This review of the EBMT and EULAR databases indicates that HSCT for vasculitis not responding to conventional treatment is feasible and may considerably alter the course of the disease. The high response rate to HSCT of these heavily pretreated patients is promising and the responses seem to be durable in most patients. Most centres used a conditioning regimen based on cyclophosphamide. Lymphocyte depletion was often used either ex vivo (CD 34 selection) or in vivo (ATG) or both. Cyclophosphamide in combination with ATG and/or CD34 selection of the graft is now the conditioning regimen in most phase III trials for autoimmune diseases (protocols are available at http://www.ebmt.org
Autologous HSCT for vasculitis in this patient group was not associated with treatment‐related mortality. The main side effects were transient and mostly associated with aplasia after conditioning (neutropenic fever). One patient died from lung cancer 2 years after HSCT. The established risk factor was smoking; in addition, he had previously received a cumulative cyclophosphamide dose >100 g. It remains possible that HSCT‐associated immunosuppression played an additional part in the development of malignancy. However, an increased incidence of secondary malignancies is reported in patients who had received a HSCT for malignant diseases.18
To our knowledge, two patients with a relapse of ANCA‐associated vasculitis after G‐CSF administration are described in the literature.19
In this series, however, no exacerbation of ANCA‐associated vasculitis after G‐CSF administration for stem cell mobilisation was observed.
It is important to mention that the median age of patients with ANCA‐associated vasculitis reported herein (37 years) was lower than that of patients reported previously in the literature.
Three patients received allotransplantation either primarily, or because of relapses after autologous HSCT; one of them died. Allogeneic HSCT in patients with autoimmune diseases has the potential to change genetically determined factors influencing “autoimmunity” through establishing a potentially “non‐autoimmune” new donor‐type immune system. This may be associated with a so‐far poorly understood graft versus autoimmune effect. Long‐term remissions in patients with autoimmune diseases after allogeneic HSCT are reported in case reports.20
Further experience comes from patients with haematological malignancies and concomitant autoimmune diseases. Remission of Wegener's disease after allogeneic HSCT with reduced conditioning for leukaemia is reported.21
Allogeneic HSCT with reduced intensity conditioning is therefore now discussed as treatment for severe autoimmune disease.22
Concerns regarding toxicity through acute or chronic GvHD remain.
The effect of autologous HSCT on the immune system in patients with autoimmune diseases is not clear, and is possibly not only due to a more intense immunosuppression but also to a “resetting” of the autoaggressive immune response. In patients with multiple sclerosis, durable remissions were seen despite a full return of normal protective immunity.23
This report is limited by the heterogeneity of the data, the post hoc definition of response to HSCT, and the incomplete nature of data on 9 of 25 patients reported to the database. We lack uniformly accepted criteria that define “response to treatment” for each of the (rare) diseases that led to HSCT in this report. In the setting of this retrospective analysis, a relatively robust read‐out for response to treatment is the intensity of post‐HSCT immunosuppression needed to control each disease. This read‐out is clearly limited and is influenced by the individual judgement of the physician.
Larger and prospectively designed clinical trials are thus needed to confirm these preliminary data. The key question remains the selection of patients for HSCT. The initial results with HSCT have shown that patients with reduced organ function and autoimmune disease have considerable treatment‐related mortality.24
In subsequent trials, therefore, those patients were excluded. The patients most likely to benefit from a more intense treatment must be identified as early as possible in the disease course. Established prognostic factors must be used to select patients at risk. For the group of necrotising vasculitidies, the Birmingham Vasculitis Activity Score and the five‐factor score are correlated with long‐term outcome and mortality.25
Patients with renal, hepatic, cerebral and gastrointestinal involvement have a higher mortality in small‐vessel vasculitides.26
These risk factors should be considered for patient selection for HSCT.
We think that before proceeding to HSCT, standard treatment should be proved ineffective. This includes cyclophosphamide either orally or in monthly intravenous pulses. Owing to promising initial results and low toxicity, rituximab may also be considered before HSCT in patients with ANCA‐associated vasculitis and cryoglobulinaemia.
In conclusion, patients with preserved organ function and poor prognostic factors failing to respond to cyclophosphamide should be considered for autologous HSCT. Owing to the limited number of patients treated so far, further experience with autologous HSCT for patients with vasculitis should be gained by developing a common international protocol. Such a protocol is in preparation.