Haemolysis, pure red cell aplasia and red cell antibody formation associated with major and bidirectional ABO incompatible haematopoietic stem cell transplantation

Abstract

Background

Acute and delayed haemolysis, alloimmunisation and pure red cell aplasia (PRCA) are potential complications after ABO incompatible haematopoietic stem cell transplantation (HSCT). The aims of this study were to investigate acute and delayed red blood cell (RBC) antibody-associated complications, including haemolysis, PRCA and alloimmunisation in major and bidirectional ABO incompatible HSCT.

Materials and methods

We retrospectively examined the transplant courses of 36 recipients of bone marrow or peripheral blood stem cells from ABO incompatible donors and evaluated the current practice of performing plasmapheresis in patients with higher isoagglutinin titres. We investigated the role of ABO incompatibility in haematopoietic recovery, transfusion requirements, alloimmunisation and PRCA.

Results

Laboratory signs of acute haemolysis were noted in five (14%) patients, one (3%) of whom had clinically overt haemolysis. Patients with haemolysis had IgM titres ≥1:8 and received >16 mL of RBC in the HSCT. In patients with higher titres, plasmapheresis performed prior to the transplant prevented acute haemolysis. Delayed haemolysis was not recorded in the follow up. Haematopoietic recovery and transfusion requirements did not differ notably between patients with and without haemolysis. De novo RBC antibodies were detected in two (5.5%) patients after HSCT, and PRCA was noted in one (3%) patient.

Discussion

Carried out with adequate graft processing, plasmapheresis and blood component support, haemolysis is not a common complication after HSCT. Our results confirm that the occurrence of haemolysis depends on larger RBC volumes and higher isoagglutinin titres. Despite the reduction of patients’ isoagglutinin titres by plasmapheresis, we still noted a critical combination for the development of laboratory signs of haemolysis (IgM titre ≥1:8 and RBC volume >16 mL). De novo immunisation to RBC antigens and PRCA are rare events following ABO incompatible HSCT.

Introduction

ABO incompatibility occurs in 30–40% of patients undergoing haematopoietic stem cell transplantation (HSCT), and is classified into three different groups: major, minor and bidirectional ABO incompatibility¹ with incidence of 15–20%, 15–20% and up to 5%, respectively. In major ABO incompatibility recipient anti-A/B isoagglutinins are directed against donor ABO antigens, while in minor ABO incompatibility donor plasma contains isoagglutinins directed against recipient red blood cells (RBC). Bidirectional ABO incompatibility represents a combination of both major and minor ABO incompatibility.

Major ABO incompatibility between a donor and recipient can lead to acute haemolysis or prolonged destruction of donor-derived RBC, with pure red cell aplasia (PRCA) and prolonged transfusion requirements^2,3. Immune haemolysis following major ABO incompatible HSCT is caused by immunological destruction of donor RBC in the transplant. Haemolysis is usually acute and is caused by recipient-derived anti-A/B isoagglutinins, or rarely delayed, caused by anti-A/B isoagglutinins produced after HSCT⁴. The risk of acute haemolysis is higher if the stem cell transplant contains more RBC. The risk can be lowered by processing the graft in order to remove incompatible RBC^5,6. In patients with higher isoagglutinin titres, plasmapheresis or antibody immunoadsorption^7–10 can be performed prior to HSCT to remove anti-A/B isoagglutinins.

Recipient haematopoiesis may persist following HSCT, and haematopoietic and immune function may be both recipient and donor in origin (mixed chimerism) for a prolonged period of time². Although increased transfusion requirements and clinical signs of haemolysis can occur, the overall impact of ABO incompatibility in HSCT is generally considered low, provided that appropriate transfusion practices are followed^2,11.

We performed a retrospective analysis of acute and delayed complications related to RBC antibodies, including haemolysis, PRCA and alloimmunisation. We evaluated the influence of isoagglutinin titre and infused RBC volume on the occurrence of acute and delayed haemolysis after ABO incompatible HSCT and assessed the impact of the current practice of plasmapheresis in patients with higher isoagglutinin titres and graft processing in order to remove incompatible RBC and plasma.

Materials and methods

This retrospective study was carried out in a Department of Internal Medicine, University Hospital Centre Zagreb during a 2-year period (2012–2013). This study included 36 patients who underwent ABO incompatible allogenic HSCT. Of these, 29 (80.5%) patients received major and seven (19.5%) bidirectional ABO incompatible transplants from related or unrelated donors. The patients’ characteristics are shown in Table I.

Table I.

Patients’ characteristics, diagnoses, conditioning regimens and Graft-versus-Host disease prophylaxis.

Patients’ characteristics
Age, years (median, range)	47 (4–59)
Gender (male/female)	22/14
Diagnosis
Acute myeloid leukaemia	14
Acute lymphoblastic leukaemia	10
Myelodysplastic syndrome	4
Non-Hodgkin’s lymphoma	3
Chronic myeloid leukaemia	2
Chronic lymphoblastic leukaemia	1
Multiple myeloma	1
Chronic granulomatous disease	1
Conditioning regimens
Bu/Cy	14
Bu/Cy + ATG	9
Flu/Bu*	3
Flu/Bu + ATG*	8
Flu/Cy*	2
Graft-versus-Host disease prophylaxis
CsA + Mtx	23
CsA + MMF	13

Bu/Cy: busulfan, cyclophosphamide; Bu/Cy+ATG: busulfan, cyclophosphamide, antithymocyte globulin; Flu/Bu: fludarabine, busulfan; Flu/Bu+ATG: fludarabine, busulfan, antithymocyte globulin; Flu/Cy: fludarabine, cyclophosphamide; CsA+Mtx: cyclosporine, methotrexate, CsA+MMF: cyclosporine, mofetil-mycophenolate;

^*reduced intensity conditioning.

Harvesting and processing the haematopoietic stem cell components

High-resolution matching for human leucocyte antigens (HLA)-A, -B, -C, -DRB and -DRQ was done for all the donor/recipient pairs (10 out of 10 antigens matched). The grafts were obtained from 19 (53%) related HLA-identical and 17 (47%) unrelated HLA-matched donors. Twenty-two patients were transplanted with peripheral blood stem cells (PBSC) and 14 with bone marrow (BM). All unrelated donors donated PBSC, whereas related donors donated PBSC or BM, depending on the patient’s diagnosis and transplantation schedule. PBSC donors were mobilised with recombinant human granulocyte colony-stimulating factor (GCS-F; Neupogen; Hoffman-La Roche Ltd., Basel, Switzerland) at a dose of 10 μg/kg of bodyweight (BW) per day, starting 4 days before leucapheresis. Leucaphereses were performed with a COBE Spectra cell separator (MNC program; Gambro BCT, Lakewood, CO, USA). All leucapheresis products were infused into patients without additional processing. Total infused apheresis product volume, infused RBC volume, infused RBC volume/kg BW and CD34+ cell count were assessed before infusion. BM was harvested from the posterior iliac crest. All BM transplants were processed with a COBE 2991 cell separator (Gambro BCT) and incompatible RBC and/or plasma were removed, while buffy coat was collected. Total infused buffy coat volume, infused RBC volume and total nucleated cell count were assessed before infusion. RBC volumes in PBSC and BM transplants were calculated according to transplant volume and haematocrit in the transplant.

Isoagglutinin titres

Anti-A/B isoagglutinin titres (IgM and IgG types) were tested before the HSCT. In patients who underwent plasmapheresis, titres were also tested following every procedure. Isoagglutinin titres were determined by incubating 3–5% suspension of type A and B RBC with 2-fold serial dilutions of the patient’s serum in saline at room temperature followed by centrifugation. The agglutinating activity of IgG isoagglutinin was determined by incubating the patient’s serum with 0.01 M dithiothreitol at 37 °C for 2 hours. Afterwards, a standard titre procedure was performed in an indirect antiglobulin test. The reciprocal of the highest dilution of serum that gives a 1+ reaction is referred to as the titre. Titre score was determined according to a standard scale for grading serological reaction.

Plasmapheresis

Plasmaphereses were indicated according to the European Group for Blood and Marrow Transplantation (EBMT) guidelines in the case a host anti-donor IgM ABO isoagglutinin titre was 1:32 or greater¹². Plasmaphereses were performed using a membrane-based technique (continuous flow) on a Diapact CRRT system device (B. Braun Medical Inc., Betlehem, PA, USA), with human albumin 5% solution as a replacement fluid and unfractionated heparin systemic anticoagulation. Calcium and potassium were given during the procedures to prevent hypocalcaemia and hypokalaemia.

Laboratory testing

Patients were screened for laboratory and clinical signs of haemolysis. Complete blood count, direct (conjugated), and total (conjugated and unconjugated) bilirubin, lactate dehydrogenase, haptoglobin, and methaemalbumin were tested daily following HSCT. Cut-off values were: lactate dehydrogenase >241 U/L, direct bilirubin >5 μmol/L, total bilirubin >20 μmol/L, haptoglobin 0.3–2 g/L and methaemalbumin <0.05 g/L. Haemolysis was noted as positive in patients with two or more positive laboratory signs of haemolysis.

Clinical haemolysis was defined as the abrupt onset of several specific signs: anxiety, dyspnoea, fever, shaking, flushed face, chest and/or lumbar pain, hypotensive shock, tachycardia, nausea and/or vomiting, cold and clammy skin, dark urine (haemoglobinuria), and disseminated intravascular coagulation.

ABO blood group testing, as well as direct and indirect antiglobulin tests were performed prospectively, every 2 weeks after HSCT. Testing was performed until RBC engraftment was demonstrated by a positive serological test for donor RBC. Haematopoietic recovery of leucocytes and neutrophils was monitored after HSCT and expressed as number of days required to reach a white blood cell count of 1×10⁹/L, and an absolute neutrophil count of 0.5×10⁹/L. RBC and platelet transfusion requirements were assessed by recording the day of the last RBC and platelet transfusion after HSCT. Routine serological testing for blood group typing was performed by a column agglutination method every time transfusion of a RBC unit was requested to demonstrate establishment of donor ABO type. A positive reaction in a serological test, no matter how weak, is evidence of donor RBC in a recipient’s blood sample, and is interpreted as a donor RBC engraftment.

Transfusion support

In case of a major ABO incompatibility, patients were transfused with recipient RBC type components and donor type platelets and fresh-frozen plasma, while in bidirectional incompatibility O type RBC and AB type platelets and fresh-frozen plasma were used. All transfused RBC and platelet components were leucoreduced and irradiated. Specific transfusion support was maintained until donor ABO blood type was confirmed twice from two different blood samples, with no discrepancy in serological ABO typing.

Alloimmunisation

The frequency and specificity of antibodies prior to HSCT and de novo immunisation after HSCT were determined during routine serological testing for ABO type and antibody screening, as mentioned above.

Pure red cell aplasia

The presence of prolonged erythroid aplasia with myeloid, lymphoid and megakaryocyte engraftment in BM biopsies and prolonged transfusion dependency for more than 90 days after HSCT in the absence of relapse, infections or drug-related toxicity was noted to determine the occurrence of PRCA after HSCT. Isogglutinins of the recipient’s ABO group directed against donor RBC had to be detectable.

Statistics

Descriptive statistics, including numbers (percents) and medians (ranges) are used to present the data. All analyses were performed using SPSS 21.0 for Windows (IBM Corp., New York, USA). The differences in total infused RBC volume and infused RBC volume/kg BW between patients transplanted with PBSC or BM were compared using the Mann-Whitney U test. The differences in hematopoietic recovery and transfusion requirements between patients with and without noted haemolysis were also compared using the Mann-Whitney U test. The differences in the occurrence of haemolysis between patients who received PBSC with the RBC volume/kg BW above or below the recommended residual volume were compared using Fisher’s exact test calculator. The difference in the numbers of patients who underwent plasmapheresis between the group with positive signs of haemolysis and the group without haemolysis were compared using Fisher’s exact test calculator. The level of statistical significance was set at 0.05 for all analyses.

Results

Haematopoietic stem cell grafts

HSCT was performed using PBSC in 22 (61%) patients and BM in 14 (39%) patients. In PBSC grafts, the median CD34+ cell count was 6.1 (range, 3.4–7.7)×10⁶/kg BW. The median total nucleated cell count in BM grafts was 2.1 (range, 0.85–5.2)×10⁸/kg BW. The median total PBSC and BM transplant volumes were 228 mL (range, 35–420) and 204 mL (range, 140–234), respectively. Patients transplanted with BM received significantly larger total RBC volumes and larger RBC volumes/kg BW (Table II) in comparison to patients transplanted with PBSC (p<0.001, Mann-Whitney U test), even after removing incompatible RBC from the BM, with the median RBC reduction being 81% (range, 69–91%).

Table II.

Characteristics of the haematopoietic stem cell transplants.

Haematopoietic stem cell source	Median (range)
Peripheral blood stem cells (n=22)
BW, kg	67 (16–102)
Total infused RBC volume, mL	7.9 (1.33–30.6)
Infused RBC volume, mL/kg BW	0.105 (0.03–0.44)
Bone marrow (n=14)
BW, kg	79 (37–119)
Total infused RBC volume, mL	73.75 (41–108.2)
Infused RBC volume, mL/kg BW	1.09 (0.39–1.99)

BW: body weight; RBC: red blood cells

Isoagglutinin titres and plasmapheresis

The median initial isoagglutinin titres for IgM and IgG were 1:16 (range, 1:2–1:256) and 1:64 (range, 1:2–1:256), respectively. Plasmaphereses were performed in eight out of 22 (36%) patients prior to PBSC transplantation and in ten out of 14 (71%) patients prior to BM transplantation. Thus, a total of 18 (42%) patients underwent plasmapheresis procedures and no serious side effects were observed. Although indicated according to the EBMT guidelines, plasmapheresis was not performed in four patients because of the medical conditions. In three patients with a titre <1:32, plasmapheresis was performed in an attempt to avoid any risk of haemolysis after infusion of a large volume of RBC in the BM graft. The median number of plasmaphereses performed per patient was 1 (range, 1–3). More than one plasmapheresis was performed in patients who had isoagglutinin titres higher than 1:32 after the first plasmapheresis. The median of isoagglutinin titres for IgM and IgG prior to HSCT were 1:4 (range, 1:2–1:64) and 1:8 (range, 1:2–1:128), respectively. Haemolysis was noted in patients with an IgM titre prior to HSCT ≥1:8 (Table III).

Table III.

Isoagglutinin titres prior to haematopoietic stem cell transplantation.

Haematopoietic stem cell source
Peripheral blood stem cells (n=22)
IgM titre prior to transplantation	1:2	1:4	1:8	1:16	1:32	1:64
Patients with haemolysis (n=2)	0	0	2	0	0	0
Patients without haemolysis (n=20)	3	6	1	6	1	3
Bone marrow (n=14)
IgM titre prior to transplantation	1:2	1:4	1:8	1:16	1:32	1:64
Patients with haemolysis (n=3)	0	0	1	1	1	0
Patients without haemolysis (n=11)	6	3	1	1	0	0

IgM: immunoglobulin M.

Haemolysis

The majority of the ABO incompatible HSCT were well tolerated. Laboratory signs of haemolysis were noted in four (14%) out of 29 major incompatible HSCT and one (14.3%) out of seven bidirectional ABO incompatible HSCT. Clinical signs of haemolysis manifested after major ABO incompatible BM transplantation in one (3%) patient. Haemolysis manifested as fever, chest and lumbar pain. The clinical signs of haemolysis disappeared shortly after infusion of the graft and the patient recovered completely. The details of the patients with positive signs of haemolysis (n=5) are shown in Table IV. Haemolysis occurred in patients with an IgM titre ≥1:8 who received more than 16 mL of RBC in their transplant. There was no statistically significant difference in the number of patients who underwent plasmapheresis between the group with positive signs of haemolysis (3/5) and the group without signs of haemolysis (15/31) (p=0,999, Fisher’s exact test). Regarding the four patients who had an indication for plasmapheresis but did not go through the procedure because of medical reasons, only one developed laboratory signs of haemolysis (patient 2 in Table IV.) Four (18%) out of 22 patients transplanted with PBSC received >0.2 mL of RBC/kg BW, which is above the recommended residual volume. A statistically significant higher frequency of laboratory signs of haemolysis was observed in the group of patients who received >0.2 mL of RBC/kg BW. Haemolysis was noted in two out of these four patients (50%) and in none among the remaining 18 patients (p=0.026, Fisher’s exact test).

Table IV.

Patients with positive laboratory (n=4) and clinical (n=1) signs of haemolysis.

Patient	HSC source	ABO incompatibility	RBC volume (mL) in HSC graft	Infused RBC volume kg/BW	Titre of IgM before HSCT	Titre of IgG before HSCT	LDH	Direct bilirubin	Total bilirubin	Haptoglobin	Methaemalbumin	Clinical signs of haemolysis	Last RBC transfusion (days)	Transfused RBC units (n)
1	BM	Major	102	1.1	8	8	↑	↑	↑	↓	↑	No	113	33
2	BM	Major	104	1.2	32	32	↑	↑	↑	norm.	↑	no	39	11
3*	BM	Major	73	1.9	16	32	↑	↑	↑	↓	↑	yes	624*	59*
4	PBSC	major	26	0.4	8	32	↑	/	↑	↓	norm.	no	60	10
5	PBSC	Bidirectional	16	0.2	8	64	↑	↑	↑	↓	↑	no	216	54

HSC: haematopoietic stem cell; RBC: red blood cells; BW: body weight; HSCT: haematopoietic stem cell transplantation; LDH: lactate dehydrogenase; BM: bone marrow; PBSC: peripheral blood stem cells; norm.: normal values.

^*Patient with pure red cell aplasia.

There were no signs of delayed hemolysis in the laboratory follow-up.

Transfusion requirements

There was no difference in hematopoietic recovery and platelet transfusion requirements between patients with and without haemolysis, while RBC transfusion requirements were higher in the group of patients with signs of haemolysis (p=0,009, Mann-Whitney test). Haematopoietic recovery and transfusion requirements following HSCT are shown in Table V.

Table V.

Haematopoietic recovery and transfusion requirements following haematopoietic stem cell transplantation.

	Patients without haemolysis (n=28)	Patients with haemolysis (n=5)	p*
Hematopoietic recovery time	Median (range)
Neutrophils >0,5×109/L, days	14 (9–30)	16 (15–18)	0,391
WBC >1×109/L, days	16 (11–31)	16 (13–21)	0,827
Transfusion requirements	Median (range)
Last RBC transfusion, days	27 (0–175)	113 (31–624)	0,009
Last platelet transfusion, days	16 (4–150)	17 (16–194)	0,448

^*Mann-Whitney test;

WBC: white blood cells; RBC: red blood cells.

Engraftment

ABO typing was not performed in patients (n=3) who died within 15 days after HSCT due to septic shock with multiorgan failure. In all other patients (n=33) RBC engraftment was confirmed in serological ABO typing. Recipient A/B isoagglutinins were either not detected in the blood test or were weakly positive but gradually disappeared after HSCT in 35 (97%) patients. One patient (3%) diagnosed with PRCA had continuously positive anti-A isoagglutinins and donor RBC were detected only after the PRCA had been cured.

Alloimmunisation

In two (5.5%) patients, de novo RBC antibodies were detected after HSCT. Anti-C, anti-D and anti-E antibodies in the first and anti-C, anti-E and anti K antibodies in the second patient were detected 5 and 3 months after HSCT, respectively. A RBC antibody with anti-D specificity was detected in one (20%) out of five RhD+ patients who received RhD− grafts after reduced intensity conditioning. No signs of haemolysis were detected, except an increase in lactate dehydrogenase noted in the second patient at the time of de novo immunisation.

Pure red cell aplasia

PRCA was diagnosed in one (3%) O+ patient treated for chronic granulomatous disease, 8 months after receiving an A+ unrelated BM transplant. IgM and IgG isoagglutinin titres prior to transplantation were 1:16 and 1:32, respectively. The patient’s conditioning regimen had been fludarabine, busulfan and antithymocyte globulin, and Graft-versus-Host disease prophylaxis, with cyclosporine and methotrexate, had been given. Laboratory, as well as clinical signs of haemolysis were observed after HSCT. The patient’s isoagglutinin titres, transplant characteristics and signs of haemolysis are shown in Table IV (patient 3). Leucocyte and platelet engraftment occurred on days +14 and +16 after HSCT, respectively. The patient failed to recover reticulocyte count and remained dependent on RBC transfusions, although complete donor chimerism was demonstrated by molecular analysis. BM biopsies showed sustained engraftment of megakaryocytes and granulocytes, but subtotal absence of erythroid precursors. Nine months following HSCT, iron overload was treated with oral deferasirox. PRCA was treated with prednisone (1 mg/kg BW once daily) and intravenous immunoglobulins (1 g/kg BW) from day +470 and day +500 after HSCT, respectively. The rise in reticulocyte count and cessation of transfusion requirements occurred 624 days following HSCT. In total, the patient received 59 RBC products. Isoagglutinins were determined periodically until they disappeared from the patient’s serum. At the time of reticulocyte count recovery, anti-A isoagglutinins against donor A RBC disappeared. The patient subsequently stayed transfusion independent all the time during follow up until day +880 after HSCT.

Discussion

ABO compatibility is not considered critical in the selection of potential haematopoietic stem cell donors, because pluripotent and early committed haematopoietic progenitor cells do not express ABO antigens¹³. Nevertheless, ABO incompatible HSCT does present a challenge because it may be associated with acute and delayed complications, such as: acute and delayed haemolysis, PRCA and alloimmunisation. The majority of these complications can be prevented with ex vivo processing of the graft and in vivo preparation of the donor.

Most authors confirm the need for BM graft processing to prevent haemolysis of donor RBC^8,13,14, although some authors define critical titres of IgM or IgG of 1:16 or less, when graft processing is not necessary¹⁵. BM graft processing before HSCT decreases RBC volume drastically, as demonstrated in our study. However, despite the reduction of RBC in BM transplants, patients still received significantly greater volumes of RBC, which is acknowledged to be important in the development of haemolysis.

As for PBSC grafts, it has been recommended that the residual RBC content in PBSC should be below a maximum accepted level (usually 0.2 mL/kg BW)¹⁶. We observed a statistically significant difference in the occurrence of haemolysis between our patients who received a PBSC graft containing above or below the recommended residual volume of RBC, which verifies that the maximum accepted level of RBC in PBSC should be taken into account.

Adequate preparation of the recipient prior to HSCT is also important for reducing haemolysis. There is inconsistency regarding both the critical isoagglutinin titre before HSCT and the necessity of plasmapheresis prior to transplantation. Although some authors suggest a critical isoagglutinin titre before HSCT, for example, Rowley recommends IgM and IgG isoagglutinin titres of 1:16 or less¹⁷, there is no consensus with respect to strategies for isoagglutinin reduction. Whenever a patient’s medical condition allowed, we followed EBMT policy: when the IgM isoagglutinin titre prior to HSCT is less than 1:64, unmodified BM or PBSC grafts may be infused¹². Although this policy is followed in Europe, it is not a standard procedure in other parts of the world, and is not routinely performed in the USA. In the case of higher titres, RBC should be removed from the graft. Although such a practice may be considered overcautious, we showed that signs of haemolysis may occur even with low initial titres if a critical amount of RBC was infused. This is also reflected in EBMT recommendations; RBC contamination should be reduced to less than 20 mL when the host anti-donor ABO isoagglutinin titre is 1:32 or greater¹².

Besides ex vivo processing of the haematopoietic stem cell graft and in vivo preparation of donor, haemolysis can be ameliorated by a proper blood component support before and following ABO incompatible HSCT, which is considered one of the most complex activities in Transfusion Medicine¹⁸. Transfusion of type O group RBC will provide compatible RBC components until late engraftment and the change in the recipient’s blood group occur. Transfusion of compatible plasma, depending on the type of ABO incompatibility, will dilute the quantity of recipient isoagglutinins¹⁹.

Increased RBC transfusion requirements after ABO incompatible HSCT have frequently been reported. In this study, patients showed prolonged RBC transfusion dependency (median, 31 days; minimum-maximum, 0–624) which is comparable with findings of previous studies^20–22. Furthermore, in our study, patients with signs of haemolysis showed statistically significant prolonged RBC transfusion dependency, compared to patients without signs of haemolysis.

Although rare, PRCA may occur in up to 20–30% of the patients after major ABO incompatible HSCT^3,18,23–25. It is presumed that anti-donor A/B isoagglutinins, released for up to 1 year after HSCT from recipient plasma cells, are responsible for delayed RBC engraftment. Although the mechanisms for these effects have not been fully understood, they may involve interactions of anti-donor isoagglutinins with incompatible ABO antigens expressed on donor erythropoietic cells²⁶. IgM and IgG isoagglutinin titres tested prior to BM transplantation in the patient who experienced PRCA were 1:16 and 1:32, respectively. It is still unclear whether high isoagglutinin titre is a risk factor for PRCA. According to Stussi et al., PRCA is observed exclusively in patients with isoagglutinin titres greater than 1:16²¹. On the other hand, Worel et al. suggest that there is no correlation between isoagglutinin titres and the occurrence of PRCA²². That is confirmed in our study, with lower isoagglutinin titres in the patient who experienced PRCA. PRCA was caused by an anti-A isoagglutinin (patient blood type O+, donor blood type A+), as found in the studies by Lee et al.²⁴. According to their study, patients with anti-A isoagglutinins against donor RBC develop PRCA more frequently than patients with anti-B. In an effort to reduce the transplant-related mortality, dose-reduced conditioning regimens have been developed. Being predominantly immunosuppressive, these protocols result in persistently elevated anti-donor ABO antibody levels²⁷.

Veelken et al. showed that slower disappearance of anti-donor ABO antibodies via the graft-versus-plasma cell effect explains the higher incidence of PRCA in patients who received cyclosporine and methotrexate, compared to that in patients given cyclosporine and mofetil-mycophenolate for the prevention of Graft-versus-Host disease²⁸. Consistent with this study, our patient developed PRCA after a conditioning regimen consisting of fludarabine, bulsulfan and anti-thymocyte globulin and Graft-versus-Host disease prophylaxis with cyclosporine and methotrexate.

In our study, de novo immunisation to RBC antigens was detected in two (5.5%) patients. Anti-C and anti-D antibodies in the first, as well as anti C and anti-E in the second patient, were probably caused by the exposure of donor lymphocytes to patient RBC antigens, while anti-E and anti-K antibodies in the first and second patients, respectively, were probably caused by immunisation to transfused RBC. RBC antibodies with anti-D specificity were detected in one (20%) out of five RhD+ patients who received RhD− grafts after reduced intensity conditioning. These results are in accordance with studies by Asfour, Worel and Mijovic who assert that de novo D immunisation in RhD+ immunosuppressed patients transplanted with RhD− grafts is a rare event, with an incidence ranging from 10–20%^29–32. The infrequent production of de novo RBC antibodies in these patients did not have an effect on the occurrence of haemolysis.

Conclusions

Acute or delayed clinically evident haemolysis, PRCA and alloimmunisation are not common after either BM or PBSC transplantation. Regarding the significance of isoagglutinin titre prior to HSCT and RBC volume in the transplant, our results are consistent with the fact that the occurrence of haemolysis depends on greater RBC volumes and higher isoagglutinin titres. In these circumstances, therefore, BM processing and plasmapheresis - although a financial burden and demanding with possible side-effects - seem to be justified. We found an association with prolonged or increased RBC transfusion requirements for patients who had positive laboratory signs of acute haemolysis. The consequences of major and bidirectional ABO incompatibility on late outcome of HSCT remain to be further investigated.