Objectives: To evaluate the utility of a centralized transfusion service model in preventing the transfusion of incompatible units in patients with sickle cell disease (SCD).
Methods: The serologic records of transfused patients with SCD were reviewed. The index hospital was where an alloantibody was initially detected.
Results: In total, 150 patients with SCD were evaluated; 66 (44.0%) of 150 were alloimmunized. In 42 (63.6%) of these patients, 1 or more antibodies evanesced. The median number of hospitals visited by patients with SCD for RBC transfusion with 1 or more evanesced antibodies was three (range, one to eight); the median number of nonindex hospitals was two (range, one to seven). Of the patients with evanesced antibodies, 28.6% received transfusions at various nonindex hospitals 20 or more times after the antibody evanesced.
Conclusions: A centralized database can help identify patients with SCD who have evanesced alloantibodies and prevent issuing incompatible RBC units.
Centralized transfusion service
Red blood cell
Delayed hemolytic reaction
Patients with sickle cell disease (SCD) are transfused with RBCs over the course of their life with simple transfusions and manual or automated RBC exchange procedures as the mainstay of treatment. They frequently become alloimmunized due to receipt of a large number of RBCs over their lifetime, their underlying inflammatory state, and the ethnic mismatch that typically exists between donors and recipients.1–3 The prevalence of RBC alloimmunization in patients with SCD who have received at least one RBC transfusion is reported to range between 7% and 30%.1,3–6 Up to 37% of these RBC alloantibodies “evanesce” or decrease in titer to below detection thresholds at some point in their life.1 A patient with an antibody that has undergone evanescence is potentially in danger of receiving RBC units that express the corresponding antigen, which could lead to an anamnestic antibody response and a delayed hemolytic transfusion reaction (DHTR). These DHTRs are also dangerous in this population, since they may provoke other clinical sequelae and/or hyperhemolysis.7 Thus, it is critical to maintain accurate documentation of the recipient’s alloantibody history and review it prior to selecting an RBC unit for a compatible transfusion.
Patients with SCD benefit from special manipulations and enhanced testing of their blood products (eg, leukoreduction and hemoglobin S negativity), additional steps that might not be performed outside of SCD treatment protocols. Thus, knowledge of the patient’s diagnosis, special blood product processing requirements, and serologic history in a database that is accessible at the location where RBCs are prepared for transfusion would facilitate selection of the safest unit for transfusion.
The Institute for Transfusion Medicine in Pittsburgh, PA, and the Puget Sound Blood Center in Seattle, WA, each manages regional centralized transfusion services (CTS), which service a total of 35 hospitals and health care facilities covering most of the population in their large metropolitan areas. Both CTS systems maintain an electronic database of the serologic and component modification requirements of all recipients who have had pretransfusion testing performed at a hospital within their networks. The entire database is accessible in the transfusion service laboratories at all CTS member hospitals within the two networks.8 Some of the patient safety benefits of having each recipient’s complete transfusion record available at every hospital within a CTS network have been described and include enhanced detection of both “wrong blood in tube” errors and increased detection of orders for blood products that are inconsistent with the patient’s component modification requirements.9,10
The goal of this study was to quantify the frequency at which a patient with SCD was at risk of experiencing a DHTR because of an evanesced antibody that was previously detected during pretransfusion testing at a facility other than the one currently being accessed.
Materials and Methods
From the electronic database of these two CTS systems, a list of patients who were 18 years or older with the diagnosis of SCD and transfused with at least one unit of RBCs during the calendar year of 2012 (Pittsburgh, CTS 1) and 2011 through 2012 (Seattle, CTS 2) was generated. The diagnosis of SCD for the patients on this list was compared with their hospital electronic medical record to verify its accuracy. For each patient with SCD, a complete list of every RBC transfusion that he or she had received at any CTS member hospital since 1999 (CTS 1) and 1997 (CTS 2), as well as the name and date(s) of the hospital(s) where the patient was transfused, was also generated from the CTS records. Given the large geographic distance between CTS 1 and CTS 2, it was assumed that no patient existed in both databases.
For all patients with SCD, the following data were collected: the date of their first and most recent RBC transfusions at a CTS member hospital, the total number of RBCs that they had received at hospitals within the CTS network, the number of different hospitals within the CTS network at which they had received at least one RBC transfusion, and the number of times and dates that they were admitted to each hospital for an RBC transfusion. Each patient’s serologic profile in the CTS database was evaluated to determine if any alloantibody, excluding anti-A1 in A subtype individuals, had ever been detected. Warm and cold autoantibodies were not included. Over the years, a variety of antibody detection methods have been used at each CTS, including manual tube methods, as well as automated and manual gel and solid-phase systems. Furthermore, if a patient presented at a CTS member hospital with a card or note from an out-of-network hospital indicating that an antibody had been detected, then its specificity was honored regardless of whether the antibody screen was negative at the time of transfusion at a CTS hospital. In such a case, the hospital at which the patient informed the transfusion service about his or her preexisting antibody was considered the index hospital.
For patients who had become alloimmunized, the date of antibody detection, the antibody’s specificity, and the name of the hospital at which it was detected (the index hospital) were recorded. If the alloimmunized patient presented for an RBC transfusion at a CTS member hospital on any subsequent date(s), the antibody’s activity status (ie, whether it was detectable) was also recorded. In this way, each alloantibody’s testing characteristics (presence or absence) was tracked from the initial detection through December 31, 2012. An antibody was defined as having undergone evanescence if it had been detected at the index hospital but had become undetectable (ie, the patient had a negative antibody screen) when the patient subsequently presented for pretransfusion testing at either the index hospital or a different hospital (a nonindex hospital). For those patients whose alloantibodies had evanesced and had not reappeared by the time the databases were accessed for this study, the maximum number of admissions for RBC transfusions at nonindex hospitals was capped at 20 for the purposes of this study.
For more than 20 years at each CTS, RBCs have been provided to patients with SCD using the following strategy: on the first presentation for an RBC transfusion, an extended serologic phenotype is obtained (CTS 1: D, C, c, E, e, K, Fya, Fyb, Jka, Jkb, M, N, S, s; CTS 2: D, C, c, E, e, K). More recently, the recipient’s RBC genotype also has been obtained (BioArray; Immucor, Rodermark, Germany) at each CTS. Patients who are not alloimmunized receive RBCs that are matched to their Rh and Kell phenotype. For patients with SCD who become alloimmunized, antigen-negative RBCs are provided, with every effort made to ensure that these RBCs are also fully matched to the patient’s extended phenotype.
Descriptive statistics were used to analyze continuous variables. This protocol was approved by the University of Pittsburgh’s Quality Improvement Board, a division of the Institutional Review Board, and the medical record review was approved by the University of Washington’s Human Subjects Division.
Of the 150 patients with SCD evaluated, 66 (44.0%) were alloimmunized Table 1. A total of 177 alloantibodies were identified in these 66 patients, of which 90 (50.8%) of 177 evanesced over the duration of this database review Figure 1. Of the 66 alloimmunized patients, 5 (7.6%) were sensitized to –D, 17 (25.8%) were sensitized to –C, 1 (1.5%) was sensitized to –c, 22 (33.3%) were sensitized to –E, no recipients were sensitized to –e, and 24 (36.4%) were sensitized to –K.
The median number of hospitals at which all patients (n = 150) received an RBC transfusion was two (range, one to eight) Figure 2. The median total number of hospitals at which patients with alloantibodies (n = 66) received an RBC transfusion was two (range, one to eight), and the median total number of hospitals at which alloimmunized patients with SCD who had evanesced antibodies (n = 42) were transfused was three (range, one to eight). The median number of nonindex hospitals (ie, not the hospital at which the antibody was first detected) at which alloimmunized patients with one or more evanesced antibodies received an RBC transfusion was two (range, one to seven).
Alloantibodies identified in this study and the associated proportion that evanesced. Antibodies within the same system are grouped together unless specified. The “Other” group contains 12 antibodies with undetermined specificity, 5 anti-Le, 3 anti-Di, and 1 each of anti-Lu, anti-Mg, anti-Sda, and anti-I.
Of the alloimmunized patients with at least one evanesced antibody, 38 (90.5%) of 42 were transfused at one or more nonindex hospitals Figure 3. Furthermore, 28.6% (12/42) of these patients had 20 or more separate admissions for transfusion at nonindex hospitals after the alloantibody had evanesced, placing them at the highest risk for a DHTR Figure 4. Four of 42 patients with at least one evanesced antibody were transfused at the index hospital only and thus would not have potentially benefitted from the area-wide patient database in avoiding a DHTR.
Alloimmunized patients with SCD with evanesced antibodies who receive RBC transfusions at hospitals that do not have access to their entire serologic history are at risk for DHTRs and subsequent consequences such as hyperhemolysis, an adverse event that can be life-threatening.7 In this study, most patients with SCD (90.5%) with evanesced alloantibodies visited multiple hospitals for RBC transfusions, and 28.6% were admitted for transfusions at various nonindex hospitals at least 20 times after the antibody evanesced. Therefore, having access to the recipient’s complete transfusion record at every site where a patient with SCD might receive a transfusion is important in ensuring the serologic safety of his or her RBC transfusion. A centralized database that contains the serologic history of each patient with SCD can help prevent exposure to antigen-incompatible RBC units and thus the potential for DHTRs, and it serves as a model for enhancing transfusion safety in a patient population that requires numerous RBC transfusions, potentially at different hospitals, over their lifetime.
In the absence of a CTS model, one strategy to decrease the risk of a DHTR is to have the patient with SCD carry a card or letter that lists his or her RBC phenotype and any alloantibodies produced.11 A bracelet or necklace inscribed with this information could also be used to alert the physician ordering a transfusion that the patient has RBC alloantibodies. However, cards, letters, and bracelets can be easily lost or damaged, thereby putting the patient in danger of receiving an RBC unit that is not compatible with his or her historic alloantibody profile and increasing the risk of a DHTR.
Number of nonindex hospital(s) at which patients with sickle cell disease with 1 or more evanesced alloantibodies received RBC transfusion(s). Note that patients who received all of their RBC transfusions at the same hospital at which the evanesced antibody or antibodies were first detected are not included.
Number of hospital admissions for RBC transfusion in alloimmunized patients with sickle cell disease patients with 1 or more evanesced alloantibodies after the antibody evanesced. The index hospital is that where the antibody was first detected; a nonindex hospital is one at which the alloimmunized patient was transfused but the antibody was not detected.
In addition to the CTS systems in Pittsburgh and Seattle, Florida Blood Services also operates on this model, and several other regional blood centers operate the transfusion medicine service at some of their local hospitals.12 Another strategy to decrease the risk of DHTRs in patients with SCD is to develop a regional registry of patients’ alloantibody information that is accessible at multiple hospitals where these patients can get transfused.13,14 In all cases, the patient safety benefit of the registry is proportionate to the number of hospitals that participate. Creation of regional or national registries may allow hospital transfusion services the ability to access the patient’s transfusion and serologic history in a manner similar to a CTS and effectively expand the number of hospitals at which their clinical and antibody information is available. In addition, several European countries such as Sweden and Holland have national registries that augment the safety of transfusions by supporting the sharing of recipient information across the country. However, the effectiveness of these types of registries in preventing DHTRs depends on the individual hospital transfusion services’ active participation in accessing, updating, and checking the database at the time an RBC unit is issued or whenever a new alloantibody specificity is identified. This latter event might not trigger a posting to the registry if the diagnosis of the patient or his or her listing on the registry were not known by the laboratory. The advantage of a CTS model is that patients’ alloantibody history, transfusion history, product modification requirements, and current type and screen results are all present in a single database used for routine compatibility testing that is accessible at each hospital within the CTS network they may visit for transfusion. Thus, no additional registries need to be built to ensure that a RBC unit issued for transfusion is compatible with the patients’ current and historic alloantibody profile as long as they are transfused only at CTS hospitals. This feature is particularly useful when pediatric patients with SCD grow older and are required to seek care at adult hospitals—since the pediatric hospitals in both cities are CTS member hospitals, the patients’ transfusion and serologic histories will be available when they are admitted for transfusion to an adult CTS member hospital later in life. Another strategy to help prevent DHTRs is to provide RBCs matched to recipients’ extended phenotype or genotype from their first transfusion, thereby reducing the potential for alloimmunization. However, genotyping technology might not readily available at all hospitals where patients with SCD receive transfusions and the inventory of genotyped donor units might not be sufficient to fully meet their demands.
Many of the antibodies reported in this study are to the common Rh and K antigens, despite the long-standing practice at these two CTS systems of providing Rh- and K-matched RBCs to the recipient. The most likely explanation for this unexpected finding is that these patients were transfused at a hospital outside of the CTS network that either does not routinely provide Rh- and K-matched RBCs to nonalloimmunized patients with SCD, or the hospital was unaware of the patients’ SCD diagnosis and provided RBCs as per their routine protocol. In order for prophylactic matching programs to successfully prevent alloimmunization, all hospitals that treat patients with SCD have to follow the same matching protocol; it is a potentially wasteful practice for one hospital to match RBCs for nonalloimmunized recipients with SCD if these patients are also transfused at hospitals that do not match. Another potential explanation for the large number of Rh and K antibodies is that some patients with SCD might have been urgently transfused with nonmatched RBCs in an emergency situation. Since patients with SCD often require urgent transfusions, preventing alloimmunization in this population is very important; in the absence of antibodies, RBCs can be issued very quickly in emergency situations. If the recipients had become alloimmunized, additional time would have been required to find antigen-negative compatible units. Other, probably minor, explanations for the anti-Rh and anti-K antibodies are the inclusion in this study of some patients who were transfused before Rh and K matching became routine at these two CTS networks and RBC unit or recipient phenotyping errors.
This study has several limitations. Although encompassing large geographic areas serving many hospitals, it is possible that patients with SCD received RBC transfusions outside of the networked hospitals of these two CTS systems. Likewise, it is possible that some of the older patients in this study received RBC transfusions in Seattle and Pittsburgh prior to the date of the earliest accessible records currently available at the CTS in each city. Thus, it is possible that the transfusion history of these patients, as well as the number of hospitals and admissions for transfusion, is incomplete. However, since this study demonstrated that many patients with SCD in this study were transfused at multiple sites, any such undocumented transfusions would actually further support the safety benefits of not only maintaining but also enlarging and continuously updating any preexisting centralized databases. Also, because the patients were not necessarily followed regularly with serial antibody screens, the exact time at which the antibody evanesced cannot be determined from this study, only that it did or did not evanesce. Furthermore, receipt of an RBC unit bearing an antigen to which a patient with SCD has an evanesced antibody does not guarantee that a DHTR will occur, and thus the claim that DHTRs are prevented by using the historical CTS record to provide RBC units that lack antigens to evanesced antibodies in patients with SCD is an assumption. It is currently not possible to predict which patients will experience a DHTR when reexposed to antigens. The prevention of DHTRs by a record review before RBCs are issued, as mandated by accrediting bodies such as the AABB, is a well-known and critical step in preventing patient morbidity. No DHTRs were reported for any patients in this study, although reaction reporting at the hospitals serviced by each CTS is passive and relies on recognition and reporting by the clinical team.
This study demonstrated that patients with SCD in Pittsburgh and Seattle frequently sought medical care at multiple hospitals and that a significant number of these patients had antibodies that underwent evanescence, thereby putting them at risk of a DHTR. The CTS model can help prevent DHTRs in alloimmunized patients with SCD with evanesced antibodies by providing access to the patients’ serologic history at every CTS member hospital that they could potentially visit for transfusion, thereby reducing the probability of issuing an RBC unit that is incompatible with an evanesced antibody.
This article is based on an abstract presented at the AABB annual meeting; October 2013; Denver, CO.
Sarah K.Harm, Mark H.Yazer, Grace F.Monis, Darrell J.Triulzi, James P.AuBuchon, MeghanDelaneyAm J Clin Pathol(2014)141 (2):
256-261DOI: http://dx.doi.org/10.1309/AJCP47QAAXTOZEKJFirst published online: 1 February 2014 (6 pages)