Skip to Content


May-August 1998
Thrombocytopenia Following Allogeneic Bone Marrow Transplantation: Discussion and Analysis of a Case

By Nora J. Morgenstern and Benjamin Lichtiger

Copyright 1993-2010 The University of Texas MD Anderson Cancer Center, Houston, Texas. All rights reserved.

Thrombocytopenia is still a major cause of morbidity and mortality in patients undergoing allogeneic bone marrow transplantation (BMT). In fact, such patients often become refractory to platelet transfusions for reasons that may vary from patient to patient. The case described here illustrates how this can be and suggests an appropriate approach to the diagnosis and work-up of the thrombocytopenic patient undergoing BMT.

Clinical Case History

A 43-year-old white female with acute myelogenous leukemia (AML) M4 that evolved from a myelodysplastic syndrome received an allogeneic bone marrow transplant (BMT) from an HLA-matched sibling. The patient’s blood was A (Rh) negative. Past surgical history revealed splenectomy at age 37 for idiopathic thrombocytopenic purpura (ITP). Screening for red blood cell (RBC) alloantibodies in the patient was negative. Screening for lymphocytotoxic antibodies in the patient showed no complement-fixing anti-HLA or anti-platelet antibodies. The donor was B (Rh) positive. Screening for RBC alloantibodies in the donor was negative.

The RBC phenotypes for both patient and donor were on file. Serologic screening for transfusion-transmitted diseases was unremarkable for both patient and donor. Screening for CMV antibodies by the latex method was nonreactive for both patient and donor. Pre-BMT restriction fragment length polymorphism (RFLP) studies on donor and patient blood revealed differences. This made it possible to look for the donor’s and the patient’s DNA in samples from the patient after BMT and to assess and monitor engraftment and chimerism.

Before transplantation, the donor’s bone marrow was processed and found to contain less than 10 cc of RBCs and no plasma. After a conditioning regimen chemotherapy and total body irradiation, the donor's processed bone marrow was infused into the patient. Following the transplant, the patient received a number of transfusions, IVIgG, G-CSF (granulocyte colony-stimulating factor), FK506 (for prevention of graft-versus-host disease [GVHD]), ganciclovir, and prophylactic antibiotics. On post-BMT day 5, creatinine levels had increased from 1.1 mg/dL pre-BMT to 3.5 mg/dL. The patient required dialysis for acute renal failure, which was attributed to the nephrotoxicity induced by cisplatin and FK506. A skin biopsy from an area with skin rash revealed acute grade 1 GVHD involving only the skin. FK506 was replaced with cyclosporine.

The results of the patient’s type and screen and direct Coombs' tests on post-BMT days 7, 14, 21, and 42 are shown in Table 1. On post-BMT day 42, despite considerable GVHD, the patient was transfusion independent with peripheral blood counts suggesting engraftment (Table 1) and RFLP results showing only donor bands and suggesting 100% engraftment.

Table 1. Laboratory Data* (PDF)

*WBC, white blood cells; FSP, fibrin split products; PT, prothrombin time; PTT, partial thromboplastin time; mf, mixed field.

Around post-BMT day 60, the GVHD appeared to have worsened, and the patient had intermittent lower gastrointestinal (GI) tract bleeding. Skin and GI biopsies showed no evidence of acute or chronic GVHD. RBC and platelet counts started dropping (Table 1) and required transfusions. Platelet counts were especially hard to increase and maintain, despite numerous random donor platelet transfusions and 4 platelet transfusions from the bone marrow donor in 7 days.

Multiple cultures from blood, urine, catheters, and skin were negative. Because the patient was afebrile and on prophylactic antibiotics, an infection was unlikely. The coagulation profile was stable. Although 3-5 cc of a heparin solution (10 U/mL) every 8 hours had been used to flush the dialysis catheters in the early post-BMT period, the patient had not received any heparin for more than a month. Repeat screening for lymphocytotoxic antibodies showed no complement-fixing HLA or platelet antibodies. A platelet suspension immunofluorescent test of the patient’s plasma showed no evidence of antibodies. Studies on the patient’s platelets to detect platelet-bound IgG or complement were negative (1,2). Cytomegalovirus (CMV) antibody tests remained negative, and a CMV antigen assay was negative.

On post-BMT day 67, the patient presented in a confused state. Initially, this was thought to be due to cyclosporine-induced neurotoxicity and electrolyte imbalances. Electrolyte levels showed minor deviations, which were corrected. Cyclosporine was replaced with methotrexate. The patient’s neurologic status improved partially but not completely. She soon developed renal failure and was restarted on dialysis. Hemoglobin levels dropped from 8.5 to 6.0 g/dL, LDH increased, and haptoglobin decreased. A direct Coombs' test was negative. A peripheral smear demonstrated many helmet cells, triangular fragments and other schistocytes, some Howell-Jolly bodies, target cells, and burr cells.

The patient was treated by plasma exchange via cryopoor plasma. After 15 plasma exchange procedures, hemoglobin and platelet counts stabilized. However, an episode of pulmonary edema followed by bilateral pulmonary hemorrhage required mechanical ventilation. Despite all these complications, the patient recovered and was discharged from the hospital on post-BMT day 97. The patient has remained stable and free of recurrences for the last 6 months since discharge from the hospital.


Our patient’s platelet count started dropping on post-BMT day 60 and was hard to increase and maintain despite numerous transfusions of random donor platelets and 4 single-donor platelet transfusions in 7 days. In looking for the cause of this post-BMT thrombocytopenia, we considered that thrombocytopenias can be classified by pathogenetic mechanisms (Table 2).

Table 2. Thrombocytopenias Classified by Pathogenetic Mechanism

Decreased production


    • Toxic agents (e.g., chemotherapy drugs)
    • Radiation
    • Infection
    • Constitutional disorders (e.g., Fanconi's anemia)
    • Aplastic anemia
    • Paroxysmal nocturnal hemoglobinuria
    • Myelopthisis (e.g., tumor, fibrosis)

Ineffective thrombopoiesis

    • Megaloblastic anemia
    • Di Guglielmo's syndrome (erythroleukemia or M6)

Abnormal distribution

Congestive splenomegaly
Myeloid metaplasia, lymphoma
Gaucher's disease

Dilutional loss

Massive blood transfusion

Abnormal destruction

Nonimmune mechanisms

    • Disseminated intravascular coagulation (DIC)
    • Thrombotic thrombocytopenic purpura (TTP)
    • Hemolytic-uremic syndrome (HUS)
    • Vasculitis
    • Intravascular malignancy

Immune mechanisms

    • Idiopathic thrombocytopenic purpura
    • Drug-induced thrombocytopenia
    • Chronic lymphocytic leukemia (CLL), lymphoma
    • Neonatal isoimmune thrombocytopenia
    • Posttransfusion purpura (PTP)

First, we considered nonimmune causes, such as delayed megakaryocyte reconstitution, acute GVHD, bleeding, disseminated intravascular coagulation (DIC), infection, fever, splenomegaly, amphotericin B toxicity, antibiotic toxicity, and veno-occlusive disease (VOD). Second, we considered immune causes, such as HLA alloimmunization, which can be identified in 25-30% of leukemia patients who undergo BMT and platelet transfusion and in as many as 80% of patients with aplastic anemia who undergo BMT (8,9). Other less common causes of post-BMT considered were posttransfusion purpura (PTP) (10), heparin-induced thrombocytopenia (HIT) (11), delayed engraftment due to major ABO incompatibility (12), chronic GVHD (9), and graft failure (13). Regardless of the pathogenetic mechanism, the best initial approach for platelet transfusions in patients with post-BMT thrombocytopenia is shown in Table 3 (9).

Table 3. Strategy for Platelet Transfusion in Patients with Post-BMT Thrombocytopenia

  1. Consider use of leukocyte depletion filters if
    • Patient not previously transfused or pregnant
    • Patient previously transfused or pregnant with no platelet-associated antibodies (lymphocytotoxic antibody screen negative)
  2. Consider autologous platelet collection and freezing, when appropriate
  3. HLA A and B type the patient
  4. Transfuse patient with platelets as needed
  5. Obtain posttransfusion platelet counts every 18-24 hours as long as counts remain satisfactory
  6. If 18- to 24-hour counts are poor, obtain posttransfusion platelet counts every hour and determine the corrected count increment (CCI)*
  7. Test for platelet-associated antibodies when appropriate and identify antibody specificity, if present

*CCI = (posttransfusion count - pretransfusion count) x body surface area (m2) / no. platelets transfused (x 1011). A CCI of 7,500-10,000 is considered satisfactory.

In our patient, delayed megakaryocytic reconstitution, delayed engraftment due to major ABO incompatibility or chronic GVHD, and graft failure were ruled out as causes because the patient had had complete engraftment before post-BMT day 42, as evidenced by the data in Table 1 and from post-BMT RFLP studies. Acute GVHD was a possible cause, but skin and GI biopsies showed no evidence of GVHD. DIC was not present since PT, PTT, fibrinogen, fibrin split products (FSP), and D-Dimmer values were normal. Infection was unlikely since no fever, leukocytosis, leukopenia, or positive cultures were noted and there was coverage with prophylactic antibiotics. However, occult infection could not be ruled out. Moreover, the use of antibiotics was not a likely cause because the patient had been on the same antibiotics since the transplant and had experienced no previous thrombocytopenia since engraftment.

Splenic sequestration was unlikely since the patient had been splenectomized and showed no evidence of the presence of any accessory splenic tissue. This patient also had no clinical evidence of VOD of the liver.

HLA alloimmunization was an unlikely cause since repeat screening for lymphocytotoxic antibodies showed no complement-fixing anti-HLA or anti-platelet antibodies and since platelet-suspension immunofluorescent tests and studies on the patient’s platelets to detect IgG or complement were negative for HLA and platelet antibodies. Also, the patient did not respond to platelet transfusions from the HLA-matched donor.

PTP after BMT (10) is a rare condition characterized by delayed-onset thrombocytopenia that appears 10 days to 2 weeks after blood transfusion. It is most often caused by platelet-specific anti-HPA-1a but can also be caused by anti-HPA-1b and HLA antibodies. Platelet antibodies can be detected in cases of PTP (10) as well as in cases of autoimmune thrombocytopenia (14,15). The pathogenesis of PTP is thought to involve an initial immunization against a platelet antigen and a later platelet destruction, even if the platelets are negative for the antigen in question. This delayed destruction is thought to occur through an "innocent bystander" mechanism by which soluble, previously transfused antigenic material adsorbs to antigen-negative platelets and ultimately causes their destruction. Treatment options for PTP include plasmapheresis, IVIgG, steroids, transfusion of antigen-negative blood components, and column immunoadsorption.

Since allogeneic RBC antibodies can cause hemolysis and delayed erythropoiesis in major ABO-incompatible BMT (12) and anti-PMN antibodies can cause neutropenia and delay engraftment of PMNs (16), it seems likely that a high platelet antibody titer can cause thrombocytopenia and delay engraftment of megakaryocytes (8). As with major and minor ABO-incompatible post-BMT hemolysis, examples of post-BMT immune thrombocytopenia due to recipient (17) or donor (18) anti-platelet antibodies have also been reported. Moreover, post-BMT autoimmune-like thrombocytopenia has been associated with acute or chronic GVHD (14) and with viral (especially CMV) infections (15).

Heparin-induced thrombocytopenia (HIT) (11) occurs in 5-20% of heparin-treated patients. The syndrome is thought to occur when heparin-dependent platelet antibodies attach to the platelet Fc receptor and esterically block the platelet GP1b receptor. It is most commonly seen in patients receiving therapeutic doses of heparin, but has also been seen in patients receiving very small doses of heparin to maintain the patency of intravascular catheters. Diagnosis of HIT is based on the occurrence of thrombocytopenia during heparin administration with no other apparent cause and resolution after cessation of heparin within 2 weeks to 2 months. Two important complications of this syndrome, besides thrombocytopenia, are arterial and venous thrombosis. Because thrombosis is more likely with higher platelet counts, withholding prophylactic platelet transfusions as well as heparin is part of this syndrome’s treatment. As mentioned earlier, our patient had received heparin in the early post-BMT period to flush her dialysis catheters, but had not received any heparin for more than a month. Therefore, HIT probably did not cause our patient’s thrombocytopenia.

When the patient presented on post-BMT day 67 in a confused state, it was initially thought that this was due to cyclosporine-induced neurotoxicity. Cyclosporine-induced changes in the central nervous system (CNS) in patients undergoing BMT include seizures and cortical blindness, with parietal and occipital abnormalities on MRI (19). Hypertension and electrolyte (especially magnesium) imbalance may also be present. Cyclosporine-induced neurotoxicity can be treated by replacing cyclosporine with alternative agents of GVHD prophylaxis, correcting hypomagnesemia, and controlling hypertension. Cyclosporine-induced changes in the CNS usually resolve within 48 hours of discontinuing cyclosporine, although the abnormalities often persist on MRI for some time. However, despite our replacing cyclosporine with methotrexate and correcting electrolyte imbalances, the patient's neurologic impairment did not resolve. Therefore, the cyclosporine-induced neurotoxicity was clearly not the only cause of our patient’s confusion.

Our patient’s history of splenectomy altered the baseline peripheral smear picture (Howell-Jolly bodies, target cells, and burr cells are expected), but this does not explain the presence of schistocytes. Schistocytes in the peripheral smear are suggestive of microangiopathic hemolytic anemia (MHA), whose major causes are listed in Table 4. A peripheral blood smear suggestive of MHA and combined with thrombocytopenia, neurologic symptoms, and renal failure in the absence of DIC, vasculitis, collagen vascular disease, malignant hypertension, disseminated carcinomatosis, or intravascular malignancy, is suggestive of thrombotic microangiopathy (TMA). TMA includes thrombotic thrombocytopenic purpura (TTP) and hemolytic-uremic syndrome (HUS).

Table 4. Major Causes of Microangiopathic Hemolytic Anemia (MHA)

  • Disseminated intravascular coagulation (DIC)
  • Collagen vascular disease
  • Renal vascular disorders (malignant hypertension, preeclampsia, renal transplant rejection)
  • Disseminated carcinomatosis
  • Thrombotic thrombocytopenic purpura (TTP) and hemolytic-uremic syndrome (HUS)

TTP has been associated with autoimmune diseases, pregnancy, viral and nonviral infections, solid tumors, various drugs, chemotherapeutic agents (e.g., cisplatin), spider bites, and other miscellaneous insults (20). However, the etiology remains unknown.

Four possible explanations of the pathophysiology are that (1) TTP plasma lacks a normally occurring immunoglobulin inhibitor; (2) unusually large circulating von Willebrand factor multimers promote platelet agglutination; (3) a plasma factor directly injures the endothelium promoting platelet agglutination; and (4) the endothelium has a deficiency of prostacyclin.

HUS may be difficult to distinguish from TTP. However, compared with TTP, HUS is marked by more common and more severe renal failure, a less prominent fluctuating neurologic deficit, equally severe MHA, slightly less severe thrombocytopenia, childhood predilection, and better prognosis. The uncertain and likely heterogeneous pathogenesis of TTP/ HUS precludes a dogmatic approach to therapy. The first successful use of whole-blood exchange transfusion in TTP was reported by Rubinstein and coworkers (21) following their observation of transient improvement in a patient receiving transfusion support for massive bleeding. It remains unclear whether the therapeutic effect is mediated by removal of toxins or by replacement of a protective factor. Current treatment modalities include therapeutic plasma exchange with fresh frozen plasma, cryosupernatant, corticosteroids, immunosuppressants (vincristine), IVIgG, splenectomy, and anti-platelet agents such as aspirin and dipyridamole (20). Immunoadsorption on a staphylococcal protein A column is under evaluation and may help patients refractory to therapeutic plasma exchange with cryosupernatant (22). The response to plasma exchange decreases with delay in initiating this therapy and with initially higher creatinine levels (23). TTP associated with human immunodeficiency virus (HIV), chemotherapy, occult infections (20), or BMT (26-28) is particularly refractory to therapy. Overall, TTP carries a mortality of 10-20% (20), but refractory TTP, such as transplantation-associated microangiopathy, appears to be more aggressive and lethal.

Transplantation-associated microangiopathy or TTP/HUS is a well-recognized entity (19-28). The postulated physiopathology is endothelial cell damage (24). It has been variously attributed to cytotoxic chemotherapy, total body irradiation, CMV infection, GVHD, and cyclosporine treatment. The diagnosis may be difficult in the BMT setting because most BMT patients experience some degree of renal impairment, altered sensorium, thrombocytopenia, or fever attributable to causes other than TTP/HUS (19). While up to 98% of BMT patients may have some degree of microangiopathy, TTP/HUS shows "significant" fragmentation, defined as 5.0% or more fragmented RBCs (19). It has been shown to occur more often in patients taking cyclosporine and methylprednisolone for GVHD prophylaxis than in patients on cyclosporine and methotrexate (25). TTP/HUS in the BMT setting is generally more resistant to treatment than the classical TTP/HUS (26-28). Most centers institute early and aggressive plasma exchange. When significant RBC fragmentation is present , the possibility that other coexisting complications, such as GVHD, CMV infection, VOD, or diffuse alveolar hemorrhage (DAH), might explain the symptoms and signs should not delay a presumptive diagnosis of TTP. Because early treatment of TTP/HUS may lead to an improved response and because treatment has few complications, it is prudent to err in the direction of overdiagnosis rather than delay treatment and increase the risk of death (19).

Schriber and Herzig (19), in a review of the literature including 172 cases of transplantation associated TTP/HUS, recognized 2 prognostically different groups: a high-mortality group (85%) and a low-mortality group (12%). This stratification was based on the presence of 3 adverse risk factors: (1) presentation before post-BMT day 120; (2) treatment with cyclosporine and/or FK506; and (3) the presence of neurologic symptoms at onset. The authors were not able to analyze the influence on prognosis of risk factors previously associated with TTP/HUS, such us GVHD, CMV infection, or total body irradiation.

Patients who had none of these risk factors belonged to the good-prognosis group (12% mortality with supportive care or plasma exchange with FFP), and those with 2 or 3 of these risk factors belonged to the poor-prognosis group (85% mortality with supportive care or plasma exchange with FFP). Patients with only one of these risk factors belonged to an intermediate-prognosis group.

Patients with high-risk TTP (multifactorial fulminant thrombotic microangiopathy) are typically acutely ill and often have other coexisting disorders, such as infections, GVHD, VOD, or DAH. Patients with low-risk TTP (conditioning regimen-associated TTP) typically have received total body irradiation as part of their conditioning regimens and suffered some degree of radiation nephritis. Such patients manifest renal failure, edema, and hypertension and in up to a third of cases may require chronic dialysis after TTP has resolved.

Stratifying post-BMT TTP patients into prognostic groups allows early institution of aggressive alternate treatment (different from the conventional plasma exchange with FFP) for those patients with a poor prognosis. In particular, treatment alternating between immunoadsorption on a staphylococcal protein A column and plasma exchange with cryosupernatant should increase response and decrease mortality rates in the poor-prognosis patients.


  1. von dem Borne AEG, Verheugt FWA, Ooterhof F, Von Riesz E, Brutel de la Riviere A, Engelfriet CP. A simple immunofluorescence test for the detection of platelet antibodies. Br J Haematol 39:195-207, 1978.
  2. Cines DB, Schreiber AD. Immune thrombocytopenia: Uses of a Coombs’ antiglobulin test to detect IgG and C3 on platelets. N Engl J Med 300:106-111, 1979.
  3. Sokol RJ, Booker DJ, Stamps R, Windle JA. Autoimmune haemolysis and red cell autoantibodies with ABO blood group specificity. Haematologia 26:121-129, 1995.
  4. Greeno EW, Perry EH, Ilstrup SJ, Weisdorf DJ. Exchange transfusion the hard way: Massive hemolysis following transplantation of bone marrow with minor ABO incompatibility. Transfusion 36:71-74, 1996.
  5. Gajewski JL, Petz LD, Calhoun L, et al. Hemolysis of transfused group O red blood cells in minor ABO incompatible unrelated donor bone marrow transplants in patients receiving cyclosporine without post-transplant methotrexate. Blood 79:3076-3085, 1992.
  6. Berkman EM, Caplan SN. Engraftment of Rh-positive marrow in a recipient with Rh antibody. Transplant Proc 9:215-218, 1977.
  7. Lopez A, de la Rubia J, Arriaga F, et al. Severe hemolytic anemia due to multiple red cell antibodies after an ABO-incompatible allogeneic bone marrow transplant. Transfusion 38:247-251, 1998.
  8. Gillis S. The thrombocytopenic purpuras: Recognition and management. Drugs 51:942-953, 1996.
  9. Benson K, Fields K, Hiemenz J, et al. The platelet-refractory bone marrow transplant patient: Prophylaxis and treatment of bleeding. Semin Oncol 20(suppl 6):102-109, 1993.
  10. Evenson DA, Stroncek DF, Pulkrabek S, et al. Posttransfusion purpura following bone marrow transplantation. Transfusion 35:688-693, 1995.
  11. Tezcan AZ, Tezcan H, Gastineau DA, Armitage JO, Haire WD. Heparin induced thrombocytopenia after bone marrow transplantation: Report of two cases. Bone Marrow Transplant 14:487-490, 1994.
  12. Sniecinski IJ, Oien L, Petz LD, Blume KG. Immunohematologic consequences of major ABO-mismatched bone marrow transplantation. Transplantation 45:530-534, 1988.
  13. Petz LD, Swisher SN, Kleinman S, Spence RK, Strauss RG. Clinical Practice of Transfusion Medicine, 3d edn, pp 757-782. New York: Churchill Livingstone, 1996.
  14. Anasetti C, Rybka W, Sullivan KM, Banaji M, Slichter SJ. Graft-v-host disease is associated with autoimmune-like thrombocytopenia. Blood 73:1054-1058, 1989.
  15. Chan JY, Chabot J, Esperou H, Flesch M, Plouvier E, Herve P. Autoimmune-like thrombocytopenia after bone marrow transplantation. Blood 74:2771, 1989.
  16. Stroncek DF, Shapiro RS, Filipovich AH, Plachta LB, Clay ME. Prolonged neutropenia resulting from antibodies to neutrophil-specific antigen NBI following marrow transplantation. Transfusion 33:158-163, 1993.
  17. Bierling P, Pignon JM, Kuentz M, et al. Thrombocytopenia after bone marrow transplantation caused by a recipient origin Bra allo-antibody: Presence of mixed chimerism 3 years after the graft without hematologic relapse. Blood 83:274-279, 1994.
  18. Taaning E, Jacobsen N, Morling N. Graft derived anti-HPA-2b production after allogeneic bone marrow transplantation. Br J Haematol 86:651-653, 1994.
  19. Schriber JR, Herzig GP. Transplantation-associated thrombotic thrombocytopenic purpura and hemolytic uremic syndrome [review]. Semin Hematol 34:126-133, 1997.
  20. Creager AJ, Brecher ME, Bandarenko N. Thrombotic thrombocytopenic purpura that is refractory to therapeutic plasma exchange in two patients with occult infection. Transfusion 38:419-423, 1998.
  21. Rubinstein MA, Kagan BM, MacGillviray MH, Merliss R, Sacks H. Unusual remission in a case of thrombotic thrombocytopenic purpura syndrome following fresh blood exchange transfusions. Ann Intern Med 51:1409-4019, 1959.
  22. Zeigler ZR, Shadduck RK, Nath R, Andrews DF III. Pilot study of combined cryosupernatant and protein A immunoadsorption exchange in the treatment of grade 3-4 bone marrow transplant-associated thrombotic microangiopathy. Bone Marrow Transplant 17:81-86, 1996.
  23. Pereira A, Mazzara R, Monteagudo J, et al. Thrombotic thrombocytopenic purpura/hemolytic uremic syndrome: A multivariate analysis of factors predicting the response to plasma exchange. Ann Hematol 70:319-323, 1995.
  24. Zeigler ZR, Rosenfeld CS, Andrews DF III, et al. Plasma von Willebrand factor antigen (vWF:AG) and thrombomodulin™ levels in adult thrombotic thrombocytopenic purpura and hemolytic uremic syndromes (TTP/HUS) and bone marrow transplant-associated thrombotic microangiopathy (BMT-TM). Am J Hematol 53:213-220, 1996.
  25. Kalhs P, Brugger S, Schwarazinger I, et al. Microangiopathy following allogeneic marrow transplantation: Association with cyclosporine and methylprednisolone for graft versus host disease prophylaxis. Transplantation 60:949-957, 1995.
  26. Sarode R, McFarland JG, Flomenberg N, et al. Therapeutic plasma exchange does not appear to be effective in the management of thrombotic thrombocytopenic purpura/hemolytic uremic syndrome following bone marrow transplantation. Bone Marrow Transplant 16:271-275, 1995.
  27. Llamas P, Romero R, Cabrera R, Sanjuan I, Fores R, Fernandez MN. Management of thrombotic microangiopathy following allogeneic transplantation: What is the role of plasma exchange? Bone Marrow Transplant 20:305-306, 1997.
  28. Chown SR, Goulden N, Cornish JM, Oakhill A, Pamphilon D, Potter MN. The role of plasma exchange for TTP/HUS post-bone marrow transplant [letter]. Bone Marrow Transplant 17:898-899, 1996.

© 2015 The University of Texas MD Anderson Cancer Center