The role of PBSCT in treatment of AL amyloidosis is far from settled

We would like to thank Dr Jaccard and colleagues and Dr Goodman and colleagues for their cogent comments on our recent paper.¹  We agree that there is no substitute for a sufficiently powered, well-designed, randomized controlled trial to answer questions pertaining to optimal treatment. In a disease as heterogeneous in its manifestations and in its severity as primary systemic amyloidosis (AL), the challenges of making treatment decisions and comparing phase 2 results are further compounded.

We too were impressed by the results of Merlini's group (Palladini et al,²  from Pavia, Italy) and eagerly await the results of the ongoing randomized controlled trial of Jaccard et al. The Pavia group has recently reported an overall hematologic response rate of 66%, a complete response rate of 33%, and an organ response rate of 48% using oral melphalan and dexamethasone.²  With a median follow-up of only 20 months, 9 patients (20%) have died, the majority (7 patients) of whom were nonresponders. It will be most interesting to see the long-term progression-free and overall survival rates from the Pavia group study. The Intergroupe Francais du Myelome 94 trial has shown that comparable complete response rates in patients with myeloma do not translate directly into comparable overall survival rates.³  In that study, despite equivalent complete response rates in the single and tandem transplantation arms, the myeloma patients undergoing tandem transplantation enjoyed a 7-year overall survival, double that of the single transplantation arm.

Drs Goodman and Hawkins suggest that there are other promising therapeutic options that “hold promise of greater and more rapid response rates than melphalan and prednisone without the substantial treatment related mortality and morbidity.. .associated with PBSCT.” We remind the authors and the readers that these “promising” alternatives have only been tested in the context of observational and phase 2 trials. Moreover, thalidomide has been shown to be quite toxic in this group of patients.^4,5  Although reported to be associated with hematologic response rates as high as 60%,⁶  vincristine-adriamycin-dexamethasone, in theory, could also be quite toxic to patients with AL given the known risk of vincristine and adriamycin in patients with neuropathy and cardiomyopathy, respectively. Lachmann et al's report of cyclical intermediate-dose intravenous melphalan (25 mg/m²) in 24 patients is also quite encouraging, with more than half of the patients achieving at least a 50% reduction in serum immunoglobulin-free light chain and more than a third demonstrating regression of amyloid load by serum amyloid P component (SAP) scintigraphy.⁶  However, only a phase 3 trial will definitively clarify its utility.

The discussion of phase 2 trials brings us to the concept of “response,” which is replete with difficulties. What types of responses are most meaningful in AL: hematologic or clinical? In a disease with low tumor burden, like AL, what is a hematologic response? Is it 50% reduction of the M-spike in the serum or in the urine (the latter frequently contaminated by nonspecific proteinuria)? Or is it disappearance of the monoclonal protein and bone marrow plasmacytosis by immunofixation and immunostains? The introduction of the serum immunoglobulin free light chain assay (The Binding Site, Birmingham, United Kingdom) has revolutionized our ability to assess hematologic response in patients with a low tumor burden, but there is not yet a consensus on hematologic response criteria using this tool. While only one third of amyloid patients have serum M-spikes of more than 0.1 g/L (1 g/dL) (measurable disease), more than 90% of patients have measurable disease with this new assay.^6,7  Moreover, what is an organ response? How much clinical improvement is required to code a response? There has been an international consensus opinion proposed by the 10th International Symposium on Amyloid and Amyloidosis, which attempts to standardize definitions of organ involvement and treatment response (M.A.G., Ray Comenzo, Rodney H. Falk, Jean Paul Fermand, Bouke P. Hazenberg, Philip N. Hawkins, Giampaolo Merlini, Philippe Moreau, Pierre Ronco, Vaishali Sanchorawala, Orhan Sezer, Alan Solomon, and Giles Grateau, manuscript submitted August 2004).

Certainly several of these issues will be attenuated in the context of a randomized clinical trial. However, only meticulous stratification of patients will allow for interpretable results. Once again, there is no consensus on the best stratification factors in patients with AL, although dominant organ involvement and number of organs involved have been used. Recent data would suggest that using serum levels of cardiac biomarkers may be powerful tools to stratify patients with AL.^8-10 

So what is the best treatment for AL? We agree with Jaccard and colleagues and Goodman and colleagues that for now, the best choice for patients is a well-designed, randomized controlled trial. Until the results are available from such trials, what are hematologists with newly diagnosed AL patients to do? Although not a phase 3 trial, our data suggest that offering high-dose chemotherapy with peripheral blood stem cell transplantation is an appropriate ethical option with potential for improved survival. Referral to an amyloid treatment center for an opinion regarding transplantation is recommended since excellent long-term survival in patients (albeit selected patients) receiving this therapy has been repeatedly demonstrated.^1,11,12 

Jaccard et al's caution against high-dose therapy and autologous hematopoietic stem cell transplantation (HSCT) for AL amyloidosis¹  may well prove right if the group's ongoing study, after completion of accrual and adequate follow-up, eventually does show equivalent outcomes with conventional-intensity melphalandexamethasone and high-dose melphalan. While their rejoinder implies good outcomes with melphalan-dexamethasone that are comparable to high-dose chemotherapy in the study, exactly what the results are and how they compare with other published data will remain unknown until the study matures and numbers replace descriptions. Methodologic issues may well influence the relevance of the findings. For example, if the dose of melphalan for HSCT in the French cooperative study has been less than 200 mg/m² (a point not clarified in Jaccard et al's comments), based on the Boston University data showing superior outcome with the higher dose of melphalan,²  its findings may well be considered questionable even before the accrual target is met.

There is a well-defined center effect influencing the results of allogeneic HSCT: larger volumes (and correspondingly, experience) being associated with superior outcomes.³  It is very likely that such an effect exists for autologous HSCT AL amyloidosis too because of the unusual and serious nature of the complications patients with this disease tend to develop after high-dose therapy. If the French experience does suggest higher treatment-related mortality with HSCT than with melphalan-dexamethasone, it is likely to be at least partially attributable to lack of adequate experience since, on an average, each center appears to be recruiting fewer than 1 patient on study per year—and thus autografting 1 patient every 2 years. It is debatable whether such a concern would be applicable to an equal extent to centers with much greater experience in treating patients with amyloidosis.^2,4  If the outcome of autotransplantation in the French study is inferior to that reported by the Boston University and Mayo Clinic groups largely due to higher treatment-related mortality, I would argue that the interests of French patients with amyloidosis would be far better served by receiving high-dose therapy at a small number of specialist centers around France rather than abandoning high-dose therapy altogether.

A number of groups, including ours, have adopted the interesting approach reported by Palladini et al⁵  in patients ineligible for HSCT because of the ineffectiveness of other treatment approaches. However, just like the Palladini group,⁵  we continue to offer high-dose therapy with autotransplantation to all patients who are eligible for it. If longer follow-up of the data presented by Palladini et al shows that the outcome of melphalan-dexamethasone-treated patients ineligible for HSCT appears similar to that of HSCT recipients treated optimally (200 mg/m² melphalan), that would indeed signal a need for a prospective, randomized study comparing melphalan-dexamethasone and high-dose melphalan.

While the retrospective experience of London's National Amyloidosis Centre⁶  quoted by Goodman and Hawkins may appear to suggest that the extent of response rather than the type of therapy is important, the paper does not provide adequate information on biologically relevant characteristics to show whether the patient populations were comparable. The study also does not clearly depict event-free and overall survival by the type of therapy provided. Additionally, after initial assessment at the Centre, patients were treated locally by referring physicians, thus bringing up the question of experience influencing outcome just as with the French study.

The Palladini study⁷  quoted by Goodman and Hawkins is an evaluation of prognostic indicators rather than specific treatment options or useful alternatives to HSCT. While HSCT may not turn out to be the best approach for AL amyloidosis in the future if newer options evolve, the only substantial peer-reviewed evidence currently available in patients who have been studied in depth and treated relatively homogeneously comes from Boston University and the Mayo Clinic—and favors high-dose melphalan with HSCT.

Following McElwain and Powles's report of the extraordinary activity of high-dose melphalan in plasma cell dyscrasias,⁸  it took 15 to 20 years for randomized studies to be published showing that high-dose therapy was indeed beneficial^9,10 —time that Barlogie et al^11,12  spent studying important second-generation dose-intensity questions. The Jaccard study evaluating HSCT in amyloidosis should of course continue. However, should we be initiating additional comparative studies without learning from the French study and the Mayo and Boston experience? I think not.