Thrombotic microangiopathy (TMA) syndromes – the hemolytic-uremic syndrome (HUS) and thrombotic thrombocytopenia purpura (TTP) – can complicate bone marrow transplantation (BMT). There is no specific treatment. Plasma exchange (PE) has been used to treat TTP/HUS, but the response rate is generally below 50% and the median mortality rate is as high as 79% in the BMT population.1

Rituximab, an anti-CD20 antibody, has been used to treat recurrent TTP,2, 3, 4, 5 but there are no reports of its use in TMA associated with BMT (BMT-TMA) and only one report of its use in HUS.6

We report a patient who developed life-threatening HUS 7 days after undergoing allogeneic BMT (allo-BMT) for acute lymphoblastic leukemia (ALL). He improved after rituximab administration.

In May 2001 a 26-year-old man was diagnosed with ALL, when he presented with fatigue and progressive lymphadenopathy. He was treated with the Group for Treatment of Childhood Acute Lymphocytic Leukemia (GBTLI)-85 protocol7 and entered first complete remission (CR). Maintenance chemotherapy consisted of methotrexate and purinethol. In January 2005, he had a marrow and testicular relapse. Second CR was achieved with the same GBTLI-85 protocol plus 3.4 Gy of testicular irradiation.

In September 2005, he underwent a matched sibling allo-BMT. Conditioning included 12 Gy of fractionated total body irradiation and cyclophosphamide 120 mg/kg. The graft consisted of 3.86 × 106 CD34+ cells/kg. Graft-versus-host disease (GVHD) prophylaxis included cyclosporine 2 mg/kg/day and methotrexate 10 mg/m2 on D+1, +3 and +6.

On D+7, he developed nausea, vomiting and gingival bleeding. The hemoglobin level was 9.8 g/dl, platelet count 19 000/mm3, total bilirubin 2.59 mg/dl, indirect bilirubin 1.91 mg/dl (normal 0–1.0 mg/dl), lactic dehydrogenase (LDH) 229 IU/l (n=120–200 IU/l), urea 34 mg/dl and creatinine 2.1 mg/dl (baseline creatinine level 0.6 mg/dl). Coagulation studies were normal and Coombs test (direct and indirect) was negative. Peripheral blood and bone marrow smears contained +3 schistocytes per high-power field. Haptoglobin levels were undetectable. There was no evidence of infection. A diagnosis of BMT-associated HUS was established (Figure 1).

Figure 1
figure 1

Hemoglobin, lactic dehydrogenase, creatinine and indirect bilirubin after plasma exchange (→) and rituximab (↑).

Full size image

Cyclosporine was discontinued immediately and methylprednisolone 2 mg/kg/day was started, together with PE. PE consisted of daily exchange of 1.0 total body plasma volumes with fresh frozen plasma replacement. Median duration of each session was 129 min (range 90–172 min).

Three days later his clinical condition and hemolysis worsened, and hematuria occurred. Transfusion of eight packs of filtered irradiated red cells and 7 U of platelets four times daily had no apparent benefit. Laboratory studies showed persistence of schistocytes, hemoglobin 6.5 g/dl, LDH 683 IU/l and bilirubin 12.4 mg/dl (indirect 6.95 mg/dl). Neurological status remained normal. Despite 6 consecutive days of PE his clinical condition continued to deteriorate, with severe anemia leading to heart failure, pulmonary edema and respiratory insufficiency, necessitating mechanical ventilation. The hemolysis continued to worsen, the hemoglobin dropped to 3.2 g/dl and platelet count to 23 000/mm3, and the LDH increased to 1853 IU/l. His renal function worsened, the creatinine rising to 3.4 mg/dl. Rituximab 375 mg/m2 was administered, starting on D+17, with prompt clinical and biological improvement (graphs). The same dose of rituximab was given for a further 3 days (total four doses).

Rituximab was well tolerated, with no apparent adverse effects. Renal function recovered, hemolysis and thrombocytopenia resolved, and no schistocytes were found in peripheral blood 1 week after the first rituximab dose. HUS has not recurred.

Neutrophil engraftment occurred on D+9 (neutrophil 990/mm3). No signs of GVHD occurred. Prednisone 1 mg/kg every other day was given until D+180 and cyclosporine was discontinued. He has good performance status and normal biological status 17 months after allo-BMT, and is in continuous marrow and testicular remission.

The reported incidence of BMT-TMA ranges widely, from 0.5 to 76%,7, 8, 9, 10, 11 owing partly to the use of different diagnostic criteria. Recently, the Blood and Marrow Transplant Clinical Trials Network Toxicity Committee (BMT CNT) published an operational definition for BMT-TMA, designed to make future studies more comparable. The present case meets these diagnostic criteria, which include RBC fragmentation, increased LDH, concurrent renal and/or neurologic dysfunction (without other explanations) and negative direct and indirect Coombs test results.12

The prognosis is poor and there is no consensus treatment,7 apart from cyclosporine discontinuation. PE is the treatment most widely used for BMT-TMA, but there are no randomized trials, the response rate is below 50%, and mortality remains higher than 80% despite PE.9, 10, 11

There are numerous reports on the successful use of rituximab in recurrent idiopathic TTP.2, 3, 4 De novo TTP is attributed to unusually large von Willebrand factor multimers (ul-vWF) that lead to platelet clumping and microvascular thrombosis. Ul-vWF are normally cleaved into smaller protein units by a metalloprotease called von Willebrand factor-cleaving protease (ADAMTS13). Impaired ADAMTS13 function, whether hereditary or acquired (autoantibodies), leads to excessive ul-vWF accumulation and to TTP.13 The use of rituximab in de novo TTP was based on the premise that rituximab might deplete B-cell clones producing anti-ADAMTS13 antibodies.14

The pathogenesis of BMT-TMA is poorly understood. Several potential triggers of microvascular damage have been suggested, including cyclosporine, FK506, TBI, infection and GVHD.7 Although elevated vWF levels have been reported after BMT-TMA, they are associated with normal metalloprotease activity and a normal vWF multimer pattern.15 Unrelated BMT is a risk factor for BMT-TMA.9, 10, 16 An immunological reaction owing to donor-recipient genetic disparity and subsequent cytokine release might trigger the microangiopathy seen in BMT-TMA. Rituximab may act in TMA-BMT by depleting CD20+ precursors of B cells, thereby attenuating T-cell activation and reducing cytokine release.

The only other reported use of rituximab in HUS involved a renal graft recipient with familial HUS. The patient was a 36-year-old woman who received over 40 PE sessions for recurrent HUS in the transplanted kidney, with no significant improvement. Courses of two doses of rituximab 375 mg/m2 were administered 1 week apart, with very rapid recovery of hematological status and graft function.5

Clinical remission from TTP has been reported as early as 1 week after rituximab administration.3, 4 In both reported cases of HUS, the clinical and hematological responses to rituximab were almost immediate.5 Though TTP and HUS are considered part of a spectrum of the same disorder, the activity and level of von Willebrand factor-cleaving protease are normal in HUS. This suggests in part a different pathogenesis of these two entities.13 The rapid response to rituximab observed in both cases of HUS, however, suggests significant B-cell involvement in the pathogenesis of this disease.

Most patients with severe BMT-TMA die and TMA rarely resolves, despite a variety of PE procedures.8 Our patient had severe BMT-TMA and it is unlikely that PE was responsible for his recovery, as the hemolysis and renal function continued to deteriorate despite six PE sessions. PE has previously been reported to provide no benefit in BMT-TMA.1 The prompt hematological and clinical recovery associated with rituximab administration thus warrant further study.