The selection and preparation of red cell components for intrauterine transfusion: A national survey

The practice regarding the selection and preparation of red blood cells (RBCs) for intrauterine transfusion (IUT) is variable reflecting historical practice and expert opinion rather than evidence‐based recommendations. The aim of this survey was to assess Canadian hospital blood bank practice with respect to red cell IUT.


INTRODUCTION
Intrauterine transfusion (IUT) of red blood cells (RBCs) for the treatment of foetal anaemia enables the delivery of a healthy baby as close to term as possible [1,2].These treatments are typically performed at tertiary care centres, by high-risk foetal maternal specialists and involve careful coordination between the clinical team, the hospital transfusion medicine laboratory and blood provider.Indications include foetal anaemia secondary to parvovirus infection, alpha thalassemia major, fetomaternal haemorrhage and foetal-foetal transfusion syndrome; however, most IUTs are performed in the setting of maternal alloimmunization to prevent complications of haemolytic disease of the foetus/newborn (HDFN) [3].With recent advances in the prevention of HDFN, this procedure is performed less frequently [4][5][6].
Literature surrounding optimal product selection and preparation of transfused components, pre-transfusion testing and posttransfusion monitoring in the context of IUT is limited.Practice tends to reflect historical tendencies and expert opinion with a limited grounding in evidence-based recommendations.Although various groups have included recommendations for IUT as part of their neonatal transfusion guidelines [7,8], there is significant variation in international practice [9].In addition, post-transfusion traceability is problematic as the transfusion event is typically documented in the maternal blood bank file and must be transferred to the neonatal record post-delivery, which may occur in a different health care facility.
The goal of the present study is to better understand the scope and variation of IUT transfusion practice in Canada related to the selection and preparation of red cells for the procedure, foetal pretransfusion testing and processes for post-transfusion traceability.

METHODS
In Canada, IUT is performed in nine highly specialized regional or provincial health care centres serviced by one of two blood suppliers (Héma-Québec in the province of Quebec and Canadian Blood Services for the rest of the country).Referral patterns across the country vary with some provinces having a single specialized centre (British Columbia, Alberta, Manitoba, Canada), whereas the more populous provinces such as Ontario and Quebec have two or three sites (respectively).The Atlantic region of eastern Canada is served by a single centre in Halifax.Referrals may be made to the nearest geographic centre even when this crosses provincial boundaries.
See site-specific data in Figure 1.

RBC selection
All sites select RBCs that are (1) allogeneic (although several noted that they would consider maternal autologous red cells in certain rare blood situations), (2) Group O, RhD negative (with the exception of cases involving maternal anti-c antibodies, see below), (3) CMV seronegative and (4) fresh (<5, ≤5 or <7 days depending on the site, Table 1).Most sites (8/9) test and select for haemoglobin S negative units.In Canada, all cellular components undergo pre-storage leukoreduction by the manufacturer.No sites assess or select components based on RBC isohaemagglutinin titres.

Prophylactic antigen matching to the maternal phenotype
All sites select units that are cognate antigen negative but the degree

Preparation of RBC for IUT
Processes employed in RBC preparation are variable (Table 1).All sites used the same method of RBC preparation regardless of indication, and RBCs were processed by the hospital transfusion service.Preparation of RBC requires the removal of plasma and additive solution to concentrate the final haematocrit of the transfused component.One site used an automated method (Terumo Cobe 2991), whereas the remainder (n = 8) employed a manual process that involved centrifugation of the red cell unit and manual extraction of the supernatant to the final target haematocrit (Hct).All sites mandate a target Hct >0.7 though the range and acceptable upper limit varied with two sites not specifying an upper limit.These two sites indicated that they do not routinely check the final haematocrit after concentrating on the red cells, but note that the upper limit is often as high as 0.97.Two sites noted that a wash step was employed in component preparation: the site using the automated method for RBC preparation noted that a wash step is integral to the processing method and one other site routinely performs a manual wash of the red cell component.Four sites reconstitute with saline following concentration, whereas five sites only concentrate the RBC without the use of any additional solution for reconstitution.No sites use plasma for reconstitution.
All sites irradiate the unit as close to transfusion as possible and always less than 24 h before infusion as per Canadian guidelines [11].
Four of nine centres did not have an irradiator on site and relied on the blood supplier to perform irradiation.Two sites filter the component (standard blood filter) prior to issuing the unit from the blood bank.

Foetal sample testing
Testing on the foetal sample varies across sites (Table 2).No sites perform additional foetal red cell phenotyping beyond the cognate antigen and none routinely performs genotyping on a foetal sample.Two sites do not perform ABO/RhD and cognate antigen typing on the foetal sample.Testing for HbF is performed at five sites by various methods and for various reasons including confirmation of foetal origin of the specimen and measurement of HbF on a post-transfusion sample to help predict future requirements for IUT.

Foetal RBC administration
Two of nine sites indicated that a blood warmer was used at the time of infusion.One site noted that this would be considered in select cases and one site specified that the unit was warmed to room temperature before infusion.The volume of blood transfused to the T A B L E 1 Selection and preparation of red blood cells (RBCs) for intrauterine transfusion (IUT) across nine Canadian hospital transfusion laboratories (sites anonymized).For the method of RBC concentration, most sites use a manual method involving unit centrifugation and supernatant reduction.
foetus was determined by the clinical team; 6/9 sites supplied a formula or reference for the calculation method (Appendix S2).A variety of methods based on the estimated placental blood volume, foetal blood volume and gestational age were used to calculate the volume transfused including the Mandelbrot technique [12], the Giannina coefficients based on gestational age [13] and a web-based calculator (Perinatology.com)[14].No transfusion reactions in either mother or foetus were reported to the blood bank following an IUT during the time period surveyed.

Traceability of red cell components for IUT
For transfusion traceability, six of nine sites stated that the IUT was documented in the maternal file in the hospital electronic laboratory information system (LIS) at the time of transfusion with manual transfer of this information or linking of the maternal file to the neonatal file in the LIS after delivery (Table 3).Review of the maternal blood bank file at the time of neonatal transfusion testing was noted to be standard practice at these sites.In three of nine sites, the IUT was documented in both a maternal and foetal electronic LIS blood bank file at the time of the IUT, and the foetal file was electronically merged with the neonatal file after birth.Two sites provided information regarding IUT to the patient as a wallet card or patient letter.
Two laboratories from hospitals performing IUT provided direct communication to the laboratory at the birth hospital or local hospital where the neonate may receive care.All other sites relied on the clinicians at the hospital performing IUT to communicate with the clinicians at the birth hospital.T A B L E 3 Post-intrauterine transfusion traceability and communication strategies to ensure optimal neonatal transfusion management after birth.

Site Transfusion traceability Communication
A IUT documented in maternal LIS blood bank file.Results of testing on the foetal sample documented on a paper file card that is kept with the maternal blood bank file card for linkage after delivery.

DISCUSSION
The results of this study highlight the significant variation in red cell IUT practice across nine Canadian centres.Practice variation has similarly been noted in other recent publications [9,15].The lack of evidence-based research regarding the optimal selection and preparation of red cell components for IUT and the lack of standardized national guidelines are contributing factors.The survey reveals that prophylactic antigen matching to the extended maternal phenotype has not been widely adopted in Canada when selecting red cells for IUT, despite evidence that this may be beneficial to prevent further maternal alloimmunization [16][17][18].This study highlights some of the challenges in post-IUT transfusion traceability and presents strategies used by Canadian hospitals to address this.
The number of IUT performed at each site varied significantly with higher numbers being performed in more populous centres (Figure 1).IUT is a highly specialized procedure that requires operator expertise and advanced transfusion medicine support [19].In a geographically large country such as Canada, this consolidation of expertise must be balanced against accessibility for patients, especially as the need for IUT can arise suddenly and sometimes unpredictably.
Requirements for RBC selection and preparation along with foetal pre-transfusion testing tend to be based on local historical practices, and given the low frequency of IUT events at many sites, there may be limited scope to critically appraise and revise practices.National guidelines for optimal IUT transfusion practice do not exist in Canada.
Standardization of practice through the development of national guidelines would be especially valuable for centres that do not frequently perform this procedure and would assist the blood supplier in planning for IUT transfusion needs.
Several studies have documented the risk of additional maternal alloimmunization following RBC IUT with rates of new antibody formation ranging from 22% to 26% [16-18, 20, 21].Given the fact that maternal alloimmunization is the most common indication for IUT [22] and these transfusion recipients have declared themselves to be immune responders [23], prophylactic extended maternal phenotype matching of transfused red cells (Kidd, Duffy and Ss in addition to Kell and Rh) has been advocated as a strategy to reduce this risk.While the relative contribution of paternally inherited foetal antigens is difficult to quantify, in a retrospective review of 63 mothers with foetuses requiring IUT (173 units transfused, 18 newly identified antibodies post-IUT), researchers noted that in 12 of 18 (67%) of cases, the donor units were antigen positive for maternal antibodies that subsequently developed [16].Schonewille et al. noted that despite Rh and Kell antigen matching, in 33 cases of immunization to non-Rh/Kell antigens where both the foetus and donor phenotypes were known, 17 (65%) were attributed to donor stimulation [17].The presence of newly identified non-Rh/Kell antibodies complicates the ability to source antigen-negative units for subsequent IUTs, may impact the severity of foetal haemolysis and transfusion interval in the current pregnancy and contribute to the risk and severity of HDFN in future pregnancies.
In the current survey, only two of nine Canadian sites documented a deliberate strategy for extended antigen matching to the maternal phenotype.One site routinely attempted to provide extended maternal phenotype-matched units (Rh, Kell, Kidd, Duffy and Ss), whereas the second considered this in cases with more than one maternal antibody.The remaining sites provide D negative, Kell negative and cognate antigen negative units.In contrast, 6 of 10 respondents in an international survey indicated that they would try to provide additional maternal phenotype-matched units beyond Rh and Kell if possible [9].In Canada, there is no guideline recommending extended phenotype matching of RBC for IUT; however, guidelines recommending the use of CMV seronegative components for IUT are well established [24].This may inadvertently lead to the prioritization of CMV seronegativity over prophylactic extended antigen matching in the hierarchy of RBC IUT selection criteria.In a recent International Forum [9] focused on IUT red cell selection practice, half of responding sites indicated that leukoreduced components are considered CMV-safe for IUT.Since the time of this Canadian survey, at least one site has abandoned the requirement for CMV seronegative units in the setting of IUT.
Information about the history of foetal transfusion is essential to avoid misinterpretation of the neonatal blood group and ensure the need for irradiated cellular components after birth is documented.
Given that IUT are performed in only nine Canadian centres, it is not uncommon for patients to deliver at a hospital different from where the IUT was provided.The use of patient letters, wallet cards and direct lab-to-lab communication helps to ensure continuity of care in the absence of linked LISs.However, even for deliveries occurring within the same institution, one mechanism for accurately transferring the foetal transfusion history to the neonate is through the creation of a unique foetal identifier in the LIS, as employed by three of nine institutions.In the post-survey discussion, many barriers to the creation of a foetal blood bank file in the LIS were highlighted including technological barriers in the LIS, ethical and legal concerns.Traceability is a key concern for blood bankers and is increasingly relevant for other specialties as foetal medical testing and procedural interventions are increasingly employed [25].
Strengths of this study include its ability to capture the breadth of Canadian practice and the collaboration and discussion that this survey fostered.Comparison was made between Canadian and international practice using the same questions employed in the international forum [9].By identifying gaps in best practices, this study paves the way for the development of national standards and benchmarking between transfusion services.Limitations of this survey include constraints in the topics assessed and inherent limitations of the survey process.Data regarding the number of IUTs performed in each pregnancy and the number of IUTs performed for different indications were not captured.
In conducting this survey, centres from across Canada compared intrauterine RBC transfusion practices highlighting the need for standardization and a national guideline.This study has prompted a re-examination of priorities for RBC selection for IUT and highlighted possible strategies for transfusion traceability in this unique setting.

A
survey was electronically distributed to the transfusion service medical directors from all hospital sites known to perform IUT in Canada in November 2020 with follow-up reminders sent monthly until March 2021.A virtual follow-up meeting was conducted to clarify responses.Respondents had the opportunity to review the final extracted data for accuracy.Formal ethics review was waived given the aggregate nature of the data collected.The survey addressed key areas of blood bank practice around intrauterine RBC transfusion including (1) the number of RBC IUT events performed annually for the years 2018, 2019 and 2020; (2) selection criteria for RBC including ABO/RhD blood group, phenotype requirements including Kell status and the degree of matching to the maternal phenotype, Cytomegalovirus (CMV) and sickle cell status; (3) RBC processing for IUT including washing, concentration, target haematocrit and irradiation; (4) administration details including transfusion volume calculations and blood warmer requirement; (5) pre-transfusion testing performed on the foetal sample; (6) history of reported transfusion reaction following an IUT within the last 3 years; (7) laboratory documentation and post-transfusion traceability.Traceability refers to the way that the maternal data were linked to the neonatal blood bank record after birth.The final survey included 23 questions (Appendix S1).The survey was piloted for validity by a national network of transfusion specialists and adjusted accordingly.Responses to all survey questions were de-identified and summarized using descriptive statistics.Analysis was conducted using Microsoft Excel 2016.Data regarding the national and provincial annual live birth rates for 2018, 2019 and 2020 were obtained from the Statistics Canada website [10].
of prophylactic antigen matching varied.Six sites routinely provide D neg, C neg, E neg and Kell neg units (with the exception of cases involving anti-c, in which case group O, D pos, c neg and E neg RBC are chosen).Red cell units would be matched to the Rh phenotype of the mother in most patients who are RhD negative but would not necessarily be an Rh phenotype match to a mother who is RhD positive with anti-Kell, for example.The response from three sites implied a more considered approach to maternal Rh phenotype matching.One site indicated that they provide 'Rh matched' Kell neg units.Two sites F I G U R E 1 The distribution and median number (interquartile range) of intrauterine red cell transfusions performed at each centre across Canada over 3 years.Typical geographic referral regions indicated by shading.AB, Alberta; BC, British Columbia; MB, Manitoba; NB, New Brunswick; NL, Newfoundland; NS, Nova Scotia; NT, Northwest Territory; NU, Nunavut Territory; ON, Ontario; PEI, Prince Edward Island; QC, Quebec; SK, Saskatchewan; YT, Yukon Territory.selectCcEe-matched Kell neg units with one of these routinely matching for Kidd, Duffy and Ss, and the other indicating that they may consider an extended maternal phenotype match in cases with more than one maternal antibody.
Abbreviations: Hb S, haemoglobin S; Hct, haematocrit; N, no; RBC, red blood cell; Y, yes. a Four hospitals did not have an on-site irradiator and irradiation was performed by the blood supplier before issuing to the hospital.b

T A B L E 2 7 Foetal red cell cognate antigen typing a 6 Direct antiglobulin test a 7 Identification of maternal antibody in foetal sample a 1 Haemoglobin F detection b 5
Abbreviation: IUT, intrauterine transfusion.a Performed on the first IUT only.b Methods of haemoglobin F detection included flow cytometry (n = 1), point of care (POC) blood gas analyzer (n = 1), Kleihauer Betke test (n = 2) and sodium hydroxide (NaOH) test (n = 1).