Defining ultrahigh-risk AL amyloidosis with VWF

Nelson Leung

In this issue of Blood, Kastritis et al report the results of using von Willebrand factor antigen (VWF:Ag) to define an ultrahigh-risk group of patients with immunoglobulin amyloid light-chain (AL) amyloidosis.1 The accurate identification of these patients could help design future clinical trials desperately needed to define the best treatment of these patients.

Median survival of patients with Mayo stage IIIb disease and a VWF:Ag level <230 U/dL was 6 months vs 2 months in those who had a VWF:Ag level >230 U/dL (P = .006). See Figure 2C in the article by Kastritis et al that begins on page 405.

One of the most remarkable achievements in medicine I got to witness is the advances made in AL amyloidosis. The diagnosis that once struck fear and trepidation into patients and clinicians is now viewed more as a worthy foe. In just 1 decade, the median overall survival (OS) went from being measured in months to being measured in years (6.3 years for all patients and 13.2 years in complete responders).2 New therapies provided the ammunitions, but it was the biomarkers that directed the attack. The serum-free light chain assay improved diagnostics, provided a target for therapy, and prognosticated.3 Cardiac biomarkers (cardiac troponin T [cTnT] and N-terminal pro-brain natriuretic peptide [NTproBNP]) assess risk and guide initial therapy. This is most pertinent in autologous stem cell transplant (ASCT), where the risk of treatment-related mortality (TRM) is many times higher in AL amyloidosis than in multiple myeloma. The Mayo Cardiac Staging System, which uses cTnT and NTproBNP, effectively reduced TRM from 10.5% to 1.1% over a 15-year period at 1 center.4,5 Patients with stage III (elevation of both cTnT and NTproBNP) disease have the highest risk of TRM and should be offered alternative options.

Despite these advances, treatment of the patients with the most advanced disease remains undefined. This was illustrated in a study of IV melphalan and dexamethasone (MDex) where 46% of patients were classified as Mayo stage III and 59% as New York Heart Association (NYHA) stage III or higher. The study had a TRM of 28% and a median OS of 18 months, which were far inferior to the results of the original MDex study, which reported a median OS of 5.1 years and TRM of 4.3% with 70% of patients with cardiac involvement, and the randomized trial comparing oral MDex to ASCT (median OS, 4.8 years; TRM, 14%), where 46% of the MDex-treated patients had cardiac involvement and 14% were at NYHA stage III or higher.6-8 The same treatment was producing vastly different results, which first raised the question whether MDex is the right treatment of the highest-risk patients and second, what is the best way to identify them? Bortezomib-based therapy was supposed to be the answer and it appeared so when a study of cyclophosphamide, bortezomib, and dexamethasone showed a 2-year OS of 97.7% in a cohort with 46% Mayo stage III patients.9 Unfortunately, an up-front study with a comparable cohort with 49% Mayo stage III patients found the OS was 55% at 3 years.10 A subgroup analysis of the latter study found Mayo stage IIIb patients (NTproBNP >8500 ng/L) had the worst prognosis with a 19% 3-year median OS.10 TRM was 12.6% and was only reported in patients with Mayo stage III (18% of IIIa and 42.8% of IIIb). Interestingly, the prognosis of the IIIb patients was similar to that of patients with stage II and IIIa disease if they survived the first 12 months. Clearly, not even all IIIb patients are the same.

In the current study, the authors explored the utility of a disintegrin and metalloproteinase with thrombospondin type-1 repeats 13 (ADAMTS-13) and VWF:Ag as markers of disease severity. They found elevated VWF:Ag levels in 76% of their patients. Although not associated with any organ involvement, elevated levels were associated with the presence of hypotension, an independent poor prognostic risk factor. More importantly, VWF:Ag was independently associated with shorter OS (P = .011). The median OS of patients in the highest, middle two, and lowest quartiles was 14, 28, 29, and 85 months, respectively (P = .002). Furthermore, a VWF:Ag level >230 U/dL was associated with early death. This effect was independent of Mayo stage IIIb. Stage IIIb patients with VWF:Ag levels >230 U/dL had a median OS of 2 months vs 6 months in those with lower levels (see figure; P = .006). On the other hand, no difference in ADAMTS-13 Ag levels was found among healthy controls and it had no independent prognostic effects despite an inverse relationship with NTproBNP.

Other biomarkers have been described with prognostic significance in AL amyloidosis but what sets VWF:Ag apart is its potential to detect the ultrahigh-risk patients that even Mayo stage IIIb cannot clearly define. These are the patients we have yet to learn how to treat. It remains one of the major challenges we have to overcome in this disease. Having the ability to identify them is the crucial first step to developing better treatment strategies specifically for their needs.


  • Conflict-of-interest disclosure: The author declares no competing financial interests.