Predictors of survival in patients with systemic light-chain amyloidosis and cardiac involvement initially ineligible for stem cell transplantation and treated with oral melphalan and dexamethasone


Raymond L. Comenzo, Howard 802, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021, USA. E-mail:


The treatment of systemic light-chain (AL) amyloidosis with symptomatic cardiac involvement at diagnosis remains a challenge. We report the results of 40 consecutive newly diagnosed AL cardiac patients who were not candidates for stem cell transplant and therefore received monthly oral melphalan and dexamethasone. Median survival was 10·5 months and baseline predictors of survival included gender, troponin I and interventricular septal thickness. The most significant predictor of survival was response to therapy. The haematological response rate was 58% (23/40) with 13% (5/40) complete responses; most responses were noted in <3 cycles. Achievement of a rapid response to therapy extends survival.

Systemic light-chain (AL) amyloidosis is an uncommon plasma cell dyscrasia in which fibrillar deposits of clonal immunoglobulin light chains cause progressive dysfunction of the heart, kidneys and other organs. (Falk et al, 1997; Pepys, 2006) Improvements in survival in AL have depended on the elimination of clonal plasma cells by chemotherapy leading to a reduction in the production of clonal light chains. The achievement of a complete haematological response is associated with organ responses and extended survival.(Merlini & Stone, 2006) Treatment options for patients with AL are in flux as the novel agents developed for myeloma are being tested in AL. Clinical studies utilizing lenalidomide and bortezomib are on-going and reports of activity in AL are encouraging. (Dispenzieri et al, 2007; Kastritis et al, 2007) The activity of standard therapy with oral melphalan and dexamethasone for AL patients with high-risk disease, such as older patients with multiple organ involvement or significant cardiac amyloid, has not been well described. Accordingly in this report, we describe our experience using oral melphalan and dexamethasone as standard therapy for AL patients with cardiac involvement who were not candidates for stem cell transplantation (SCT). These data demonstrated that gender, troponin I, and interventricular septal thickness were baseline predictors of survival, and that subsequent response to treatment was also a significant predictor of survival. These results highlight that response to therapy translates into a limited improvement in survival for high-risk AL patients, and also provide a point of reference for the design of future clinical trials with novel agents. Achievement of a timely response to therapy is the initial goal and extends survival.

Patients and methods


Between December 2003 and February 2007, patients seen in our center with untreated AL with cardiac involvement who were not candidates for SCT received oral melphalan and dexamethasone as standard therapy. All had a histological diagnosis of amyloidosis, clonal plasma cell dyscrasia, and symptomatic organ involvement. Patients did not receive SCT because they were more than 60 years old, or had ≥3 organ systems involved, or had advanced cardiac amyloid as previously defined. (Comenzo et al, 1999) Hereditary amyloidosis was excluded by genetic testing.


Oral melphalan was given at 0·22 mg/kg and dexamethasone at 20 mg/m2 for four consecutive days of a 28-d cycle. For patients with creatinine > 221 μmol/l, melphalan was given at 0·18 mg/kg. For patients over 70 years of age, doses were reduced by 50%. Prophylactic medications included ciprofloxacin and fluconazole, each given for the first 10 d of the cycle at doses adjusted for renal function, and a daily proton-pump inhibitor. Patients were seen monthly prior to each cycle. Those who exhibited clinical improvement were re-evaluated for SCT.

Response criteria

Haematological responses were scored as previously described but also took into account the free light chain (FLC) level. (Cohen et al, 2007) The response was described as partial if the pathological light chain level was reduced by >50% from baseline with appropriate change in the ratio, and as complete if the FLC values and ratio normalized. Evaluations in patients with complete haematological responses included repeat bone marrow studies at the end of therapy.

Statistical analysis

We analyzed survival as of 1 November 2007 as a function of the baseline characteristics age, gender, number of organs involved, New York Heart Association (NYHA) class, brain natriuretic peptide (BNP), troponin I, interventricular septal thickness, left-ventricular ejection fraction, and FLC levels. Survival was also assessed as a function of response to therapy. All analyses were performed with PRISM (GraphPad, San Diego, CA). Survival curves were plotted with the method of Kaplan and Meier using the logrank test for comparisons. Analyses were two-tailed using < 0·05 significance level.

Results and discussion

Forty consecutive newly diagnosed patients received oral melphalan and dexamethasone during this 38-month period (Table I). The median ages of men and women were 64 (48–78) and 73 (51–82) years, respectively (= 0·04, Mann-Whitney). All had evidence of cardiac involvement, determined by history, examination and non-invasive testing, and were considered too high risk for SCT. Eighty-five percent of patients were NYHA class II and higher. Twenty-percent (8/40, five men and three women) had endomyocardial biopsies demonstrating amyloidosis. The median interventricular septal thickness was 14 mm and in 39% of patients was greater than 15 mm.

Table I.   Patient characteristics.
Baseline characteristic (n = 45)Value
  1. NYHA, New York Heart Association; GI, gastrointestinal; FLC, serum free light chains; LVEF, left ventricular ejection fraction; BNP, brain natriuretic peptide.

Age, years, median (range)67 (48–82)
No. male/female26/14
Time from diagnosis to enrollment, months, median (range)1·5 (1–12)
NYHA Class ≥ II, no. patients (%)34 (85)
Monoclonal light chain, no. patients (%)
 λ33 (83)
 Κ7 (17)
Marrow plasma cells, %, median (range)10 (1–53)
No. organs involved, %
 18 (20)
 218 (45)
 39 (22)
 45 (13)
Organ involvement, no. patients (%)
 Heart40 (100)
 Kidney25 (63)
 Liver/GI 11 (28)
 Peripheral nervous system 15 (38)
Albumin, g/l, median (range) 35 (14–48)
Alkaline phosphatase, IU/l, median (range)121 (51–2656)
Creatinine, μmol/l, median (range)123·76 (70·72–344·76)
24-h proteinuria, g/24 h, median (range)1·6 (<0·15–13)
Abnormal serum FLC κ:λ ratio, no. patients (%) 38 (95)
Interventricular septal thickness, mm, median (range)14 (11–26)
LVEF by echo, %, median (range) 55 (19–79)
Serum BNP, ng/l, median (range) 1000 (100–4510)
Serum troponin I, μg/l, median (range) 0·12 (0–1·19)

The median number of cycles of oral melphalan and dexamethasone was 4 (range, 1–16). Based on intention-to-treat, the haematological response rate was 58% (23/40) including 13% (5/40) complete responses. The median number of cycles to achieve response was 2 (range, 1–6) and 70% (16/23) of responses occurred in 3 months or less. After two to five cycles of therapy, four patients a median of 59·5 years old (range, 54–64 years) showed clinical improvement (three had haematological responses) and had stem cells mobilized and collected. Three had reductions in BNP of 47%, 56% and 81% from baseline and improvement to NYHA class IIA, and one had the 24-h urine protein drop by greater than 50% to less than 500 mg. Three were mobilized with granulocyte-colony stimulating factor (G-CSF) alone, and one with melphalan and G-CSF after not mobilizing with G-CSF. Two underwent stem cell transplant, while one received two more cycles of oral therapy, achieved a complete response and entered observation, and one died of disease progression.

Seventy-six percent of the patients who died (22/29) succumbed to sudden death or disease progression. There was no statistically significant difference in baseline characteristics between the nine patients who died within 3 months of starting therapy and others. Those who died 3 to 12 months after starting therapy received a median of four cycles (range, 1–6) and seven had haematological responses. Of those who died more than 12 months after starting therapy, four had haematological responses and two received additional therapy with thalidomide. One of them died of treatment-related acute myeloid leukemia. Distribution and causes of death and overall survival are shown in Fig 1A and B. Median survival was 10·5 months from diagnosis.

Figure 1.

 Distribution of deaths and predictors of survival. In panel (A) the distribution and causes of deaths are noted and in (B) the overall survival of these 40 patients is shown; median survival is 10·5 months after diagnosis. In (C) significantly longer survival is noted for women because women had a higher response rate (71% vs. 50%) and because women responders lived longer (median 37 vs. 8 months). In (D) patients with troponin I levels less than 0·12 had superior survival as did those with interventricular septal thickness less than 14mm as shown in (E). There is no difference in survival between those greater than 70 years of age and their younger counterparts (F). In (G) the significant difference in survival for responders is shown. Responders lived a median of 22 months and non-responders 9 months from diagnosis (= 0·005). The curves with and without the two patients who underwent stem-cell transplant, only one of whom had a haematological response prior to SCT, are shown. PR, partial response; CR, complete response; NR, no response.

Baseline characteristics associated with differences in survival were gender, troponin I and interventricular septal thickness (Fig 1C–E). Troponin I and septal thickness have been previously identified as prognostic markers. (Dispenzieri et al, 2004) Differences in other baseline characteristics did not predict survival. There was no difference in survival between patients aged 70 years and older who received dose reductions and younger patients who did not (Fig 1F). Subsequent response to therapy was the most significant predictor of survival, with responders living a median of 22 and non-responders 9 months (Fig 1G). When response was taken into account, women fared better than men as the result of a higher response rate [71% (10/14) vs. 50% (13/26), = 0·19 by χ2] and because women responders lived longer (median 37 vs. 8 months, = 0·10).

In addition to oral melphalan and dexamethasone, therapies for AL include high-dose melphalan with autologous SCT combined with adjuvant thalidomide and dexamethasone; oral thalidomide, cyclophosphamide and dexamethasone; and the new agents bortezomib and lenalidomide. (Cohen et al, 2007; Dispenzieri et al, 2007; Kastritis et al, 2007; Wechalekar et al, 2007) Although this report highlights the limited benefit provided by oral melphalan and dexamethasone to AL patients with advanced cardiac involvement, it is also important to note that in a randomized phase II trial the survival of newly diagnosed patients with symptomatic cardiac involvement treated with SCT without adjuvant therapy was a median of 10 months (Sanchorawala et al, 2004), comparable to the survival we report herein. In contrast, we recently reported a 63% two-year survival in cardiac patients treated with risk-adapted melphalan SCT and adjuvant thalidomide and dexamethasone, a result that supports combination approaches early in the course of disease (Cohen et al, 2007).

In this series, patients with less infiltrated left-ventricular septa and lower troponin I levels at baseline had better survival. The most significant predictor of survival was response to therapy and, among those achieving complete response, reductions in BNP were seen (median 499 ng/ml at baseline vs. 180 ng/ml at 12 months, = 0·06, Wilcoxon matched pairs). Our findings highlight the need for continued pursuit of improvements in treatment, supportive care, anti-coagulation and anti-arrhythmic management, and use of cardiac transplant followed by SCT. (Maurer et al, 2007) Moreover, the safe effective application of the new agents bortezomib and lenalidomide will require identification of the optimal dose-schedule and supportive measures to limit toxicity and interruptions of therapy. Finally, our data highlight the importance of developing novel approaches, such as CD32B monoclonal antibody therapy, for these patients whose survival depends on a prompt hematological response to treatment (Zhou et al, 2008).


This work was supported by NIH grant CA05826, the Amyloidosis Research Fund, the Amyloidosis Foundation, the Werner and Elaine Dannheiser Fund for Research on the Biology of Aging of the Lymphoma Foundation, and the Demarest Lloyd, Jr. Foundation.