Diagnostic Value of Hemostatic Parameters in Bone Marrow Transplant- Associated Thrombotic Microangiopathy

Diagnostic value of hemostatic parameters in bone marrow transplant- associated thrombotic microangiopathy

H Kanamori1, A Maruta2, S Sasaki1, E Yamazaki1, S Ueda1, K Katoh1, T Tamura1, M Otsuka-Aoba1, J Taguchi1, H Harano1, K Ogawa1, H Mohri1, T Okubo1, M Matsuzaki3, S Watanabe4, H Koharazawa2, H Fujita2 and F Kodama2

1First Department of Internal Medicine, Yokohama City University School of Medicine; 2Department of Hematology/Chemotherapy, Kanagawa Cancer Center; 3Division of Blood Transfusion; 4Division of Laboratory Medicine, Yokohama City University Hospital, Yokohama, Japan

Summary: by the conditioning regimen with radio/chemotherapy con- tributes to veno-occlusive disease (VOD) of the liver,7,8 We investigated hemostatic parameters in a prospective thrombotic microangiopathy (TMA)9–11 and graft-versus- study of 16 patients who received bone marrow trans- host disease (GVHD).12,13 TMA is a clinical diagnosis plants (BMT). We found a significant rise in the levels which includes hemolytic uremic syndrome (HUS) and ␣ of fibrinogen, plasmin- 2 antiplasmin inhibitor com- thrombotic thrombocytopenic purpura (TTP); both are plex, tissue-plasminogen activator·plasminogen acti- characterized by microangiopathic hemolytic anemia, vator inhibitor complex (t-PA·PAI), von Willebrand fac- thrombocytopenia, microvascular thrombosis and multi- tor antigen, and thrombomodulin on day 14 after organ failure.14 Although several factors may contribute to transplant compared with values before transplant. the onset of this complication, the underlying pathophysiol- Protein C and thrombin-antithrombin III levels did not ogy in TMA consists mainly of microvascular endothelial change significantly. No significant changes in pro- damage, platelet aggregation, and abnormalities of the thrombin time ratio, activated partial thromboplastin coagulation system. Recently it has been reported that time, or protein S were detected. Patients who had thrombomodulin is a useful marker to assess endothelial grades II–IV graft-versus-host disease (GVHD) (n = 6) cell injury not only in patients with HUS/TTP but also in showed a significantly higher level of t-PA·PAI on day BMT-TMA.15,16 In this study, we measured hemostatic 14 compared with those with grades 0–I GVHD (n = 10) parameters, including thrombomodulin, during BMT in a (P = 0.0062). Three patients with grades II–IV GVHD prospective study to find early predictive factors for the developed thrombotic microangiopathy (TMA) on days development of TMA. 19, 19 and 62. In these patients, we noted significantly lower levels of fibrinogen (P = 0.0383), and significantly higher levels of t-PA·PAI (P = 0.0008) and thrombomod- Patients and methods ulin (P = 0.0001) on day 14 compared with those patients who did not develop TMA. These results sug- Patients gest that prothrombotic states and endothelial damage may be caused by the conditioning regimen and/or acute Sixteen consecutive patients who received a BMT between GVHD during BMT; thrombomodulin values on day 14 October 1994 and February 1996 were analyzed. These post BMT may be useful in surveillance for TMA patients were transplanted because of hematologic malig- because of endothelial cell injury. nancies (four with AML, three with ALL, one with acute Keywords: bone marrow transplantation; thrombotic undifferentiated leukemia and eight with CML). Twelve microangiopathy; hemostatic parameters; thrombomodulin; patients were at low risk (first remission or first chronic endothelial injury phase of leukemia), and four patients were at high risk (leukemias beyond first remission or chronic phase). Their median age was 29 years (range 16–48 years) (Table 1). Fifteen patients received marrow transplants from an HLA- Allogeneic bone marrow transplantation (BMT) is an effec- identical donor and one from an identical twin. There were tive strategy in patients with a variety of hematologic dis- nine related donors and seven unrelated donors. eases.1–3 However, BMT-related complications are serious and may be associated with high mortality.4–6 In various BMT-related complications, endothelial cell injury caused BMT procedures Four conditioning regimens were employed: thiotepa (TT, 180 mg/m2 for 2 days), cyclophosphamide (CY, 2250 Correspondence: Dr H Kanamori, First Department of Internal Medicine, 2 Yokohama City University School of Medicine, 3–9 Fukuura, Kanazawa- mg/m for 2 days) and total body irradiation (TBI, 12 Gy) ku, Yokohama 236, Japan in nine patients; TBI, TT and CY in three patients; busulfan Received 27 June 1997; accepted 24 October 1997 (4 mg/kg for 4 days), cytosine arabinoside (Ara-C 1400 Hemostatic parameters in BMT angiopathy H Kanamori et al 706 Table 1 Patient characteristics and clinical course

Patient no. Age/Sex Diagnosis/ BMT-type Conditioning regimen Acute GVHD TMA Outcome/Cause of death Status at BMT

1 46/M CML/CP MUD BU/CA/CY II ϩ Dead/CMV pneumonia 2 29/M Ph-ALL/CR 1 MUD TBI/VP16/CY III Ϫ Alive 3 23/M APL/CR 1 Allo TT/CY/TBI 0 Ϫ Alive 4 21/F CML/CP MUD BU/CA/CY IV ϩ Dead/Cerebral infarction 5 34/M AML/AP MUD BU/CA/CY II Ϫ Alive 6 29/M Ph-ALL/CR 1 Allo TBI/TT/CY 0 Ϫ Alive 7 46/F CML/CP Allo TBI/TT/CY II Ϫ Alive 8 48/F CML/CP Allo TBI/TT/CY II ϩ Dead/Multiorgan failure 9 29/M AML/CR 7 MUD TT/CY/TBI I Ϫ Dead/Heart failure 10 29/M CML/CP Syn TT/CY/TBI 0 Ϫ Alive 11 33/M CML/CP Allo TT/CY/TBI I Ϫ Alive 12 19/F AML/CR 2 MUD TT/CY/TBI 0 Ϫ Alive 13 16/F AML/CR 1 Allo TT/CY/TBI 0 Ϫ Alive 14 28/M CML/CP Allo TT/CY/TBI I Ϫ Alive 15 21/M ALL/CR 2 MUD TT/CY/TBI I Ϫ Alive 16 19/F AUL/CR 1 Allo TT/CY/TBI I Ϫ Alive

CML = chronic myelogenous leukemia; Ph-ALL = Philadelphia chromosome-positive acute lymphoblastic leukemia; AML = acute myelogenous leukemia; AUL = acute undifferentiated leukemia; CR = complete remission; CP = chronic phase; AP = accelerated phase; MUD = matched unrelated donor; Allo = allogeneic; Syn = syngeneic; BU = busulphan; Ara-C = cytosine arabinoside; CY = cyclophosphamide; TBI = total body irradiation; VP16 = etoposide.

mg/m2 ϫ 2 for 2 days) and CY in three patients; and TBI, peripheral blood smear. Hemolysis was defined if the serum etoposide (1600 mg/m2 for 1 day), and CY in one patient LDH was elevated with a decrease in hemoglobin and (Table 1). increased reticulocytes, as well as a drop in the platelet Thirteen patients received methotrexate (MTX) on days count. Autoimmune hemolytic anemia was excluded by a 1, 3, 6 and 11 and cyclosporin A (CsA) (3 mg/kg daily as negative Coombs’ test. a continuous infusion) for prevention of GVHD. Two patients received CsA only and one patient with an identical Statistical analysis twin donor received neither MTX or CsA. The results are expressed as mean and standard error. The Laboratory methods difference between values pre-conditioning and post-BMT was determined using Student’s t-test for paired data. Mean Measurement of hemostatic parameters was performed pre- values between groups of patients were evaluated with conditioning and on day 14 after BMT. Prothrombin time Student’s t-test for unpaired data. ratio (PT ratio), activated partial thromboplastin time (aPTT) and plasma fibrinogen were measured on the Sys- mex CA-5000 (TOA, Kobe, Japan) (normal ranges 0.87– Results 1.15 for PT ratio, 25.3–39.3 s for aPTT and 147–325 mg/dl for fibrinogen). Protein C was determined using a latex Clinical data photometric immunoassay (LPIA) (LPIA ace; DIA- IATRON, Tokyo, Japan) (normal range 70–150%). Protein Engraftment of the bone marrow was obtained in all S was determined using an enzyme immunoassay (EIA) patients. No acute GVHD occurred in five patients. Eleven (Asserachrom total protein S; Boehringer Mannheim, patients developed acute GVHD between days 8 and 26 Germany) (normal range 65–135%). Thrombin–antithrom- after BMT (grade I, five patients; grade II, four patients; bin III complex (TAT) was measured using EIA (TAT grade III, one patient and grade IV, one patient). No SRL; SRL, Tokyo, Japan) (normal Ͻ3.0 ng/ml). Plasmin- patients developed VOD of the liver. Laboratory findings ␣ 2 antiplasmin inhibitor complex (PIC) was determined showed no disseminated intravascular coagulation (DIC) on using LPIA (normal Ͻ0.8 ␮g/ml). Tissue-plasminogen day 14. activator·plasminogen activator inhibitor complex (t- There were three patients with TMA with onset on days PA·PAI) was evaluated by EIA (tPAI-C test; Teijin, Tokyo, 19, 19 and 62 (Table 1). Clinical courses in these patients Japan) (normal Ͻ17 ng/ml); von Willebrand factor antigen were as follows: patient 1 developed acute GVHD (grade (vWF:Ag) was tested by EIA (Asserachrom vWf) (normal II) limited to the skin on day 15 after BMT; he received range 50–155%). Thrombomodulin (TM) was measured by prednisolone. On day 36, the patient presented with macro- EIA (TM Parassera, Fujirebio, Tokyo, Japan) (normal scopic hematuria caused by adenovirus. He had a high- Ͻ4.5 FU/ml). grade fever, progressive confusion without focal neurologic signs, and renal dysfunction associated with microangiopathy on day 62. At that time, laboratory data revealed a Criteria for TMA creatinine of 4.3 mg/dl, a LDH of 1075 U/l, a drop in the TMA was diagnosed by the presence of intravascular hemoglobin level from 9.0 to 5.3 g/dl, and a decrease in hemolysis in association with red cell fragmentation on the platelet count from 39 to 6 ϫ 109/l. The CsA level was Hemostatic parameters in BMT angiopathy H Kanamori et al 707 280 ng/ml. Hemodialysis with filtration was performed for Table 3 Comparison of hemostatic parameters on day 14 in groups acute renal failure and TMA, but his condition deteriorated according to acute GVHD grade progressively. He developed cytomegalovirus (CMV) pneu- GVHD 0–I GVHD II–IV P value monia and died on day 93. In patient 4, hematuria appeared = = following preconditioning treatment without evidence of (n 10) (n 6) viral infection. Skin GVHD developed on day 13 and pro- PT ratio (INR) 1.1 Ϯ 0.04 1.04 Ϯ 0.06 0.2027 gressed to grade IV despite treatment with CsA and predni- aPTT (s) 34.6 Ϯ 2.3 31.9 Ϯ 1.5 0.4172 solone. Massive red cell fragmentation on the peripheral Fibrinogen (mg/dl) 429 Ϯ 23 356 Ϯ 38 0.1006 blood smear was seen on day 19. Serum LDH was 1579 Protein C (%) 74 Ϯ 688Ϯ 15 0.3357 U/l, platelet count was 15 ϫ 109/l, and creatinine was 1.9 Protein S (%) 80 Ϯ 468Ϯ 7 0.1331 TAT (ng/ml) 4.2 Ϯ 1.1 6.7 Ϯ 1.2 0.1674 mg/ml. The serum CsA was evelated to 430 ng/ml. From PIC (␮g/ml) 0.8 Ϯ 0.1 0.7 Ϯ 0.1 0.3954 day 37, FK506 was administered instead of CsA for acute t-PA·PAI (ng/ml) 14.2 Ϯ 3.0 34.3 Ϯ 6.4 0.0062 GVHD treatment. The patient developed generalized seiz- vWF:Ag (%) 218 Ϯ 14 217 Ϯ 39 0.9894 ures on day 47. Cranial computed tomography revealed TM (FU/ml) 3.5 Ϯ 0.3 5.1 Ϯ 1.0 0.0788 multifocal infarction, and she expired after cardiac arrest on day 51. Patient 8 developed TMA on day 19 post BMT Values are expressed as mean Ϯ s.e. with hemolytic anemia. She had grade II acute GVHD on day 26. On day 35, serum LDH was 1231 U/l and the plate- (FU/ml) let count was 17 ϫ 109/l. Plasma exchange with fresh- frozen plasma was started on day 37, and vitamin E was 8 administered with a subsequent decrease in serum LDH and bilirubin. However, microangiography hemolysis and 6 thrombocytopenia recurred on day 110. In addition, CMV infection developed, followed by multiorgan failure, and 4 she died on day 152. 2

0 Changes in hemostatic parameters Pre-conditioning Day 0 Day 14

The levels of t-PA·PAI in patients treated with non-TBI Figure 1 Alterations in thrombomodulin levels in patients with/without regimens were higher than those with TBI regimens (43.2 TMA. (᭺) GVHD, 0; (ᮀ) GVHD, I; (᭛) GVHD, II; (᭝) GVHD, III; (̃) v ̄ vs 16.8 ng/ml, P = 0.0357). GVHD, IV; ( , ) TMA. Fibrinogen, PIC, t-PA·PAI, vWF:Ag, and TM on day 14 after transplant were significantly increased compared with shown). Only t-PA·PAI increased among groups classified before treatment (Table 2). PT ratio, aPTT and protein S by acute GVHD grade (P = 0.0062) (Table 3). did not show significant changes between pre-conditioning Hemostatic parameters on day 14 in the three patients and day 14 post-BMT. The trend towards a decrease in who developed TMA were compared to patients without protein C among all patients on day 14 compared with pre- TMA. TAT increased on day 14, especially in patients with conditioning (P = 0.0691) was due to a significant decrease TMA, but without statistical significance. We noted sig- in patients with grades 0–I GVHD (P = 0.0357) (data not nificantly lower levels of fibrinogen and higher levels of t- PA·PAI and TM in patients who later developed TMA Table 2 Alterations of hemostatic parameters in patients who (Table 4). TM levels at pre-conditioning, day 0 and day 14 underwent BMT are indicated individually in Figure 1. A marked increase in TM levels on day 14 was demonstrated in patients with Pre-conditioning Day 14 P value (n = 16) after BMT (n = 16) Table 4 Comparison of hemostatic parameters on day 14 in patients with/without TMA PT ratio (INR) 1.1 Ϯ 0.04 1.1 Ϯ 0.03 0.9421 aPTT (s) 32.5 Ϯ 1.5 33.6 Ϯ 1.6 0.1572 Fibrinogen (mg/dl) 269 Ϯ 22 402 Ϯ 21 0.0002 With TMA Without TMA P value = = Protein C (%) 96 Ϯ 580Ϯ 7 0.0691 (n 3) (n 13) Protein S (%) 78 Ϯ 476Ϯ 4 0.6833 TAT (ng/ml) 3.7 Ϯ 5.7 5.2 Ϯ 8.6 0.1407 PT ratio (INR) 1.04 Ϯ 0.06 1.12 Ϯ 0.04 0.3848 PIC (␮g/ml) 5.4 Ϯ 6.0 7.8 Ϯ 8.0 0.033 aPTT (s) 33.5 Ϯ 1.7 33.6 Ϯ 1.9 0.9723 t-PA·PAI (ng/ml) 10.0 Ϯ 1.7 21.2 Ϯ 3.9 0.0042 Fibrinogen (mg/dl) 312 Ϯ 35 423 Ϯ 22 0.0383 vWF:Ag (%) 122 Ϯ 9 218 Ϯ 16 0.0001 Protein C (%) 99 Ϯ 26 75 Ϯ 6 0.1524 TM (FU/ml) 2.5 Ϯ 0.2 4.1 Ϯ 4.6 0.0074 Protein S (%) 68 Ϯ 11 76 Ϯ 4 0.3093 TAT (ng/ml) 8.5 Ϯ 1.6 4.4 Ϯ 0.9 0.0552 ␮ Ϯ Ϯ Values are expressed as mean Ϯ s.e. PIC ( g/ml) 0.8 0.2 0.8 0.1 0.7402 Ϯ Ϯ BMT = bone marrow transplantation; PT = prothrombin time; aPTT = t-PA·PAI (ng/ml) 45.1 7.2 16.4 2.8 0.0008 Ϯ Ϯ activated partial thromboplastin time; TAT = thrombin-antithrombin III vWF:Ag (%) 269 66 206 13 0.1352 = ␣ = TM (FU/ml) 7.4 Ϯ 0.6 3.3 Ϯ 0.2 0.0001 complex; PIC plasmin- 2 antiplasmin inhibitor complex; t-PA·PAI tissue plasminogen activator-plasminogen activator inhibitor complex; vWF:Ag = von Willebrand factor antigen; TM = thrombomodulin. Values are expressed as mean Ϯ s.e. Hemostatic parameters in BMT angiopathy H Kanamori et al 708 TMA, despite their levels at day 0 being within the normal contribute to the pathogenesis of several complications range. In patients with and without TMA, serum levels of post-BMT, including GVHD and TMA. CsA and creatinine on day 14 were not significant (250 The most telling finding in our study was the increased ng/ml in TMA vs 313 ng/ml in non-TMA for CsA and 0.63 TM level on day 14 post BMT in patients who later vs 0.75 mg/dl, respectively, for creatinine). developed TMA. TM is a high-affinity thrombin receptor on the surface of vascular endothelial cells and a cofactor in the activation of protein C.23 There is a good correlation Discussion between TM levels and clinical complications related to endothelial cell injury in septicemia, VOD and GVHD.16 BMT-TMA is a serious complication following BMT.9–12,16 Increased TM levels in patients with BMT-TMA have been The frequency in this study was 19%, and all patients who reported by Zeigler et al.15 They tested a ratio of TM:creati- developed TMA died. It is likely that the clinical course nine to provide a measure of the renal contribution to the of our patients was affected by preceding or simultaneous TM elevation. We similarly evaluated this ratio in our ser- complications of GVHD and/or viral infection. In previous ies. There was no difference bewteen TM and TM:creatin- reports of BMT-TMA,9,10 there is a wide clinical spectrum ine values among patients with or without TMA. Thus, an from subclinical lesions to fulminant disease; lethal cases influence of renal function was excluded in this study. The often have simultaneous complications such as GVHD or level of t-PA·PAI on day 14 may be associated with various infection. Zeigler et al10 found many causes of thrombocy- factors during BMT; the significant rise of t-PA·PAI on topenia in BMT patients; BMT-associated TMA is depen- day 14 was seen in patients with non-TBI regimens, severe dent on finding evidence of microangiopathic hemolytic GVHD and TMA. Moreover, the individual values of t- anemia. We diagnosed BMT-TMA on the appearance of PA·PAI were normal or already increased on day 0 (data red cell fragmentation, hemolytic anemia and thrombocyto- not shown). In contrast, TM levels on day 0 were normal penia. No patients with VOD were observed, and no labora- in all patients. Therefore, we believe that TM values are tory findings indicated DIC on day 14. According to the more strongly correlated with TMA than t-PA·PAI. classification by Pettitt and Clark,9 our three patients had It has been reported that the severity of acute GVHD and multifactorial fulminant TMA. Although the pathogenesis CsA prophylaxis is highly associated with microangiopathy of BMT-TMA is still not well understood, it is likely that post BMT.12 Kalhs et al11 have suggested that the combined many factors including the conditioning regimen, CsA, use of CsA and corticosteroids following BMT is an GVHD and infection contribute to the onset of TMA. important trigger for the development of microangiopathy. We found significant hemostatic abnormalites in patients They noted increased levels of vWF:Ag in patients who at an early phase post BMT. Others have reported signifi- received CsA and corticosteroids compared with CsA and cant declines in protein C and protein S after BMT17 and MTX. In particular, patients with severe microangiopathy decreased protein C during acute GVHD.18 Gordon et al19 showed the highest values of vWF:Ag in the CsA and demonstrated that protein C deficiency correlated with corticosteroids group. Although the mean difference was thrombotic complications in patients undergoing BMT. not significant, vWF:Ag was higher in our patients with Nevertheless, we detected no significant changes in protein TMA, yielding more evidence of endothelial damage in C or protein S, except for decreased protein C levels on these patients. Two of three patients with TMA in this day 14 in patients with grades 0–I GVHD. This discrepancy report received FK506 instead of CsA after the onset of may be influenced by patient heterogeneity in the different TMA. Adverse effects of FK50624,25 may contribute to the studies. Although a low protein C level before conditioning progression of TMA. therapy may predispose to VOD,20 we could not evaluate In conclusion, alterations in hemostatic parameters were this relationship because there were no patients with VOD. more common in the early phase following BMT, although In this study, we noted a significant rise in the fibrinogen, the causes of these abnormalities are unclear. Our results PIC, tPA-PAI, vWF:Ag, and TM on day 14 post BMT com- suggest a hypercoagulable, prothrombotic state with endo- pared with pre-conditioning. These changes may be caused thelial injury induced by conditioning therapy, acute by the conditioning regimen or by acute GVHD. The GVHD, and other complications. The measurement of TM marked increase in t-PA·PAI in patients with severe GVHD levels on day 14 may be useful for early detection of BMT- suggests that endothelial injuries are related to GVHD TMA, although a direct causal relationship between TM grade. In previous investigations by Holler et al,12 endo- and TMA has not been shown. thelial damage during GVHD was the central lesion of microangiopathy on the basis of analysis of hemostatic parameters such as vWF:Ag. They also reported a highly References significant association of microangiopathy with severity of acute GVHD and CsA prophylaxis. Furthermore, increased 1 Santos GW, Tutschka PJ, Brookmeyer R et al. Marrow trans- levels of t-PA and PAI were associated with microangiopa- plantation for acute nonlymphocytic leukemia after treatment thy following BMT and increased levels of tumor necrosis with busulfan and cyclophosphamide. New Engl J Med 1983; 309: 1347–1353. factor alpha (TNF␣).21 Recently, a correlation between ␣ 2 Thomas ED, Clift RA, Fefer A et al. Marrow transplantation TNF level and complications after BMT, especially acute for the treatment of chronic myelogenous leukemia. Ann 22 GVHD, has been reported. TNF␣ is involved in endo- Intern Med 1986; 104, 155–163. thelial injury and causes increased capillary permeability. 3 Appelbaum FR, Sullivan KM, Buckner CD et al. Treatment Therefore, TNF␣-mediated endothelial cell damage may of malignant lymphoma in 100 patients with chemotherapy, Hemostatic parameters in BMT angiopathy H Kanamori et al 709 total body irradiation, and marrow transplantation. J Clin (TM) levels in adult thrombotic thrombocytopenic Oncol 1987; 5: 1340–1347. purpura/hemolytic uremic syndromes (TTP/HUS) and bone 4 Wingard JR, Piantadosi S, Vogelsang GB et al. Predictions of marrow transplant-associated thrombotic microangiopathy death from chronic graft-versus-host disease after bone mar- (BMT-TM). Am J Hematol 1996; 53: 213–220. row transplantation. Blood 1989; 74: 1428–1435. 16 Testa S, Manna A, Porcellini A et al. Increased plasma level 5 Forman SJ, Zaia JA. Treatment and prevention of cytomegalo- of vascular endothelial glycoprotein thrombomodulin as an virus pneumonia after bone marrow transplantation: where do early indicator of endothelial damage in bone marrow trans- we stand? Blood 1994; 83: 2392–2398. plantation. Bone Marrow Transplant 1996; 18: 383–388. 6 Shulman HM, Hinterberger W. Hepatic veno-occlusive dis- 17 Harper PL, Jarvis J, Jennings I et al. Changes in the natural ease-liver toxicity syndrome after bone marrow transplan- anticoagulants following bone marrow transplantation. Bone tation. Bone Marrow Transplant 1992; 10: 197–214. Marrow Transplant 1990; 5: 39–42. 7 McDonald GB, Hinds MS, Fisher LD et al. Veno-occlusive 18 Leblond V, Salehian BD, Borel C et al. Alterations in natural disease of the liver and multiorgan failure after bone marrow anticoagulant levels during allogeneic bone marrow transplan- transplantation: a cohort study of 355 patients. Ann Intern Med tation: a prospective study in 27 patients. Bone Marrow 1993; 118: 255–267. Transplant 1993; 11: 299–305. 8 Jones RJ, Lee KSK, Beschorner WE et al. Veno-occlusive 19 Gordon BG, Haire WD, Patton DF et al. Thrombotic compli- disease of the liver following bone marrow transplantation. cations of BMT: association with protein C deficiency. Bone Transplantation 1987; 44: 778–783. Marrow Transplant 1993; 11: 61–65. 9 Pettitt AR, Clark RE. Thrombotic microangiopathy following 20 Faioni EM, Krachmalnicoff A, Bearman SI et al. Naturally bone marrow transplantation. Bone Marrow Transplant 1994; occurring anticoagulants and bone marrow transplantation: 14: 495–504. plasma protein C predicts the development of veno-occlusive 10 Zeigler ZR, Shadduck RK, Nemunaitis J et al. Bone marrow disease of the liver. Blood 1993; 81: 3458–3462. 21 Seeber C, Hiller E, Holler E et al. Increased levels of tissue transplant-associated thrombotic microangiopathy: a case plasminogen activator (t-PA) and tissue plasminogen activator series. Bone Marrow Transplant 1995; 15: 247–253. inhibitor (PAI) correlate with tumor necrosis factor alpha 11 Kalhs P, Brugger S, Schwarzinger I et al. Microangiopathy (TNF␣)-release in patients suffering from microangiopathy following allogeneic marrow transplantation. Transplantation following allogeneic bone marrow transplantation (BMT). 1995; 60: 949–957. Thromb Res 1992; 66: 373–383. 12 Holler E, Kolb HJ, Hiller E et al. Microangiopathy in patients 22 Remberger M, Ringden O, Markling L. TNF␣ levels are on cyclosporine prophylaxis who developed acute graft-ver- increased during bone marrow transplantation conditioning in sus-host disease after HLA-identical bone marrow transplan- patients who developed acute GVHD. Bone Marrow tation. Blood 1989; 73: 2018–2024. Transplant 1995; 15: 99–104. 13 Sviland L, Sale GE, Myelson D. Endothelial changes in 23 Ishii H, Uchiyama H, Kazama M. Soluble thrombomodulin cutaneous graft-versus-host disease: a comparison between antigen in conditioned medium is increased by damage of HLA matched and mismatched recipients of bone marrow endothelial cells. Thromb Haemostas 1991; 65: 618–623. transplantation. Bone Marrow Transplant 1990; 5: 39–42. 24 Gharpure VS, Devine SM, Holland HK et al. Thrombotic 14 Thompson CE, Damon LE, Ries CA et al. Thrombotic micro- thrombocytopenic purpura associated with FK506 following angiopathies in the 1980s: clinical features, response to treat- bone marrow transplantation. Bone Marrow Transplant 1995; ment, and the impact of the human immunodeficiency virus 6: 715–716. epidemic. Blood 1992; 80: 1890–1895. 25 Mach-Pascual S, Samii K, Beris P. Microangiopathic hemo- 15 Zeigler ZR, Rosenfeld CS, Andrews III DF et al. Plasma von lytic anemia complicating FK506 (Tacrolimus) therapy. Am J Willebrand factor antigen (vWF:AG) and thrombomodulin Hematol 1996; 52: 310–312.