Prospective study on the behaviour of the metalloprotease ADAMTS13 and of von Willebrand factor after bone marrow transplantation

Authors

  • F. Peyvandi,

    1. Angelo Bianchi Bonomi Haemophilia and Thrombosis Centre, University of Milan, Luigi Villa Foundation, Department of Medicine and Medical Specialities, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, Milan
    Search for more papers by this author
  • S. M. Siboni,

    1. Angelo Bianchi Bonomi Haemophilia and Thrombosis Centre, University of Milan, Luigi Villa Foundation, Department of Medicine and Medical Specialities, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, Milan
    Search for more papers by this author
  • D. Lambertenghi Deliliers,

    1. Organ and Tissue Transplant Immunology Unit, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, Milan, Italy
    Search for more papers by this author
  • S. Lavoretano,

    1. Angelo Bianchi Bonomi Haemophilia and Thrombosis Centre, University of Milan, Luigi Villa Foundation, Department of Medicine and Medical Specialities, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, Milan
    Search for more papers by this author
  • N. De Fazio,

    1. Organ and Tissue Transplant Immunology Unit, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, Milan, Italy
    Search for more papers by this author
  • B. Moroni,

    1. Angelo Bianchi Bonomi Haemophilia and Thrombosis Centre, University of Milan, Luigi Villa Foundation, Department of Medicine and Medical Specialities, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, Milan
    Search for more papers by this author
  • G. Lambertenghi Deliliers,

    1. Organ and Tissue Transplant Immunology Unit, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, Milan, Italy
    Search for more papers by this author
  • P. Mannuccio Mannucci

    1. Angelo Bianchi Bonomi Haemophilia and Thrombosis Centre, University of Milan, Luigi Villa Foundation, Department of Medicine and Medical Specialities, IRCCS Maggiore Hospital, Mangiagalli and Regina Elena Foundation, Milan
    Search for more papers by this author

Flora Peyvandi MD, PhD, Angelo Bianchi Bonomi Haemophilia and Thrombosis Centre, IRCCS Maggiore Hospital, University of Milan, Via Pace, 9 – 20122, Milan, Italy. E-mail: flora.peyvandi@unimi.it

Summary

Thrombotic microangiopathies (TMAs) are rare but serious complications of bone marrow transplantation (BMT). Clinical manifestations are similar to those of thrombotic thrombocytopenic purpura (TTP), but prognosis is generally poorer despite plasma exchange. The enzymatic activity of the plasma metalloprotease ADAMTS13, which cleaves ultralarge thrombogenic multimers of von Willebrand factor (VWF) derived from activated endothelial cells, is very low or undetectable in patients with classic TTP, and protease deficiency is thought to play a mechanistic role in the formation of platelet thrombi in the microcirculation. This is the first prospective study to evaluate the incidence of TMA in 46 consecutively recruited patients undergoing autologous or allogeneic BMT and explore in parallel the behaviour of ADAMTS13, VWF antigen and VWF multimer size. The incidence of post-BMT TMA was 6% (three of 46); all cases occurred after allogeneic BMT. Compared with baseline values plasma ADAMTS13 activity was significantly reduced in patients undergoing BMT, particularly after the conditioning regimen (mean values: 50 ± 22 vs. 77 ± 32%; P < 0·0001). In the three patients who developed TMA, ADAMTS13 decreased after conditioning, but was very low in one case only (8%). VWF antigen levels progressively increased after the conditioning regimen (228 ± 75 vs. 178 ± 76% at baseline, P = 0·002). The mean proportion of high-molecular weight VWF multimers did not change in the various stages of BMT, even though ultralarge multimers were transiently found in same cases with and without TMA. Hence, the measurements evaluated in this study are not clinically useful to predict the occurrence of post-BMT TMA.

Thrombotic microangiopathies (TMAs) are rare disorders characterised by microvascular platelet thrombi leading to microangiopathic haemolytic anaemia, consumptive thrombocytopenia and ischaemic manifestations in several organs (Elliott & Nichols, 2001; George & Vesely, 2003; Vesely et al, 2003). The identification of the von Willebrand factor (VWF) cleaving protease ADAMTS13 has contributed to elucidate the pathogenesis of thrombotic thrombocytopenic purpura (TTP), the prototype of TMAs (Furlan et al, 1996; Tsai, 1996; Levy et al, 2001). ADAMTS13 is a plasma metalloprotease that cleaves ultralarge, thrombogenic VWF multimers secreted from activated endothelial cells. In many patients with TTP the enzymatic activity of ADAMTS13 is undetectable or very low (usually <10%; Furlan et al, 1997, 1998; Tsai & Lian, 1998; Vesely et al, 2003; Peyvandi et al, 2004), so that ultralarge VWF multimers remain uncleaved in circulating blood. As a results of this, circulating platelets adhere massively, via glycoproteins (GP)Ibα and IIb–IIIa, to ultralarge VWF multimers anchored to P-selectin on endothelial cells and ultimately form occlusive thrombi in the microcirculation (Dong et al, 2002, 2003; Padilla et al, 2004). ADAMTS13 deficiency may be congenital or acquired. The rare congenital deficiency is an autosomal recessive disorder due to mutations in the ADAMTS13 gene (Levy et al, 2001). The more frequent, acquired deficiency is often due to autoantibodies against ADAMTS13, which develop in association with systemic autoimmune diseases, pregnancy, infections, the intake of some drugs and other conditions (Furlan et al, 1998; Tsai & Lian, 1998). TMA is a relatively frequent and serious complication after bone marrow transplantation (BMT), particularly occurring in patients receiving allogeneic BMT. TMA is often associated with graft-versus-host disease (GVHD) and with ciclosporin therapy for GVHD prophylaxis, and usually occurs within the first 150 d after BMT. In contrast to classic TTP, post-BMT TMA is generally unresponsive to plasma exchange (Fuge et al, 2001; Sarkodee-Adoo et al, 2003). The reported incidence of this complication is extremely variable (from 0·5% to 63·6%), perhaps due to the variety of diagnostic criteria and diseases leading to BMT (George et al, 2004). Risk factors include acute GVHD, ciclosporin prophylaxis for GVHD, total body irradiation, intensive conditioning chemotherapy, systemic infections and female sex (Guinan et al, 1988; Holler et al, 1989; Pettitt & Clark, 1994; Iacopino et al, 1999; van der Plas et al, 1999; Fassas et al, 2001; Ruutu et al, 2002; Sarkodee-Adoo et al, 2003).

To date, the mechanistic role of ADAMTS13 in post-BMT TMA had been evaluated only in a small number of retrospective studies (van der Plas et al, 1999; Arai et al, 2001; Elliott & Nichols, 2001; Allford et al, 2002; Vesely et al, 2003; George et al, 2004), including a total of 40 patients, and in no instance was a severe deficiency of the protease found. Only one study (Arai et al, 2001) also measured plasma VWF antigen (VWF:Ag), i.e. the substrate of the protease and a marker of endothelial cell activation: in all 10 post-BMT TMA patients, VWF:Ag was higher than normal and three of them had ultralarge VWF multimers detectable in plasma. We prospectively investigated the post-BMT behaviour of ADAMTS13 activity, VWF:Ag and VWF multimer size in a cohort of 46 patients with haematological malignancies, three of whom developed TMA.

Materials and methods

Patients

Forty-six consecutive patients (27 men and 19 women) with haematological malignancies who had a median age of 46 years (range: 19–70) and underwent allogeneic (19 cases) or autologous (27 cases) BMT between January 2004 and January 2005 were included in this study. It was conducted with the patients’ consent and approval by the Institutional Review Board. The characteristics of the 46 patients reported in Table I include age, gender, underlying disease leading to BMT, GVHD prophylaxis in recipients of allogeneic grafts and conditioning chemotherapy regimens. Acute GVHD was diagnosed and graded from I to IV according to consensus criteria (Przepiorka et al, 1995; Cahn et al, 2005). The clinical and laboratory parameters used for staging and grading acute GVHD included the percentage of body surface area with skin rash, total bilirubin, severity of diarrhoea and Karnofsky's performance status. Acute GVHD was treated with corticosteroids according to published criteria (Hings et al, 1994). Standard institutional guidelines for supportive care were adopted for all patients, including plasma and red blood cells transfusions and prophylaxis against bacterial, fungal and viral infections. The diagnosis of TMA was made on the basis of the concomitant presence of megakaryocytic thrombocytopenia (platelet count <100 × 109/l), Coombs’-negative haemolytic anaemia (haemoglobin <10·5 g/dl), absent or reduced plasma haptoglobin levels, fragmented erythrocytes and increased serum levels of lactate dehydrogenase (LDH) with no other identifiable cause for those abnormalities (Elliott & Nichols, 2001; George & Vesely, 2003; Vesely et al, 2003). More specifically, thrombocytopenia and anaemia because of bone marrow ablation were excluded on the basis of a white blood cell (WBC) count higher than 1·5 × 109/l and the presence of signs of mechanical haemolysis (increased bilirubin, reduced haptoglobin, fragmented erythrocytes). Disseminated intravascular coagulation (DIC) was excluded by finding normal values of the prothrombin time (PT), fibrinogen and D-dimer and autoimmune haemolysis was excluded by a negative direct antiglobulin test (Arai et al, 2001; Teruya et al, 2001).

Table I.   Patients characteristics and transplant procedures.
Patient numberAge (years)SexDiseaseTransplant typeGVHD prophylaxisConditioning regimen
  1. AML, acute myeloid leukaemia; PNH, paroxysmal nocturnal haemoglobinuria; NHL, non-Hodgkin lymphoma; ALL, acute lymphoblastic leukaemia; CML, chronic myeloid leukaemia; HD, Hodgkin disease; AA, aplastic anaemia; MM, multiple myeloma; Allo-BMT, allogeneic-related donor bone marrow transplantation; Allo-PBSC, allogeneic-related donor peripheral blood stem cell transplantation; Auto-PBSC, autologous peripheral blood stem cell transplantation; MUD-PBSC, matched unrelated donor peripheral blood stem cell transplantation; MUD-BMT, matched unrelated donor bone marrow transplantation; Auto-BMT, autologous bone marrow transplantation; CsA, ciclosporin; MTX, methotrexate; MMT, micophenolate; AraC, cytosine arabinoside; TBI, total body irradiation; BCNU, carmustine; ATG, antithymocyte globulin.

143MAMLAllo-BMTCsACyclophosphamide/AraC/TBI
259MMyelofibrosisMini Allo-PBSCCsAThiotepa/cyclophosphamide
352MAMLAllo-PBSCCsAThiotepa/cyclophosphamide
439MPNHAllo-BMTCsABusulphan/cyclophosphamide
538FNHLAllo-PBSCCsAVepesid/cyclophosphamide/TBI
622FHDAuto-PBSC BCNU/vepesid/cyclophosphamide
748MNHLMini MUD-PBSCCsA/MMTMelphalan/fludarabine/Campath/TBI
848MNHLMini MUD-BMTCsA/MMTMelphalan/fludarabine/Campath/TBI
932MALLMUD-BMTCsA/MtxCyclophosphamide/TBI
1053MAMLAllo-PBSCCsA/MtxThiotepa/cyclophosphamide
1170MNHLAuto-PBSC Melphalan
1226FCMLAllo-PBSCCsA/MtxBusulphan/cyclophosphamide
1343FNHLMini Allo-PBSCCsA/MMTPentostatine/TBI
1454MNHLAuto-PBSC BCN/vepesid/cyclophosphamide
1551MNHLAuto-PBSC Mitoxantrone/melphalan
1636FHDAuto-PBSC BCNU/AraC/etoposide/melphalan
1764MNHLAuto-PBSC BCNU/vepesid/cyclophosphamide
1846FAMLAuto-BMT AraC/cyclophosphamide/TBI
1956MNHLAuto-PBSC Mitoxantrone/melphalan
2065FNHLAuto-PBSC Melphalan
2150FNHLAllo-PBSC Singenic Vepesid/cyclophosphamide/TBI
2231FHDMini MUD-PBSCMMTMelphalan/fludarabine/Campath/TBI
2337FAAMUD-PBSCCsA/MtxCyclophosphamide/fludarabine/ATG/TBI
2460FMMAuto-PBSC Melphalan
2564MMMAuto-PBSC Melphalan
2661FNHLAuto-PBSC BCNU/vepesid/cyclophosphamide
2728MAMLAuto-PBSC AraC/cyclophosphamide/TBI
2822MALLMUD-BMTCsAAraC/cyclophosphamide/TBI
2921FNHLAuto-PBSC BCNU/cyclophosphamide/vepesid
3063MMMAuto-PBSC Mitoxantrone/melphalan
3143FMMAuto-PBSC Melphalan
3229FHDAuto-PBSC BCNU/vepesid/cyclophosphamide
3360FMMAuto-PBSC Melphalan
3465FNHLAuto-PBSC Mitoxantrone/melphalan
3551MNHLAuto-PBSC Melphalan
3630MNHLAuto-PBSC Thiotepa/melphalan
3724MNHLAllo-BMTCsA/MtxBusulphan/cyclophosphamide
3860MAMLAllo-PBSCCsA/MtxThiotepa/cyclophosphamide
3956MNHLAuto-PBSC Melphalan
4062MNHLAuto-PBSC BCNU/vepesid/cyclophosphamide
4156MHDAuto-PBSC BCNU/vepesid/cyclophosphamide
4251MNHLAuto-PBSC BCNU/vepesid/cyclophosphamide
4352FNHLAuto-PBSC BCNU/vepesid/cyclophosphamide
4456MMMMini Allo-PBSCCsA/MMTTBI
4524MNHLAuto-PBSC BCNU/AraC/etoposide/melphalan
4619FALLMUD-Cord bloodCsAThiotepa/cyclophosphamide/ATG

Laboratory methods

Sample collection.  Venous blood samples were collected into evacuated tubes containing sodium citrate as anticoagulant for the measurement of VWF:Ag and ADAMTS13 activity and into ethylenediaminetetraacetic acid (EDTA)-anticoagulated tubes for VWF multimeric analysis. Blood was then centrifuged at 1000 g for 15 min at 4°C and plasma samples were stored at −80°C. Plasma ADAMTS13 activity, VWF:Ag and VWF multimeric size were determined at seven different time points: before starting the conditioning regimen (baseline), after conditioning at day −1 before BMT and at days +1, +7, +15, +30 and +60 after BMT. At the same time points, blood counts and haemolysis indices (haptoglobin plasma levels, fragmented erythrocytes, LDH levels, serum bilirubin) were obtained.

Assays.  ADAMTS13 activity was measured by enzyme immunoassay as described by Gerritsen et al (1999), with some modifications. The source of VWF used as substrate for the protease was a purified VWF concentrate (Facteur Willebrand Humain Tres Haute Purite, provided by Laboratoire Francais du Fractionnement et des Biotechnologies, Lille, France). The product, which lacks endogenous ADAMTS13 activity, was reconstituted to a VWF concentration of 100 U/ml, aliquoted and stored at −30°C until use. Prior to digestion the concentrate was thawed, diluted 1 in 33 with 5 mol/l urea in 5 mmol/l Tris-HCl, pH 8, and incubated for 10 min at room temperature. Subsequently, 50 μl of concentrate dilution were added to 100 μl of each test plasma as a source of ADAMTS13 diluted in 5 mmol/l Tris-HCl, pH 8 containing 12·5 mmol/l Pefabloc SC (Roche, Mannheim, Germany) and digested overnight at 37°C. The most recent evaluation of the reproducibility of this assay yielded a within-assay coefficient of variation of 9% and a between-assay coefficient of variation of 12%. The lower limit of assay sensitivity was 6·25% of the normal protease levels. The lower value of the normal range (46%) was calculated on the basis of the fifth percentile of the distribution of the values obtained in 200 healthy individuals.

VWF:Ag was measured in plasma by an enzyme immunoassay using rabbit antihuman VWF polyclonal antibodies as first and second antibodies (Dako, Glostrup, Denmark). The normal range was calculated as for ADAMTS13.

von Willebrand factor multimer analysis was carried out using low-resolution sodium dodecyl sulphate-agarose gel electrophoresis with 0·9% low gelling temperature agarose in order to optimally resolve ultralarge VWF multimers. After electrophoresis, gels were exposed to human anti-VWF antibodies labelled with I125 (Amersham, UK) and then transferred to intensifier cassettes for autoradiography (Ruggeri & Zimmerman, 1981). VWF multimers were scanned with a densitometer (Scanjet 5200 C; Hewlett Packard, Piscataway, NJ, USA), which resolved the multimers in peaks (OD × mm) and densitometric analysis was performed using a software program (Image Master; Amersham Pharmacia Biotech, Piscataway, NJ, USA). As recommended by Budde et al (2002), multimers were defined as low-molecular weight (corresponding to bands 1–5), intermediate-molecular weight (bands 6–10) and high-molecular weight (bands >10). The corresponding areas were computed and expressed as a percentage of the total area. Ultralarge VWF multimers were considered to be present when the percentage of high-molecular weight multimers was >34%, i.e. 2 SD above the mean values found in the plasma from 50 healthy individuals (28·3 ± 2·7%).

Statistical analysis

Arithmetic mean values, standard deviations and standard errors of the variables considered were calculated. Multiple comparisons of continuous variables were carried out using one-way anova (t-test for parametric data). Values of P < 0·05 were considered significant. All the statistical analyses were performed using the sas statistical package (SAS Institute, version 8, Cary, NC, USA).

Results

Patients

Graft-versus-host disease developed in three of 19 patients who underwent allogeneic transplantation (16%): two patients developed grades I–II acute cutaneous GVHD and the remaining patient had grade III acute intestinal GVHD. The mortality reported as a consequence of BMT was 17% (eight of 46). Causes of fatal outcome were respiratory failure due to interstitial pneumonia (four cases), multiorgan failure, septic shock, gastric bleeding and respiratory insufficiency in one case each. Death occurred at post-BMT days +122, +56, +24 and +70 in the four cases of respiratory failure, and days +44, +478, +126 and +100 in the remaining four cases.

Three of 46 patients (6·5%) developed TMA after BMT, two females and one male (N5, N12 and N28). In these patients the underlying diseases were non-Hodgkin lymphoma (N5), chronic myeloid leukaemia (N12) and acute lymphoblastic leukaemia (N28). Two patients had received BMT from phenotypically identical related donors (N5 and N12) and one from a HLA-A-, HLA-B- and DRB1-matched unrelated donor (N28). All the patients had received immunosuppressive therapy for GVHD prophylaxis. Conditioning regimens for each patient were vepesid/cyclophosphamide/total body irradiation, busulphan/cyclophosphamide and cytosine arabinoside/cyclophosphamide/total body irradiation respectively. The main clinical and laboratory findings of these three patients at the time of TMA are given in Table II. Marrow ablation, DIC and autoimmune haemolysis were excluded as indicated above. TMA developed between 18 and 44 d after BMT. Only one patient (N5) developed clinically symptomatic TMA, with mild dyspnoea, confusion, epistaxis, menorrhagia and haematuria. She developed grade III acute intestinal GVHD 46 d after BMT and 24 d after the diagnosis of TMA. Laboratory findings showed thrombocytopenia, microangiopathic haemolytic anaemia with undetectable serum haptoglobin, increased serum LDH and signs of moderate renal failure with normal hepatic function (Table II). After a transient early improvement the patient deteriorated again on day +120 after BMT and developed septic shock with fatal outcome on day +126. The remaining two patients had laboratory signs of TMA without overt clinical manifestations or GVHD. Patient N12, the only one who survived, had thrombocytopenia, microangiopathic haemolytic anaemia with undetectable haptoglobin and slightly increased levels of LDH 18 d after BMT (Table II). The third patient (N28) had thrombocytopenia, microangiopathic haemolytic anaemia with undetectable haptoglobin and increased LDH 44 d after BMT; he died on day +56 because of interstitial pneumonia (Table II). Bacterial infections were documented in all the three patients at days +29, +2 and +12 after TMA diagnosis, but at the time of diagnosis culture studies performed on these patients were negative. All patients were treated with defibrotide 20 mg/kg by daily intravenous infusions (Corti et al, 2002).

Table II.   Clinical and laboratory findings in patients who developed TMA following allogeneic bone marrow transplantation.
Patient numberTMA diagnosis after BMT (d) Hb (g/dl)* Platelets (×109/l)† Haptoglobin (g/l)‡ LDH (IU/l)§GVHD grade (day after BMT)Systemic infection (day after BMT)FeverNeurological signsRenal signsBleeding symptomsTMA treatment (mg/kg)Days from BMT to deathCause of death
  1. Reference range: *12–16 g/dl; †140–440 × 109/l; ‡0·4–1·9 g/l; §230–460 IU/l.

  2. TMA, thrombotic microangiopathy; BMT, bone marrow transplantation; Hb, haemoglobin; LDH, lactate dehydrogenase; GVHD, graft-versus-host disease.

5229·18<0·21592Intestinal III (+46)Bacterial (+51)NoConfusionSerum creatinine 1.6 mg/dl, haematuriaEpistaxis, menorrhagiaDefibrotide 20126Septic shock
12189·319<0·2543NoBacterial (+20)NoNoNoNoDefibrotide 20
284410·124<0·2686NoBacterial (+56)YesNoNoOral mucosal bleedingDefibrotide 2056Multifocal interstitial pneumonia

ADAMTS13 activity

As there was no statistically significant difference between patients who had undergone allogeneic versus autologous BMT in terms of ADAMTS13, VWF:Ag and VWF multimers (Table III), the data are presented together. In the whole group of 46 patients ADAMTS13 activity had already decreased after conditioning chemotherapy when assessed on the day before BMT (day −1; mean values: 50 ± 22 vs. 77 ± 32 at baseline; P < 0·0001). ADAMTS13 remained lower than before conditioning until +30 d after BMT (mean values: 62 ± 31 vs. 77 ± 32; P = 0·02; Fig 1). Only six of 46 patients (13%) had ADAMTS13 values below the lower limit of the normal range (46%) before conditioning, but after conditioning, on day −1 before BMT, the number of patients with low ADAMTS13 increased to 21 (46%). Of the three patients who developed TMA only one (N5) had moderately reduced ADAMTS13 levels at baseline (34%) that fell further to 28% after conditioning and remained low in the post-BMT period (between 18% and 31%; Table IV). Patient N12 had normal values at baseline (75%), which fell to 41% after conditioning and fluctuated between normal and low borderline values during post-BMT follow up (Table IV). The third patient (N28) had normal baseline values (60%) that fell to 33% after conditioning and remained consistently low (Table IV). Only once were ADAMTS13 levels lower than 10%, i.e. the values typically present in patients with classic TTP (8% in patient N28, at day +54).

Table III.   Mean values of ADAMTS13, VWF:Ag and VWF multimers values in patients who had undergone allogeneic (19 cases) versus autologous (27 cases) BMT.
 ADAMTS13VWF:AgHigh-molecular weight VWF multimers
Allo-BMTAuto-BMTAllo-BMTAuto-BMTAllo-BMTAuto-BMT
  1. All values are expressed as a percentage of normal plasma value. None of the differences was statistically significant by multivariate analysis.

  2. VWF, von Willebrand factor; BMT, bone marrow transplantation.

Baseline76 ± 3178 ± 32166 ± 64187 ± 8430 ± 329 ± 4
Day −156 ± 2346 ± 20220 ± 60233 ± 8530 ± 429 ± 3
Day +159 ± 2852 ± 25240 ± 76242 ± 8628 ± 429 ± 4
Day +753 ± 2047 ± 24275 ± 94276 ± 10128 ± 630 ± 3
Day +1550 ± 2149 ± 25274 ± 100355 ± 15727 ± 629 ± 3
Day +3064 ± 3061 ± 32309 ± 107304 ± 17628 ± 428 ± 3
Day +6070 ± 3575 ± 30261 ± 95232 ± 12828 ± 430 ± 5
Figure 1.

 Mean values (± SD) of ADAMTS13 activity, VWF:Ag and percentage of high-molecular weight (HMW) von Willebrand factor (VWF) multimers at each time point in 46 patients who underwent bone marrow transplantation (BMT). The dashed lines show the lower limit of the normal range for ADAMTS13 activity and the higher limit of the normal range for VWF:Ag and high-molecular weight VWF multimers.

Table IV.   Pre- and post-BMT levels of ADAMTS13 in 43 patients without TMA and in the three patients who developed TMA.
 BaselineDay −1Day +1Day +7Day +15Day +30Day +60
  1. ADAMTS13 activity values are expressed as percentage activity of normal plasma. Values below the lower limit of normal levels (46%) are shown in italics.

  2. *ADAMTS13 activity was measured on day +54.

  3. TMA, thrombotic microangiopathy; BMT, bone marrow transplantation.

Patients without TMA (mean ± SD)79 ± 3251 ± 2956 ± 2650 ± 2350 ± 2463 ± 3175 ± 30
Patient N534282231182026
Patient N1275414145619393
Patient N286033363430398*

VWF:Ag

Figure 1 shows that, in the whole group of 46 patients, there was a significant increase of VWF:Ag after conditioning on day −1 before BMT. High levels were sustained until day +60 after BMT, the peak being reached on day +15 (mean values: 322 ± 141 vs. 178 ± 76 at baseline; P < 0·0001). The three patients with TMA had slightly lower values of VWF:Ag than the 43 patients without TMA (Table V).

Table V.   Pre- and post-BMT levels of VWF:Ag in 43 patients without TMA and in the three patients who developed TMA.
 BaselineDay −1Day +1Day +7Day +15Day +30Day +60
  1. VWF:Ag levels are expressed as a percentage of normal plasma. Values above the upper limit of normal levels (150%) are shown in italics.

  2. *VWF:Ag was measured on day +54.

  3. TMA, thrombotic microangiopathy; BMT, bone marrow transplantation.

Patients without TMA (mean ± SD)183 ± 76231 ± 76245 ± 82280 ± 98327 ± 144311 ± 155241 ± 118
Patient N5144182172280288284216
Patient N1290130196179214234256
Patient N28111228191186222207349*

VWF multimeric analysis

The mean percentage of high-molecular weight VWF multimers remained unchanged throughout the BMT period (Fig 1), but one of the three patients who developed TMA (N28) had ultralarge multimers (more than 34%) both at baseline and after conditioning (Table VI). On the other hand, ultralarge VWF multimers were also detected in two patients who underwent BMT and did not develop TMA (N15, N16).

Table VI.   Pre- and post-BMT high-molecular weight VWF multimers in 43 patients without TMA and in the three patients who developed TMA.
 BaselineDay −1Day +1Day +7Day +15Day +30Day +60
  1. High-molecular weight VWF multimers values are expressed as a percentage of normal multimers. Values above the upper limit of normal level (34%; shown in italics) identify ultralarge multimers.

  2. *High-molecular weight VWF multimers were measured on day +54.

  3. VWF, von Willebrand factor; TMA, thrombotic microangiopathy; BMT, bone marrow transplantation.

Patients without TMA (mean ± SD)29 ± 329 ± 428 ± 429 ± 528 ± 428 ± 329 ± 5
Patient N531343228302931
Patient N1231333132173130
Patient N2835363335303334*

Discussion

ADAMTS13 activity was previously measured only retrospectively in a small number of patients with post-BMT TMA, and found to be normal or only slightly reduced (van der Plas et al, 1999; Arai et al, 2001; Elliott & Nichols, 2001; Allford et al, 2002; Vesely et al, 2003). This is the first study to investigate the behaviour of ADAMTS13, VWF:Ag and VWF multimers in parallel, in a cohort of 46 patients observed prospectively at seven different time points before and after allogeneic or autologous BMT. Three patients developed TMA after allogeneic BMT (two from related donors and one from an unrelated donor), with an incidence of 6·5%. These data confirm a higher frequency of TMA in allogeneic than autologous BMT (Pettitt & Clark, 1994; Iacopino et al, 1999) and suggest that immunosuppressive therapy for GVHD prophylaxis contributes to the occurrence of TMA. Patients who underwent BMT had lower values of ADAMTS13 after the conditioning regimen than at baseline, so that many patients received the graft at a time when ADAMTS13 was below the lower limit of normal values. The moderate decrease of ADAMTS13 activity consistently observed after conditioning chemotherapy is perhaps due to the occurrence of acute phase reactions, inflammation and impairment of liver function (Holler et al, 1990; Krenger et al, 1997), because these conditions were previously associated with moderately reduced levels of the protease (Mannucci et al, 2001). The mechanism of the decrease of protease remains speculative (cleavage by other plasma proteases? consumption becuase of ADAMTS13 docking to ultralarge VWF multimers stretched upon activated endothelial cells?). In the three patients who developed TMA, ADAMTS13 was reduced after conditioning but as many as 18 additional patients with similarly reduced levels (18–45%) did not develop TMA. Only one of the three patients who developed TMA had ADAMTS13 deficiency that was as severe (8%) as in the majority of patients with classic TTP, confirming the original data of van der Plas et al (1999) that most cases of BMT-related TMA do not exhibit severely deficient levels of ADAMTS13 activity.

A marked increase of VWF:Ag occurred after the conditioning regimen, to peak on day +15 after BMT, perhaps as a consequence of massive endothelial cell activation in these patients (Licciardello et al, 1985; Charba et al, 1993; Arai et al, 2001). In patients who did not develop TMA the multimeric distribution of VWF during BMT was similar to that of normal individuals, in agreement with van der Plas et al (1999). One of the three patients who developed TMA had ultralarge thrombogenic multimers already before BMT and after conditioning chemotherapy. On the other hand, two patients with ultralarge multimers before BMT did not develop TMA, so, at moment, there is no evidence that analysis of the multimeric composition of VWF is clinically useful to predict the occurrence of TMA.

On the whole, this prospective study demonstrated that there are significant changes in ADAMTS13 and VWF levels after BMT, reflected by lower levels of enzymatic activity of the protease and higher levels of its substrate compared with the baseline condition. This and previous studies have found that plasma levels of ADAMTS13 were not dramatically reduced after BMT, but Dong et al (2004) have recently shown that static ADAMTS13 assays, such as that used in this and other studies, may not accurately measure the ADAMTS13 activity required to cleave ultralarge, endothelium-derived VWF multimers under the flow conditions of the microcirculation. Until assays that measure this interaction are available for clinical studies, the possibility of an alteration in the balance between VWF and its cleaving protease in the mechanisms of post-BMT TMA cannot be ruled out.

Ancillary