Management of alloimmunized patients


  • 5D-S41-02

Sandra T. Nance, American Red Cross Biomedical Services, 700 Spring Garden Street, Philadelphia, PA, 19123, USA

Alloimmunization in the transfused patient is relatively common, especially in the chronically transfused population. Alloimmunization is defined as an immune disorder caused by incompatibility between recipient and donor antigens.

It has been recognized that alloimmunization may compromise the therapeutic effect of transfusion. Haemolytic transfusion reactions are thought to be a common cause of immediate life-threatening events associated with transfusion [1]. While some of the most severe are due to human error in the transfusion of ABO incompatible blood, there is a recognized subset because of antibodies other than ABO.

The risks of transfusion vary between low and high human development indexes [2]. The index is based on life expectancy, literacy, participation in higher education and per capita income. Alloimmunization (1:1600) and acute haemolytic transfusion reactions (1:13 000) and delayed haemolytic transfusion reactions (1:9000) are reported in high human development index areas and variably reported in those with low human development index areas.

Global reports of incidence of alloimmunization

In Macedonia, an evaluation of the incidence of alloimmunization and of the most common antibodies was published. Antibodies were reported in 150 transfused patients studied. The most frequent antibodies were anti-K (26), anti-E (25) and anti-c (6). Women made up 78·6% of patients with alloantibodies [1]. In Uganda, of 214 transfused patients, 13 were alloimmunized (6·1%) [3]. The most frequent antibodies identified were anti-E (6) anti-S (3). The number of units transfused was significantly associated with the alloimmunization rate. It was noted that patients with malaria were less likely to develop antibodies to red cells. In a report from Saudi Arabia, the frequency of alloimmunization is 22%. The most frequent antibodies were anti-K and anti-E. It was noted that the patients who received blood from the same ethnic group did not form antibodies [4]. From India, a group led by Thakral [5] determined that the incidence of alloimmunization was 3·5% and anti-c was the most frequent, followed by anti-E and anti-M.

In the United States of America, a study of patients with concurrent alloantibodies was reported [6]. In this study, alloimmunization to multiple antigens occurred in 96 of 443 alloimmunized patients. The incidence of alloimmunization was 2·4% [7]. The most frequent alloantibodies were anti-K (21·9%), anti-E (19·4%) and anti-D (9·1%) Anti-K and anti-E were found in the largest number of antibody pairs. Anti-C or c was most commonly linked to another antibody, while anti-P1 (3/22) and anti-M (3/18) were least likely. The patients studied were almost exclusively military veterans, and thus, the antibodies were transfusion-induced instead of pregnancy related.

Diagnosis and risk of alloimmunization

A recent study reported that certain disease states seemed to be risk factors for alloimmunization [8]. Disease states that were associated were diabetes mellitus, solid malignancy and previous allogeneic peripheral blood progenitor cell transplantation. In addition, women were more likely than men to be alloimmunized.

A study in Blood detailed a mathematical model for immunogenicity of blood group antigens in transfusion in the United States [9]. In this report, the authors used only men and excluded antibodies reactive at room temperature. There were 462 of the total 540 antibodies directed at clinically important antigens and 10 were reactive at room temperature, leaving 452 antibodies to be analysed. Based on this data, the authors revised Giblett’s [10] model for interracial transfusion to include disappearance of antibodies and pregnancy-induced antibodies. In Tormey and Stack’s work, the immunogenicity is dependent on number of antibodies of a specificity, probability of exposure (patient being antigen negative and donor being antigen positive) and persistence of antibody. The analysis led to the immunogenic order that follows (D is excluded): K>Cw>Lua>Jka>E>Lea>P1>c>M>Leb>Fya>C>S in transfused men having IgG alloantibodies. A report from the Netherlands related to disappearance of antibody suggests that antibody screening testing at set times after transfusion would reduce the risk of not detecting an antibody later because of loss of reactivity which may be associated with a subsequent delayed haemolytic transfusion reaction [11].

There are certain populations of patients for whom alloimmunization has extreme consequences, pregnant women and patients with sickle cell disease (SCD). For these, consequences can be fetal demise or a clinical state of untransfusability.

Alloimmunization and management in pregnancy

A number of publications have covered alloimmunization in pregnancy. A study in New York, USA, in 1997 to determine the frequency of RBC alloimmunization capable of causing haemolytic disease of the fetus and newborn (HDFN) showed that anti-K was the most frequent antibody (22%), followed by anti-D (18%) and anti-E (14%) [12]. The authors promoted the prenatal determination of fetal antigens.

Reddy et al. [13], in a report on an international consensus on stillbirth classification that for red cell alloimmunization to be considered as a cause of stillbirth, the following should be present: maternal antibody against fetal antigen, antibody titre of ≥ 1:8 for anti-D, for all others ≥ 1:16, clear evidence of fetal anaemia with hydrops and evidence of fetal extramedullary hematopoiesis.

Alloimmunization in pregnancy in Sweden from years 1992 to 2005 was the subject of a report in 2008 [14]. In the study, 335 IgG alloimmunizations were reported, 41 prophylactic anti-D were also detected. Anti-D accounted for 120 (40%), other Rh antibodies 128 (40%) with 77 being anti-E, anti-K 28 (3%) and other antibodies 59 (19%). Maternal plasma exchange and/or IVIG in 12 cases, 10 of which were severe anti-D and two anti-E were reported. There were no deaths, two hydropic newborns who survived, 29 newborns were transfused and 23 received phototherapy alone. All infants were purposefully delivered 2 weeks before due date. Swedish guidelines call for newborn exchange transfusion with haemoglobin of < 120 g/l and/or bilirubin > 110 umol/l as well as rising bilirubin or active haemolysis at birth.

Two publications, one from Tunis and one from Brazil [15,16], reinforced the importance of Rh Immune Globulin in pregnant D-negative women, use of Doppler ultrasound studies of the fetal middle cerebral artery and use of intrauterine transfusion.

Another report from Brazil focused on non-anti-D antibodies in alloimmunization in pregnancy [17]. This study compared perinatal outcomes in 15 non-D cases (seven were Lewis antibodies) and 55 D-negative alloimmunized patients. The rate of intrauterine transfusion was the same in both groups. The recommendations focused on the management of patients such as avoiding oxytocin, performing amniotomy at the proper time, avoiding manual removal of the placenta and uterine massage and unclamping the umbilical cord while awaiting placental detachment. A comprehensive report from Moise [18], the United States, suggested that K and k immunization be treated more conservatively than other non-D antibodies. This report also detailed the change in alloimmunization in pregnancy in the United States over 30 years.

A review article by Moise [19] discussed management of Rh alloimmunization in pregnancy and focused on Doppler ultrasonography, the use of DNA testing to predict phenotype, intrauterine transfusion and suppression of erythropoiesis associated with repeated neonatal transfusions. Two cases were reported in the same issue of the journal that focused on profound anaemia caused by maternal anti-K and anti-D [20].

From Canada, a report [21] reviewed the equation for fetal transfusion needs because of alloimmunization with emphasis on improvement on prediction of haematocrit in those fetuses younger than 32 weeks gestational age.

From the Netherlands, a report of high additional alloimmunization after RH- and K-matched intrauterine transfusions is concerning for haemolytic disease of the fetus [22]. The conclusion was a recommendation to match more completely for maternal antigens (RH, K, FY, JK, S) to prevent further alloimmunization through intrauterine transfusion. In a report from Denmark, there was a suggestion that combinations of antibodies could be potentially more harmful through a synergistic effect [23].

Alloimmunization and management sickle cell disease

Chronically transfused patients have increased exposures to non-self antigens. In sickle cell disease, treatment options are limited for many patients, and maintenance of haemoglobin A by transfusion is helpful in abating clinical sequelae. Some patients receive chronic intermittent transfusion. These transfusions are not without risk of transfusion acquired infections, transfusion reactions, sickle cell events, platelet refractoriness and alloimmunization. Many reports have focused on transfusion of leucoreduced phenotypically matched or racially matched blood. That and careful monitoring of the patients is recommended to make transfusion therapy safer [24]. At the minimum, the recommendation is for limited phenotype matching for C E and K antigens in the United States.

Data from Saudi Arabia in a review of sickle cell cases from 1996 to 2004 on 350 patients receiving at least one transfusion showed an alloimmunization rate of 13·7% [25]. The most common antibodies were anti-E (19%), non-specific and inconclusive (25%), anti-K (10%), anti-c (6%).The recommendation was for leukoreduced, and haemoglobin S-negative red cells matched for Rh and K antigens.

A report from Kuwait of patients with sickle cell disease who either received ABO/Rh-matched non-leucoreduced red cells (group 1, n = 110) or received ABO/Rh leukoreduced Rh- and K-matched transfusions (group 2, n = 123) [26]. Those who were in group 1 had a 65% alloimmunization rate with 100 antibodies and those in group 2, 24% alloimmunization rate (48 antibodies). Of interest is that in group 2, there were 38 antibodies to C, c E e and K and only 10 others (in FY, JK, MNS systems and P1). It was noted that patients in group 2 were less likely to develop antibodies in the other blood group systems.

In Uganda, the alloimmunization rate in patients with sickle cell disease is 6%. It was noted that there is homogeneity between donors and patients and a low transfusion rate in Uganda [27].

In France, molecular methods were used to determine partial C status of patients with sickle cell disease [28]. Forty-nine of 177 C+ patients carried abnormal C antigen as determined by molecular studies. Eleven of 37 partial C patients who received transfused formed anti-C. A comment was made regarding the limited availability of additional C-units for partial C patients.

One case report highlighted the usefulness of intravenous immunoglobulin and steroids to correct severe anaemia caused by hyperhaemolysis [29]. A report from the United States on HLA alloimmunization in sickle cell disease showed that HLA antibodies were detected in 25/73 (34%) of paediatric patients [30]. Of those with red cell antibodies, 16/30 (53%) had HLA antibodies.

Alloimmunization and management in other diseases

Alloimmunization in other disease states has been reported. There is a report on patients with thalassemia in Taiwan, where the alloimmunization rate was reported to be 37% [31]. And in Iran, the rate was reported to be 5·7% [32]. A report from Malay showed an alloimmunization rate of 8·6% [33].

In liver transplant, in a study of 2000 patients in the United States, there has been a report of alloimmunization rate of 5·7%, with over half against Rh antigens [34].

Another report, also from the United States, the authors concluded that alloimmunization from blood transfusion along with immune modulation greatly affected the survival of patients after liver transplantation [35].

A report of hemato-oncology patients also indicated that these chronically transfused patients could also benefit from antigen matching [36]. A similar report in 2006 from the same group reviewed 20 years of retrospective data on repeatedly transfused patients and suggested that those that were alloimmunized could benefit from a greater degree of phenotype/genotype matching. These authors also presented an immunization risk per antigen specificity in patients already sensitized [37].

There is some thought that other patient populations (like those with solid malignancies, previous allogeneic hematopoietic stem cell transplant and history of diabetes) may have increased risk of alloimmunization. The number of transfusion exposure plays a role in the rate of alloimmunization. There have also been a number of studies evaluating the immune response and the role of various mediators. Inflammation may have a role in alloimmunization as reported in a mouse model [38,39]. Other thoughts include febrile reactions to platelets, splenic antigen presenting cells and storage of red cells [40–44].

Preventing alloimmunization

The articles referenced thus far have recommended that in the chronically transfused population, in mothers already alloimmunized, and for some other diagnoses, selective matching of the patients blood type may prevent alloimmunization.

For some patients, sensitization to even the easiest blood group antigen (like K) is devastating. The pregnant patient with an alloantibody to an antigen the fetus possesses is cause for alarm. Some forward thinking countries have established that women receive K- and c-negative blood until beyond childbearing years. Through unmatched transfusion, K and c can cause haemolytic disease of the newborn. Most countries match for D.

An important point in the management of the alloimmunized patient is the prevention of alloimmunization especially in the two groups previously mentioned. The ‘side-effect’ (allo-antibody) of an unmatched transfusion that results in the patient being less able to receive treatment (red cell transfusion) that other patients can receive should be considered in the selection of blood used for transfusion. Molecular and racial matching should be evaluated in the patient with sickle cell disease. Transfusion of blood negative for c and K should be required for women until their childbearing years have ended.

A step further is to utilize molecular typing test results to match the patient’s type with a suitable donor. This requires that a sufficient inventory of units with molecular typing results be available. Several centres are in the process or are using molecular results for the prevention of alloimmunization. Some recent articles are given as references [45–47].

Evaluation of alloimmunized sera

Most often, routine pretransfusion antibody detection tests are negative. Infrequently, the antibody detection test is positive. Most often, when the antibody detection test is positive, the resolution is easily performed using one or two panels. Rarely, the one or two panels do not provide a specificity(ies), and further work must be carried out. The most complex cases often involve sera reactive with all red cells tested except their own. These may be because of an antibody to a high prevalence antigen or antibodies to multiple common specificities. An article from France showed a problem solving strategy in such cases [48]. Often, in complex cases, phenotyping or genotyping of the patient’s red cells is used along with antibody identification techniques such as chemical treatments of reagent red cells and adsorption studies to separate the specificities when multiple specificities are present. Determination of the clinical significance of antibodies is important in cases with multiple antibodies or an antibody to a high prevalence antigen making compatible red cell products difficult to find. Some factors contributing to clinical relevance are: temperature of reactivity, IgG subclass and reactivity in the monocyte monolayer assay. Patient-specific factors that may contribute to alloimmunization potential are less well known. Patients with SCD present challenges in the frequency of alloimmunization and the nature of the antibodies formed. People of Black African descent more commonly have partial RH antigens (partial e and D). Testing for these antigens has become more accurate with the advent of molecular testing. There are facilities who match these patients known to have partial antigen expression with compatible type donors.

Once the antibody specificity(ies) is(are) known, the quest to provide blood products is initiated. This activity involves the primary blood supplier, regional or national donor registries and can be international. The WHO International Rare Donor Panel is available and the ISBT Working Party for Rare Donors supports the Bristol centre (International Blood Group Reference Laboratory) that manages the programme.


Efforts should be made to limit alloimmunization. The high rate seen in some patients impairs the effectiveness of the transfusion and in severe instances, may make the patient require such rare blood as to be ‘untransfusable’. Avoiding all red cell alloimmunization is likely not possible. Patients with clinically significant antibodies should receive antigen-negative blood. Sometimes blood is not available locally or regionally or even nationally. Programmes should be in place or started to provide antigen-negative blood. Often these are termed Rare Donor Programmes. Some countries have national programmes; others have regional programmes that collaborate. If there is no blood in the country, the WHO International Donor Program, operated by the International Blood Group Reference Laboratory in Bristol UK, is a resource to find rare blood all over the world. Members of the ISBT Working Party for Rare Donors are also an international resource.