Immuno-hematological monitoring after allogeneic stem cell transplantation: a single-center, prospective study of 104 patients

Abstract

Background

The impact of ABO incompatibility on the outcome of hematopoietic stem cell transplantation (HSCT) is still debated. We report the results of a prospective, single-center study evaluating the impact of ABO mismatch on the development of immediate and late immuno-hematological complications, and the efficacy of the protocol used at the “Sapienza” University (Rome, Italy) to manage ABO incompatibility in patients undergoing HSCT.

Materials and methods

From January 2013 to December 2016, we prospectively analyzed all patients undergoing HSCT. Graft manipulation or desensitization strategies were used according to ABO incompatibility, donor sex and donor transfusion history. Red blood cell and platelet transfusions were given based on immunohematological features.

Results

From January 2013 to December 2016, 104 consecutive patients underwent HSCT from a matched related donor (29.81%), matched unrelated donor (53.58%), cord blood (1.9%) or haploidentical donor (14.42%). Forty-nine patients (47%) were ABO-identical and 55 (53%) ABO-incompatible (23 major, 25 minor, 7 bidirectional). Donor engraftment, graft failure or other complications did not differ between ABO compatible or incompatible patients. ABO incompatibility did not show a significant impact on graft-versus-host disease, overall survival or disease-free survival. Factors associated with the need for prolonged red blood cell support were ABO incompatibility (p=0.0395), HLA disparity between donor and recipient (p=0.004) and the onset of hemorrhagic cystitis (p=0.015). In multivariate analysis HLA disparity was the only statistically significant condition (p=0.004).

Discussion

ABO incompatibility does not represent a barrier to allogeneic HSCT. It is, however, associated with prolonged transfusion requirements. Close immunohematological monitoring, as a shared standard procedure, allows appropriate transfusion support to be provided and limits post-HSCT immuno-hematological complications.

INTRODUCTION

Allogeneic hematopoietic stem cell transplantation (HSCT) from a human leukocyte antigen (HLA)-compatible stem cell donor is a potentially curative treatment for patients with malignant and non-malignant hematological diseases^1–3. Since the HLA system is inherited independently of blood type, approximately 40–50% of all HSCT are performed across the ABO blood group barrier^2,4,5. Although ABO incompatibility is not considered a contraindication to HSCT⁶, immuno-hematological consequences, such as immediate and delayed hemolytic complications or the onset of pure red cell aplasia must be considered carefully during HSCT planning and donor choice, in order to allow the safest possible HSCT^7–12. Moreover, published data have shown different results concerning the relationship between ABO mismatches and transplant outcomes, such as transplant-related mortality (TRM), overall survival (OS), disease-free survival (DFS) and the onset of graft-versus-host disease (GVHD)^{6,11,12,13,14}.

We report here the results of a prospective, single-center study and the strategy employed to manage ABO incompatible HSCT at the Hematology Center and Immunohematology and Transfusion Medicine Unit of the “Sapienza” University (Rome, Italy) from January 2013 to December 2016. The aim of this study was to evaluate the impact of ABO mismatch on the development of immediate and late immuno-hematological complications, DFS, OS, GVHD and TRM. Additionally, we analyzed the efficacy of the protocol used at the “Sapienza” University to manage ABO incompatibility in patients undergoing HSCT.

MATERIALS AND METHODS

From January 2013 to December 2016 we prospectively analyzed, with a complete immuno-hematological assessment, all consecutive patients undergoing an allogeneic HSCT, from family donors (sibling or haploidentical), matched unrelated donors, mismatched unrelated donors or cord blood units. We excluded patients receiving a second allogeneic HSCT. Analyses were performed within 30 days prior to the HSCT, every week up to 30 days after HSCT, then every 15 days until 100 days after HSCT or until red blood cell (RBC) engraftment.

Immuno-hematological work-up

All patients had their RBC typed for A, B, D, C, c, E, e, K and k blood group antigens, by liquid phase using reagents provided by Ortho Clinical Diagnostics (Raritan, NJ, USA) and by microplates in automated instruments (Neo-immucor, Immucor, Milan, Italy). The direct antiglobulin test (DAT) was performed with a broad-spectrum antiserum and with monospecific anti-IgG, -IgA, -IgM, -C3d, and -C3b antisera, in liquid phase and by column agglutination tests (reagents supplied by Ortho Clinical Diagnostics, Johnson & Johnson, Medical S.p.A., Milan and DiaMed, Milan, Italy). The elution test, in case of a positive DAT, was performed with a low pH glycine buffer using a commercially available kit (Elu-Kit, Immucor, Norcross, GA, USA) and incubated with an O RBC panel and A or B RBC panel to check for its specificity on the basis of the patient’s RBC phenotype. The indirect antiglobulin test, including untreated and treated (ficin/papain) autologous and homologous RBC (Resolve C, Ortho Clinical Diagnostics and ID-Dia-med Panel, DiaMed), was performed with sera from the patients and donors. When RBC alloantibodies were detected, the specificity was identified using O RBC extended panels (Ortho Clinical Diagnostics and Dia- Med). Cryo-agglutinins were detected by testing patients’ serum samples with autologous and homologous RBC at 4°C and 20°C. ABO iso-hemagglutinin (IHA) titration was performed in liquid phase; serial doubling dilutions of the serum sample were performed and incubated for 1 hour at 20°C for the detection of anti-A/-B IgM. The patients’ sera used for the IgG titer were treated with Neutr-AB (Grifols Diagnostics AG, Düdingen, Switzerland) in order to neutralize the IgM interference. The titer was recorded as the reciprocal of the dilution. Anti-A/-B immune hemolysins were detected by testing patients’ fresh serum samples with an equal volume of fresh A/B 5% RBC at 37°C for 10 minutes.

Management of ABO incompatibility

Graft manipulation or desensitization strategies were decided on the basis of ABO incompatibility and the donor’s transfusion history. All female and previously transfused male donor grafts were depleted of plasma in order to avoid transfusion-related acute lung injury (TRALI)^15,16. RBC and platelet transfusion support was provided according to the immuno-hematological features. Immuno-hematological results and transfusion data were prospectively collected in a database, as was information about age, sex, diagnosis, donor hematopoietic stem cell source, conditioning regimen and post-HSCT complications such as cytomegalovirus or Epstein-Barr virus infections, venous-occlusive disease, hemorrhagic cystitis, hemolytic-uremic syndrome and GVHD. Informed consent to participation in the study was obtained from all patients, or their parents or legal guardians in the case of minors.

Definitions

ABO incompatibility was defined as major if IHA directed against donor RBC antigens were detected in the recipient plasma, minor if the donor had IHA directed against recipient RBC, or bidirectional in the case of both major and minor incompatibility^17,18.

Erythroid engraftment was defined as the achievement of a peripheral reticulocytosis >1%^19,20 and the conversion of a recipient to donor ABO phenotype, in the case of ABO incompatibility. Neutrophil recovery was defined as the achievement of an absolute neutrophil count (ANC) ≥ 0.5×10⁹/L for 3 consecutive days and a platelet recovery ≥ 20×10⁹/L, with no platelet transfusions in the preceding 7 days^18,21.

Graft failure was defined as primary in the case of no initial donor cell engraftment, a peripheral ANC <0.5×10⁹/L by day +28 after HSCT with peripheral blood stem cells or bone marrow progenitors without relapse and by day +42 after a cord blood unit transplant. Secondary graft failure was defined as the loss of donor cells after an initial engraftment and a recurrent ANC <0.5×10⁹/L²¹.

Graft manipulation according to the immuno-hematological evaluation

Major incompatibility

In the case of a major ABO mismatch, manipulation was not required if IHA ≤64; for IHA >64, IHA titration was re-evaluated 10 days before the infusion; for IHA ≥ 128 and RBC contamination < 15 mL, manipulation was not required; if RBC contamination was >15 mL, RBC were depleted from the graft; for IHA ≥ 256, recipient IHA titers were reduced by plasma exchange or immune-absorption. If anti-A/-B immune-hemolysins were detected and the source of stem cells was the bone marrow, RBC depletion was required, even with IHA ≤256^2,5,22.

Minor incompatibility

In the case of a minor ABO mismatch, for IHA ≤64 manipulation was not required; for IHA >64 plasma depletion during the stem cell harvest or from the graft was required. If anti-A/-B immune-hemolysins were detected, plasma depletion was required regardless of other factors^2,5,22.

Bidirectional incompatibility

The combination of both plasma or RBC depletion from the graft or recipient plasma exchange was planned according to the above-mentioned criteria for major and minor incompatibility^{2,5,22, 23}.

If antibodies directed against other antigens were detected, such as those of the Rh system or other blood group systems, the necessity of interventions involving the recipient, donor or graft was considered case by case, based on the antibody titer against the recipient’s or donor’s RBC antigens.

For hematopoietic stem cell units (from bone marrow or peripheral blood) that required RBC depletion, HES 6% (Grifols) was added (twice the volume of RBC as estimated by hematocrit). After gravity sedimentation for 60 minutes, the RBC layer was manually removed and the hematopoietic stem cells underwent further centrifugation (1,800 rpm for 15 minutes) to remove supernatant (HES and plasma). The final buffy-coat was suspended in saline plus anticoagulant as needed, according to the recipient’s body weight. In the case of peripheral blood stem cells, the hematopoietic stem cells were collected using a Com.Tec cellular separator (Fresenius Kabi, Bad Homburg, Germany) to obtain a final product with reduced contamination by RBC and/or decreased plasma volume, without further manipulation. Plasma-depletion of hematopoietic stem cells was performed with semi-automated centrifugation (1,800 rpm for 10 minutes at 10°C).

Graft-versus-host disease prophylaxis and grading

All patients received standard cyclosporine (3 mg/kg, target blood levels of 200–400 ng/mL). A short course of methotrexate²⁴ (15 mg/m² on day +1, 10 mg/m² on days +3, +6, and +11 ) was added in the case of a sibling or unrelated donor. In the latter case, anti-thymocyte globulin (2.5 mg/kg/ day for 3 days; total dose, 7.5 mg/kg) was given from day −4 to −2 before the transplant. In the case of a haploidentical donor, mycophenolate mofetil (15 mg/kg every 12 hours) was added to cyclosporine and post-transplant cyclophosphamide (50 mg/kg on days +3 and +5) was infused. Cord blood transplant recipients received anti-thymocyte globulin (total dose 7.5 mg/kg) as well as 6-methylprednisolone (1 mg/kg) and cyclosporine (3 mg/kg, drug levels maintained between 200–400 ng/mL)²⁵. Acute and chronic GVHD were defined according to consensus criteria of the National Institutes of Health^26,27.

Transfusion policy

The recommendations for transfusion support for ABO-incompatible HSCT are summarized in Table I^28,29. All blood components underwent pre-storage leukocyte reduction and irradiation with 25 Gray. When available, single donor apheresis platelet transfusions were preferred in order to reduce the risk of anti-HLA alloimmunization³⁰.

Table I.

Transfusion support

ABO incompatibility	Recipient	Donor	RBC	Platelets		Plasma
			1	1	2
Major	O	A	O	A	AB, B*, O*	A or AB
	O	B	O	B	AB, A*, O*	B or AB
	O	AB	O	AB	A*, B*, O*	AB
	A	AB	A or O	AB	A*, B*, O*	AB
	B	AB	B or O	AB	B*, A*, O*	AB
Minor	A	O	O	A	O*	A or AB
	B	O	O	B	O*	B or AB
	AB	O	O	AB	A*, B*, O*	AB
	AB	A	A or O	AB	A*, B*, O*	AB
	AB	B	B or O	AB	B*, A*, O*	AB
Bidirectional	A	B	0	AB	A*, B*, O*	AB
	B	A	0	AB	B*, A*, O*	AB

^*Plasma-reduced platelet concentrates or suspended in additive solution. RBC: red blood cells.

Statistical analysis

A descriptive analysis, including clinical and demographic characteristics of the patients, was performed through the median and range for continuous variables; absolute values and relative frequencies were used for categorical variables. The association between variables was tested by Pearson’s chi-square test or Fisher’s exact test, when appropriate. Parametric or non-parametric tests were used to analyze the differences between the groups, when appropriate. Survival curves were estimated according to the Kaplan-Meier method; the log-rank test was used to assess differences between subgroups. OS and DFS were defined as the time from HSCT until death and relapse/progression, respectively. Univariate and multivariable analyses were performed using the Cox proportional hazard model. A multivariable Cox regression model was selected by minimizing the Akaike information criterion, using a forward stepwise selection strategy.

Times to ANC, RBC and platelet engraftment were evaluated by competing risk analyses. Cumulative incidence was estimated from competing risk representation, comparing groups through Gray test statistics. Statistical significance was fixed at the 0.05 level. All analyses were performed using Open Source R software (version 3.3.2, The R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

From January 2013 to December 2016, we prospectively analyzed and monitored 104 consecutive patients. As of November 2021, the median follow-up period was 4.9 years (range, 0.5–6.3). Sixty-four patients were male (61.54%), 40 were female (38.46%). The median age at HSCT was 38 years (range, 1–66). Most patients had hematological malignancies. The median time from diagnosis to HSCT was 25 months (range, 2–252).

Seven patients (6.7%) received a non-myeloablative conditioning regimen, 33 (31.7%) received reduced intensity conditioning and 64 patients (61.5%) were given myeloablative conditioning, the most commonly used being thiotepa (5 mg/kg, on days −7 and −6), fludarabine (50 mg/m²) and busulfan (3.2 mg/kg on days −5, −4, and −3)^31,32. The donor was a matched sibling in 31 cases (29.81%), an unrelated donor in 56 cases (53.58%) (10/10 HLA matched in 31, 9/10 in 21 and 8/10 in 4), a haploidentical donor in 15 cases (14.42%), while two patients (1.9%) received a cord blood unit.

Sixty-one donors (58.65%) were male and forty-one were female (39.42%); their median age was 34 years (range, 6–68). The stem cell source was bone marrow for 52% of the patients, peripheral blood for 46% of the patients and a cord blood unit for 2%. The median CD34⁺ cell dose was 4.23×10⁶/kg (range, 1.07–10.33), the median total nucleated cell dose was 5.29×10⁸/kg (range, 1.07–10.33), and the median CD3⁺ cell dose was 0.56×10⁸/kg (range, 0.13–7.9). Forty-nine (47%) grafts were ABO-identical and 55 (53%) ABO-incompatible: 23 major, 25 minor, and seven bidirectional. The characteristics of the patients, diseases and transplants are summarized in Table II.

Table II.

Patient and transplant characteristics

Characteristics	N = 104 (%)
Gender (male)	64 (61.54)
Age
Median [IQR]	38 [23–54]
Diagnosis
Aplastic anemia	7 (6.73)
Acute lymphoblastic leukemia	27 (25.96)
Acute myeoloid leukemia	34 (32.69)
Non Hodgkin’s lymphoma	12 (11.54)
Hodgkin’s lymphoma	3 (2.88)
Chronic lymphocytic leukemia	6 (5.76)
Chronic myeloid leukemia	2 (1.92)
Multiple myeloma	1 (0.96)
Myelofibrosis	6 (5.77)
Myelodysplastic syndrome	3 (2.88)
Other	3 (2.88)
Advanced disease stage at HSCT	51 (49.04)
Not advanced disease stage at HSCT	53 (50.96)
Pre-HSCT lines of treatment
0	3 (2.88)
1 – 2	63 (60.58)
≥3	38 (36.54)
MRD	31 (29.81)
MUD	56 (53.85)
MUD 10 loci /10	31
MMUD 9 loci /10	21
MMUD 8 loci /10	4
Haploidentical/MMRD	15 (14.42)
Cord blood unit	2 (1.92)
Donor gender (male)	61 (58.65)
Donor age
Median [IQR]	34 [24 – 45]
IgG CMV R/D
Pos /Pos	24 (23.08)
Pos/Neg	8 (7.69)
Neg/Pos	18 (17.31)
Neg/Neg	54 (51.92)
ABO match
Matched	49 (47.12)
MRD	17
MUD	15
MMUD	12
MMRD/Haplo	5
Cord blood unit	0
Major mismatch	23 (22.12)
MRD	9
MUD	5
MMUD	5
MMRD/Haplo	3
Cord blood unit	1
Minor mismatch	25 (24.03)
MRD	7
MUD	10
MMUD	5
MMRD/Haplo	3
Cord blood unit	0
Bidirectional mismatch	7 (6.73)
MRD	0
MUD	5
MMUD	2
MMRD/Haplo	0
Cord blood unit	0
Conditioning regimen
Non-myeloablative	7 (6.73)
Reduced intensity	33 (31.73)
Myeloablative	64 (61.54)
Stem cell source
Bone marrow	54 (51.92)
Peripheral blood	48 (46.15)
Cord blood	2 (1.92)
Cellularity
TNC median	5.29×108/kg (range 1.97–10,33)
CD34+	4.23×106 /kg (range 0.92–11,9)
CD3+	0.56×106/kg (range 0–7,9)
Graft manipulation
No	39 (37.50)
Plasma-depletion	61 (58.65)
RBC depletion	2 (1.92)
Plasma + RBC depletion	2 (1.92)
Adverse reactions during HSC infusion	29 (27.88)
Hypertension	22 (21.15)
Hypotension	3 (2.88)
Fever	2 (1.92)
Headache	1 (0.96)
Shivers	1 (0.96)

IQR: interquartile range; HSCT: hematopoietic stem cell transplantation; MRD: matched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; MMRD: mismatched related donor; CMV: cytomegalovirus; R. recipient; D: donor; Pos: positive; Neg: negative; Haplo: haploidentical; TNC: total nucleated cells; RBC: red blood cells.

Pre-transplant immuno-hematological work-up

The pre-HSCT donor immuno-hematological assessment showed that the median titers of natural and immune anti-A/B IHA were 512 (range, 4–1024) and four (range 1–256), respectively. It was possible to analyze immune-hemolysins in 15/32 ABO-incompatible donors, according to the availability of fresh blood samples, and was found to be negative in all of them. In 10/104 donors, alloantibodies were detected: anti-Wra in eight patients (IgM in 7, IgG in 1), anti-P1 IgM in one patient, and anti-M IgM in one patient.

Recipient immuno-hematological assessment showed a median titer of natural and immune anti-A/B IHA of 32 (range, 2–512) and 16 (range, 1–256), respectively. Immune-hemolysins were observed in 10/30 cases of ABO major or bidirectional incompatibility.

In nine of the 104 patients allo-antibodies were detected: anti-Wr^a IgG in seven, anti-K IgG in one (titer 64), and anti-E IgG in one (titer 1).

Graft manipulation

The graft was depleted of plasma in 61 cases (58.7%) (with a 99% median CD34⁺ cell recovery; range, 91–99.98%), taking into account the immuno-hematological characteristics of the donor and recipient (9 cases), due to the risk of TRALI (32 cases) or to limit cardiocirculatory overload (20 cases). RBC were removed from the graft in two cases, due to an ABO incompatibility (CD34⁺ cell average recovery of 93%; range, 64–99.4%). In two cases, both plasma and RBC were removed from the graft, due to the immuno-hematological characteristics of both the donor and the recipient, and to reduce the risk of TRALI. Two patients were treated with plasmapheresis: one because of an ABO bidirectional incompatibility and one because of an ABO major incompatibility together with high titer anti-K alloantibody (donor phenotype: K⁺). In both cases, no alternative donors were available. Both patients obtained a reduction of IHA, that enabled a safe infusion of hematopoietic stem cells.

Adverse reactions

Twenty-nine patients had adverse reactions during the hematopoietic stem cell infusion, which was hypertension in 22 cases (in 19/22 cases with bone marrow as the stem cell source); a significant correlation between the hematocrit value (median 17.42%; range, 7.5–65.0) of the hematopoietic stem cell product and the onset of hypertension (p=0.0310) was observed. Other reactions were hypotension (3 patients, peripheral blood as the stem cell source), fever (2 patients, peripheral blood stem cells), headache (1 patient, bone marrow stem cells) and shivers (1 patient, peripheral blood stem cells). Overall, 17 out of 29 patients showing adverse reactions received ABO mismatched grafts (7/17 received unmanipulated graft).

Laboratory signs of intravascular hemolysis (e.g. increasing levels of free hemoglobin and lactate dehydrogenases, reduction of hematocrit and haptoglobin) were observed in five of 30 patients (16.7%) with ABO major and bidirectional incompatibility within 8 days after HSCT, with a median titer of anti-A/-B IHA of eight (range, 2–128), anti-A/-B hemolysin in three cases and median erythrocyte contamination of 344 mL (range, 15–495).

Donor cell engraftment

Engraftment of ANC and platelets was observed at a median of 21 days (range, 12–37) and 20 days (range, 8–150) after HSCT, respectively. RBC engraftment (reticulocytosis >1%) was observed at 22 days (range, 16–183) after HSCT. Fifteen patients died without trilinear engraftment, most of them due to the onset of sepsis (13 of 15), at a median of 30 days after HSCT (range, 6–119). Two patients (8.7%) developed graft failure (secondary in both cases).

In our experience, donor engraftment of ANC, platelets and RBC (Table III and Online Supplementary Content, Figure S1A–C) or the onset of graft failure did not differ between patients with or without ABO incompatibility.

Table III.

Engraftment

ABO incompatibility	Engraftment, days: median (range)
	ANC	Platelets	RBC
ABO matched	21 (12–33)	22 (10–150)	24 (16–96)
ABO major MM	14 (14–37)	20 (13–60)	23 (19–62)
ABO minor MM	21 (15–30)	21 (12–48)	22 (18–44)
ABO bidirectional MM	23 (14–27)	21 (10–51)	22 (18–70)

ANC: absolute neutrophil count; RBC: red blood cells; MM: mismatch.

In both cases of ABO major and minor mismatch, the conversion of the ABO phenotype occurred at a median of 100 days (range, 28–235); a reduction of ABO IHA was observed at a median of 31 days (range, 21–200) for ABO major mismatch; in the presence of a bidirectional mismatch, the conversion of the ABO phenotype occurred at a median of 75 days (range, 28–274), with a reduction of ABO IHA observed at a media of 24 days (range, 7–50).

One patient developed pure red cell aplasia (0.96% of the total population, 3.33% of major/bidirectional ABO mismatched HSCT).

Transfusion requirements

RBC and platelet transfusion requirements persisted until 30 days (range, 0–741) and 21 days (range, 1–380) after HSCT, respectively. Factors associated with the need for prolonged RBC support were ABO incompatibility (p=0.0395, Online Supplementary Content, Figure S2A), HLA disparity between donor and recipient (p=0.004) and the onset of hemorrhagic cystitis (p=0.015). In multivariate analysis, HLA disparity was the only statistically significant condition (p=0.004). ABO incompatibility did not show a significant impact on platelet transfusion requirements (Online Supplementary Content, Figure S2B). A comparison between different hematopoietic stem cell sources showed a statistical impact (p=0.001) on the total number of blood transfusions within 100 days after HSCT (Online Supplementary Content, Figure S2C).

The total numbers of RBC and platelet transfusions within 100 days after HSCT were five (range, 0–54) and nine (range, 1–91), respectively; a statistical correlation with the onset of hemorrhagic cystitis was observed (p=0.0234).

Acute and chronic graft-versus-host disease

Twenty-eight patients (26.9%) developed acute GVHD (≥ grade II) at a median of 41 days after HSCT ( range, 13–188). One patient showed acute GVHD at 188 days after HSCT, following donor lymphocyte infusion. Fifteen patients (14.42%) suffered from chronic GVHD at a median of 185 days after HSCT (range, 127–461); the source of stem cells was bone marrow and peripheral blood for six and nine patients, respectively. No statistical correlation between ABO mismatch (major, minor or bidirectional) and the onset of acute GVHD (p=0.069) or chronic GVHD (p=0.67%) was observed.

Other complications after hematopoietic stem cell transplantation

Two patients (1.92%) developed venous-occlusive disease and 16 (15.38%) had hemorrhagic cystitis. More in detail, the conditioning regimens in the patients who developed hemorrhagic cystitis were myeloablative, reduced intensity and non-myeloablative in 12, three and 11 cases, respectively; donors were matched unrelated (5 cases), mismatched unrelated (6 cases), haploidentical (4 cases) and a cord blood unit (1 case), with ABO incompatibility in six cases (2 major, 2 minor and 2 bidirectional).

All post-HSCT complications, including infections (bacterial, fungal and viral), were taken into account in terms of impact on the transfusion support. As previously discussed, the only complication with a significant impact on transfusion requirements was hemorrhagic cystitis (p=0.0234). No impact was observed between the onset of post-HSCT complications and ABO mismatch.

Overall outcomes

At a follow-up of 2 years (range, 0.5–4.3), 61 patients (58.65%) are alive, 59 in complete remission and two with disease relapse. The 3-year OS, DFS and TRM were 56.7% (95% confidence interval [95% CI]: 48–67.1], 49.8% (95% CI: 41–60.4) and 24.2%, respectively. Twenty-three patients (22.12%) relapsed at a median of 189 days after HSCT (range, 52–900). At a median follow-up of 4.9 year (interquartile range, 0.5–6.3), OS and DFS were respectively 55.8% (95% CI: 47–66.2] and 48.7% (95% CI: 40–59.4). ABO mismatch did not show any statistical influence on OS, DFS (Online Supplementary Content, Figure S3A, S3B) or TRM.

Autoimmune hemolytic anemia

Two female patients (1.9%) developed mixed autoimmune hemolytic anemia (AIHA), both following ABO matched HSCT. They had severe steroid-resistant anemia (hemoglobin 3.6 g/dL and 7.1 g/dL). The first patient showed a positive DAT (IgG, without conventional specificity, IgA and C3d) on day 299 after HSCT; the serological evaluation confirmed the presence of an IgG antibody detectable in the serum, directed against the E antigen and pan-reactive, as well as a cold IgM active up to 37°C, with a titer of 256 at 4°C. The donor and recipient RhD phenotypes were CcDee and ccDEE, respectively. The woman was treated with rituximab, followed by intravenous immunoglobulins, obtaining resolution of the AIHA. In the second case, the patient showed a positive DAT 600 days after HSCT. An IgG directed against the C antigen was identified; the serological evaluation showed IgG and IgA antibodies with anti-C and anti-e specificity. Both donor and recipient had a CCee RhD phenotype. The patient was treated with rituximab, obtaining clinical and serological resolution.

Post-transplant immune-hematological monitoring

Among the 49 ABO-matched HSCT recipients, two experienced AIHA. Moreover, one patient (recipient O Rh⁺/donor A Rh⁻) developed an antibody (IgG) directed against the RhD antigen 294 days after HSCT. Two patients (in both cases, donor and recipient were CcDee) showed an anti-E alloantibody (IgG), 100 and 235 days after HSCT, respectively; in one of the two cases, the anti-E alloantibody (IgG), detected in the pre-HSCT work-up, persisted until 240 days after the HSCT. Two patients developed pan-reactive IgG antibodies, without evidence of AIHA.

Among the 23 patients with ABO major incompatibility, one had a positive DAT for an anti-A1 IgG (donor phenotype A) 80 days after HSCT. In one patient, affected by aplastic anemia, with anti-K alloimmunization detected at the pre-HSCT work-up, the presence of an IgG anti-K was confirmed in the post-HSCT monitoring, after an initial reduction in antibody (0 at day 60 after HSCT) with prolonged erythroid aplasia, followed by a rebound that started at day 100. The patient developed graft failure (donor chimerism of 33.8% at 130 days after HSCT). Six years after HSCT, he is alive, in good clinical condition, albeit requiring monthly transfusion support.

One of the 25 patients with minor ABO incompatibility showed an anti-A IgG antibody between 21 and 35 days after HSCT. One patient developed an anti-B antibody 21 days after the HSCT (negative DAT). One patient had a positive DAT (C3d⁺) 14 days after HSCT. All patients showed laboratory signs of mild hemolysis.

Two of the seven patients with bidirectional ABO incompatibility showed a positive anti-C3d DAT, from days 21 to 80 and from days 21 to 28.

DISCUSSION

Data on the incidence of immediate and delayed hemolysis, autoimmune complications, time of engraftment, pure red cell aplasia, GVHD, OS, DFS, and TRM in ABO mismatched HSCT recipients are still unclear. Many studies include a limited number of patients, various graft sources, different conditioning and GVHD prophylaxis regimens and various transfusion strategies. At present, a general consensus about the best strategies to manage ABO incompatible HSCT is still lacking, even if it is now well established that risks of immediate and delayed hemolytic consequences in HSCT can be partially avoided by graft manipulation and appropriate transfusion support^2,5,28,29,33.

The aim of this study was to evaluate the impact of ABO mismatch on immediate and late immune-hematological complications, on donor engraftment, DFS, OS, GVHD and TRM, considering patients, donors, diseases and transplant characteristics. Another purpose was to validate the efficacy of the procedure to manage ABO incompatible HSCT at our center.

An ABO incompatibility in patients undergoing HSCT is still an open issue, despite advances in knowledge and development of new technologies, probably because of the heterogeneity of published studies and preventive strategies employed³⁴.

In particular, data regarding the impact of ABO incompatibility on donor cell engraftment are controversial. In 2008, Kimura et al. analyzed the impact of ABO mismatch on the outcome of recipients of bone marrow HSCT from unrelated donors in the Japan Marrow Donor Program, showing that major ABO incompatibility induced delayed neutrophil, platelet, and RBC engraftment. In the following years, several large registries failed to demonstrate slower engraftment rates in case of ABO mismatching. In particular, in a recent experience, Caanani et al.³⁵ on behalf of the European Society for Blood and Marrow Transplantation, described comparable engraftment rates between ABO matched and mismatched HSCT in the setting of mismatched unrelated HSCT for acute myeloid leukemia.

According to these data, in our experience, based on both matched and mismatched, related and unrelated HSCT, donor platelet and neutrophil engraftment was not affected by ABO mismatch.

As expected, we observed prolonged transfusion requirements following mismatched ABO HSCT^2,9.

In our prospective study, we did not observe any impact of ABO disparity on OS, DFS, TRM or on the onset of GVHD, probably thanks to the enhanced GVHD prophylaxis employed over the years. As recently shown, the effects of ABO mismatch may become more evident in terms of survival and GVHD, if analyses are limited to haploidentical transplantation^35,36. Hence, the impact of ABO mismatch may vary according to the different HSCT settings³⁶. In our study, the incidence of pure red cell aplasia was 3.33% in major/bidirectional ABO mismatched HSCT, compared to the 29% described in the literature². This difference may be due to the low median IHA titer and its rapid reduction after HSCT observed in our experience. To date, the impact of immunomodulatory drugs employed in conditioning regimens and GVHD prophylaxis platform on pre-HSCT IHA titers is still unknown³⁷. Hemolysis occurred within 8 days after HSCT, without acute complications or a worse outcome: 16.7% of patients had major or bidirectional ABO incompatibility, with a median anti-A/-B IHA titer of eight (range, 2–128), anti-A/-B hemolysin in three cases and a median erythrocyte contamination of 344 mL (range, 15–495). The incidence of AIHA in our population was 1.9%, comparable with data published in the literature (2–6%)³³. These data highlight the efficacy of our strategies.

Finally, we point out the role of anti-K antibody, characterized by the ability to cause bone marrow suppression. One of our patients, who underwent HSCT from an ABO major incompatible and K⁺ donor, because of a lack of other donors, experienced prolonged erythroid aplasia followed by graft failure. Early expression of Kell glycoprotein in committed progenitor cells (burst-forming unit erythroid and colony-forming unit erythroid) suggests a critical role in early stages of erythropoiesis, as demonstrated in the setting of anti-K-mediated disease of the fetus and the newborn^38,39. Anti-K causes direct suppression via an apoptotic process, or by inhibiting normal formation of the erythrocyte cytoskeleton⁴⁰. Since Kell glycoprotein is a metallo-endopeptidase that cleaves endothelin-3 to produce bioactive endothelin-3, anti-K was posited to modulate peptide growth factors on the cell surface^38,41.

The immuno-hematological assessment in HSCT requires a detailed evaluation of donor and recipient characteristics, such as ABO and Rh mismatch and evidence of alloimmunization to RBC antigens, taking into account the patient’s diagnosis, disease stage, conditioning regimen, stem cell source and HLA compatibility. Moreover, we emphasize the importance of close clinical and laboratory monitoring, as well as correct transfusion support²⁸, with the aim of limiting hemolysis triggered by the infusion of an incompatible inoculum.

CONCLUSIONS

In this prospective study, we confirmed that ABO incompatibility does not represent a barrier to allogeneic HSCT. It is, however, associated with prolonged transfusion requirements. Strict immuno-hematological monitoring, as a shared standard procedure, allows for appropriate transfusion support and the efficient treatment of patients, effectively limiting post-HSCT immuno-hematological complications.