Introduction

Beta thalassemia is an inherited disorder resulting from mutation in the beta globin gene leading to impaired beta globin production, ineffective erythropoiesis, reduced red blood cell (RBC) survival, and chronic hemolytic anemia. Patients with severe form of the disease, i.e., beta thalassemia major (TM), are transfusion-dependent (TD) since early childhood and for life [1, 2]. Survival of TD TM patients has improved significantly over the past few decades as better treatment became available [3].

Nevertheless, development of anti-RBC alloantibodies and autoantibodies has been increasingly recognized as a major challenge of chronic transfusion therapy [4,5,6], as it significantly shortens in vivo survival of transfused RBC and limits availability of further safe transfusions [4, 7, 8].

Alloimmunization and autoimmunization rates observed in TD TM patients are highly variable and are in the range of 2.9–37.0% [9,10,11,12,13,14,15] and 0.0–40.0% [11, 16,17,18], respectively, whereas in the general population, the rates are in the range of 0.4–7.0% [5, 19,20,21,22,23], respectively. However, most reports are based mostly on TD TM pediatric patients, and less data is available on the rates in adult TD TM patients.

The aim of this study was to assess RBC alloimmunization and autoimmunization rates among the patients with TD TM followed and treated in our Comprehensive Center of Adult Thalassemia, Hemoglobinopathies and Rare Anemias, in a 1300-bed major quaternary care and referral teaching medical center.

Materials and methods

Patients

The study cohort included TD TM patients, currently treated in our center. All patients were > 18 years of age, regularly transfused every 2–3 weeks. All patients received ABO and RhD compatible RBC transfusions collected by Magen David Adom (MDA, Israeli Red Cross organization) National Blood Services from voluntary donors according to national guidelines. Since the year 2004, all RBC transfusions have been leuko-reduced.

Patients’ files and the blood bank database were screened for clinical data and transfusion records since the beginning of follow-up at early childhood. Data included date of birth, gender, ethnicity, ABO-D blood group, history of splenectomy, history of pregnancies, age at first transfusion, lifelong cumulative number of blood units transfused, and detected alloantibodies and autoantibodies, including specificities and age at first detection.

Laboratory investigations

Using standard blood bank methods [24], plasma was analyzed prior to each transfusion for detection of new antibodies to RBC antigens. The plasma was tested with commercially available screening cells using automated systems (IH1000 Bio-Rad, Hercules, California, USA, and AutoVue Ortho Clinical Diagnostics, Raritan, NJ, USA). A 3-cell screen was used for the antibody screening procedure, with low ionic strength solution (LISS)-IgG cards. If the antibody screen was positive, commercial panels (Bio-Rad, Hercules, CA, USA, and Ortho Clinical Diagnostics, Raritan, NJ, USA) including papain- and/or ficin-treated cells were performed to identify the antibodies as well as a direct anti globin test (DAT). Saline indirect anti-globulin test (SIAT) and polyethylene glycol test (PEG IAT) were used to further investigate and resolve the antibody reactivity. Adsorption methods were employed in patients presenting with a new autoantibody. In cases of a positive DAT, further investigation using specific reagents to detect IgG, or complement, was carried out. Eluates were prepared when IgG was detected on RBC and tested against commercial available panel cells. Cases that could not be resolved by the hospital blood bank laboratory, such as cases with multiple alloantibodies, those with a combination of alloantibodies and autoantibodies, or those that were recently transfused, were referred to MDA National Blood Group Reference Laboratory (MDA - NBGRL) for further investigation and confirmation. In case an alloantibody was identified, the information was documented in the blood bank records, and all subsequent RBC transfusions were negative for the corresponding antigen and cross-matched using LISS-IgG cards (Bio-Rad, Hercules, CA, USA). A notification letter with RBC antibodies details was sent to the patient.

Statistical analysis

Statistical analysis was performed using the SPSS Software, Version 26. Categorical variables were presented by (n, %) and analyzed using chi-square test, Fisher’s exact test, or logistic regression with the Wald test. Continuous variables were presented by mean and SD, and were analyzed using nonparametric Wilcoxon two-sample test. p values < 0.05 were considered statistically significant.

Results

Forty patients were included in the study, 21 (52.5%) males and 19 (47.5%) females. There were 29 (72.5%) Muslim Arab patients and 11 (27.5%) Jewish. The median age was 31.0 years (range 20–47 years). The median age at first transfusion was 1.07 years (range 0.25–4.0 years). The median number of lifelong cumulative blood units transfused was 809 (range 439–1363). Thirty (75.0%) patients were splenectomized (SPX) and the median age at splenectomy was 10.0 (range 4.0–20.0). Patients’ demographic and clinical characteristics are shown in Table 1.

Table 1 Demographic and clinical characteristics of 40 TD TM patients
Full size table

Alloimmunization was found in 17 (42.5%) patients. Thirty-four alloantibodies were identified, with the most frequent being RH related (12 of 34, 35.3%) followed by those of the Kell system (8 of 34, 23.5%). Specificities of alloantibodies are shown in Table 2. Eleven of the 17 alloimmunized patients (64.7%) had multiple antibodies. Clinical characteristics of alloimmunized patients are shown in Table 3.

Table 2 Specificity and prevalence of alloantibodies detected in 17 TD TM patients
Full size table
Table 3 Clinical characteristics of 17 alloimmunized TD TM patients
Full size table

Autoimmunization was demonstrated in 2 (5.0%) patients; both had warm antibodies and had also alloantibodies.

Age at first transfusion was positively related to the probability of developing alloantibodies (chi-square test; p = 0.02), as shown by binary logistic regression analysis (odds ratio 2.476; 95% CI [1.040–5.895]; Wald test; p = 0.04). While starting transfusions at the age of 3 months led to a 26% lifelong probability of developing alloantibodies, starting transfusions at the age of 1 year led to a 41% lifelong probability of developing alloantibodies. Moreover, starting transfusions at the age of 3 years led to an 81% lifelong probability of developing alloantibodies.

Splenectomy was highly correlated with a lifelong development of alloantibodies, and the incidence of alloimmunization was significantly higher in splenectomized patients, compared with those with no splenectomy (chi-square test; p = 0.016).

Furthermore, logistic regression analysis of the lifelong probability of developing alloantibodies on the age at first transfusion and splenectomy demonstrated a strong positive relationship (Wald test; p = 0.002). The later transfusions started the higher the probability of developing alloantibodies (odds ratio 2.745; 95% CI [1.144–6.589]; Wald test; p = 0.024), and the performance of splenectomy greatly increased that same probability (odds ratio 6.957; 95% CI [1.497–32.325]; Wald test; p = 0.013). Starting transfusions at the age of 1 year and not having splenectomy at a later age led to a lifelong probability of 24% of developing alloantibodies, while having splenectomy at a later age led to a lifelong probability of 68% of developing alloantibodies.

In a series of logistic regressions, the incidence of alloimmunization found not to be influenced by patients’ age, gender, ethnicity, duration of transfusion therapy, and cumulative number of blood units transfused (p > 0.05).

Discussion

This study is the first report from Israel showing a substantially high rate of alloimmunization among adult TD TM patients. Data analysis showed that alloimmunization rate was higher in patients who started transfusion later in life and had been splenectomized.

Anti-Rh and anti-Kell blood group system alloantibodies are the most frequently detected in the general population [19,20,21]. In our study, the most prevalent alloantibodies were against antigens of RH system (12 antibodies) and of Kell system (8 antibodies), similar to reports from other centers [4,5,6,7, 11, 16, 18, 23, 25].

Age at first transfusion was positively related to the probability of developing alloantibodies as starting transfusions at a later age were associated with higher alloimmunization rates. It has been previously shown that young age at the time of initial RBC exposure is associated with a significant lower likelihood of alloimmunization [16, 18, 23, 26]. The immunologic mechanisms underlying these observations are not fully understood. It was suggested that patients at a very early age manifest immunological immaturity in many aspects, including humoral unresponsiveness to stimulations by carbohydrate antigens, hence resulting in immune tolerance to allogeneic RBC antigens and consequently reduced alloimmunization risk [23, 26, 27].

Splenectomy has been reported in several studies as a significant risk factor for alloimmunization [7, 23, 28]. The mechanism by which removal of the spleen increases alloantibody formation is not clear. It has been suggested that removal of the spleen led to non-filtering of antigens from the blood stream which results in higher alloantibodies production in splenectomized subjects [28]. It has also been hypothesized that post-splenectomy conformational changes in RBC membranes may enhance immunomodulation [6, 28]. In our study, splenectomy was highly correlated with alloantibodies development. Furthermore, logistic regression analysis of age at transfusion initiation and splenectomy found that splenectomy has a highly significant impact on the probability of developing alloantibodies.

All patients included in our study received RBC transfusions matched only for ABO and Rh(D). Matching RBC transfusions beyond ABO and Rh(D) antigens is currently considered as a preferred way to reduce alloimmunization in TD TM patients, although no standardized policies or consensuses have been widely implemented [5, 23, 29, 30]. Phenotypically matched RBC transfusions for Rh and Kell systems were reported to be associated with alloimmunization rates as low as 2.8–3.7% compared with rates as high as 22.6–33.0% observed in earlier studies [26, 28]. Moreover, it has been shown that RBC genotyping can improve matching of donor blood to patient, hence further reducing alloimmunization risk [27, 31]. However, in a recent study, despite showing that patients may benefit from receiving RBC transfusions based on genotyping, it was concluded that phenotypically matched RBC transfusions for ABO, Rh(D), and Kell systems seem to be already sufficient for RBC transfusions in thalassemia patients, preventing the development of alloantibodies [32].

Donor leukocytes within RBC units have been implicated in stimulating alloimmunization, and consequently, the use of leuko-reduced blood has been widely implemented in effort to reduce alloimmunization. Currently, most blood transfused in the USA, in Canada, and in many other developed countries is pre-storage leuko-reduced [23, 28, 33]. In our center, all adult patients were exposed to non-leuko-reduced blood, as universal leuko-reduction has been implemented in Israel since 2004. This might have contributed to our high rates of red cell alloimmunization.

The incidence of alloimmuniuzation was not influenced by patients’ age, gender, ethnicity, duration of transfusion therapy, and cumulative number of blood units transfused, in contrast to what is described in the literature [21, 23, 34], probably due to the limitation of this retrospective observational study of a relatively small cohort size.

In conclusion, we have found a substantially high rate of alloimmunization in adult TD TM patients, which might be attributed to donor/recipient heterogeneity and to multi-transfusions over many years without RBC antigen matching beyond ABO and RhD, and without leuko-reduction. Alloimmunization rate was higher in patients who started transfusion later in life and had been splenectomized. Early diagnosis of TD TM as well as early initiation of blood transfusions, avoidance of splenectomy, and implementation of Cc, Ee, and K antigen matching may reduce the incidence of alloimmunization in TD thalassemia patients.