Haemolytic transfusion reactions is a well-known complication in multiply transfused patients. This reaction generally results from a conflict between antigens expressed on donor RBCs and antibodies produced by the recipient. For ABO blood group incompatibilities, the antibodies are naturally occurring rather than being the result of previous exposure to the antigens. In such situations, applying ABO compatibility rules is sufficient to prevent transfusion reactions. In cases of error, the haemolytic accident is immediate, and frequently dramatic, involving intravascular haemolysis through complement activation (Membrane Attack Complex). More frequently, transfusion reactions are delayed. Such delayed reactions result from a secondary immune response in a previously transfused or pregnant patient in whom an alloantibody has developed. Screening tests are intended to detect these antibodies, which can then be taken into account by giving RBCs free of the particular antigen. In some cases, however, the alloantibody titre may be below detectable levels at the time of pre-transfusion testing and the antibody is therefore ignored. Upon repeat transfusion, alloantibodies are restimulated, with increased cytotoxic capacities, and a switch to IgG1 and IgG3 subclasses; the titre, affinity, and capacity to bind complement all increase. The haemolytic transfusion reaction develops within 4 to 6 days. Screening tests of patient plasma and direct antiglobulin tests (which characterize the in vivo sensitization of donor RBCs by the recipient antibodies) will then give positive results. Donor RBCs are mostly destroyed by macrophages expressing immunoglobulin (Ig)Fc receptors in the spleen and in the liver. Complement opsonization may synergise IgG-mediated haemolysis through a C3b receptor on macrophages. Prevention of accidents of this type is currently based on taking any previously detected antibodies into account, and avoiding alloimmunization against the major immunogenic antigens carried by RBCs. Initially, these antigens are the RH and KEL antigens; when patients become immunized, they also include FY, MNS and JK antigens.

The characteristic features of DHTR in SCD are a delayed accident (more than 6 days after transfusion), a dramatic drop in post-transfusion haemoglobin (Hb) to below pre-transfusion levels caused by destruction of both donor and recipient RBCs, the presence of SCD-related manifestations, and exacerbation haemolysis on further transfusion. Profound reticulopenia is frequent and may contribute to the drop in the Hb concentration.

Various serological features may be associated with this post-transfusion hyperhaemolysis syndrome. The antigen differences between donor and recipient can lead to patients developing reactions related to alloimmunization due to production of clinically significant antibodies to donor RBC antigens. A specific feature in SCD patients in this standard situation is the involvement of antibodies which are generally not considered to be clinically significant (anti-M, autoantibodies) [11,12]. Also, rare blood groups or variant antigens carried only by SCD individuals of African descent are associated with the production of rare antibodies which can cause DHTR. Some cases are more enigmatic, and no relevant antibody can be detected. Many case reports or series of case reports have been published illustrating that in many cases of DHTR, the hyperhaemolysis syndrome is not due to a conflict between donor RBC antigens and the antibodies in recipients. For example, in a prospective study of 49 transfused SCD patients, two patients presented with DHTR without detectable antibodies [13]. This complication is thus unpredictable: additional transfusions exacerbate existing haemolysis, and induce severe complications of the disease.

Table 2 lists most of the publications since 1984 reporting cases of DHTR. All these publications, except our previous report, describe retrospective studies and do not indicate the incidence of DHTR in SCD. Nevertheless, they illustrate the substantial variability of the immuno-hematological presentation of DHTR. They include 70 cases, 70% involving antibodies, and in half of these cases, previous alloimmunization was documented. In many cases, the antibodies had been detected previously, but were not taken into account because they were not detected in the patient serum at the time of the transfusion. Also, in half of the cases, more than one antibody was detected. The main specificities were, as expected, RH, FY, JK, and Ss (68%). Antibodies against low-prevalence antigens were also found as well as autoantibodies alone. Antibodies that are not classically clinically significant (Lea, M, weak reactivity) were found in 18% of these cases. In almost one third of these case reports, no antibodies were detected.

In the following sections, we consider the appearance and clinical significance of autoantibodies in DHTR, and various possible reasons both for the severity of DHTR in SCD related to antibodies, and for DHTR without detectable antibodies.

Autoantibodies frequently develop with erythrocyte alloimmunization [10]. The incidence of autoantibodies among transfused SCD patients is about 6 to 10%. A similar phenomenon is also observed in cases of thalassemia, as a result of alloimmunization.

Studies in mice give some insight into the mechanism of autoantibody development, and T regulators have been implicated [14]. A frequent problem with autoantibodies is that it can be difficult to characterize relevant alloantibodies when performing antibody screening given that it is complex to cross match blood units with the patient plasma. Therefore, to ensure compatibility, time-consuming techniques are required to characterize alloantibodies resulting in delays before safe transfusion can be provided.

In addition to this technical problem, autoantibodies may shorten red cell survival and contribute to the hyperhaemolysis observed in patients presenting with DHTR. Castellino et al. documented haemolysis due to autoantibodies in four of 14 patients with autoantibodies. In all cases, direct antiglobulin tests (DAT) detected IgG and C3 on the RBC surface [15]. Cold-reactive autoantibodies of the IgM isotype are less frequent, and do not seem to be harmful. In some cases, autoantibodies were the major factor in the development of DHTR, inducing typical autoimmune haemolytic anaemia (AIHA) [12]. In many of these life-threatening cases of post-transfusion AHAI, the antibodies involved are panagglutinins. Panagglutination makes the choice of blood and identification of associated alloantibodies very difficult.

When autoantibodies are found in a SCD patient, the extent to which they are likely to affect the outcome of transfusion is unclear. Nevertheless, their presence may indicate that a process of autoimmunization is under way, and therefore, that close clinical and laboratory observation is required after any subsequent transfusion.

An important issue, in addition to the possible development of AHAI, is the difficulty in some settings to distinguish between autoantibodies and alloantibodies associated with a variant antigen carried by the patient. This diagnosis of partial antigen can be considered in patients producing an antibody against an expressed RH antigen, as variants are very frequent in SCD patients. The most frequent types are anti-D in D+, anti-e in e+ and anti-C in C+. In vitro adsorption techniques for patients who have not recently been transfused and/or DNA testing are required to distinguish between a partial antigen and a normal antigen in the carrier. An additional problem with autoantibodies in SCD is extended phenotyping, which is sometimes needed to identify the antigens against which the patient can produce antibodies. The implementation of molecular analysis in laboratories may allow this particular difficulty to be overcome.

Whether an antibody against blood group antigens is clinically significant is a major issue for the management of transfusion. This issue arises when the pre-transfusion screening test is positive, but also when post-transfusion haemolysis occurs with new antibodies and further transfusions are required. The antibody titre, isotype, optimal temperature, affinity for Fc gamma receptor and C1q binding, and also the density of the target antigen on RBCs all affect the likelihood of haemolysis and therefore contribute to whether or not an antibody is harmful in a recipient. The antibodies found by investigations of post-transfusion haemolysis in SCD frequently present the characteristics of clinically significant antibodies. Nevertheless, in many cases of DHTR in SCD, some antibodies are found that are not expected from these characteristics to cause haemolysis of transfused RBCs; also, antibody-dependent haemolysis reactions in SCD patients is frequently very severe. Therefore, patient status and host factors are presumably important. The inflammatory state of SCD patients can result in effector cells being activated. Hyper-reactive macrophages may be involved, and cytokines induce the expression of FcR on effector cells, such as macrophages and NK cells [16]. Another feature that is likely to contribute to dangerous incompatibility between host antibodies and transfused RBCs is the integrity of membrane of the transfused RBCs. The host environment may exert strong oxidative stress and this may have a deleterious effect on stored RBCs, facilitating their binding to antibodies and complement, and also the destruction of sensitized RBCs by macrophages. Phosphatidylserine (PS)-exposure in donor RBCs is significantly higher after incubation with pre transfusion plasma samples from patients with acute complications of SCD, than with samples from patients in steady-state [3]. Therefore, host effects on RBCs, probably due to the inflammatory status, may potentialize the “antibody effect”.

The biological presentation of DHTR falls schematically into two classes: the classical conflict between antigens on transfused RBCs and antibodies in the patient plasma against transfused RBCs; and the more enigmatic situation without detectable antibodies. Hyperhaemolysis without detectable antibodies is poorly understood, and consequently difficult to prevent and to treat. This syndrome has been attributed to macrophage activation, bystander haemolysis, reactive haemolysis, and the possible continuation of the painful episodic haemolysis observed for autologous RBCs. RBCs incubated in pre transfusion plasma from patients suffering DHTR without detectable antibodies show significant eryptosis and this has led to the suggestion of accelerated senescence of transfused RBCs in these cases. PS-exposure of RBCs from concentrates may account for this destruction in the SCD environment, as PS increases RBC adhesion, facilitating engulfment in the slow blood flow of the spleen, and recognition by macrophages. Various factors in plasma may also contribute to donor PS-RBC destruction, one example being PLA2, produced in SCD with acute chest syndrome (ACS) [17]. Transfused RBCs may be particularly sensitive to the patient environment because they already suffer from storage lesions. However, this syndrome is observed only in SCD patients. Another consistent feature of post transfusion haemolysis in SCD is the recurrence of vaso-occlusive crisis (VOC) as the main symptom. In mice, post transfusion haemolysis can induce VOC, and this effect is inhibited by inhibitors of inflammatory cytokine receptors [18]. A study in humans showed that PS-exposure of autologous RBCs increased during post-transfusion haemolysis [13]. PS-exposed sickle RBCs have increased adherence capacity [19] and this may be an additional explanation of the recurrence of VOC.