- © 2001 by American Society of Clinical Oncology
Eligibility for Hematopoietic Stem-Cell Transplantation for Primary Systemic Amyloidosis Is a Favorable Prognostic Factor for Survival
- Angela Dispenzieri,
- Martha Q. Lacy,
- Robert A. Kyle,
- Terry M. Therneau,
- Dirk R. Larson,
- S. Vincent Rajkumar,
- Rafael Fonseca,
- Philip R. Greipp,
- Thomas E. Witzig,
- John A. Lust and
- Morie A. Gertz
- From the Division of Hematology and Internal Medicine, and Section of Biostatistics, Mayo Clinic, Rochester, MN.
- Address reprint requests to Angela Dispenzieri, MD, Mayo Clinic, 200 First St SW, Rochester, MN 55905; email: dispenzieri.angela{at}mayo.edu
Abstract
PURPOSE: Based on the success of hematopoietic stem-cell transplantation (HSCT) for multiple myeloma, HSCT is being used to treat patients with primary systemic amyloidosis (AL). This article addresses the extent to which eligibility to undergo HSCT is a favorable prognostic feature and explores prognostic factors within the subset of eligible patients.
PATIENTS AND METHODS: The Mayo Clinic amyloid database was queried for all patients with AL seen at the Mayo Clinic from 1983 through 1997 who would have been eligible for peripheral-blood stem-cell transplantation. Inclusion criteria included biopsy-proven amyloid, symptomatic disease, absence of a clinical diagnosis of multiple myeloma, age ≤ 70 years, cardiac interventricular septal thickness ≤ 15 mm, cardiac ejection fraction more than 55%, serum creatinine ≤ 2 mg/dL, and direct bilirubin ≤ 2.0 mg/dL.
RESULTS: Median age was 56 years (range, 25 to 70) with 79 (34%) older than 60 years. One hundred patients had early cardiac involvement; 41, hepatic involvement; 167, renal involvement; and 39, nerve involvement. The 229 patients have had a median follow-up of 52 months, and 151 have died. The median survival was 42 months with 5- and 10-year survival rates of 36% and 15%, respectively. Important predictors of survival were size of M-component in 24-hour urine, number of involved organs, alkaline phosphatase, performance score, and weight loss.
CONCLUSION: The same patients who are eligible for HSCT are a good-risk population who do relatively well with chemotherapy (median survival, 42 months), substantially better than the expected median survival of 18 months for all patients with AL. A randomized trial is needed to assess the true effect of HSCT.
PRIMARY SYSTEMIC amyloidosis (AL) is a rare multisystem disease with an incidence of eight patients per million people per year.1 Standard therapy with melphalan and prednisone results in a median survival of 12 to 18 months for all patients, and a median survival of only 4 to 6 months for those with overt congestive heart failure.2,3 Colchicine and interferon have been tried without significant benefit.4-7 The initially promising results of dexamethasone have yet to be verified.8 To date, there have been conflicting results on the overall usefulness of 4′-iodo-4′-deoxydoxorubicin.9,10 Based on the success of hematopoietic stem-cell transplantation (HSCT) for multiple myeloma,11 many institutions have begun applying this therapeutic modality to patients with AL. The efficacy of HSCT for AL amyloidosis is currently under study. Encouraging hematologic and organ response rates as high as 62% and 44%, respectively, have been reported. Unfortunately, procedure-related mortality rates of 15% to 43% have been reported as well.12-18
The concern for selection bias arises in the case of HSCT for AL, as it does in any other nonrandomized study.19-21 The need for a prospective randomized study to address the role of HSCT in the treatment of AL is especially important given the procedure-related mortality. Such a study, however, may not be feasible. Information regarding the prognosis of these relatively healthier amyloid patients should play a role in the decision-making process for these patients and their physicians. Though the majority of patients with AL amyloid die within a few years, 5-year and 10-year survivors are not rare, which makes the decision of whether to risk HSCT even more difficult.22 Previously recognized adverse prognostic factors for survival have included presence of congestive heart failure, elevated beta 2-microglobulin, elevated plasma cell labeling index, coexistent multiple myeloma, and hepatic involvement.3,23-25 Response to alkylator therapy predicts longer survival.26 This article addresses whether meeting the functional criteria to undergo HSCT is also a favorable independent prognostic feature.
PATIENTS AND METHODS
The Mayo Clinic amyloid database was queried for all patients with the diagnosis of AL from 1983 through 1997. From the resulting 1,288 patients, those who met the Mayo Clinic eligibility criteria for HSCT but had not received a transplant were selected for our study group (N = 279). The diagnosis of AL amyloid was made on the basis of at least one positive Congo red stain, proof of a clonal plasmaproliferative disorder by immunofixation of either urine or serum, or demonstration of restricted immunoglobulin light chain staining by immunofluorescence or immunohistochemistry in bone marrow plasma cells or in the amyloid itself. Patients with secondary, familial, senile, or localized amyloidosis AL or with overt symptomatic multiple myeloma were excluded.
Inclusion criteria included biopsy-proven AL, symptomatic disease, age ≤ 70 years, cardiac interventricular septal thickness ≤ 15 mm, cardiac ejection fraction of more than 55%, serum creatinine ≤ 2 mg/dL, and direct bilirubin ≤ 2.0 mg/dL. Additionally, patients were required to have an amyloid-defined syndrome. Organ system involvement was defined as follows: (1) cardiac, interventricular septum more than 12 mm and/or infiltrative cardiomyopathy and/or diastolic dysfunction determined by echocardiogram27; (2) liver, alkaline phosphatase more than 1.5 times normal levels and hepatomegaly (> 4 cm below right costal margin); (3) renal, total urine protein more than 0.5 g per 24 hours; (4) nerve, peripheral neuropathy (other than carpal tunnel syndrome) or autonomic neuropathy.
Fifty of these patients were excluded: 10 had concurrent multiple myeloma by clinical parameters (bone marrow plasma cells > 30% and lytic bone lesions and/or hemoglobin < 10 g/dL); in 11, AL amyloid was not clearly documented either by demonstration of a clonal plasmaproliferative disorder or by immunohistochemical studies of the amyloid; 12 were treated with HSCT during their course; and 17 did not have an amyloid-defined syndrome. Fatigue, weight loss, or involvement of skin, tongue, soft tissue, joints, or muscle alone was not considered sufficient for inclusion.
Patients received various therapies, including melphalan/prednisone, melphalan/prednisone/colchicine, vincristine/carmustine/melphalan/cyclophosphamide/prednisone, dexamethasone, and cyclophosphamide. This retrospective study was approved by the Mayo Foundation Institutional Review Board according to the Declaration of Helsinki.
Overall survival was estimated using Kaplan-Meier analysis. Survival curves are presented in terms of time since diagnosis and are shown with 95% confidence intervals (CIs). Because many patients were diagnosed before their evaluation at the Mayo Clinic, the analyses were adjusted to account for left truncation. Survival rates for individual risk factors consisted of discrete variables and were estimated using Kaplan-Meier analysis. The survival impact of discrete variables was assessed using log-rank tests. Continuous variables were evaluated using Cox proportional hazards models. In addition, a multivariate Cox model was created using stepwise selection. The threshold of statistical significance was set at alpha = 0.05. All analysis was carried out using SAS version 6.12 (SAS Institute, Cary, NC) and S-Plus version 3.4 (Insightful Corp, Seattle, WA).
RESULTS
Patient Characteristics
Table 1 lists the presenting features of the 229 patients who were eligible for HSCT based on the criteria detailed in Patients and Methods. Median age was 56 years (range, 25 to 70 years), with 75 patients (34%) older than 60 years. There were 96 females and 133 males. According to the study design, no patient had severe cardiac or liver involvement. Renal involvement, which was limited to patients with creatinine levels less than 2.0 g/dL, was the most common feature at presentation. One hundred sixty-seven patients (73%) had renal involvement, 128 of whom had nephrotic-range (> 3 g/d) proteinuria. One hundred patients (44%) had early cardiac involvement; 41 (18%), hepatic involvement; 39 (17%), nerve involvement; nine, small bowel involvement; and three, pulmonary involvement. For the purpose of comparison with other published series, the systems included in the organ count are renal, cardiac, hepatic, and nervous. Patients with small bowel and/or pulmonary involvement only were included only in the one-organ-involved count so as not to omit them from the analysis. The median number of involved organ systems was one, with a range of one to four.
Clinical Features of 229 Patients With Primary Systemic Amyloidosis
Laboratory features of patients are summarized in Table 2. Using immunoelectrophoresis and/or immunofixation, we found that 146 patients (64%) had a detectable monoclonal protein in their serum; 166 (72%), in their urine; and 201 (88%), in either. Of the 28 patients without a detectable serum or urine monoclonal protein, bone marrow clonality was demonstrated in 14; the remaining 14 had kappa or lambda clonality demonstrated by immunohistochemical staining of amyloid deposits. In total, the light chain was kappa in 59 and lambda in 171. Thirteen patients had ≥ 20% plasma cells in their bone marrow. The symptom complexes of these patients, however, were consistent with AL rather than multiple myeloma in that none of these patients had lytic bone lesions or elevated calcium or creatinine levels. The average number of organ systems involved by amyloid in this group was 1.6 (range, one to four systems).
Laboratory Features
Survival
Of the 229 patients in the study group, 151 have died, with a median follow-up for surviving patients of 52 months (range, 0.2 to 186 months). Follow-up was available for 96% of patients. The median survival for the entire study group was 42 months (95% CI, 43 to 57 months) and the 2-, 5-, and 10-year survival rates were 61% (95% CI, 54% to 68%), 36% (95% CI, 30% to 43%), and 15% (95% CI, 9% to 24%), respectively (Fig 1). Factors having an adverse impact on survival on univariate analysis included 24-hour urinary monoclonal protein greater than 1 g (Fig 2), increasing number of organ systems involved (Fig 3), inferior performance status (Fig 4), more than 10 pounds of weight loss, male sex, elevated alkaline phosphatase, elevated platelet count, nerve involvement, and less than 30 days from time of diagnosis to presentation at Mayo Clinic (Table 3). There was a trend toward inferior survival in patients with increased beta 2–microglobulin as a continuous variable, though levels were available for only 142 patients. Serum creatinine, bilirubin, serum albumin, kappa or lambda clonality, C-reactive protein, size or presence of serum monoclonal protein, quantity of urinary albumin, cardiac involvement, and plasma cell labeling index were not predictive of survival.
Fig 1. Overall survival from diagnosis. Two-year survival, 61% (95% CI, 54% to 68%); 5-year, 36% (95% CI, 30% to 43%); 10-year, 15% (95% CI, 9% to 24%).
Fig 2. Survival by urinary monoclonal protein. —, urine M-spike ≤ 1 g/d (median survival, 46 months); – –, more than 1 g/d (median survival, 21 months).
Fig 3. Survival by number of organs involved. —, one organ (median survival, 46 months); – –, two organs (median survival, 44 months); ---, ≥ three organs (median survival, 8 months).
Fig 4. Survival by Eastern Cooperative Oncology Group performance status. —, 0; – –, 1; ---, 2; -–, 3. Adjusted pairwise comparison: 0 versus 1, P = .02; 0 versus 2, P < .001; all other comparisons not significant.
Univariate Survival Analysis
Specific organ involvement was prognostic for survival on univariate analysis. Median survival of patients with renal, cardiac, nerve, and liver involvement was 44, 28, 16, and 14 months, respectively. Liver involvement, as defined in Patients and Methods, was not a significant marker for adverse outcome, although elevated alkaline phosphatase was. Nerve involvement imparted a survival disadvantage, which did not persist on multivariate analysis. Involvement of three organ systems was associated with worse survival than involvement of only one or two organs (8 v 43 or 46 months, respectively; P = .001) (Fig 3). The seemingly worse survival in men is explained by the association between male sex and more advanced disease and men’s trend toward worse prognostic factors at presentation: 50% of men versus 31% of women had more than one organ involved at presentation; a performance status of 0 in 37% of men as compared with 47% of women; a higher percentage (32% v 20%) diagnosed at Mayo Clinic; 47% of men versus 29% of women with 10 pounds or more of weight loss; and a urine M-spike greater than 1 g/d in 23% of men versus 15% of women.
Multivariate analysis is demonstrated in Table 4. The three clinical parameters predictive of poor prognosis were increasing number of organs involved, worsening performance status, and at least 10 pounds of weight loss, with relative risk of dying at 2.0 (95% CI, 1.2 to 3.2), 1.3 (95% CI, 1.1 to 1.6), and 1.5 (95% CI, 1.1 to 2.1), respectively. The two laboratory parameters in the multivariate analysis predictive of poor prognosis were elevations of 24-hour urine M-protein and serum alkaline phosphatase, with relative risks of death at 2.3 (95% CI, 1.4 to 3.2) and 1.8 (95% CI, 1.2 to 2.5).
Multivariate Survival Model
As a supplement to the main analyses, we also did a formal comparison of survival in the present cohort with 39 patients who underwent HSCT for AL amyloid at the Mayo Clinic. Each was matched with two individuals chosen from the cohort of 229 patients reviewed in this article. They were matched for age (± 5 years), sex, and number of involved organs. Six-month survivals were 0.85 (95% CI, 0.74 to 0.97) and 0.83 (95% CI, 0.75 to 0.92) for the HSCT and study cohort groups, respectively. One-year survivals were 0.77 (95% CI, 0.65 to 0.91) and 0.74 (95% CI, 0.65 to 0.85), and 2-year survivals were 0.68 (95% CI, 0.53 to 0.87) and 0.60 (95% CI, 0.50 to 0.72) for the HSCT and control groups, respectively. The P values for all three analyses were more than .9. In summary, there was no significant difference observed between the two groups.
DISCUSSION
Standard therapy of AL with melphalan and prednisone produces response rates of 20% to 30%5,6,28 and median survival of 12 to 18 months.2-4 More effective treatment strategies are clearly needed. HSCT is one such strategy that has yielded encouraging results, with hematologic and organ response rates as great as 62% and 44%, respectively. Unfortunately, these initially promising results came at the expense of procedure-related mortality as great as 15% to 43%12-17 (Table 5).
Comparison With Other Studies in Response Rates and Survival
The present study demonstrates that patients with AL who are eligible for HSCT have a longer than expected median survival rate even when treated with standard chemotherapy. The 42-month median survival observed in our study population is in stark contrast to that of the average patient presenting with AL, who has a median survival of 12 to 18 months.5-7 Our findings highlight the importance of selection bias when interpreting HSCT data.19-21,30 The improved survival of patients with multiple myeloma treated with HSCT over that of those receiving standard therapy11 justifiably sets the stage for enthusiasm about HSCT as a therapeutic modality for AL. The high procedure-related mortality in patients with AL amyloid, however, underscores the distinction between these two related diseases.
Though HSCT for AL may provide both superior long-term survival and clinical responses over standard therapy, no such data are yet available. In our experience, approximately 5% of all AL patients survive for 10 years or more with standard chemotherapy,22 which makes concerns of selection bias in single-armed HSCT studies more salient. In the present study of 229 patients, the 24-month survival rate was 61% (95% CI, 54% to 68%), which compares favorably with that of the HSCT group of Comenzo et al13 of 65%. Furthermore, Moreau et al15 reported an actuarial overall survival of 59% at 48 months after HSCT, which may be superior to the present study’s 48-month survival of 43% (95% CI, 37% to 51%).
Within this highly selected cohort, which comprised less than 20% of the 1,288 amyloid patients seen at the Mayo Clinic in the 15-year study period, five characteristics were prognostic for survival on multivariate analysis: number of organs involved, performance status, elevation of the urinary M-protein, alkaline phosphatase, and weight loss (Tables 3 and 4⇑). Higher urine monoclonal protein concentrations may reflect either lower creatinine clearance or higher light chain production, whereas elevated alkaline phosphatase levels may serve as surrogate markers for early liver involvement and/or cardiac involvement with congestive hepatopathy. The apparent survival advantage of diagnosis at a medical facility other than Mayo Clinic reflects referral bias, favoring patients physically capable of traveling to a large AL treatment center and thereby excluding patients with severe cardiac disease or poor performance status.
The reported procedure-related mortality rates of 15% to 43% for patients with AL undergoing HSCT12-17 give concern, but they will likely improve as selection criteria are refined. Risk factors for HSCT-related mortality include more than two systems involved,13,15 more than 55 years of age,13 and predominant cardiac involvement.13 Interestingly, the first two factors were adverse prognostic factors for overall survival in patients in the present series. Patients with severe cardiac involvement had been excluded.
A limitation of this study is its inability to assess response rates or quality-of-life issues, which are difficult to assess accurately in the context of a retrospective study. Though quality of life and response are paramount, survival is the only reliable outcome within the confines of a retrospective study involving such a complex multisystemic syndrome. Another shortcoming of this study is its reliance on comparisons between it and other published series (Table 5). Invariably, there is lack of uniformity for protocol enrollment and even for defining organ involvement and disease severity. If one contrasts the present series with that of the Boston University’s group, the largest published series, using a regimen of melphalan 200 mg/m2,13 the eligibility criteria are comparable. For example, differences between patient characteristics are as follows: 34% of patients in this study were over age 60, whereas the Boston group’s oldest patient was 61; 10% of patients in the present study had a creatinine level of 1.5 g/dL or more, whereas the cutoff in the other study was 1.5 g/dL; and, finally, the present study required an ejection fraction of 55% as compared with 50%. Though sicker patients have received HSCT at a dosage of 100 to 140 mg/m2, there are not yet convincing data to suggest that the attenuated dose is any better than the standard melphalan regimen, with a 19% response rate (five patients of 27) on an intention-to-treat basis.29
In conclusion, this study demonstrates that eligibility for stem-cell transplant is a favorable prognostic factor leading to 31/2-year survival rates with standard chemotherapy comparable to those reported for HSCT. Similar analyses performed in patients with breast cancer and sarcoma have confirmed the dangers of selection bias.19-21,30 In contrast, the matched analysis performed by Bladé et al31 in 1996 suggested a possible role of selection bias in myeloma patients, but the positive randomized clinical trial of Attal et al11 verified the survival advantage of myeloma patients undergoing HSCT. At the Mayo Clinic, we advocate the use of HSCT as a treatment modality in selected patients with AL but are humbled by the potential morbidity and mortality of HSCT in these patients. It is therefore important to recognize that the same patients who are eligible for HSCT represent a good-risk population who do relatively well with chemotherapy (median survival, 42 months), substantially better than the expected median survival of 18 months with alkylating agent therapy.5 Standardization of classification of AL organ involvement, better characterization of risk factors for HSCT-related mortality, and a randomized clinical trial evaluating HSCT versus standard chemotherapy are all indicated. The option of HSCT for AL must be presented in the context of expected survival, including both risks and benefits, with and without transplantation. Until a randomized trial can be completed, this study offers the best available data estimating survival for patients with AL treated with standard chemotherapy who possess the favorable prognostic factor of HSCT eligibility.
Acknowledgments
Supported in part by grant no. CA62242 from the National Cancer Institute and the Quade Foundation.
- Received November 9, 2000.
- Accepted April 16, 2001.
