Thrombotic Microangiopathy Following Allogeneic Bone Marrow Transplantation Is Associated with Intensive Graft-Versus-Host Disease Prophylaxis

Home , Aplastic anemia

Bone Marrow Transplantation, (1998) 22, 351–357  1998 Stockton Press All rights reserved 0268–3369/98 $12.00 http://www.stockton-press.co.uk/bmt Thrombotic microangiopathy following allogeneic bone marrow transplantation is associated with intensive graft-versus-host disease prophylaxis

RL Paquette1, L Tran1 and EM Landaw2

Departments of 1Medicine and 2Biomathematics, UCLA School of Medicine, Los Angeles, CA, USA

Summary: is diffuse vascular involvement. The onset of TM has been associated with infection by a variety of bacterial or viral Thrombotic microangiopathy (TM), manifesting clini- pathogens.1 There is a well-documented association cally as thrombotic thrombocytopenic purpura or hem- between infection with E. coli strain O157:H7 and the olytic uremic syndrome, is an uncommon complication occurrence of HUS.2,3 Individuals infected with the human after bone marrow transplantation (BMT). A retrospec- immunodeficiency virus have an increased incidence of tive analysis of potential risk factors for TM following TTP.4 TM also occurs in relation to pregnancy; TTP tends allogeneic BMT was performed. Clinical data were ana- to occur in the ante-partum period while HUS is more fre- lyzed from seven patients diagnosed with severe TM quent post-partum.1 Administration of chemotherapy, and 409 patients who underwent BMT during the same especially mitomycin C, has been associated with the devel- time period and who survived for at least 100 days after- opment of HUS.5 These various etiologies are thought to wards. Six of the seven patients with TM received inten- share the ability to induce injury to the vascular endo- sive GVHD prophylaxis consisting of cyclosporine, thelium. Plasma exchange is highly effective therapy for methotrexate and glucocorticoids, whereas only 66 of TTP; HUS may also respond to this therapy.3,6 the 409 patients without TM received this regimen TM is an uncommon complication of allogeneic bone (P Ͻ 0.001, Fisher’s exact test). This regimen was marrow transplantation (BMT). More than 63 cases of TM administered to patients older than 40 years, or recipi- have been reported to occur in this setting.7 The first case ents of a mismatched or unrelated allograft. Univariate of TM following allogeneic BMT was observed during the analysis also revealed an increased risk of TM associa- earliest use of cyclosporine (CsA) prophylaxis for graft- ted with the use of an unrelated bone marrow donor versus-host disease (GVHD).8 Subsequently, the occur- (P = 0.02), but no significant association with patient age rence of TM was found to be significantly associated with or gender, diagnosis, amount of prior chemotherapy, the use of CsA prophylaxis and the severity of GVHD in transplant conditioning regimen or severity of GVHD. BMT patients.9 However, TM has also been observed in A multivariate exact logistic regression analysis patients undergoing autologous BMT and in allogeneic revealed that only the type of GVHD prophylaxis had BMT patients who have not received CsA, suggesting that a significant impact on the risk for TM. The combined the use of the total body irradiation (TBI) and cyclophos- use of cyclosporine, methotrexate and glucocorticoids as phamide (CY) conditioning regimen might play a critical GVHD prophylaxis may predispose to the development role in the etiology of BMT-related TM.10 of TM following BMT. We recently noted an increase in the incidence of TM in Keywords: thrombotic thrombocytopenic purpura; hemo- patients undergoing allogeneic BMT at our institution. As lytic uremic syndrome; bone marrow transplantation; graft- a result, we conducted a retrospective analysis of disease versus-host disease and treatment-related variables in order to identify factors associated with the development of this complication.

Thrombotic microangiopathy (TM) is a pathological pro- Materials and methods cess characterized by occlusion of the microvasculature by hyaline thrombi resulting in microangiopathic hemolytic anemia, thrombocytopenia and end organ dysfunction. Patients Clinically, TM manifests as hemolytic uremic syndrome Seven patients were identified as having TM due to the (HUS) when the renal vasculature is primarily affected, or presence of microangiopathic hemolytic anemia and pro- as thrombotic thrombocytopenic purpura (TTP) when there gressive thrombocytopenia without laboratory evidence of disseminated intravascular coagulation (positive fibrin degradation product or d-dimer assays). Microangiopathic Correspondence: Dr R Paquette, UCLA Division of Hematology/ Oncology, 42-121 Center for the Health Sciences, 10833 Le Conte Ave, hemolytic anemia was defined as the presence of schisto- Los Angeles, CA 90095-1678, USA cytes on the peripheral blood smear, anemia (hemoglobin Received 16 October 1997; accepted 8 April 1998 Ͻ12 mg/dl), reticulocytosis (absolute reticulocyte count Risk of TTP/HUS after allogeneic BMT RL Paquette et al 352 Ͼ114 000/␮l), and elevated serum lactate dehydrogenase adjusted to maintain serum levels between 200 and (LDH Ͼ220 U/l) persisting for more than 1 week in the 400 ng/ml. CsA was administered for 6 months to related absence of a positive direct antiglobulin test. The control BMT patients and 12 months to unrelated BMT patients, patient population included all patients who had undergone after which time the dose was tapered by 25 mg/day each allogeneic BMT at UCLA between October 1983 and June week in the absence of active GVHD. Methotrexate (MTX) 1996, received CsA for GVHD prophylaxis and survived 15 mg/m2 was given intravenously on day 1, and 10 mg/m2 at least 100 days post-BMT without developing TM. All on days 3 and 6 following BMT. Either methylprednisolone patient data were obtained from computerized records and (Mpred) or prednisone was administered twice daily at hospital charts with the approval of the UCLA Institutional 0.5 mg/kg on days +3to+42, once daily at 0.5 mg/kg on Review Board. days +43 to +56 and once daily at 0.3 mg/kg on days +57 to +100, then tapered over 1 month in the absence of GVHD. Acute GVHD was diagnosed and graded according to Prior chemotherapy standard clinical criteria.11 Treatment was individualized In order to assess the potential effect of chemotherapy depending on disease severity, but initially consisted of administered prior to conditioning and transplantation, a Mpred 2 mg/kg/day. Higher Mpred doses with or without representative sample of 30 control patients was selected antithymocyte globulin were administered depending on the so that the proportion of diagnoses was identical to that of initial response to therapy. the control population. The distribution of the subpopul- ation did not differ from that of the entire control popu- Therapeutic apheresis and response to treatments lation with respect to age, gender, donor status, GVHD prophylaxis, acute GVHD severity, or conditioning regi- Plasma exchange was performed using a Cobe Spectra men. The prior therapy of the control subgroup then was device (Lakewood, CO, USA) to replace one plasma vol- compared to that of the TM patients. ume with fresh frozen plasma on a daily basis. The staphylococcal protein A column (PROSORBA column; ´ Conditioning regimens IMRE Corp., Seattle, WA, USA) was used in the setting of unresponsive or progressive TM in spite of plasma A large number of different regimens were used to con- exchange. One liter of plasma was processed through the dition patients for BMT. In order to simplify the data analy- PROSORBA column using a Cobe Spectra machine every sis, conditioning regimens were grouped as follows: other day. (1) CY/TBI included CY 60 mg/kg for 2 days with TBI, total lymphoid irradiation (TLI) or both using various doses and schedules, or CY 90 mg/kg with etoposide 60 mg/kg Statistical analysis and TBI 1200 cGy in six fractions, or CY 60 mg/kg for 2 Nine variables were evaluated for possible association with days with TBI 1125 cGy in five fractions and vincristine TM: age (age Ͼ median vs age р median, where median 0.2 mg/kg for 2 days; (2) cytosine arabinoside (Ara- was 27 years), gender, diagnosis, donor type (related vs C)/CY/TBI included Ara-C 0.5 g/m2 i.v. every 12 h × 4 unrelated and HLA-identical vs mismatched), transplant doses with CY 60 mg/kg i.v. × 2 days and TBI 1125 cGy date (10/83–12/88 vs 1/89–6/96), conditioning regimen, use in five fractions; (3) busulfan (BU)/CY included BU of TBI, GVHD prophylaxis regimen and GVHD severity. 1 mg/kg p.o. every 6 h × 16 doses with CY 60 mg/kg/day Univariate analysis was performed using Fisher’s exact test i.v. × 2 days alone or with melphalan 140 mg/m2; (4) Ara- (two-tailed). Factors having univariate associations or bor- C/TBI included Ara-C 3 g/m2 every 12 h for 4 days with derline associations with TM were further investigated in TBI 1350 cGy in six fractions alone or with TLI 225 cGy, a logistic model to assess the additive effects of these vari- or Ara-C 2 g/m2 every 12 h for 4 days with mitoxantrone ables and their interactions on the occurrence of TM and 12 mg/m2 for 3 days and TBI 1125 cGy in five fractions; to confirm the lack of association with the remaining vari- and (5) other. ables. Multivariate logistic analysis was performed using exact methods (LogXact).12 Additional analyses were per- GVHD prophylaxis and treatment formed using, as the control group, only 296 of the 409 control patients who were over the age of 16 years (median All study patients received CsA for GVHD prophylaxis. No age 33). These results were similar to the analysis for the conventions for the selection of GVHD prophylaxis were full cohort and are not included. consistently followed during the study period. However, indications for the use of triple immunosuppressive therapy consisting of CsA, MTX and glucocorticoids included Results patients over 39 years of age, or those receiving bone marrow from an unrelated or mismatched related donor. Patients with TM following BMT Administration of CsA was begun on day −1 by giving 3 mg/kg intravenously over 12 h, followed by 3 mg/kg/day Seven patients (mean ± s.d., age, 37 ± 12 years) who for patients р35 years or 2 mg/kg/day for patients Ͼ35 underwent allogeneic BMT at UCLA between October years by continuous intravenous infusion. When patients 1983 and June 1996 were diagnosed with TM. The clinical were tolerating oral medications, the intravenous dose of and laboratory characteristics of the TM patients are shown CsA was administered orally twice daily. Doses were in Tables 1 and 2. TM was diagnosed between 47 and 105 Risk of TTP/HUS after allogeneic BMT RL Paquette et al 353 Table 1 Patients with TM following allogeneic BMT

Patient Age Sex Diagnosis Year of BMT Donor status BMT conditioning GVHD prophylaxis Stage aGVHDa

1 43 F ALL in 2nd 1985 6/6 related Ara-C, TBI CsA 0 remission 2 32 F CML, chronic phase 1991 6/6 unrelated Ara-C, CY, TBI CsA, MTX, GCs I 3 35 M Refractory AML 1991 6/6 unrelated Ara-C, CY, TBI CsA, MTX, GCs 0 4 48 M Refractory AML 1994 6/6 unrelated BU, CY CsA, MTX, GCs III 5 19 F CML, chronic phase 1995 5/6 related Ara-C, CY, TBI CsA, MTX, GCs III 6 27 F ALL, 3rd relapse 1995 6/6 unrelated Ara-C, CY, TBI CsA, MTX, GCs II 7 52 F CML, chronic phase, 1996 6/6 related BU, CY CsA, MTX, GCs III relapse after BMT aPrior to diagnosis of TM. ALL = acute lymphocytic leukemia; CML = chronic myelocytic leukemia; AML = acute myelocytic leukemia; BMT = bone marrow transplant; TBI = total body irradiation; Ara-C = cytarabine; CY = cyclophosphamide; BU = busulfan; GVHD = graft-versus-host disease; CsA = cyclosporine; MTX = methotrexate; GCs = glucocorticoids; aGVHD = acute GVHD.

Table 2 Clinical characteristics of TM patients at diagnosis

Patient Day of diagnosis MAHA Thrombocytopenia ARF Neurologic changes Fever

1d+91 ++0 ++ 2d+71 +++++ 3d+105 ++000 4d+48 ++++0 5d+47 ++0 ++ 6d+56 ++0 ++ 7d+93 +++00

MAHA = microangiopathic hemolytic anemia; ARF = acute renal failure. days after BMT. The majority of the patients had fevers and patients did not differ from that of the control group. Other neurological abnormalities but it was sometimes difficult to variables including patient age or gender, diagnosis, BMT determine whether these were due to TM, concurrent infec- conditioning regimen, date of transplant, or GVHD severity tion or liver dysfunction. Three patients had acute renal also had no association with the development of TM insufficiency at the time TM was diagnosed. (Table 3). Exact logistic regression analysis showed that after controlling for the type of GVHD prophylaxis, no other variables had a significant association with the occur- Clinical features associated with TM rence of TM. The clinical characteristics of the TM patients were com- ± pared to those of 409 control patients (mean s.d. age, Outcome of therapy for TM 26 ± 13 years) who had undergone allogeneic BMT at UCLA between October 1983 and June 1996, received CsA CsA was discontinued in all but one patient when TM was for GVHD prophylaxis and survived at least 100 days fol- diagnosed. There was no amelioration of the TM as a result lowing BMT (Table 3). Univariate analysis revealed that of CsA discontinuation. All of the patients underwent thera- the use of a three-drug regimen consisting of CsA, MTX peutic apheresis with plasma exchange; three patients also and glucocorticoids for GVHD prophylaxis (six of seven had apheresis with the use of a staphylococcal protein A patients with TM vs 66 of 409 controls; P = 0.0007, Fish- column (Table 4). None of the patients had any evidence er’s exact test), or use of an unrelated bone marrow donor of hematological or clinical improvement as a result of the (four of seven vs 68 of 409; P = 0.019) was associated with apheresis procedures. the occurrence of TM in our patient population. Use of TBI All of the patients died between 15 and 307 days follow- was similar in each group (five of seven vs 294 of 409; ing the diagnosis of TM. Although all of the patients had P = 1.0). A representative subgroup of 30 control patients active TM at the time of death, TM was not the primary was compared to the TM patients to assess the association cause of death in any of the cases. of the amount of previous chemotherapy on the risk for TM. Of the 30 control patients, 10 received no prior therapy, 10 had one to two courses of prior chemotherapy, and Discussion 10 underwent more than two courses of treatment or a previous bone marrow transplant. The numbers for the TM The GVHD prophylaxis regimen of CsA, MTX and gluco- patients for the same categories were two, three and two, corticoids was found to be associated with an increased risk respectively. Therefore, previous treatment of the TM of developing TM in our patient population. Because this Risk of TTP/HUS after allogeneic BMT RL Paquette et al 354 Table 3 Univariate analysis of patient variables

Variables Categories TM patients Control patients Association with TM (n = 7) (n = 409) (Fisher’s exact text)

Age р27 years 2 199 P = 0.28 Ͼ27 years 5 210 Gender female 5 164 P = 0.13 male 2 245 Diagnosis AML 2 127 P = 0.85 CML 3 95 ALL 2 90 AA or PNH 0 55 HD or NHL 0 15 Other 0 27 Donor type 6/6 related 2 322 P = 0.01 5/6 related 1 19 6/6 unrelated 4 57 5/6 unrelated 0 11 Date of transplant 1983–1988 1 195 P = 0.13 1989–1996 6 214 Conditioning regimen TBI/TLI+Ara-C+CY 4 153 P = 0.53 TBI/TLI+Ara-C 1 72 TBI/TLI+CY 0 92 BU+CY 2 64 Other 0 28 TBI regimen 5 294 P = 1.00 No TBI 2 115 GVHD prophylaxis CsA 1 218 P Ͻ 0.001 CsA + MTX 0 76 CsA + GCs 0 49 CsA + MTX + GCs 6 66 aGVHD severity 0 2 167 P = 0.43 1186 2172 3360 4024

AA = aplastic anemia; PNH = paroxysmal nocturnal hemoglobinuria; HD = Hodgkin’s disease; NHL = nonHodgkin’s lymphoma; TLI = total lymphoid radiation.

Table 4 Treatment of TM

Patient CsA Apheresis Response Outcome discontinued PEX SPA

1 + 5 0 None Death, relapse (d+398) 2 + 10 0 None Death, interstitial pneumonia (d+87) 3 0 4 3 None Death, relapse (d+234) 4 + 16 0 None Death, liver failure (d+68) 5 + 9 0 None Death, brain mass (d+73) 6 + 8 3 None Death, liver failure (d+71) 7 + 16 1 None Death, interstitial pneumonia (d+142)

CsA = cyclosporine; PEX = plasma exchange; SPA = staphylcoccal protein A column.

regimen was primarily employed in patients with a high were the indications for the use of triple immunosuppres- risk for developing acute GVHD it is difficult to determine sion. Our analysis, and those of previous publications, have whether the increased risk of TM was related to the not identified increased age as a risk factor for TM. Univar- immunosuppressive regimen or to the risk factors for which iate analysis did reveal that the use of unrelated donor bone the prophylaxis was given. Age over 39, or use of bone marrow was statistically associated with the development marrow from an unrelated or mismatched related donor of TM. However, the multivariate analysis failed to demon- Risk of TTP/HUS after allogeneic BMT RL Paquette et al 355 strate an additive effect of donor status on TM risk after phate (ADP), epinephrine or collagen in vitro.24 A similar controlling for GVHD prophylaxis regimen. Because these effect occurs in vivo: platelets obtained from normal indi- variables are closely linked and our sample size is small, viduals or renal allograft recipients treated with CsA are it may be difficult to exclude the use of an unrelated donor hyperaggregable in response to ADP or epinephrine.24,25 as an independent risk factor for developing TM. This CsA-treated renal allograft patients have diminished plate- association has not been reported previously, but the let ADP and ATP pools, and have persistently elevated majority of papers describing BMT-related TM preceded serum platelet factor 4 levels, suggesting that chronic plate- the use of unrelated donors. In recent reports of TM follow- let activation is occurring.25 ing allogeneic BMT, 33–89% of patients suffering this The importance of CsA in the pathogenesis of post-BMT complication received their bone marrow from an unrelated TM is supported by reports of TM occurring in patients donor.13–16 The GVHD prophylaxis regimens used by each receiving CsA following solid organ transplantation. center were not described in detail, but at the Royal Hospi- Patients treated with CsA after receiving a renal or hepatic tal for Sick Children, where T cell depletion of the graft allograft have developed HUS.26,27 Because TM developed and CsA were used for GVHD prophylaxis, 89% of the in these patients in the absence of chemotherapy or radi- patients who developed TM had undergone an unrelated ation treatment, and without the development of GVHD or transplant.15 This suggests that recipients of an unrelated graft rejection, it appears that CsA administration alone BMT may have an increased risk of developing TM com- may be sufficient to induce TM in some individuals. Our pared to those receiving bone marrow from a matched sib- findings suggest that additional treatment-related variables ling, independent of GVHD prophylaxis regimen. may increase the risk of developing TM in patients receiv- Although TM was primarily observed in patients at high ing CsA after BMT. risk for acute GVHD, the severity of acute GVHD experi- Glucocorticoids (GCs) affect vascular endothelial cells enced by the TM patients was not significantly different in a manner that could enhance the prothrombotic effects from that observed in the control patients. These findings of CsA. GCs block endothelial cell production and release differ from those of a previous study in which BMT of the vasodilators PGE2 and PGI2 in response to various patients with grade II or greater acute GVHD had a signifi- stimuli.28–34 Administration of GCs to thrombocytopenic cantly higher incidence of TM than those with grade 0 or 1 rabbits results in a dose-dependent shortening of the bleed- GVHD.9 We also failed to identify a significant association ing time which may result from inhibition of endothelial between the use of TBI conditioning regimen and the devel- PGI2 production and consequent augmentation of platelet opment of TM. This suggests that radiation does not play aggregation.29 GCs also stimulate vascular endothelial cell an essential role in the causation of TM, as was pre- synthesis of fibronectin matrix, which is capable of viously proposed.10 inducing platelet adhesion to the subendothelium by bind- CsA has many biological effects that could contribute to ing glycoprotein IIb/IIIa on platelet surface membranes.35 the causation of TM. CsA inhibits the ability of cultured The prolonged administration of both CsA and GCs may human endothelial cells to produce prostacyclin (PGI2)in result in an increased risk of developing TM. Recently, response to arachidonic acid, thrombin or calcium ionoph- allogeneic BMT patients receiving Mpred and CsA for ores.17,18 Administration of CsA to rats leads to increased GVHD prophylaxis were found to have significantly higher renal excretion of thromboxane B2 (TxB2), the stable serum LDH levels, lower platelet counts and higher plasma metabolite of TxA2, suggesting that CsA induces increased von Willebrand factor antigen levels than patients receiving 19 22 renal synthesis of TxA2. Impaired PGI2 production or MTX and CsA as GVHD prophylaxis. Three of 14 enhanced TxA2 elaboration resulting from CsA adminis- patients who received GCs and CsA in that series tration could cause vasoconstriction and enhanced platelet developed TM. All but one of our patients who developed aggregability. CsA also has been reported to have a direct TM received GCs, CsA and MTX as prophylaxis for cytotoxic effect on vascular endothelial cells in vitro.20 GVHD. The relative contribution of the GCs or MTX to Endothelial cell damage could result in the exposure of sub- the risk for TM cannot be assessed because no patient who endothelial collagen and further augment platelet aggre- received either GCs or MTX alone with CsA developed gation. There is clinical evidence that CsA can induce vas- TM. The use of GCs or MTX as single agents, even in high cular endothelial cell injury in vivo. Elevated serum levels doses, has not been associated with an increased risk of of von Willebrand factor (vWF), a protein synthesized and TM. Therefore, it is likely that the increased frequency of released by the vascular endothelium, were observed in TM in our patients derives from the combined effects of renal transplant recipients experiencing CsA nephrotoxic- several treatment-related variables. ity.21 Elevated serum vWF levels were also observed in Plasma exchange or the use of a staphylococcal protein allogeneic BMT patients receiving CsA, but not MTX, as A column was unsuccessful in treating TM in our patients, GVHD prophylaxis; patients who developed TM while all of whom died of their underlying diseases or transplant- receiving CsA had the highest vWF levels.22 Another vas- related complications. Previous experience with the use of cular endothelial protein, thrombomodulin, was present at plasma exchange or a staphylococcal protein A column for elevated levels in the plasma of patients with CsA nephro- TM following allogeneic BMT resulted in mixed success. toxicity, consistent with induction of endothelial injury There have been at least 53 such patients reported in the by CsA.23 literature to date.13–16,36–44 Nineteen of 33 patients treated In addition to its effects on endothelial cells, CsA also with plasma exchange alone had some clinical improve- affects platelet function. CsA directly increases the aggre- ment, but only seven responding patients survived. gation of normal platelets in response to adenosine diphos- Responses to plasma exchange were often incomplete and Risk of TTP/HUS after allogeneic BMT RL Paquette et al 356 almost all survivors had persistent renal insufficiency. 7 Moake JL, Byrnes JJ. Thrombotic microangiopathies associa- Although early experience with a staphylococcal protein A ted with drugs and bone marrow transplantation. Hematol column was negative, recent reports from the Western Oncol Clin North Am 1996; 10: 485–497. Pennsylvania Cancer Institute suggest that a regimen of 8 Powles RL, Clink HM, Spence D et al. Cyclosporine A to alternating plasma exchange using cryoprecipitate-poor prevent graft-versus-host disease in man after allogeneic bone- marrow transplantation. Lancet 1980; 1: 327–329. plasma and apheresis with a protein A column is effective 14,16 9 Holler E, Kolb HJ, Hiller E et al. Microangiopathy in patients in approximately half of post-BMT TM patients. In on cyclosporine prophylaxis who developed acute graft-ver- general, patients with allogeneic BMT-related TM often die sus-host disease after HLA-identical bone marrow transplan- from infection, progressive GVHD, complications of the tation. Blood 1989; 73: 2018–2024. TM or relapse of the underlying disease. Survival has been 10 Chappell ME, Keeling DM, Prentice HG, Sweny P. Haemo- shown to correlate inversely with the severity of the TM, lytic uraemic syndrome after bone marrow transplantation: an which can be determined by the level of serum lactate adverse effect of total body irradiation? Bone Marrow Trans- dehydrogenase and the percentage of schistocytes on the plant 1988; 3: 339–347. peripheral blood smear.44 The poor response of BMT-asso- 11 Thomas ED, Storb R, Clift RA et al. Bone-marrow transplan- ciated TM to plasma exchange in this series may partly be tation. New Engl J Med 1975; 292: 895–902. 12 Mehta CR, Patel NR. Exact logistic regression: theory and due to the presence of concomitant life-threatening dis- examples. Stat Med 1995; 14: 2143–2160. orders in most of the patients or the severe nature of 13 Sarode R, McFarland JG, Flomenberg N et al. Therapeutic their TM. plasma exchange does not appear to be effective in the man- Our patients had serious TM at the time that it was agement of thrombotic thrombocytopenic purpura/hemolytic appreciated. Because patients may often experience protrac- uremic syndrome following bone marrow transplantation. ted cytopenias following allogeneic BMT the diagnosis of Bone Marrow Transplant 1995; 16: 271–275. TM may be delayed. Retrospectively, most patients had 14 Zeigler ZR, Shadduck RK, Nath R, Andrews DF III. Pilot laboratory abnormalities suggestive of TM for days to study of combined cryosupernatant and protein A immunoad- weeks prior to the time that it was clinically evident. Given sorption exchange in the treatment of grade 3–4 bone marrow the often poor response of severe BMT-associated TM to transplant-associated thrombotic microangiopathy. Bone Marrow Transplant 1996; 17: 81–86. standard therapy, early intervention may be of benefit. 15 Chown SR, Goulden N, Cornish JM et al. The role of plasma BMT patients receiving an aggressive immunosuppressive exchange for TTP/HUS post-bone marrow transplant. Bone regimen should be monitored closely for the development Marrow Transplant 1996; 17: 898–899. of this complication so that treatment can be instituted early 16 Dua A, Zeigler ZR, Shadduck RK et al. Apheresis in grade 4 in the course of the disease. Alternatively, because the bone marrow transplant associated thrombotic microangio- benefit of triple immunosuppression in patients at high risk pathy: a case series. J Clin Apheresis 1996; 11: 176–184. for GVHD has not been clearly established, a prospective 17 Brown Z, Neild GH. Cyclosporine inhibits prostacyclin pro- clinical trial should be conducted to determine the impact duction by cultured human endothelial cells. Transplant Proc of this regimen on the survival of this patient population. 1987; 19: 1178–1180. 18 Voss BL, Hamilton KK, Samara ENS, McKee PA. Cyclospor- ine suppression of endothelial prostacyclin generation. Trans- plantation 1988; 45: 793–796. Acknowledgements 19 Perico N, Benigni A, Zoha C et al. Functional significance of exaggerated renal thromboxane A2 synthesis induced by This work was supported in part by Public Health Service grant cyclosporin A. Am J Physiol 1986; 251: F581–F587. K08 HL02919 awarded by the National Heart, Lung and Blood 20 Zoja C, Furci L, Ghilardi F et al. Cyclosporin-induced endo- Institute and National Cancer Institute grant CA16042. thelial cell injury. Lab Invest 1986; 55: 455–462. 21 Brown Z, Neild GH, Willoughby JJ et al. Increased factor VIII as an index of vascular injury in cyclosporine nephrotoxicity. Transplantation 1986; 42: 150–153. References 22 Kahls P, Brugger S, Schwarzinger I et al. Microangiopathy following allogeneic marrow transplantation. Association with 1 Rose M, Rowe JM, Eldor A. The changing course of throm- cyclosporine and methylprednisolone for graft-versus-host dis- botic thrombocytopenic purpura and modern therapy. Blood ease prophylaxis. Transplantation 1995; 60: 949–957. Rev 1993; 7: 94–103. 23 Yoshida M, Kozaki M. Ioya N et al. Plasma thrombomodulin 2 Boyce TG, Swerdlow DL, Griffin PM. Eschericia coli levels as an indicator of vascular injury caused by cyclospor- O157:H7 and the hemolytic-uremic syndrome. New Engl J ine nephrotoxicity Transplantation 1990; 50: 1066–1069. Med 1995; 333: 364–368. 24 Grace AA, Barradas MA, Mikhailidis DP et al. Cyclosporine 3 Siegler RL. The hemolytic uremic syndrome. Pediatr Clin A enhances platelet aggregation. Kidney Int 1987; 32: 889– North Am 1995; 42: 1505–1529. 895. 4 Ucar A, Fernandez HF, Byrnes JJ et al. Thrombotic 25 Cohen H, Neild GH, Patel R et al. Evidence for chronic plate- microangiopathy and retroviral infections: a 13-year experi- let hyperaggregability and in vivo activation in cyclosporin- ence. Am J Hematol 1994; 45: 304–309. treated renal allograft recipients. Thromb Res 1988; 49: 91– 5 Doll DC, Yarbro JW. Vascular toxicity associated with anti- 101. neoplastic agents. Semin Oncol 1992; 19: 580–596. 26 Van Buren D, Van Buren CT, Flechner SM et al. De novo 6 Rock GA, Shumak KH, Buskard NA et al. Comparison of hemolytic uremic syndrome in renal transplant recipients plasma exchange with plasma infusion in the treatment of immunosuppressed with cyclosporine. Surgery 1985; 98: thrombotic thrombocytopenic purpura. New Engl J Med 1991; 54–62. 325: 393–397. 27 Bonser RS, Adu D, Franklin I, McMaster P. Cyclosporin- Risk of TTP/HUS after allogeneic BMT RL Paquette et al 357 induced hemolytic uremic syndrome in liver allograft recipi- 37 Craig JIO, Sheehan T, Bell K. The haemolytic uraemic syn- ent. Lancet 1984; 2: 1337. drome and bone marrow transplantation. Br Med J 1987; 28 MacIntyre DE, Pearson JD, Gordon JL. Localisation and 295: 887. stimulation of prostacyclin production in vascular cells. Nat- 38 Cohen H, Bull HA, Seddon A et al. Vascular endothelial cell ure 1978; 271: 549–551. function and ultrastructure in thrombotic microangiopathy fol- 29 Blajchman MA, Senyi AF, Hirsh J et al. Shortening of the lowing allogeneic bone marrow transplantation. Eur J Haema- bleeding time in rabbits by hydrocortisone caused by inhi- tol 1989; 43: 207–214. bition of prostacyclin generation by the vessel wall. J Clin 39 Loomis LJ, Aronson AJ, Rudinsky R, Spargo BH. Hemolytic Invest 1979; 63: 1026–1035. uremic syndrome following bone marrow transplantation: a 30 Jose PJ, Page DA, Wolstenholme BE et al. Bradykinin-stimu- case report and review of the literature. Am J Kidney Dis 1989; lated prostaglandin E2 production by endothelial cells and its 14: 324–328. modulation by antiinflammatory compounds. Inflammation 40 Tschuchnigg M, Bradstock KF, Koutts J et al. A case of 1981; 5: 363–378. φ thrombotic thrombocytopenic purpura following allogeneic 31 J rgensen KA, Stoffersen E. Hydrocortisone inhibits platelet bone marrow transplantation. Bone Marrow Transplant 1990; prostaglandin and endothelial prostacyclin production. Pharm- 5: 61–63. acol Res Comm 1981; 13: 579–586. 41 Silva VA, Frei-Lahr D, Brown RA, Herzig GP. Plasma 32 Lewis GD, Campbell WB, Johnson AR. Inhibition of prosta- exchange and vincristine in the treatment of hemolytic uremic glandin synthesis by glucocorticoids in human endothelial syndrome/thrombotic thrombocytopenic purpura associated cells. Endocrinology 1986; 119: 62–69. with bone marrow transplantation. J Clin Apheresis 1991; 6: 33 De Caterina R, Weksler BB. Modulation of arachidonic acid metabolism in human endothelial cells by glucocorticoids. 16–20. Thromb Haemostas 1986; 55: 369–374. 42 Johnson PRE, Liu Yin JA, Drayson MT. Thrombotic throm- 34 Rosenbaum RM, Cheli CD, Gerritsen ME. Dexamethasone bocytopenic purpura following allogeneic bone marrow trans- inhibits prostaglandin release from rabbit coronary microves- plantation. Bone Marrow Transplant 1991; 7: 321–322. sel endothelium. Am J Physiol 1986; 250: C970–C977. 43 Juckett M, Perry EH, Daniels BS, Weisdorf DJ. Hemolytic 35 Piovella F, Giddings JC, Almasio P et al. Effects of ticlopidine uremic syndrome following bone marrow transplantation. and dexamethasone on fibronectin and factor VIII-related anti- Bone Marrow Transplant 1991; 7: 405–409. gen synthesis by cultured endothelial cells. Thromb Res 1983; 44 Zeigler ZR, Shadduck RK, Nemunaitis J et al. Bone marrow (Suppl. IV): 69–73. transplant-associated thrombotic microangiopathy: a case ser- 36 Marshall RJ, Sweny P. Haemolytic-uraemic syndrome in ies. Bone Marrow Transplant 1995; 15: 247–253. recipients of bone marrow transplants not treated with cyclosporin A. Histopathology 1986; 10: 953–962.