Author + information
- Received February 23, 2016
- Revision received March 25, 2016
- Accepted March 29, 2016
- Published online June 28, 2016.
- Brett W. Sperry, MDa,∗ (, )
- Asad Ikram, MDa,
- Rory Hachamovitch, MDa,
- Jason Valent, MDb,
- Michael N. Vranian, MDa,
- Dermot Phelan, MD, PhDa and
- Mazen Hanna, MDa
- aDepartment of Cardiovascular Medicine, Cleveland Clinic Foundation, Cleveland, Ohio
- bDepartment of Hematology, Cleveland Clinic Foundation, Cleveland, Ohio
- ↵∗Reprint requests and correspondence:
Dr. Brett W. Sperry, Division of Cardiovascular Disease, Cleveland Clinic Foundation 9500 Euclid Avenue. Cleveland, Ohio 44195.
Background Light-chain amyloidosis (AL) with cardiac involvement carries a poor prognosis; median untreated survival is <6 months. Three-drug therapy with bortezomib, dexamethasone, and an alkylating agent (BDex+AA) is associated with improved biomarker response rates in AL amyloidosis.
Objectives This study sought to evaluate the effect of BDex+AA as a first-line treatment strategy on mortality in patients with symptomatic heart failure from AL cardiac amyloidosis.
Methods Patients newly diagnosed with symptomatic New York Heart Association (NYHA) functional class ≥II heart failure due to AL amyloidosis were retrospectively studied. Initial treatment strategy was adjudicated and propensity score analysis was used to adjust for the nonrandomized allocation of treatments. Survival was assessed using a Cox proportional hazards model after adjusting for the propensity score for receiving treatment, age, NYHA functional class, and ejection fraction.
Results Among 106 treated patients (age 64.6 ± 11.3 years, 63% male, 76% lambda subtype), 40 received the 3-drug regimen and 66 received other regimens. Mortality was 65% overall, 48% in the BDex+AA cohort (median survival time 821 days), and 76% in patients who received other regimens (median survival time 223 days). Initial treatment with BDex+AA was associated with decreased mortality after multivariable adjustment (hazard ratio: 0.209; 95% confidence interval: 0.069 to 0.636; p = 0.006). This association remained after further adjustment for components of the Mayo Stage.
Conclusions Use of BDex+AA in the treatment of AL amyloidosis in patients presenting with symptomatic heart failure is associated with improved survival after adjusting for clinical variables.
Amyloidosis is a disease in which misfolded proteins aggregate into fibrils, deposit in the extracellular matrix, and result in organ dysfunction (1). Cardiac involvement almost exclusively develops from the deposition of either transthyretin or immunoglobulin light chains (AL), leading to a restrictive cardiomyopathy. AL amyloidosis, a disease of clonal plasma cells, can be rapidly fatal, with a median survival of 6 months without treatment if cardiac involvement is diagnosed (2). Heart failure at presentation carries the worst prognosis compared with other manifestations (3), yet heart failure symptom assessment is underrepresented in current analyses of treatment effects.
Therapeutic regimens for AL amyloidosis have evolved over time. The combination of prednisone and melphalan, an alkylating agent, was first established in 1997 and conferred improved survival compared with colchicine (3). High-dose dexamethasone was then substituted for prednisone due to a more rapid response rate (4). Bortezomib, the first therapeutic proteasome inhibitor, is very effective against plasma cells and has shown promising results in addition to background therapy in AL amyloidosis. Bortezomib, when added to melphalan and dexamethasone, provided a higher response rate, but no change in mortality, in a matched case-control study. However, there was a survival advantage in lower-risk patients (5). A similar effect on hematologic response was noted when bortezomib was compared with thalidomide in the setting of background therapy with cyclophosphamide (another alkylating agent) and dexamethasone (6). Bortezomib, when added to an alkylating agent and dexamethasone (BDex+AA), is thought to work synergistically in arresting the life cycle of clonal plasma cells (7).
These studies included a varied number of patients with cardiac involvement as defined by the Mayo stage, which is based upon the biomarkers troponin T, N-terminal of the prohormone of brain natriuretic peptide (NT-proBNP), and the free light-chain burden (8). However, few have assessed the presence or severity of heart failure symptoms. Although BDex+AA has emerged as the treatment of choice in all-comers with AL amyloidosis due to improved biomarker response rates, there are no direct mortality data confirming benefit in patients with symptomatic heart failure. We sought to evaluate the effect of BDex+AA as a first-line treatment strategy on mortality in patients presenting with symptomatic heart failure from AL cardiac amyloidosis.
Consecutive patients seen at our institution from 2004 to 2015 with confirmed AL cardiac amyloidosis were reviewed and included for analysis. All patients were newly diagnosed, and had New York Heart Association (NYHA) functional class II to IV heart failure symptoms at presentation. Patients who had received immunosuppression previously (i.e., for multiple myeloma) were excluded. Patients who did not undergo treatment were included in Kaplan-Meier analysis as a separate group, but not included in subsequent modeling.
A diagnosis of AL amyloidosis was based on clinical presentation and the assessment of serum immunoglobulin free light chains and immunofixation. Histological diagnosis was achieved with endomyocardial and/or bone marrow biopsy staining with Congo Red or Thioflavin S, and confirmed with immunohistochemistry and/or mass spectrometry. Patients who did not undergo endomyocardial biopsy were diagnosed with cardiac involvement on the basis of advanced imaging criteria in the setting of known extracardiac amyloidosis. Cardiac magnetic resonance imaging was considered positive for amyloid cardiomyopathy if there were morphological and structural abnormalities consistent with the diagnosis, for example, wall thickening of the left (LV) or right ventricle or interatrial septum and biatrial enlargement, coupled with abnormal gadolinium kinetics and diffuse subendocardial or transmural delayed gadolinium enhancement. Echocardiographic criteria were structural and functional changes consistent with amyloid infiltration: anteroseptal or posterior wall thickness >12 mm without another cause of LV hypertrophy, biatrial enlargement, low tissue Doppler velocities, and short deceleration time, as well as an apical sparing pattern of peak systolic longitudinal strain (9).
Extracardiac involvement was evaluated in 5 organs as defined by the most recent consensus statement (10), and required histological, laboratory, or symptomatic diagnosis. In the absence of a histological diagnosis, liver involvement was diagnosed if alkaline phosphatase was >1.5 times the upper limit of normal, renal involvement if 24-h urine protein was ≥0.5 g/day, nerve involvement if symmetric extremity sensorimotor peripheral neuropathy or autonomic disorder were present, and soft tissue involvement if tongue enlargement, arthropathy, claudication, skin changes, myopathy, or carpal tunnel syndrome attributed to amyloidosis were present. Patients with pre-existing organ dysfunction were not considered to have organ involvement due to amyloidosis. Lung and gastrointestinal involvement were only diagnosed histologically.
All patients underwent comprehensive evaluation including clinical, laboratory, electrocardiographic, and echocardiographic measurements. Electronic medical records were retrospectively reviewed, and data were adjudicated at the time of treatment initiation. Free light-chain burden was quantified using the serum immunoglobulin-free light-chain difference (dFLC), defined as the level of the affected minus the unaffected light chains. Multiple myeloma was defined as clonal bone marrow plasma cells ≥10% or biopsy-proven bony or extramedullary plasmacytoma along with a myeloma defining event using CRAB (hypercalcemia, renal failure, anemia and bone lesions) criteria per the updated International Myeloma Working Group guidelines (11).
Transthoracic echocardiography was performed using Vivid 7 or Vivid 9 (GE Medical, Milwaukee, Wisconsin) or EPIQ (Philips Medical Systems, Bothell, Washington) ultrasound systems. Echocardiographic parameters were retrospectively reviewed off-line by a reviewer blinded to clinical information, and measurements were made in standard fashion as described by the American Society of Echocardiography guidelines (12). Anteroseptal and posterior wall thickness and LV end-diastolic diameter were measured in the parasternal long-axis view in diastole at the level of the mitral valve leaflet tips and used to calculate LV mass. Cardiac magnetic resonance imaging was performed using 1.5-T or 3-T magnetic resonance scanners (Philips Achieva, Best, the Netherlands). Endomyocardial biopsy histology was reviewed by 2 cardiac pathologists experienced in cardiac amyloidosis.
Mortality was assessed by electronic medical records. Patients with incomplete follow-up data were contacted by telephone.
Therapeutic regimens included dexamethasone at varying doses once weekly in 4-week (28-day) cycles. Bortezomib was dosed at 1.3 mg/m2 via subcutaneous or intravenous injection weekly. Cyclophosphamide was given at a dose of 300 mg/m2 orally or intravenously both once weekly for 3 consecutive weeks followed by a week off. Melphalan was administered as 9 mg/m2 daily for 4 days (reduced by 25% in patients with estimated glomerular filtration rate <30 ml/min per 1.73 m2), thalidomide daily for 4 weeks, and lenalidomide daily for 3 weeks.
Categorical variables are presented as a frequency (n) and continuous variables are expressed as mean ± SD or median (interquartile range). Categorical variables were compared using the Fisher exact test, and continuous variables, using a 2-tailed Student t test or Wilcoxon rank sum test as appropriate. Frequency of BDex+AA use was measured across quartiles of year of diagnosis. Spearman rank correlation coefficients were calculated to assess the correlation between year of diagnosis as a continuous variable and baseline characteristics.
The goal of our survival analysis was to determine the association between an initial treatment strategy of BDex+AA for AL amyloidosis and risk of mortality after adjusting for potential confounders. Our approach incorporated several steps. First, we performed multiple logistic regression analysis to develop a propensity score to adjust for nonrandomized treatment allocation. Subsequently, we performed a Cox model to assess the effect of BDex+AA treatment on mortality after adjusting for particularly important covariates that were significantly associated with mortality on univariable analysis. Finally, we performed sequential Cox models by adding components of the Mayo staging system. This overall approach was selected in order to minimize model overfitting and take into account nonrandomized treatment strategies.
The purpose of this score was to represent the predictors of receiving BDex+AA as accurately as possible. Thus, all measured factors which may have influenced this decision were included: year of diagnosis, time from diagnosis to treatment, age, sex, race, multiple myeloma, lambda subtype, hypertension, atrial fibrillation, obesity, smoking, coronary artery disease, anemia, estimated glomerular filtration rate, proteinuria, NYHA functional class, diuretic use, presence of an implantable cardioverter-defibrillator, number of systems involved with amyloidosis, serum albumin level, initial dexamethasone dose, systolic blood pressure, echocardiographic ejection fraction, mitral valve deceleration time, and LV mass. The final propensity score was incorporated into the survival analysis to adjust for the determinants of therapy.
Time to event was defined as chemotherapy initiation date to event or end of follow-up. Patients were analyzed for mortality and censored at the time of last follow-up, 3 years, or at the time of heart transplantation or left ventricular assist device (LVAD) implantation. Univariable survival analyses are graphically represented using Kaplan-Meier curves and analyzed using the log-rank test.
A Cox model was developed to assess the association between BDex+AA use and mortality after adjusting for the propensity score for receiving treatment, age, NYHA functional class III to IV symptoms at presentation, and ejection fraction. We then performed secondary analyses incorporating log-transformed biomarkers in patients with available data using separate sequential models that included troponin T, NT-proBNP, and dFLC as continuous variables. Model discrimination was tested using Harrell’s C-statistic.
Patients who had no follow-up within the past 6 months were contacted by telephone to assess vital status and interim heart transplantation or LVAD. All statistical tests were 2-sided, and p values < 0.05 were considered significant. All model assumptions were examined including linearity, collinearity, additivity, and proportional hazards. Statistical analysis was performed using Stata (version 13, StataCorp LP, College Station, Texas). The study was approved by the Cleveland Clinic Institutional Review Board.
A total of 119 patients met inclusion criteria for cardiac amyloidosis with symptomatic heart failure: 71 (61%) were diagnosed on the basis of endomyocardial biopsy and 48 (39%) on the basis of extracardiac biopsy and imaging criteria. Thirteen patients did not receive treatment for AL amyloidosis due to advanced clinical status or patient preference. Of the 106 patients who received chemotherapy, all received dexamethasone, and 98 (92%) received additional therapy. There were 63 patients (59%) who received bortezomib, 47 (44%) cyclophosphamide, 16 (15%) melphalan, and 10 (9%) thalidomide or its derivative as a component of first-line therapy. Fifteen (14%) patients underwent subsequent autologous stem cell transplantation. Of those, 8 were treated with BDex+AA as first-line therapy. The initial therapeutic regimen was BDex+AA in 40 (38%) patients. Cyclophosphamide was the initial alkylating agent in 36 cases, and melphalan in 4 cases. Seven patients included in the cohort receiving other regimens later received BDex+AA after relapse, but were included in the other regimens cohort in order to mimic an intention-to-treat analysis. The use of the BDex+AA regimen has increased over time (p < 0.001) (Figure 1A).
The baseline characteristics of patients who received treatment are seen in Table 1. The average age at diagnosis was 64.6 ± 11.3 years, and 63% of patients were male. There were 81 (76%) patients identified with lambda subtype, with 19% meeting diagnostic criteria for multiple myeloma. There were no significant differences in baseline characteristics or markers of disease severity between patients in the BDex+AA cohort and those treated with other therapies. In addition, there was no correlation between any of the measured variables and year of diagnosis.
Results of propensity score analysis found that a more recent date of diagnosis (odds ratio [OR]: 1.002; 95% confidence interval [CI]: 1.001 to 1.004; p < 0.001), anemia (OR: 0.090; 95% CI: 0.009 to 0.940; p = 0.044), presence of coronary artery disease (OR: 36.899; 95% CI: 2.684 to 507.144; p = 0.007), and diuretic use (OR: 0.015; 95% CI: 0.001 to 0.551; p = 0.023) were significantly associated with receiving BDex+AA. This propensity score performed very well, correctly classifying 91% of patients with a C-statistic of 0.972.
Of patients who received any treatment, 69 died (65%) over a median follow-up of 465 days. One patient underwent LVAD implantation, and there were 3 heart transplantations. No patient was lost to follow-up. All patients who did not receive any treatment died with a median time to death of 47 days. Nineteen (48%) patients died who received BDex+AA, with a median survival of 821 days, and 50 (76%) died who received other chemotherapeutic regimens, with a median survival of 223 days. All deaths were related to AL amyloidosis (i.e., progressive heart failure, arrhythmia, autonomic dysfunction with hypotension or failure to thrive).
Survivor curves stratified by year of diagnosis quartile are illustrated in Figure 1B. The unadjusted year of diagnosis was not associated with mortality (p = 0.713). There was a decrease in 3-year mortality with increasing propensity scores for receiving BDex+AA (p = 0.029). The Central Illustration shows a Kaplan-Meier curve for mortality stratified by type of treatment, with a significant association between the type of treatment received and mortality (p < 0.001).
Table 2 shows the unadjusted and adjusted survival analyses of BDex+AA on mortality. BDex+AA was associated with decreased mortality after adjusting for the propensity score for receiving BDex+AA, age, NYHA functional class III to IV symptoms, and ejection fraction. Harrell’s C-statistic for the multivariable model was 0.675. The association between BDex+AA and survival remained in separate sequential models that adjusted for components of the Mayo Stage: troponin T (p = 0.016), NT-proBNP (p = 0.037), and dFLC (p = 0.005) (Figure 2). In addition, we assessed the incremental value of bortezomib on triple therapy with BDex+AA. When added to the multivariable Cox model as a covariate, bortezomib as initial therapy was not associated with mortality (hazard ratio: 1.182; 95% CI: 0.548 to 2.550; p = 0.671). In sensitivity analysis, the association of BDex+AA on survival persisted when censoring the 7 patients who initially received other regimens but later received BDex+AA at the time of crossover (hazard ratio: 0.291; 95% CI: 0.111 to 0.763; p = 0.012).
This study explores the use and efficacy of chemotherapeutic agents in patients with AL cardiac amyloidosis presenting with symptomatic heart failure. Treatment with bortezomib, dexamethasone, and an alkylating agent (BDex+AA) has emerged over time as the most prevalent regimen and is associated with improved survival in this population.
Symptomatic cardiac involvement in AL amyloidosis is an extremely strong predictor of adverse events and mortality. Historically, median untreated survival was 5 to 6 months at the time of cardiac diagnosis (2,3), and the median treated survival in our cohort was about 15 months. The continued poor outcomes in patients with cardiac involvement limit the type of treatment available to patients, particularly stem cell or heart transplantation (13).
The addition of bortezomib to dexamethasone and an alkylating agent as a treatment in AL amyloidosis has previously been shown to improve clinical response rate, but not overall survival (5,6,14). Secondary analyses did show subgroups in which BDex+AA proved significantly beneficial. These were predominantly lower-risk patients: those without NYHA functional class III to IV symptoms or NT-proBNP <8,500 ng/l at the time of diagnosis and those surviving to at least 6 months (5). A recent large multicenter European retrospective case series of 230 patients has also validated the success of this regimen on biomarker response, showing a 60% overall hematologic response rate. Yet, prognosis continued to be guarded in patients with stage IIIb cardiac involvement (NT-proBNP >8,500 ng/l) with an overall response rate of 42% and median survival of 7 months (15). Indeed, NT-proBNP level was also highly predictive of mortality in our cohort. Additional case series have described the response rates of BDex+AA in all-comers with AL amyloidosis (16,17), and those at high cardiac risk (18). It is important to note that in these as well as several other studies, hematologic response predicts improved survival.
There has been some concern with bortezomib in patients with cardiac involvement. There have been post-marketing reports of fluid retention in patients treated for amyloidosis, and trials in myeloma excluded patients with congestive heart failure (19,20). Our current report strengthens the recommendation for BDex+AA as first-line therapy in AL amyloidosis and adds to the body of data for its utility in patients with cardiac involvement.
The use of BDex+AA has increased over time; however, there was no difference in mortality in the current era as compared with earlier years in the study period. Despite the evolution of therapeutics, there is significant mortality in a similar pattern irrespective of year quartile. The lack of a significant temporal decrease in mortality speaks to the disease severity of this population and the extremely poor prognosis despite novel treatments. Additionally, this may be partially related to improving diagnostic methods over time.
There is a need for clinical trials in AL amyloidosis with novel agents, given the persistently poor prognosis in the setting of cardiac involvement. Treatments with the potential to clear deposited amyloid fibrils are currently being investigated in patients with cardiac involvement (VITAL Amyloidosis Study, a Global Phase 3, Efficacy and Safety Study of NEOD001 in Patients With AL Amyloidosis [VITAL Amyloidosis]; NCT02312206) and in those with persistent symptoms after treatment (The PRONTO Study, a Global Phase 2b Study of NEOD001 in Previously Treated Subjects With Light Chain (AL) Amyloidosis [PRONTO]; NCT02632786). These will be the first randomized controlled trials enrolling only patients with cardiac involvement, yet even in these studies, patients with an NT-proBNP above a certain level are excluded.
As with most studies in this rare disease, our study was limited by its retrospective design. Propensity score analysis and multiple clinical predictors were used to adjust for the nonrandomization of treatments, yet this may not fully adjust for differences between treatment arms. This was an effectiveness study, and not designed to assess treatment efficacy. Patients received different chemotherapeutic regimens over time, reflecting the variations in clinical practice patterns in AL amyloidosis. The study is not powered to assess differences between all regimens or agents. Rather, we chose to assess the most validated and contemporary regimen in this disease as compared with other treatments. In addition, because this is a study of first-line treatment strategy, we do not account for potential changes to therapy or the effect of therapy on biomarkers. Although 7 patients initially receiving other regimens subsequently received BDex+AA after relapse, their inclusion in the other regimens arm should have reduced the survival difference between arms.
The use of bortezomib, dexamethasone, and an alkylating agent (BDex+AA) in the treatment of AL amyloidosis in patients presenting with symptomatic heart failure has increased over time. This regimen is associated with improved survival after adjusting for clinical variables. Given the persistently poor prognosis, future studies of novel agents in AL amyloidosis are needed in patients with symptomatic heart failure.
COMPETENCY IN PATIENT CARE: In patients with light-chain amyloidosis and symptomatic heart failure, treatment with bortezomib, dexamethasone, and an alkylating agent is associated with improved survival.
TRANSLATIONAL OUTLOOK: Future studies of novel chemotherapeutic agents, particularly those with fibril clearing properties, are needed in patients with symptomatic heart failure due to light-chain amyloidosis given the continued poor prognosis in this disease.
Dr. Valent is a teacher and speaker for Takeda Pharmaceuticals, Celgene, and Amgen. Dr. Hanna is a consultant to Pfizer and Prothena Pharmaceuticals. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- immunoglobulin light chain
- bortezomib, dexamethasone, and alkylating agent
- confidence interval
- free light-chain difference
- left ventricle/ventricular
- left ventricular assist device
- N-terminal pro–B-type natriuretic peptide
- New York Heart Association
- odds ratio
- Received February 23, 2016.
- Revision received March 25, 2016.
- Accepted March 29, 2016.
- American College of Cardiology Foundation
- Palladini G.,
- Perfetti V.,
- Obici L.,
- et al.
- Ma M.H.,
- Yang H.H.,
- Parker K.,
- et al.
- Kumar S.,
- Dispenzieri A.,
- Lacy M.Q.,
- et al.
- Phelan D.,
- Collier P.,
- Thavendiranathan P.,
- et al.
- Gertz M.A.,
- Comenzo R.,
- Falk R.H.,
- et al.
- Palladini G.,
- Sachchithanantham S.,
- Milani P.,
- et al.
- Venner C.P.,
- Lane T.,
- Foard D.,
- et al.
- Mikhael J.R.,
- Schuster S.R.,
- Jimenez-Zepeda V.H.,
- et al.
- Jaccard A.,
- Comenzo R.L.,
- Hari P.,
- et al.
- Falk R.H.