Thrombotic Microangiopathy after Allogeneic Stem Cell Transplantation in the Era of Reduced-Intensity Conditioning: The Incidence Is Not Reduced

Transcription

1 Biology of Blood and Marrow Transplantation 10: (2004) 2004 American Society for Blood and Marrow Transplantation /04/ $30.00/0 doi: /j.bbmt Thrombotic Microangiopathy after Allogeneic Stem Cell Transplantation in the Era of Reduced-Intensity Conditioning: The Incidence Is Not Reduced Avichai Shimoni, Moshe Yeshurun, Izhar Hardan, Abraham Avigdor, Isaac Ben-Bassat, Arnon Nagler The Division of Hematology and Bone Marrow Transplantation, Chaim Sheba Medical Center, Tel-Hashomer, Israel Correspondence and reprint requests: Avichai Shimoni, MD, Department of Bone Marrow Transplantation, Chaim Sheba Medical Center, Tel-Hashomer, Israel ( Received December 16, 2003; accepted March 1, 2004 ABSTRACT Thrombotic microangiopathy (TMA) is one of the most severe complications of stem cell transplantation (SCT). Endothelial cell injury caused by the toxic effects of high-dose chemoradiotherapy is likely the primary event in pathogenesis. The incidence, clinical settings, and risk factors for TMA in the era of nonmyeloablative conditioning have not been well defined. The data on 147 consecutive SCTs in a single center were collected, and patients with TMA were identified. Patient characteristics, response to therapy, and outcome were recorded, and risk factors were determined. TMA occurred in 22 of 147 transplantations, with a projected incidence of 20% 4%. TMA occurred in 3 clinical settings: classic multifactorial TMA, TMA associated with severe hepatic graft-versus-host disease (GVHD), and TMA associated with second SCT, with a projected incidence of 8% 3%, 73% 14%, and 70% 16% of patients at risk, respectively. TMA occurred after 23% 6% of nonmyeloablative and 16% 5% of myeloablative conditioning regimens (not significant). Univariate analysis determined SCT from unrelated donors, SCT during advanced or active disease, second SCT within 6 months of a prior SCT, and acute GVHD as risk factors for TMA. The last 2 factors remained significant in a multivariate model. Thirty-two percent of patients responded to therapy. The peri-tma mortality rate was 68% 10%. Six patients had diffuse alveolar hemorrhage complicating TMA. SCT-associated TMA is a relatively common complication with unsatisfactory therapy and grim prognosis. Fludarabine-based nonmyeloablative conditioning does not confer a lesser risk for TMA. This observation may relate to the selective use of these regimens in elderly and heavily pretreated patients or to the lack of reduction of GVHD with these regimens, and fludarabine itself may be involved in causing endothelial damage. Further exploration of novel preventive and therapeutic measurements is required in high-risk settings American Societyfor Blood and Marrow Transplantation KEY WORDS Thrombotic microangiopathy Stem cell transplantation Nonmyeloablative INTRODUCTION Thrombotic microangiopathy (TMA) is characterized by the classic clinical pentad of thrombocytopenia, microangiopathic hemolytic anemia, neurologic abnormalities, renal abnormalities, and fever [1] resulting from the formation of widespread platelet thrombi within the microvasculature. Classic idiopathic TMA is associated with severe deficiency or inhibition of the plasma metalloprotease ADAMTS13 [2-4]. This protease specifically cleaves von Willebrand factor, thus reducing its multimeric size. Idiopathic TMA is characterized by inhibition of ADAMTS13 by an autoantibody, thus causing accumulation of unusually large von Willebrand factor multimers, which are implicated in the formation of platelet aggregates [4]. TMA is one of the most severe and devastating complications of stem cell transplantation (SCT) [5-17]. SCT-associated TMA is not related to ADAMTS13 deficiency [7,18-19]; rather, endothelial cell injury is likely the primary event in the pathogenesis [6,20]. Multiple factors contribute to endothelial damage after SCT. The initial endothelial injury is induced by 484

2 Posttransplantation TMA the toxic effects of high-dose chemotherapy and total body irradiation given during pre-sct conditioning. However, TMA is much more common after allogeneic SCT, occurring on average after 5% to 15% of SCTs, compared with less than 1% after autologous SCT [5-17]. The clinical severity is also higher after allogeneic SCT, for which TMA-associated mortality is on the average 50%, compared with 26% in the autologous setting [6]. Factors associated with allogeneic SCT such as the use of cyclosporin A, acute graft-versus-host disease (GVHD), cytokine release syndromes, and infections such as with cytomegalovirus (CMV) or fungi [21] may perpetuate endothelial cell injury and enhance the appearance of TMA. In particular, the use of cyclosporin A may have a central role in the pathogenesis of SCT-associated TMA. The importance of cyclosporine is documented by the observation of TMA after solid organ transplantation in association with its use, although this occurs with a lower incidence of approximately 5% [22]. Cyclosporine may cause direct endothelial injury [6,20,23] and may also have procoagulant activity because of increased platelet aggregation and vasoconstriction caused by a variety of effects on vascular endothelium and platelets [7]. Inflammatory cytokines involved in GVHD may mediate vascular endothelial cell injury or procoagulant activity, and the endothelial cell itself may become a target of GVHD [7]. Reduced-intensity or nonmyeloablative conditioning regimens have been recently introduced into clinical practice. These regimens were designed to allow some tumor cytoreduction, as well as sufficient immunosuppression to promote allograft engraftment and the induction of a graft-versus-malignancy effect as the primary curative therapeutic goal [24,25]. Nonmyeloablative SCT (NST) is less toxic, thus allowing the treatment of older patients and patients with comorbidities not eligible for standard ablative conditioning. The relative toxicities, GVHD rates, and overall outcome with these regimens still need to be defined. Because toxic effects of the conditioning regimen have a central role in the pathogenesis of SCTassociated TMA, it is conceivable that TMA risk after NST will markedly decrease. This study was designed to evaluate the risk factors and clinical settings of TMA in the era of NST, and the results showed that TMA risk did not decrease. PATIENTS AND METHODS Patient Identification Data on all allogeneic transplantations from July 1, 2000, in a single transplant center were prospectively recorded, and a data set for patients with SCTassociated TMA was established. A total of 147 allogeneic transplants were given to 132 patients with BB&MT various hematologic malignancies during this period. Fifteen patients treated for posttransplantation relapse with intensive chemotherapy (either ablative or nonmyeloablative) and mobilized donor lymphocytes with stem cells were considered as having a second SCT. Patients given nonmobilized donor lymphocyte infusions were not considered as having a second SCT. Thirty-three patients had at least 1 prior SCT (autologous or allogeneic). The diagnostic criteria for TMA included thrombocytopenia; either a decreasing platelet count or a failure to achieve platelet engraftment; and microangiopathic hemolysis, as evidenced by increased lactate dehydrogenase (LDH) and fragmented red blood cells observed in peripheral blood film. Neurologic symptoms and renal abnormalities supported the diagnosis but were not required. At the time of writing, the data set included 22 patients diagnosed with TMA after SCT. Conditioning Regimens Sixty-three SCTs were considered to follow myeloablative conditioning: busulfan/cyclophosphamide (n 37); cyclophosphamide/total body irradiation (n 15); carmustine, etoposide, cytarabine, and melphalan (n 7); or high-dose melphalan (n 4). The conditioning regimen was selected on the basis of the underlying malignancy. Patients not eligible for standard ablative SCT because of advanced age, comorbidities, or extensive prior therapy (including prior SCT) and patients with chronic myeloid leukemia (CML) in first chronic phase were eligible for nonmyeloablative conditioning. Eighty-four SCTs followed a reduced-intensity regimen that consisted of a combination of fludarabine and intravenous busulfan (n 29) or melphalan (n 41) or other combinations (n 14) based on the underlying malignancy. The allograft source was peripheral blood stem cells in most SCTs (n 144) and bone marrow in a minority of SCTs (n 3). GVHD prophylaxis consisted uniformly of cyclosporine and a short course of methotrexate. Patients with a matched unrelated or mismatched related donor SCT were given antithymocyte globulin during conditioning. Ex vivo T-cell depletion was not used. Standard institutional supportive care guidelines were followed. Acute and chronic GVHD was diagnosed, staged, and graded on the basis of standard criteria. Chimerism was assessed approximately 1 month after SCT by fluorescent in situhybridization analysis for X and Y markers in sex-mismatched SCT and by microsatellite analysis in sex-matched SCT [26]. Definitions Disease status was determined before SCT according to standard criteria. Early disease status included acute leukemia in first complete remission, 485

3 A. Shimoni et al. CML in first chronic phase, multiple myeloma, and lymphoma in first remission. All other disease phases were considered advanced disease. Patients were determined as having disease in remission if they had no evidence of disease by standard criteria. In patients with acute leukemia, this required less than 5% blasts in marrow aspirate and normal blood counts. For the purpose of this analysis, patients with CML in chronic phase were included with those in remission. Complete response to therapy of TMA required resolution of symptoms attributed to TMA, increase of platelet count to /L, and normalization of LDH. Partial response required improvement in these parameters that was not sufficient to determine complete response. Statistical Analysis The incidence of TMA was calculated and plotted by using Kaplan-Meier analysis [27]. Patients were censored at the time of last follow-up or death or at the start of conditioning for a second SCT. Categorical risk factors for TMA incidence were compared by using the log-rank test. Variables found significant in the univariate analysis were included in a Cox proportional hazard model. Overall survival was calculated from the day of diagnosis of TMA and was plotted with Kaplan-Meier analysis. RESULTS Patient Characteristics at Diagnosis TMA was diagnosed in 22 patients after allogeneic SCT. The disease and patient characteristics at the time of diagnosis of TMA are outlined in Table 1. The median age at diagnosis was 43 years (range, years). There were 17 men and 5 women with a diagnosis of acute myeloid (n 10) or lymphoid (n 5) leukemia, non-hodgkin lymphoma (n 3), multiple myeloma (n 3), or CML (n 1). The donors were HLA-matched siblings (n 12) or matched unrelated donors (n 10). Fourteen patients had TMA after NST, and 8 had TMA after ablative SCT. TMA was diagnosed a median of 30 days after SCT (range, days). All patients were thrombocytopenic at the time of diagnosis, with a median platelet count of /L (range, /L). Ten patients had rapidly decreasing platelet counts, and in 12 patients, TMA was diagnosed before platelet engraftment. All patients had evidence of microangiopathic hemolysis, as evidenced by fragmented red blood cells in peripheral blood film. All patients had increased LDH (median, 716 IU/L [range, IU/L] and 967 IU/L [range, IU/L] at diagnosis and maximally during the course, respectively [normal institutional level, IU/L]). The median bilirubin level was 3.3 mg/dl (range, mg/dl) and 16.1 mg/dl (range, mg/dl) at diagnosis and maximally during the course, respectively. The median creatinine level was 1.3 mg/dl (range, mg/dl), and only 3 patients had a creatinine level 2 mg/dl. Six patients had excessive cyclosporine levels within 1 week before diagnosis, and 4 patients were off cyclosporine therapy at the time of diagnosis of TMA. Eleven patients had neurologic symptoms during the course, such as confusion or changes in consciousness. However, most of these patients had multiple other potentially contributing factors (in particular, liver failure), and in only 2 patients (Table 1; patients 2 and 7) was neurologic dysfunction thought to be predominantly related to TMA. Fever was not considered a diagnostic criterion. Most patients had infections at or before diagnosis of TMA and were treated with antibiotics, but only 1 patient (patient 15) had a documented disseminated fungal infection. CMV reactivation was common but was not related in timing to TMA diagnosis. Clinical Settings and Predicting Factors TMA occurred in 22 (15%) of 147 SCTs, with a projected incidence of 20% 4% (Figure 1). TMA was diagnosed in 3 overlapping clinical settings. Nine patients had classic multifactorial TMA (Table 1; patients 1-9). The projected incidence was 8% 3%. In 9 cases, TMA occurred in patients previously diagnosed with severe (stage III-IV) hepatic acute GVHD (bilirubin 6 mg/dl) at the time of diagnosis of TMA (Table 1; patients 10-18). All of these patients, as well as 2 additional patients with less than stage III hepatic GVHD at diagnosis (patients 8 and 9, who belonged to the classic posttransplantation TMA group), had progressive jaundice (bilirubin 15 mg/dl, believed to predominantly result from GVHD) during the course. Overall, TMA was diagnosed during the course of severe hepatic GVHD in 11 of 15 patients, or at a projected risk reaching 73% 14%. Seven patients had TMA during the course of a second allogeneic SCT for post-sct relapse, all from the original donor (Table 1; patients 16-22). Three of these patients also had severe hepatic GVHD that overlapped with the hepatic GVHD-associated TMA variant (patients 16-18). TMA occurred during 7 of 15 second allogeneic SCTs, with a projected risk as high as 70% 16%. When all SCTs were considered, the univariate analysis identified SCT from an unrelated donor, SCT during advanced or active disease, a prior ablative SCT (allogeneic, autologous, or both) within 6 months before the current SCT (Figure 2), and acute GVHD as factors associated with the occurrence of SCT-associated TMA (Table 2). Male sex had borderline significance. A multivariate Cox regression model determined that acute GVHD and a second 486

4 Posttransplantation TMA Table 1. Patient Characteristics at Diagnosis of TMA Patient No. TMA Type* Age (y)/ Sex Disease Status Donor Prior SCT Conditioning Regimen Acute GVHD PLT# LDH Diag/Max (IU/L)** Bilirubin Diag/Max** (mg/dl) Cr CSA Neurologic (mg/dl) ( g/l) Findings 1 Classic 49/M ALL ref rel Sib No NST FM 2 Classic 44/F AML MUD No NST 1 ref FB 3 Classic 53/M MM Sib Auto >6 NST 4 Classic 50/F MDS Sib No Ablative BuCy 5 Classic 57/M MM Sib Auto >6 NST 6 Classic 22/M AML MUD Auto <6 NST ref rel 7 Classic 56/M NHL ref MUD Auto <6 NST 8 Classic/hepatic 50/M NHL Sib No Ablative ref BEAM 9 Classic/hepatic 52/M NHL MUD Auto >6 NST ref 10 Hepatic 45/M ALL MUD Auto <6 NST CR3 mo F/low TBI 11 Hepatic 40/M 2AML MUD No Ablative untreat BuCy 12 Hepatic 34/M CML MUD No Ablative CP BuCy 13 Hepatic 33/M MM MUD Auto <6 NST ref 14 Hepatic 37/M AML Sib No Ablative CR2 BuCy 15 Hepatic 28/M AML MUD No NST CR2 FM 16 Hepatic/2 allo 51/M AML MUD Allo <6 NST pt-relapse mo FB 17 Hepatic/2 allo 45/M AML Sib Allo <6 Ablative pt-relapse mo HD mel 18 Hepatic/2 allo 33/F AML Sib Allo <6 NST pt-relapse mo FA/ida 19 2 allo 21/M ALL Sib Allo <6 NST pt-relapse mo Mit/hidac 20 2 allo 42/M AML Sib Allo <6 NST pt-relapse mo Mit/hidac 21 2 allo 23/F ALL Sib Allo <6 Ablative pt-relapse mo HD mel 22 2 allo 35/F ALL Sib Allo <6 Ablative pt-relapse mo HD mel GrII / / No GI No / / Yes GrII Gl / / No GrIII / / No Gl No / / No No 7 730/ / No GrI s / / Yes GrIV 7 570/ / Yes s, GrIV / / Yes GrIV / / Yes GrIV 9 670/ / Yes GrIV 3 507/ / Yes s, GrIV / / Yes GrIV / / No s, GrIV / / Yes liv GrIV / / Yes GrIV / / Yes GrIV / / No s, No / / No GrIII Gl GrII s, Gl GrIII Gl, liv / / No 4 752/ / No / / No *TMA indicates thrombotic microangiopathy; 2allo, second allogeneic transplantation. ALL, acute lymphatic leukemia; AML, acute myeloid leukemia; MDS, myelodysplastic syndrome; MM, multiple myeloma; NHL, non- Hodgkin lymphoma; ref, refractory; rel, relapse; CR, complete remission; CP, chronic phase, 1 ref, primary refractory; pt-relapse, post transplantation relapse; untreat, previously untreated. Sib, HLA-matched sibling; MUD, matched unrelated donor. SCT, stem cell transplantation; auto, autologous; allo, allogeneic, / 6 mo, more/less than 6 mo before current transplantation. NST, nonmyeloablative stem cell transplantation; F, fludarabine; M, melphalan; Bu, busulfan; Cy, cyclophosphamide; BEAM, carmustine, etoposide, cytarabine, and melaphan; low TBI, total body irradiation 200 cg; A, cytarabine; ida, idarubicin; Mit, mitoxantrone; hidac, high-dose cytarabine; HD mel, high-dose melaphan. Gr, GVHD grade; Gl, gastrointestinal; liv, liver; s, skin. #PLT, platelet count at diagnosis ( 10 9 /L). **Diag/max, at diagnosis/maximally during the course. Cr, creatinine. CSA, maximal cyclosporin A serum level during 1 wk before diagnosis. BB&MT 487

5 A. Shimoni et al. onset of TMA. The Kaplan-Meier survival analysis predicts an overall survival of 9% 8% (Figure 3). However, the peri-tma mortality rate, where TMA may have caused or contributed to mortality, was 68% 10%. Ten patients died of complications of acute GVHD. In 4 patients, the primary cause of death was diffuse alveolar hemorrhage (DAH), and 1 died of central nervous system bleeding. Two patients with severe hepatic GVHD also had DAH as a contributing cause of death. One patient died of cerebral TMA. Two additional patients died of relapse. For the purpose of this analysis, patients who had a second allogeneic SCT for post-sct relapse and who died in remission were not considered as dying of relapse. Figure 1. Cumulative incidence of thrombotic microangiopathy (TMA) after 147 allogeneic stem cell transplantations. SCT within 6 months of a prior ablative SCT were the only predicting factors (Table 3). The same factors remained significant even when second allogeneic SCTs were not included (data not shown). Comparison of NST and Myeloablative SCT The risk of TMA after NST and myeloablative conditioning was 23% 6% and 16% 5%, respectively. Nonmyeloablative conditioning was therefore not associated with a lesser risk for TMA, and there was even a non statistically significant trend for a higher risk. Table 4 summarizes the differences in patient characteristics and outcomes between the 2 cohorts. As expected by the treatment strategy, patients allocated to NST were older, more of them had lymphoma or multiple myeloma rather than acute leukemia, and more had had a recent prior SCT. Acute GVHD rates and severity and chimerism analysis were similar between cohorts. Treatment and Outcome The outcomes after diagnosis of TMA are outlined in Table 5. The initial treatment strategy was dose reduction or complete withdrawal of cyclosporine therapy (in 18 patients treated with cyclosporine at the time of diagnosis) and transfusion of cryoprecipitate-poor plasma. Four patients had a complete response, and 1 had a partial response, for an overall response rate of 23%. Four patients had therapeutic plasma exchange (TPE) after the failure of prior therapy. One had complete response and 1 had a partial response; 2 did not respond. One patient was treated with defibrotide, with no response. The overall response rate to all lines of therapy was 32%. Four patients (18%) are currently alive and disease free. Eighteen patients died 2 to 227 days after the DISCUSSION TMA is a severe and relatively common complication after allogeneic SCT. It occurred in our study with an estimated risk of 20% 4% of all transplantations. The reported incidence varies from 0.5% to 21.4% in different studies [5-17] but is on average 5% to 15% [17]. This broad range of incidence is related in part to lack of a uniform definition of SCT-associated TMA, small numbers of patients included in each study, a wide range of differences in patient and SCT characteristics included in different studies, and, probably, underdiagnosis of this syndrome. The incidence of TMA in this study was in the high reported range for several reasons. First, we used a less strict definition of TMA. The presence of thrombocytopenia and microangiopathy, as evidenced by the markedly increased LDH and fragmented red blood cells observed in peripheral blood film, in the absence of Figure 2. Cumulative incidence of thrombotic microangiopathy (TMA) after 147 allogeneic stem cell transplantations and a comparison of the incidence among patients with a history of recent ( 6 months) allogeneic or autologous stem cell transplantation and those with no recent transplantation. 488

6 Posttransplantation TMA Table 2. Clinical Factors for Prediction of TMA* Variable n No. with TMA Projected Incidence P Value All SCT (15%) 20 4% Age <40 y 53 9 (17%) 22 7% NS >40 y (14%) 28 5% Sex Male (20%) 26 6%.06 Female 60 5 (8%) 11 5% Disease type Acute leukemia (20%) 27 7% NS Chronic leukemia 20 1 (5%) 6 6% MM/NHL/others 51 6 (12%) 14 5% Disease status at SCT Advanced (17%) 24 5%.05 Early 38 3 (8%) 10 5% Active disease (19%) 26 6%.03 Remission 57 5 (9%) 12 5% Prior SCT within 6 mo Yes (52%) 75 13% <.0001 No (9%) 12 4% Prior SCT >6 mo Yes 33 3 (9%) 11 6% NS No 93 8 (9%) 12 4% Donor Matched unrelated (26%) 38 11%.02 Matched sibling (11%) 9 3% Conditioning NST (17%) 23 6% NS Ablative 63 8 (13%) 16 5% TBI Yes 16 1 (6%) 11 10% NS No (16%) 21 4% Acute GVHD grade II-IV (33%) 38 8% I 92 4 (4%) 6 3% NS indicates not significant; TBI, total body irradiation; NHL, non-hodgkin lymphoma; MM, multiple myeloma. *Survival rates were estimated by the Kaplan-Meier method and categorical values were compared by using the log-rank test. Early disease: acute leukemia in first remission, CML in first chronic phase, and MM/NHL in first remission; advanced disease; all the rest. Patients with prior SCT within 6 mo were not included. Five patients with 1 antigen mismatched related donors were included with the matched unrelated group. alternative apparent etiologies, sufficed to make the diagnosis of TMA. With the advent of effective therapy, this definition, rather than the classic pentad, is now widely accepted for the diagnosis of idiopathic TMA [28]. The diagnosis of TMA in SCT patients is difficult and requires a high level of clinical suspicion because its presentation overlaps with other post-sct complications. Renal failure and neurologic symptoms may be related to other post-sct conditions and are often absent at the time of diagnosis of TMA. In our series, renal failure was uncommon, and neurologic symptoms were thought to be related to TMA in only 2 patients; in most other patients they were attributed BB&MT Table 3. Proportional Hazards Regression Model for the Occurrence of Thrombotic Microangiopathyafter Stem Cell Transplantation* Variable TMA Incidence Hazard Ratio (95% Confidence Interval) P Value Sex (male) 1.9 ( ) NS Disease status Advanced 1.1 ( ) NS Active disease 1.8 ( ) NS Prior SCT <6 mo 6.4 ( ).001 Donor (MUD) 1.6 ( ) NS Acute GVHD (grade III-IV) 14.5 ( ).01 NS indicates not significant; MUD, matched unrelated donor. *The factors found significant in the univariate analysis (Table 2) were included in a Cox regression model. to other conditions, predominantly hepatic failure. Fever is also most often associated with concomitant infections rather than TMA. Even the elements of the less strict definition may be related to other complications. Thrombocytopenia is common after SCT for Table 4. Comparison of Patient Characteristics and Outcomes between Patients after NST and Myeloablative Conditioning Variable NST Myeloablative P Value n Median age, y (range) 50 (16-66) 37 (18-65).003 Sex Male 53 (63%) 34 (54%) NS Female 31 (37%) 29 (46%) Disease type Acute leukemia 30 (36%) 46 (73%).001 Chronic leukemia 12 (14%) 8 (13%) MM/NHL/others 42 (50%) 9 (14%) Disease status at SCT Advanced 66 (79%) 43 (68%) NS Early 18 (21%) 20 (32%) Active disease 51 (61%) 39 (62%) Remission 33 (39%) 24 (38%) Prior SCT within 6 mo Yes 16 (19%) 5 (8%).06 No 68 (81%) 58 (92%) Donor Matched unrelated 22 (26%) 22 (26%) NS Matched sibling 62 (74%) 62 (74%) Acute GVHD grade II-IV 32 (38%) 24 (38%) NS III-IV 19 (23%) 16 (25%) Chimerism Complete 50 (60%) 39 (62%) NS 95%-99% donor 7 (8%) 7 (11%) 50%-94% donor 4 (5%) 4 (6%) Not evaluable* 23 (27%) 13 (21%) NS indicates not significant; MM, multiple myeloma; NHL, non- Hodgkin lymphoma. *Chimerism was evaluated 1 mo after SCT. Not evaluable indicates patients dying or with disease progression before chimerism testing and patients with nonengraftment or missing data. 489

7 A. Shimoni et al. Table 5. Patient Treatment and Outcome Treatment Patient No. TMA Type CSA Plasma Infusion TPE Response Outcome 1 Classic Reduced Yes No CR Died of chronic GVHD day Classic DC Yes Yes PR Died of relapse day 69 3 Classic DC Yes No NR Died of DAH day 58 4 Classic Reduced No No NR Died of DAH day 29 5 Classic Reduced Yes No CR Alive and well day Classic DC Yes No CR Died of relapse day Classic DC Yes Yes NR Died of cerebral TMA day 84 defib NR 8 Classic/hepatic DC Yes Yes NR Died of acute GVHD day 29 9 Classic/hepatic Reduced Yes No NR Died of acute GVHD and DAH day 2 10 Hepatic Reduced Yes No NR Died of acute GVHD and DAH day 3 11 Hepatic Reduced Yes No NR Died of acute GVHD day 5 12 Hepatic Reduced Yes No NR Died of acute GVHD day 4 13 Hepatic Reduced Yes No NR Died of acute GVHD day Hepatic DC Yes No NR Died of CNS bleeding day Hepatic DC Yes No NR Died of chronic GVHD day Hepatic/2allo Off at diagnosis Yes No NR Died of acute GVHD day Hepatic/2allo Reduced Yes No NR Died of acute GVHD day Hepatic/2allo Off at diagnosis Yes Yes CR Alive day allo Off at diagnosis Yes No NR Died of DAH day allo Reduced Yes No CR Alive day allo DC Yes No NR Died of DAH day allo Off at diagnosis Yes No PR Alive day 92 DC indicates discontinued; TPE, therapeutic plasma exchange; defib, defibrotide; 2allo, second allogeneic transplantation; CR, complete remission; PR, partial response; NR, no response; CSA, cyclosporin A. multiple reasons, such as delayed engraftment, GVHD, infections, CMV infection, and drugs. Delayed platelet engraftment is common after unrelated donor SCT. In most studies, as also observed in our study, as many as 50% of all patients with TMA did not have platelet engraftment at the time of onset of TMA, such that declining platelet count could not be documented. Similarly, TMA may overlap with cyclosporine toxicity, and fragmented red blood cells are often observed in blood films of patients treated with cyclosporine [20,29]. However, a markedly increased LDH level is more specific and should always raise suspicion for TMA [9]. Figure 3. Overall survival after the diagnosis of thrombotic microangiopathy after allogeneic stem cell transplantation. In this study, we identified 3 clinical settings for the occurrence of SCT-associated TMA with some overlap: a classic SCT-associated TMA, TMA associated with severe hepatic GVHD, and TMA associated with a second allogeneic SCT. We determined a separate clinical setting for patients who had a second allogeneic SCT. These patients were given intensive chemotherapy followed by granulocyte colony-stimulating factor mobilized lymphocytes and stem cells. The major difference from the classic variant is that cyclosporine was not administered after SCT. There was some overlap with the hepatic GVHD variant, because some patients in this group had severe GVHD before the diagnosis of TMA. The classic multifactorial TMA occurred with an estimated risk of 8% 3%, which is very similar to the risk in other large case series. The excess TMA risk in this study was attributed to the other 2 clinical settings. The identification of risk factors for the occurrence of TMA will conceivably allow early diagnosis and, perhaps, the timely introduction of therapeutic or preventive measurements. In this study, multivariate analysis determined that the occurrence of acute GVHD and a second SCT within 6 months were the major predictors of TMA, with a hazard ratio of 14.5 and 6.4, respectively. Univariate analysis also suggested that SCT from an unrelated donor and during advanced or active disease might predict TMA (Table 490

8 Posttransplantation TMA 2), but these factors did not remain statistically significant in the multivariate model (Table 3). Most studies have shown that SCT from unrelated donors carries a higher risk for TMA [8,9,11,12]. Some studies [8,9,14,15] have shown that female sex is a risk factor, but other studies did not support this association [5,11,13,15]. Similarly, older age was found as a risk factor in some [9], but not all [11-13], studies. Acute GVHD is considered a major risk factor for TMA in some studies [5,7,8,12,15,20], but not others [8,11,13]. Corticosteroid treatment increases the risk of TMA [13,16]. Some researches have suggested not including patients with severe GVHD in TMA studies [12] because they usually do not respond to TPE. However, most studies have included these patients because GVHD may have a role in the pathogenesis of TMA. These patients may not benefit from current available therapy [12], but they may benefit from future therapies and therefore should not be excluded. In this series, patients with severe hepatic GVHD had a 73% 14% risk of developing TMA. TMA is probably underdiagnosed in this setting because of the complicated and terminal status of patients in this condition. Only a few reports suggest an association of second SCT with TMA [7,9]. Our study included a relatively large number of allogeneic SCTs after failure of an autologous or allogeneic SCT (Tables 1 and 2). When this SCT was given within 6 months of the current allogeneic SCT, the projected risk of TMA was 75% 13% (Table 2). The pathogenesis of SCT-associated TMA is complicated and multifactorial. An intense conditioning regimen is considered to have a role in the pathogenesis of TMA by inducing endothelial cell damage. Total body irradiation during conditioning was found to be a risk factor for TMA in some studies [30,31] but not in others [8,9,11,13]. There are very few data on the occurrence of TMA after nonmyeloablative conditioning [7,32,33]. Elliot et al. [7] reported 2 cases of TMA among 13 NST recipients (15.4%); both had a prior autologous SCT. This was higher than after standard ablative conditioning in that study. Similarly, in this study, 14 of 84 NSTs were complicated with TMA (Table 2). The multivariate analysis did not determine that the use of less intensive conditioning conferred a lesser risk for TMA. If one assumes that TMA is initiated by the toxicity of the conditioning regimen, this unexpected observation may occur because the beneficial effects of reduced toxicity are offset by other factors that increase that risk. This relates in part to the selective use of NST in older and more debilitated patients, in patients with a second SCT or extensive prior therapy, and after unrelated donor SCT all of these are risk factors for TMA. GVHD is another major risk factor for TMA. When reduced-intensity regimens were introduced, it was hoped that GVHD rates would decrease [25]. The BB&MT pathogenesis of GVHD involves cytokine release that is associated with the tissue injury induced by the conditioning regimen [34], and it was assumed that limitation of tissue injury would limit GVHD. Also, the mixed chimerism often detected after NST was thought to be able to limit GVHD [26]. With time it was appreciated that GVHD rates were not decreased; rather, GVHD may be delayed throughout the course [35]. Also, with the use of reduced-intensity regimens such as the combination of fludarabine and busulfan or melphalan, as we used in most NSTs in this study, most patients achieve complete chimerism early after SCT (similarly to ablative SCT) and are not protected from GVHD [26]. Indeed, in our study there was no difference in chimerism and acute GVHD rates between the 2 cohorts (Table 4). Because GVHD may have a major role in TMA pathogenesis, the inability to limit its incidence and severity with NST may explain why TMA is not reduced. Eissner et al. [36] reported that fludarabine induces apoptosis, activation, and alloreactivity of human endothelial cells and also causes damage to dermal and alveolar epithelial cells. Fludarabine is a major constituent of nonmyeloablative conditioning, and the effect of fludarabine on endothelial cells may contribute to the pathogenesis of TMA after NST. As discussed previously, acute GVHD is probably as common after NST as after standard ablative conditioning, and as reported by Eissner et al. [36], the endothelial cells may become a common target for GVHD after fludarabine therapy. It is interesting to note that defibrotide was found to be protective against fludarabine-related damage in that study. There are initial reports that defibrotide, a fibrinolytic agent used for the treatment of veno-occlusive disease of the liver (VOD) [37], may be a promising treatment for SCT-associated TMA [38]. There is a rationale for exploring the use of defibrotide in SCT with a high risk of TMA, such as second SCTs and those from unrelated donors, especially after fludarabine-containing conditioning regimens. SCT-associated TMA has been associated with disseminated fungal infections or viral infections, most frequently CMV [5,7,9,12]. In this study, only 1 patient had documented fungal infection at the time of diagnosis. Although many patients had CMV reactivation during the course, it could not be related to the time of diagnosis. The introduction of TPE dramatically changed the prognosis of patients with idiopathic TMA, with very high response rates. However, the treatment of SCT remains controversial. In particular, there are controversial data as to whether TPE is effective in this setting [39-41]. In this study, the first measurement was usually a reduction of cyclosporine dose and plasma infusion, with a response rate of 23%. Four patients of the nonresponders had TPE, and 2 had 491

9 A. Shimoni et al. partial or complete responses. Overall, 32% responded. It can be concluded that TPE may have a role in SCT-associated TMA. However, because of the complexity of the patients and questionable efficacy, TPE may not necessarily be introduced as the first therapeutic strategy. The occurrence of TMA carried an ominous prognosis. Only 4 patients remain alive, and the Kaplan-Meier estimate of long-term survival is less than 10%. There was a wide variability in the reported mortality in patients with SCT-associated TMA, ranging from 31% to 100%. However, the estimated mortality was 84% for patients with 2 or 3 of the following criteria: TMA within 120 days of SCT, use of cyclosporine, and renal or neurologic symptoms [6]. Most patients in our study belonged to this high-risk group. Similarly to other studies, most patients died of unrelated causes, such as GVHD and relapse. TMA was the principal or contributing cause of death in 8 patients (36%), 1 patient died of cerebral TMA and refractory convulsions, and 5 patients died of bleeding complications (Table 5). It is interesting to note that 6 patients with TMA also had DAH. The association of TMA with VOD has been reported in a few studies [5,8,15], suggesting that both syndromes, which may involve small-vessel injury, share the same pathogenesis. The incidence of VOD in our study was very low because of the use of nonmyeloablative conditioning and intravenous busulfan in most patients [42], therefore not allowing documentation of this association. The association of TMA and DAH has been reported before in only 1 study [43]. Four patients with TMA and DAH were given fludarabine-containing conditioning, and, as suggested previously, fludarabine can cause endothelial and pulmonary epithelial damage [36]. The association of GVHD, VOD, TMA, and DAH may be related to the cytokine release cascade triggered by endothelial damage [7,10]. The increased level of inflammatory cytokines and T-cell activation in unrelated donor SCT explains the higher risk for these complications in this setting. In conclusion, SCT-associated TMA remains a relatively common complication with unsatisfactory therapy and a grim prognosis. The use of fludarabinebased reduced-intensity conditioning does not confer a lesser risk of TMA. Further exploration of novel therapeutic interventions and use of preventive measurements in high-risk patients are required. REFERENCES 1. Amorosi EL, Ultmann JE. Thrombotic thrombocytopenic purpura: report of 16 cases and review of the literature. Medicine. 1966;45: Moake JL. Thrombotic microangiopathies. N Engl J Med. 2002;347: Tsai HM. Deficiency of ADAMTS13 and thrombotic thrombocytopenic purpura. Blood. 2002;100: Tsai HM, Lian EC. Antibodies to von Willebrand factorcleaving protease in acute thrombotic thrombocytopenic purpura. N Engl J Med. 1998;339: Pettitt AR, Clark RE. Thrombotic microangiopathy following bone marrow transplantation. Bone Marrow Transplant. 1994; 14: Schriber JR, Herzig GP. Transplantation-associated thrombotic thrombocytopenic purpura and hemolytic uremic syndrome. Semin Hematol. 1997;34: Elliott MA, Nichols WL Jr, Plumhoff EA, et al. Posttransplantation thrombotic thrombocytopenic purpura: a single-center experience and a contemporary review. Mayo Clin Proc. 2003; 78: Ruutu T, Hermans J, Niederwieser D, et al. Thrombotic thrombocytopenic purpura after allogeneic stem cell transplantation: a survey of the European Group for Blood and Marrow Transplantation (EBMT). Br J Haematol. 2002;118: Fuge R, Bird JM, Fraser A, et al. The clinical features, risk factors and outcome of thrombotic thrombocytopenic purpura occurring after bone marrow transplantation. Br J Haematol. 2001;113: Takatsuka H, Takemoto Y, Okamoto T, et al. Thrombotic microangiopathy following allogeneic bone marrow transplantation. Bone Marrow Transplant. 1999;24: Uderzo C, Fumagalli M, De Lorenzo P, et al. Impact of thrombotic thrombocytopenic purpura on leukemic children undergoing bone marrow transplantation. Bone Marrow Transplant. 2000;26: Roy V, Rizvi MA, Vesely SK, George JN. Thrombotic thrombocytopenic purpura-like syndromes following bone marrow transplantation: an analysis of associated conditions and clinical outcomes. Bone Marrow Transplant. 2001;27: Paquette RL, Tran L, Landaw EM. Thrombotic microangiopathy following allogeneic bone marrow transplantation is associated with intensive graft-versus-host disease prophylaxis. Bone Marrow Transplant. 1998;22: Iacopino P, Pucci G, Arcese W, et al. Severe thrombotic microangiopathy: an infrequent complication of bone marrow transplantation. Gruppo Italiano Trapianto Midollo Osseo (GITMO). Bone Marrow Transplant. 1999;24: Zeigler ZR, Shadduck RK, Nemunaitis J, Andrews DF, Rosenfeld CS. Bone marrow transplant-associated thrombotic microangiopathy: a case series. Bone Marrow Transplant. 1995;15: Kalhs P, Brugger S, Schwarzinger I, et al. Microangiopathy following allogeneic marrow transplantation. Association with cyclosporine and methylprednisolone for graft-versus-host disease prophylaxis. Transplantation. 1995;60: Allford SL, Bird JM, Marks DI. Thrombotic thrombocytopenic purpura following stem cell transplantation. Leuk Lymphoma. 2002;43: van der Plas RM, Schiphorst ME, Huizinga EG, et al. von Willebrand factor proteolysis is deficient in classic, but not in bone marrow transplantation-associated, thrombotic thrombocytopenic purpura. Blood. 1999;93: Vesely SK, George JN, Lammle B, et al. ADAMTS13 activity in thrombotic thrombocytopenic purpura-hemolytic uremic syndrome: relation to presenting features and clinical outcomes in a prospective cohort of 142 patients. Blood. 2003;102: Holler E, Kolb HJ, Hiller E, et al. Microangiopathy in patients 492

10 Posttransplantation TMA on cyclosporine prophylaxis who developed acute graft-versushost disease after HLA-identical bone marrow transplantation. Blood. 1989;73: 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. 1995;16: Dzik WH, Georgi BA, Khettry U, Jenkins RL. Cyclosporineassociated thrombotic thrombocytopenic purpura following liver transplantation successful treatment with plasma exchange. Transplantation. 1987;44: Zoja C, Furci L, Ghilardi F, Zilio P, Benigni A, Remuzzi G. Cyclosporin-induced endothelial cell injury. Lab Invest. 1986; 55: Slavin S, Nagler A, Naparstek E, et al. Nonmyeloablative stem cell transplantation and cell therapy as an alternative to conventional bone marrow transplantation with lethal cytoreduction for the treatment of malignant and nonmalignant hematologic diseases. Blood. 1998;91: Champlin R, Khouri I, Shimoni A, et al. Harnessing graftversus-malignancy: non-myeloablative preparative regimens for allogeneic haematopoietic transplantation, an evolving strategy for adoptive immunotherapy. Br J Haematol. 2000;111: Shimoni A, Nagler A. Non-myeloablative stem cell transplantation (NST): chimerism testing as guidance for immune-therapeutic manipulations. Leukemia. 2001;15: Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53: George JN. How I treat patients with thrombotic thrombocytopenic purpura-hemolytic uremic syndrome. Blood. 2000;96: Kanamori H, Takaishi Y, Takabayashi M, et al. Clinical significance of fragmented red cells after allogeneic bone marrow transplantation. Int J Hematol. 2003;77: Carlson K, Smedmyr B, Hagberg H, Oberg G, Simonsson B. Haemolytic uraemic syndrome and renal dysfunction following BEAC (BCNU, etoposide, ara-c, cyclophosphamide) TBI and autologous BMT for malignant lymphomas. Bone Marrow Transplant. 1993;11: Chappell ME, Keeling DM, Prentice HG, Sweny P. Haemolytic uraemic syndrome after bone marrow transplantation: an adverse effect of total body irradiation? Bone Marrow Transplant. 1988;3: Corradini P, Tarella C, Olivieri A, et al. Reduced-intensity conditioning followed by allografting of hematopoietic cells can produce clinical and molecular remissions in patients with poor-risk hematologic malignancies. Blood. 2002;99: van Besien K, Bartholomew A, Stock W, et al. Fludarabinebased conditioning for allogeneic transplantation in adults with sickle cell disease. Bone Marrow Transplant. 2000;26: Antin JH, Ferrara JL. Cytokine dysregulation and acute graftversus-host disease. Blood. 1992;80: Mielcarek M, Martin PJ, Leisenring W, et al. Graft-versus-host disease after nonmyeloablative versus conventional hematopoietic stem cell transplantation. Blood. 2003;102: Eissner G, Multhoff G, Gerbitz A, et al. Fludarabine induces apoptosis, activation, and allogenicity in human endothelial and epithelial cells: protective effect of defibrotide. Blood. 2002;100: Richardson PG, Elias AD, Krishnan A, et al. Treatment of severe veno-occlusive disease with defibrotide: compassionate use results in response without significant toxicity in a high-risk population. Blood. 1998;92: Corti P, Uderzo C, Tagliabue A, et al. Defibrotide as a promising treatment for thrombotic thrombocytopenic purpura in patients undergoing bone marrow transplantation. Bone Marrow Transplant. 2002;29: Milone J, Napal J, Bordone J, Etchegoyen O, Morales V. Complete response in severe thrombotic microangiopathy post bone marrow transplantation (BMT-TM) after multiple plasmaphereses. Bone Marrow Transplant. 1998;22: Teruya J, Styler M, Verde S, Topolsky D, Crilley P. Questionable efficacy of plasma exchange for thrombotic thrombocytopenic purpura after bone marrow transplantation. J Clin Apheresis. 2001;16: Kolker O, Sviri S, Linton DM, Or R, Nagler A, Shapira MY. Plasmapheresis for thrombotic thrombocytopenic purpura following bone marrow transplantation. Acta Haematol. 2001;105: Shimoni A, Bielorai B, Toren A, et al. Intravenous busulfanbased conditioning prior to allogeneic hematopoietic stem cell transplantation: myeloablation with reduced toxicity. Exp Hematol. 2003;31: Srivastava A, Gottlieb D, Bradstock KF. Diffuse alveolar haemorrhage associated with microangiopathy after allogeneic bone marrow transplantation. Bone Marrow Transplant. 1995;15: BB&MT 493