken into account when determining policy. In addition, protocols and surveillance for follow-up of D-negative females of CBP who receive D-positive whole blood or red cells are required.