Bone Marrow Transplantation, (1997) 19, 813–817
 1997 Stockton Press All rights reserved 0268–3369/97 $12.00
De novo chronic graft-versus-host disease presenting as hemolytic
anemia following partially mismatched related donor bone marrow
transplant
K Godder, AR Pati, SH Abhyankar, LS Lamb, W Armstrong and PJ Henslee-Downey
University of South Carolina and Richland Memorial Hospital, Center for Cancer Treatment and Research, Columbia, SC, USA
Summary:
Chronic graft-versus-host disease (cGVHD) is a disease
of immune dysregulation that resembles an autoimmune
disease. It usually involves the skin, mucosal and serosal
surfaces and, less commonly, the hematopoietic system.
We report hemolytic anemia (HA) as the primary manifestation of de novo cGVHD in recipients of partially
mismatched related donor transplants. Five of 40 eligible patients developed HA at a median of 168 days
post-transplant. Recipients were mismatched for one to
three major HLA antigens. Conditioning therapy consisted of total body irradiation, etoposide, Ara-C, cyclophosphamide and steroids. GVHD prophylaxis included
partial T cell depletion, using anti a/b CD3 antibody
(T10B9) and complement, in addition to post-transplant
immunosuppression. At presentation, all patients were
receiving cyclosporine with or without low-dose steroids. Along with a mean Hb of 7.1g%, patients had an
increased reticulocyte count, a mild raised lactic
dehydrogenase and a positive Coombs’ test (in 2/5
patients). Four patients had also demonstrated a
decrease in platelet count. Treatment was initiated with
high-dose steroids and intravenous gamma globulin and
response was observed within 1 week. Awareness of this
presentation of cGVHD and early therapeutic intervention can result in successful reversal of presumed
immune-mediated red cell and platelet destruction.
Keywords: chronic graft-versus-host disease; hemolytic
anemia; bone marrow transplantation; T cell depletion;
thrombocytopenia
Chronic graft-versus-host disease (cGVHD) occurs in more
than 50% of long-term survivors of BMT and is thought to
be due to proliferation of mature donor T cells in response
to host allo-antigens. Risk factors are HLA disparity, acute
GVHD (aGVHD), older host age, and male recipients of
grafts from alloimmune female donors.1,2 Symptoms
develop most commonly immediately following aGVHD or
after a quiescent period, and only 20% of patients will
Correspondence: Dr K Godder, Center for Cancer Treatment and
Research, Seven Richland Medical Park, Columbia, South Carolina,
29203, USA
Received 13 May 1996; accepted 23 December 1996
develop cGVHD de novo. Clinical presentation of chronic
disease differs from the acute in that it resembles collagen
vascular disease which is associated with significant morbidity post-BMT.3 The introduction of GVHD prophylaxis
which combines T cell depletion4 with post-transplant
immunosuppression,5 has significantly lowered the incidence of acute and subsequent cGVHD in patients receiving HLA-matched4,5 or -mismatched allogeneic transplant,6
but may also have changed its presentation. Although the
association between cGVHD and thrombocytopenia is well
documented,7–10 few cases of other immunohematologic
complications post-BMT have been described. 11–15 In this
study, we report a presumably immune-mediated hemolytic
anemia (HA) as the principal manifestation of cGVHD in
patients following partially mismatched related donor
(PMRD)-BMT, discuss its possible mechanism, and
emphasize the importance of early intervention with
immunosuppressive therapy.
Material and methods
Patients and clinical protocol
In this retrospective analysis 40 consecutive patients who
survived beyond 120 days following PMRD-BMT for hematological malignancies or marrow failure disorders were
evaluated for cGVHD. Pre-transplant conditioning therapy
included TBI and high-dose chemotherapy combining etoposide (not given to marrow failure patients), Ara-C, cyclophosphamide and steroids, as previously describe.16 Marrow grafts were T cell depleted by incubation with an antia/b TCR T10B9.1A-31 monoclonal antibody (provided by
John Thompson, University of Kentucky, under a FDA
approved IND for investigational use) and baby rabbit
complement.17 Additional GVHD prophylaxis included
post-transplant immunosuppression using cyclosporine
(levels maintained between 113 and 168 ng/ml whole blood
using a monoclonal antibody assay), a course of antithymocyte-globulin (10 mg/kg on day +5 to +16) and methylprednisolone (30 mg/m2 daily until after day +16 with a subsequent taper of 10% weekly). Patients’ complete blood
count was checked at least once weekly until day 100 posttransplant, and at least every other week after they returned
to their referring institution.
cGVHD presented as hemolytic anemia
K Godder et al
814
Data collection
Patients’ charts were reviewed for demographics, complications following BMT, occurrence of acute and chronic
GVHD, diagnosis and treatment of anemia and response to
treatment. Referring institutions were contacted when
patients were treated locally.
Evaluation of hemolytic anemia
HA was suspected when an acute fall in hemoglobin
occurred in a patient who had otherwise stable engraftment,
was transfusion independent, and had no evidence of blood
loss. Patients were evaluated using the following tests: peripheral blood cell morphology, reticulocyte count, lactic
dehydrogenase (LDH), haptoglobin, direct and indirect
Coomb’s test and platelet antibodies, when available. For
the Coombs’ (direct antiglobulin) test, patients’ red cells
were incubated with a polyspecific antiglobulin reagent. In
patients whose Coombs’ test was positive, cells were eluted
and further testing was done with a panel of antibodies
directed against human immunoglobulins or complement
component, to determine specificity. Indirect Coombs’ was
performed by incubating patients’ sera with pooled
screened red cells. As most of the patients were back at
their referring institutions, data on further characterization
of the hemolysis was not always available. Appropriate
investigations were performed to exclude the possibility
of infection.
Immunophenotype analysis by flow cytometry
Peripheral blood was routinely collected from donors prior
to BMT and from patients pre-BMT and on days +14, +28,
+60, +100 and +180. The cells were stained with MAbs to
the T lymphocyte markers CD3, CD4, CD8, TCR-ab, and
TCR-gd; the natural killer markers CD16, CD56, and
CD57; the B lymphocyte marker CD19; the activation markers CD25, IL-2Rp75, and HLA-DR, and IgG1 and IgG2a
control MAbs conjugated with the appropriate fluorochrome. MAbs were obtained from Becton Dickinson
Immunocytometry Systems (BDIS, San Jose, CA, USA),
except for anti-TCR-ab which was obtained from T Cell
Diagnostics (Cambridge, MA, USA).
When the peripheral WBC count was less than
0.5 × 106/ml, flow cytometric analysis was performed on
mononuclear
cells
enriched by
Ficoll–Hypaque
(Histopaque; Sigma, St Louis, MO, USA) density gradient
separation. All other samples were prepared by erythrocyte
lysis using FACS-Lyse from BDIS. Cells were stained as
described previously.18 At least 10 000 events were collected on a Becton Dickinson FACSort flow cytometer. Cell
subpopulations in the lymphocyte scatter gate were quantitated and normalized to the actual percentage of lymphocyte
events within this gate using Lysis II software. Cell subpopulation percentages were expressed as a percentage of the
total lymphocyte population.
ited involvement. Of those developing cGVHD, five
patients (28%) presented with HA, with or without thrombocytopenia on days 112–272 (median 168) post-transplant
(Table 1). Patients with HA were HLA mismatched with
their donors for one antigen (n = 1), two antigens (n = 3)
and three antigens (n = 2) and were all sex mismatched.
The donors were mothers for three patients, and a father or
sister for one patient each. Four patients were ABO compatible with their donors but had a variety of minor antigen
mismatches. None of the patients developed aGVHD within
3 months of BMT, however, all reported limited skin manifestations of cGVHD that were visible at the time of diagnosis of HA.
At time of presentation, hematologic findings included a
mean Hb level of 7.1g%, reticulocyte count of 10.7%
(range 5.8–12.9%), a peripheral blood smear free from evidence of microangiopathy, elevated LDH in 3/5 patients
(Table 2) and Coombs’ test positive in 2/5 patients (direct
in two patients and indirect in one). No specificity to red
cell antigen was demonstrated in the direct (following
elution) or indirect Coombs’ positive patients. Haptoglobin
was measured in two patients (CS and CK) and was found
to be low. In addition to a decline in hemoglobin level, 4/5
patients presented with a decrease in their platelet count to
a median of 50 000/mm 3 (range from 3000–141 000). In one
patient (CK), platelet antibody was tested and found to be
weakly positive.
All five patients had been maintained on GVHD prophylaxis with cyclosporine; however, four were taken off steroid therapy within 2 weeks prior to presentation. Treatment
of HA was initiated with pulse high dose steroids (500–
1000 mg/kg × 2 doses) and daily intravenous gammaglobulin × 3–5 doses and continued as shown in Table 3.
Response was observed in 4/5 patients within 1 week and
complete recovery that allowed tapering of therapy within
4 weeks. However, one of these patients (DH) required retreatment for a transient recurrence of HA. In the remaining
patient (HR), delay in diagnosis and treatment was associated with a slow response (3 weeks) and the need for continuous high-dose intensive immunosuppressive therapy
using four drugs (pulse high-dose steroids, IVIG, cyclosporine and azathioprine). Three patients continued to manifest
limited skin cGVHD symptoms once immunosuppressants
were tapered and had to be maintained on cyclosporine
alternating with low-dose steroids.19
The proportion of CD3+ T lymphocytes, CD19+ B lymphocytes and CD56+ NK cells from patients who developed
HA was within the observed range for non-affected patients
post PMRD-BMT, and thus no association was seen
between the distribution of lymphocyte subsets and the
development of HA or thrombocytopenia (Figure 1). All
patients became full donor hematopoietic chimeras by
RFLP, and remain in hematological remission from their
primary malignancy between 17 and 32 months post-transplant.
Results
Discussion
Eighteen of 40 (45%) eligible patients following PMRDBMT developed cGVHD, most of them (15/18) with lim-
In this retrospective study we reported on HA, a relatively
uncommon manifestation of cGVHD,3 as a major manifes-
cGVHD presented as hemolytic anemia
K Godder et al
Table 1
Patient
DH
MH
HR
CS
CK
815
Patient characteristics
Age
onset
Diagnosis
Donor
Sex
mismatched
ABO/
incompatibility
Minor RBC
incompatibility
HLA
mismatched
1
3
2
24
3
ALL
ALL
ALL
CML
MDS
mother
sister
mother
father
mother
y
y
y
y
y
0+/n
A+/n
0+/n
0+/n
0+/y
Jka, K, LeB, N
E, Jka
FyA, s
FyA, M
M
2
3
1
2
1
Acute
GVHD
Day of
n
n
n
n
n
median
184
150
168
112
272
168
RBC = red blood cells.
Table 2
Hemological manifestations
Patient
HgB
(g%)
DH
MH
HR
CS
CK
5.9
7.4
5.5
7.7
8.9
Reticulocyte
count (%)
LDH
(mg%)
12.2
NA
11.8
5.8
12.9
Coombs’
244
221
2052
1562
1170
Platelets
(× 1000/ml)
Direct
Indirect
+
−
+
−
−
−
−
+
−
−
141
112
34
3
50
Hgb = hemoglobin; LDH = lactate dehydrogenase; NA = not assessed.
Table 3
Patient
DH
MH
HR
CS
CK
Treatment and response
Cyclosporine
IVIG
Steroid
Response
+
+
+
+
+
× 5 days, qod, then monthly
× 3 days, then weekly
TIW × 12 w, BIW × 1 w then weekly
× 3 days × 2 w, BIW × 2 w then weekly
BIW × 3 w, weekly × 2 w then qow
HD-MPD
HD-MPD
HD-MPD + Imuran
HD-MPD
HD-MPD
1 week
1 week
3 weeks
1 week
1 week
HD-MPD = methylprednisolone 500–1000 mg/m 2 followed by prednisone 0.5–2 mg/kg; TIW = three times a week; w = week; BIW = twice weekly;
qod = every other day; qow = every other week.
tation of the disorder among patients who received T celldepleted PMRD-BMT. HA, often accompanied by thrombocytopenia, presented de novo in approximately one-third
of patients who developed cGVHD. Patients had no history
of aGVHD and had otherwise limited skin manifestation of
the disorder. Risk factors that may have contributed to the
development of cGVHD in our patients included HLA disparity, both known major and unknown minor HLA differences, and sex mismatched donor–recipient pairs. Nevertheless, our patients did not have other risk factors
including aGVHD and older age as previously reported in
cGVHD.1,2 In spite of limited evidence of an immunemediated process (as suggested by the positive Coombs’
test and platelet antibody, when tested) all patients
responded to intensive immunosuppressive therapy.
Other conditions that may give rise to HA post-BMT
include infection or medication. Our patients maintained
stable engraftment with no evidence of recurrent disease or
infection. Other than cyclosporine, patients did not receive
medications known to cause HA. Cyclosporine is reported
to be associated with hemolytic-uremic syndrome,20
although the disorder was also described in patients postBMT who did not receive the medication.21 None of our
patients developed renal insufficiency (data not shown),
hypertension or any other clinical or morphologic manifestation of thrombotic microangiopathy.
The mechanism of immune HA post-BMT is not completely understood, although several investigators have suggested that T cell suppression may lead to predominance of
B cell function, resulting in production of auto-antibodies.
Hazelhurst et al,22 speculated that enhanced B cell activity
secondary to lack of CD8 T cell suppression was the cause
of the HA in nine recipients of T cell-depleted marrows.22
Others have suggested that cyclosporine post-transplant
may permit antibody formation due to its selective effect
on T cell function with sparing of B cells.23 In our study, we
were not able to demonstrate a difference in B cell number
between patients who developed HA and those who did
not, and moreover, B cell recovery in all post-PMRD-BMT
patients was relatively delayed compared to that reported
in recipients of HLA-matched BMT (following both T celldepleted and non-depleted grafts).24
cGVHD presented as hemolytic anemia
K Godder et al
816
CD3+ T lymphocytes
normal recovery
% Lymphocytes
100
80
60
40
20
0
CK
0
CS
50 100 150 200 250 300 350 400
50
MH
DH
CD19 B lymphocytes
normal recovery
% Lymphocytes
HR
40
30
20
10
0
0 50 100 150 200 250 300 350 400
Days post-BMT
Figure 1
T and B cell recovery in hemolytic anemia patients compared to non-hemolytic anemia patients post-PMRD-BMT.
As to the origin of the cells causing immune-mediated
HA, persistence of residual host cells may produce antibodies directed against the newly developed donor red
cells. This usually occurs in the early post-transplant period
and accounts for the HA seen in ABO mismatch.23,25 Most
of our patients were ABO compatible and, although they
all had different minor group incompatibilities, those are
not usually associated with hemolysis. Immune HA may
also occur as a consequence of a mixed hematopoietic
chimera where donor antibodies are formed against recipient hematopoietic cells. This is unlikely to be the mechanism in our patient population, since all of them were full
donor hematopoietic chimeras by cytogenetic analysis and
RFLP with no evidence of residual hematopoietic host
cells.
Finally, post-transplant immune system dysregulation,
whereby donor cells triggered by host tissue produce antibodies against the engrafted hematopoietic system, giving
rise to an ‘auto’-immune (donor-anti-donor) HA, is consistent with the concept of cGVHD as an autoimmune disorder.3 The relatively infrequent detection of antibodies
(Coombs’ test) in our patients could be related to low numbers of B cells resulting in a low production level of an
aberrant antibody. This antibody is rapidly adsorbed and
the formed antibody-coated red blood cells are instantly
destroyed. Although immune thrombocytopenia post-BMT
has been characterized7–10 and reported to be associated
with cGVHD, 7–9 only three cases of true autoimmune hemolytic anemia are described in the literature. The first was
in a child with SCIDS who received a marrow from the
mother (similar to our findings in 3/5 patients using
maternal donors).11 In the other two reported patients, HA
was part of pancytopenia of immune etiology postBMT.12,14 The reason for the high prevalence of HA in
patients post-PMRD-BMT is unknown, however, factors
such as T cell depletion or its methodology may play a
role. Recent data from our center show that using a different antibody (OKT3) for T cell depletion, has not been
associated with HA.
As more patients receive grafts which are HLA disparate,
referring and transplant physicians need to increase awareness for HA as a late complication of BMT. Recognition
of HA post-BMT, followed by prompt and aggressive
immunosuppressive therapy is important in achieving control of red cell and platelet destruction; waiting for
additional laboratory findings before establishing treatment
may interfere with timely and effective therapy.
Acknowledgements
We wish to thank Ms Jan Cannon and Ms Gail Wingo for their
assistance in data collection and Ms Joan Parnell for her assistance
in preparation of the manuscript.
References
1 Atkinson K, Horowitz MM, Gale RP et al. Risk factors for
chronic graft-versus-host disease after HLA-identical sibling
bone marrow transplantation. Blood 1990; 75: 2459–2464.
2 Ochs LA, Miller WJ, Filipovich AH et al. Predictive factors
for chronic graft-versus-host disease after histocompatible sibling donor bone marrow transplantation. Bone Marrow Transplant 1994; 13: 455–460.
cGVHD presented as hemolytic anemia
K Godder et al
3 Graze PR, Gale RP. Chronic graft versus host disease: a syndrome of disordered immunity. Am J Med 1979; 66: 611–620.
4 Marmont AM, Horowitz MM, Gale RP et al. T-cell depletion
of HLA-identical transplants in leukemia. Blood 1991; 78:
2120–2130.
5 Ringden O, Horowitz MM, Sondel P et al. Methotrexate,
cyclosporine, or both to prevent graft-versus-host disease after
HLA-identical sibling bone marrow transplants for early leukemia. Blood 1993; 81: 1094–1101.
6 Henslee-Downey PJ, Gee AP, Godder K et al. Minimal risk
of graft versus host disease (GVHD) following haploidentical
but partially mismatched related donor (PMRD) bone marrow
transplantation (BMT). Exp Hematol 1994; 22: 147 (Abstr.).
7 Anasetti C, Rybka W, Sullivan KM et al. Graft-v-host disease
is associated with autoimmune-like thrombocytopenia. Blood
1989; 73: 1054–1058.
8 First LR, Smith BR, Lipton J et al. Isolated thrombocytopenia
after allogeneic bone marrow transplantation: existence of
transient and chronic thrombocytopenia syndromes. Blood
1985; 65: 368–374.
9 Cahn JY, Chabot J, Esperou H et al. Autoimmune-like thrombocytopenia after bone marrow transplantation. Blood 1989;
74: 2771 (letter).
10 Bierling P, Cordonnier C, Fromont P et al. Acquired autoimmune thrombocytopenia after allogeneic bone marrow
transplantation. Br J Haematol 1985; 59: 643–646.
11 Neithammer D, Bienzle U, Rodt H et al. Rhesus incompatibility and aplastic anemia the consequence of split chimerism
after bone marrow transplantation post bone-marrow transplantation for severe combined immuno-deficiency. Thymus
1980; 2: 75–82.
12 Bashey A, Owen I, Lucas GF et al. Late onset immune pancytopenia following bone marrow transplantation. Br J Haematol 1991; 78: 268–274.
13 Minchinton RM, Waters AH. Autoimmune thrombocytopenia
and neutropenia after bone marrow transplantation. Blood
1985; 66: 752 (letter).
14 Klumpp TR, Caliguri MA, Rabinowe SN et al. Autoimmune
pancytopenia following allogeneic bone marrow transplantation. Bone Marrow Transplant 1990; 6: 445–447.
15 Koeppler H, Goldman JM. ‘Auto’-immune neutropenia after
allogeneic bone marrow transplantation unresponsive to conventional immunosuppression but resolving promptly after
splenectomy. Eur J Haematol 1988; 41: 182–185.
16 Henslee-Downey PJ, Parrish RS, Romond EH et al. Combined
ex vivo and in vivo T-lymphocyte depletion for the control
of graft-versus-host disease following haploidentical marrow
transplant. Transplant 1995; 61: 738–745.
17 Lee C, Brouillette M, Lamb L et al. Use of a closed system for
Vab-positive T-cell depletion of marrow for use in partially
mismatched related donor (PMRD) transplantation. Advances
in Bone Marrow Purging and Processing; Fourth International Symposium. Orlando, FL, 1994; 389: 523–532.
18 Gee A, Lamb L. Preparation of samples for flow cytometric
analysis: recommendations from the ISHAGE course in San
Jose. In: Wunder E, Sovalat H, Henon PR, Serke S (eds).
Hematopoietic Stem Cells. AlphaMed Press: Dayton, 1994,
p 67.
19 Sullivan KM, Witherspoon RP, Storb R et al. Alternating-day
cyclosporine and prednisone for treatment of high-risk chronic
graft-v-host disease. Blood 1988; 72: 555–561.
20 Atkins K, Biggs JC, Hayes J et al. Cyclosporin A associated
nephrotoxicity in the first 100 days after allogeneic bone marrow transplantation: three distinct syndromes. Br J Haematol
1983; 54: 59–67.
21 Rabinowe SN, Soiffer RJ, Tarbell NJ et al. Hemolytic-uremic
syndrome following bone marrow transplantation in adults for
hematologic malignancies. Blood 1991; 77: 1837–1844.
22 Hazelhurst GR, Brenner MK, Wimperis JZ et al. Haemolysis
after T-cell depleted bone marrow transplantation involving
minor ABO incompatibility. Scand J Haematol 1986; 37: 1–3.
23 Haas RJ, Rieber P, Helmig M et al. Acquired immune haemolysis by anti A 1 antibody following bone marrow transplantation. Blut 1986; 53: 401–404.
24 Keever CA, Small TN, Flomenberg N et al. Immune reconstitution following bone marrow transplantation: comparison of
recipients of T-cell depleted marrow with recipients of conventional marrow grafts. Blood 1989; 73: 1340–1350.
25 Hows J, Beddow K, Gordon-Smith E et al. Donor-derived red
blood cell antibodies and immune hemolysis after allogeneic
bone marrow transplantation. Blood 1986; 67: 177–181.
817