Allogeneic stem cell transplantation (SCT) with myeloablative conditioning is potentially curative therapy for myeloma, but is reportedly associated with a high risk of nonrecurrence mortality (NRM). Reduced-intensity conditioning (RIC) allows for the reduction of NRM, but the recurrence rate is increased. The role and timing of allogeneic SCT in the disease course remains controversial. To the authors' knowledge, there are limited data regarding the long-term outcome of RIC in the recurrent/refractory setting.
A retrospective analysis was conducted of SCT outcomes in 50 patients who received RIC for recurrent/refractory myeloma between the years 2001 and 2004. All patients were given fludarabine-melphalan based conditioning and stem cell grafts from a related (n = 27) or unrelated donor (n = 23).
The median age was 53 years. Forty-seven patients failed a prior autologous SCT. Thirty patients were in disease remission at the time of SCT and 20 had stable or progressive disease. With a median follow-up of 6.4 years (range, 5-7.9 years), the overall and progression-free survival (PFS) rates were 34% and 26%, respectively. The NRM rate was 26%. Adverse prognostic factors for survival included SCT not in remission, long duration of disease (>5 years from diagnosis), and transplantation from a female donor to a male recipient. The 7-year PFS in 19 patients with none of these adverse prognostic factors was 47%. Chronic graft versus host disease and the achievement of complete remission after SCT were associated with improved outcome.
Allogeneic hematopoietic stem cell transplantation (SCT) is a potentially curative approach in patients with multiple myeloma. However, the role and timing of allogeneic SCT in the disease course are still controversial.1-3 With modern therapy, which includes novel agents (thalidomide, bortezomib, and/or lenalidomide) for induction and maintenance combined with a single or tandem autologous SCT, the survival of myeloma patients markedly improved, and survival beyond 10 years has become more common.4-7 Allogeneic SCT with myeloablative conditioning has been reported to result in long-term progression-free survival (PFS), with a plateau in survival curves suggesting possible cure.8-13 However, it was associated with excessively high rates of nonrecurrence mortality (NRM), reaching 30% to 50%, leading to a nearly complete abandonment of this approach. The introduction of reduced-intensity conditioning (RIC) regimens, which are associated with an improved safety profile, renewed the interest in allogeneic SCT for myeloma.14-16 Cytoreductive autologous SCT followed by RIC allogeneic SCT (tandem auto-allo SCT) has been tested in a few studies, with a relatively long follow-up in newly diagnosed patients in comparison with a tandem autologous SCT.17-22 However, results are conflicting, and autologous SCT is still considered the standard of care for newly diagnosed patients. The vast majority of these patients are destined to develop disease recurrence after SCT. Although patients who develop disease recurrence after autologous SCT can still achieve durable responses to novel agents, prognosis is unfavorable, with a median survival of only 13 to 32 months.6, 23 Allogeneic SCT has been used in this setting; however, to the best of our knowledge, there are only sparse data regarding the long-term outcome because most studies have a limited follow-up of up to 3 years.14-16, 24, 25 Long-term data are essential to evaluate the outcome of RIC SCT as a plateau in survival curves may only become apparent years after SCT. In this study, we evaluated the outcome of 50 patients with recurrent/refractory myeloma after RIC SCT. We demonstrated that, with a median follow-up of 6.4 years, 26% remain progression-free. Thus, long-term remission after allogeneic SCT in the salvage setting is possible in a subset of myeloma patients.
MATERIALS AND METHODS
We retrospectively analyzed the long-term results of 50 patients with refractory or recurrent multiple myeloma given allogeneic SCT between January 2001 and June 2004 in the 2 participating centers. This time period was selected to allow a minimum of 5 years of follow-up for surviving SCT recipients. Patients were eligible for this analysis only if they failed a single or multiple prior autologous transplantations or if they failed prior lines of chemotherapy and autologous transplantation was not feasible due to failed stem cell collection. Patients treated in a planned tandem autologous-allogeneic SCT were not included in this analysis. All patients were required to meet standard institutional eligibility criteria for RIC, including a Karnofsky performance status of 0 to 2, no active opportunistic infection, creatinine <2 mg/dL, bilirubin <3 mg/dL, hepatic enzymes <3 times of the upper limit of normal, cardiac ejection fraction >40%, and a diffusion capacity of lung for carbon dioxide (DLCO) >40% of predicted. There was no age limitation. Patients had to have a human leukocyte antigen (HLA)-compatible related or unrelated donor willing to donate granulocyte–colony-stimulating factor (G-CSF)-mobilized peripheral blood stem cells (the preferred requested option) or bone marrow. HLA compatibility was based on high-resolution DNA typing of class I and class II antigens. One antigen mismatch was allowed. All patients were treated on institution review board-approved protocols and provided written informed consent. Some of the patients in this analysis were included in an earlier report.14
All patients were scheduled for RIC, and myeloablative conditioning was not considered for myeloma patients during the study period. A planned autologous-allogeneic SCT was not used as a salvage approach in any of the patients. Patients were given a conditioning regimen comprised of fludarabine at a total dose of 90 to 150 mg/m2 and melphalan at a total dose of 100 to 150 mg/m2. Patients with an unrelated or mismatched donor and some of the patients with a sibling donor were also given antithymocyte globulin (ATG; total dose 15-60 mg/kg). The actual doses given were determined by the attending physician according to patient status and prior therapy. Prophylaxis against graft versus host disease (GVHD) was comprised of cyclosporine A and a short course of methotrexate (15 mg/m2 on Day 1 and 10 mg/m2 on Days 3 and 6). GVHD prophylaxis was administered for 3 months and tapered afterward in patients with no active GVHD. G-CSF was administered routinely from Day +7 until engraftment. A standard regimen of antibiotic prophylaxis was used to prevent bacterial, viral, fungal, and Pneumocystis carinii infections. Routine maintenance therapy was not given after the transplant but was allowed at the discretion of the attending physician. Patients were eligible for treatment with novel agents and/or donor lymphocyte infusion if not achieving complete remission (CR) or when progressing after SCT.
Evaluation of Response
Neutrophil and platelet engraftment were defined as the first of 3 days with an absolute neutrophil count (ANC) >0.5 × 109/L and the first of 7 days with an untransfused platelet count >20 × 109/L, respectively. Acute and chronic GVHD were graded and staged by standard criteria. Disease response was assessed on Day +100 according to European Group for Blood and Marrow Transplantation (EBMT)/International Bone Marrow Transplant Registry criteria.26 Briefly, CR required the disappearance of monoclonal gammopathy in serum and urine as determined by immune-fixation analysis and <5% plasma cells in a bone marrow aspirate. A partial response (PR) was defined when there was a >50% reduction, but not fulfilling criteria for CR. Recurrence from CR was defined as recurrence of the monoclonal globulin or bone marrow plasmacytosis, and disease progression was defined when there was an increase of at least 25% in paraprotein or development of new bone lesions. Serum light chain analysis was not routinely available at the time of study.
Overall survival (OS) was calculated from the day of SCT until death or last follow-up. PFS was calculated from the day of SCT until disease recurrence or disease progression, death from any cause, or last follow-up. The probabilities of OS and PFS was estimated using the Kaplan-Meier method.27 Disease recurrence and NRM rates were estimated using cumulative incidence analysis and considered as competing risks.28 In the analysis of the cumulative incidence of GVHD, disease recurrence was considered a competing risk. Landmark analysis was performed for chronic GVHD and disease response whereby only patients surviving free of disease progression for >100 days after SCT were included. The effect of various patient and disease categorical variables on survival probabilities was studied with the log-rank test. A Cox proportional hazards model was used to determine the significance of multiple pretransplantation variables in determining these outcomes. Variables found to be significant at the level of P ≤ 0.2 were entered into the model. A time-dependent model was fitted to include post-transplantation factors, such as GVHD and disease response for determining survival.
Patient and Donor Characteristics
The study group included 50 patients with multiple myeloma treated with allogeneic SCT at 2 institutions (Chaim Sheba Medical Center, Tel-Hashomer, Israel; and University Hospital Hamburg, Germany) between January 2001 and June 2004. The 2 institutions shared similar eligibility and treatment protocols. Patient and donor characteristics are outlined in Table 1. The median age was 53 years (range, 32-64 years). This was a relatively heavily pretreated patient group, with a median of 3 years from diagnosis (range, 6 months-14 years). All patients had developed recurrent or refractory myeloma. Forty-seven patients developed disease recurrence after 1 (n = 31) or 2 (n = 16) prior autologous transplants, whereas 3 patients did not have a prior autologous transplant due to failure to collect autologous stem cell grafts, but developed disease recurrence after conventional chemotherapy. The median time from the last autologous SCT was 2 years (range, 3 months-6.5 years). The total number of lines of therapy was ≥3 in 17 patients (34%). The donors were HLA-matched siblings (n = 27) or matched unrelated (n = 23). Among the unrelated donors, 18 were a 10/10 HLA match, whereas 5 had 1 antigen mismatch (Ag DQ, 2 patients; Ag C, 3 patients). At the time of allogeneic SCT, 30 patients had responded to prior therapy, whereas 20 patients did not achieve at least a PR. All grafts were of peripheral blood stem cells.
Forty-seven patients engrafted. The median time to an ANC <0.5 × 109/L was 14 days (range, 10-27 days), and the median time for a platelet count >20 × 109/L was 14 days (range, 9-40 days) from the day of SCT. Three patients died before engraftment from multiorgan toxicity (n = 2) or infection (n = 1). Twenty-five patients had acute GVHD (grade 1 in 7 patients, grade 2 in 11 patients, grade 3 in 4 patients, and grade 4 in , 3 patients). The cumulative incidence rates of acute GVHD grade 2 to 4 and grade 3 to 4 were 51% (95% confidence interval [95% CI], 39-67%) and 19% (95% CI, 10-37%), respectively. Chronic GVHD occurred in 20 of 40 evaluable patients with an overall cumulative incidence of 63% (95% CI, 47-84%).
In all, NRM occurred in 13 patients with a 5-year cumulative incidence rate of 26% (95% CI, 17-42%). Three patients died from organ toxicities, 2 from multiorgan failure before engraftment, and 1 from thrombotic microangiopathy. Six patients died of acute GVHD and 1 of chronic GVHD with or without related infections. Three patients died of infections not related to active GVHD (1 before engraftment).
Disease Response and Recurrence
Disease response was assessed at Day 100. Forty patients were evaluable for assessment; 7 had died by Day 100, and 3 developed disease recurrence before Day 100. Twenty-three patients achieved a CR, and 17 patients achieved a PR. Seven of the patients with PR fulfilled the criteria for very good partial remission (VGPR). Patients were not given routine maintenance therapy after SCT. However, 9 patients were given thalidomide after SCT to convert PR to CR. Five of these patients achieved CR and were analyzed with those achieving CR with no maintenance because their outcome was similar. It is interesting to note that only 1 of these patients later developed disease progression. In all, 24 patients had disease progression after SCT, with a cumulative incidence of 48% (95% CI, 36-64%). The median time to disease progression in these patients was 9 months (range, 1 month-2.4 years). There were no late recurrences beyond 3 years from SCT in this series.
OS and PFS
At a median follow-up of 6.4 years (range, 5-7.9 years), 16 patients were alive and 34 had died. Thirteen patients died of treatment-related complications and 21 of disease recurrence. Three additional patients developed disease recurrence, but were alive with further therapies at the time of last follow-up. The median survival was 2.3 years, and the estimated 7-year OS was 34% (95% CI, 21-47%) (Fig. 1). The estimated 7-year PFS was 26% (95% CI, 14-38%). All 13 patients who were alive and free of disease progression had had >5 years of follow-up, and 5 of them were free of disease progression for >7 years from SCT at the time of last follow-up. The median survival after disease progression was 1 year (range, 1 month-6 years). Table 2 outlines the univariate analysis of pretransplantation factors predicting survival rates. OS and PFS were better in patients with sensitive myeloma at the time of SCT (to either chemotherapy or novel agents) (Fig. 2), in patient/donor combination other than female donor to male recipients, and in patients not having a prior autologous SCT (although only 3 patients in this group). There was a tendency toward worse PFS in patients with a long duration of disease (> 5 years before allogeneic SCT). There was a nonstatistically significant trend toward worse outcome with matched unrelated donors. We could not assess the prognostic effect of deletion 13 accurately due to missing data. When these factors were included in a Cox proportional hazards model, the independent factors found to be predictive of worse OS were refractory disease (hazard ratio [HR], 2.5; 95% CI, 1.4-4.6% [P = .003]) and SCT from a female donor to a male recipient (HR, 5.5; 95% CI, 2.5-12.5% [P = .001]). The pretransplantation factors found to be predictive of worse PFS were refractory disease (HR, 3.6; 95% CI, 1.4-4.6% [P = .001]), SCT from a female donor to a male recipient (HR, 4.1; 95% CI, 1.7-9.6% [P = .001]), and a disease duration of >5 years (HR, 2.8; 95% CI, 1.3-6.1% [P = .01]). The 7-year PFS in 19 patients with none of these adverse prognostic factors was 47% (95% CI, 25-70%).
Table 2. Analysis of Pretransplantation Factors Predicting for OS and PFS After SCT
To further analyze post-transplantation events, such as chronic GVHD and achievement of a CR, we fitted a time-dependent Cox model to patients who were alive and free of disease progression at Day 100. Table 3 outlines the multivariate analysis of factors included in the model in predicting outcome. In addition to the prior factors, chronic GVHD was associated with improved OS and PFS (HR, 0.3; 95% CI, 0.1-0.9% [P = .02]), and achievement of a CR at Day 100 was also associated with improved PFS (HR, 3.8; 95% CI, 1.5-9.9% [P = .005 for no CR after SCT]). Achievement of VGPR was associated with a longer PFS in comparison with PR, median PFS times of 1.8 years and 0.5 years, respectively (Fig. 3). However, long-term PFS was poor in both groups, and only patients with CR after transplant were able to achieve long-term remission.
Table 3. Multivariable Analysis of Pretransplantation and Post-Transplantation Factors Predicting for OS and PFS After SCTa
HR (95% CI)
HR (95% CI)
OS indicates overall survival; PFS, progression-free survival; SCT, stem cell transplantation; HR, hazard ratio; 95% CI, 95% confidence interval; SD, stable disease; PD, progressive disease; GVHD, graft versus host disease; F→M, SCT from female donor to male recipient; CR, complete remission; NS, not significant.
Pretransplantation and post-transplantation factors reaching a statistical significance of ≤.2 in the univariate analysis were included in a time-dependent Cox model.
SD/PD status at SCT
No CR after SCT
Time from diagnosis of >5 y
Novel agents and autologous transplants changed markedly the natural history of multiple myeloma; however, the disease remains an incurable malignancy with conventional therapy. Allogeneic SCT with standard myeloablative conditioning has been shown by multiple studies to be potentially curative in myeloma. Plateau in the range of 30% to 40% of patients was consistently shown in these studies, years after SCT.8-13 In the Southwest Oncology Group SWOG9321 randomized trial, patients with a donor received myeloablative SCT. The 7-year OS and PFS rates were 39% and 22%, respectively.8 However, the allogeneic arm was closed prematurely after enrolling only 36 patients due to a 1-year NRM of 53%. Similar NRM rates of 30% to 50% have been reported in other studies.9, 10 Although disease recurrence rates were shown to be lower with allogeneic SCT than with autologous SCT, these high rates of NRM resulted in OS that was better, in an earlier EBMT study, with autologous SCT.9 Later EBMT studies demonstrated improvement in NRM rates in more recent years, mostly due to better patient selection, but possibly due to improvement in supportive care, yet this has not led to wider use of this approach.11, 15
RIC regimens were developed to reduce NRM rates and to allow allogeneic SCT in elderly and medically infirm patients.29 RIC rapidly became a promising approach to investigate in myeloma, for which NRM was historically a major limiting factor. RIC has reduced the NRM to 10% to 20%.14-21, 24, 25 A retrospective large study of the EBMT compared outcome after RIC (320 patients) and myeloablative SCT (196 patients).16 RIC was found to reduce NRM (24% vs 37%), but this rate was offset by an increase in disease recurrence rates (54% vs 27%), resulting in similar OS. Similar effects of dose intensity were seen in acute leukemia and in other hematologic malignancies, especially when the malignancy is not in a good remission.30, 31 However, there is an inherent bias in studies with this methodology related to the difference in patient characteristics between patients allocated to myeloablative conditioning and RIC.
Two approaches were developed to use RIC in the course of the disease. RIC could be used in the initial treatment program for newly diagnosed patients or as a salvage approach for patients failing prior therapy, which in modern therapy most often includes single or tandem autologous transplants. The therapy for newly diagnosed patients most often uses an autologous-allo approach. The autologous SCT is used for cytoreduction. The allogeneic part is administered to provide the possibly curative graft versus myeloma (GVM) effect. The temporal separation between the 2 parts allows reduction in NRM. Three prospective studies explored this approach. The French Intergroup Francophone du Myelome (IFM) 99-03/99-04 studied high-risk patients (with high β2-microglobulin and deletion13). It showed a trend for better OS in the tandem auto over the auto-allo approach, and there were no progression-free survivors at 5 years after transplant in both arms.20 The Italian study demonstrated an advantage in survival rates for the auto-allo arm,18 whereas the Spanish PETHEMA study indicated no difference, but a trend toward a plateau in the auto-allo arm.19 Similarly, an EBMT study recently showed a trend for better PFS and a plateau in the curve.21 These conflicting results may be explained by different patient population studied, different regimens, and different GVHD prophylaxis regimen used. Collectively, they demonstrate, with a long follow-up of >5 years, an NRM of 11% to 18% and similar survival rates.22 Until there are definitive answers from larger studies, this approach can only be recommended within clinical studies.
We have used a different approach by using allogeneic SCT as a salvage approach for patients failing prior therapies, which included autologous SCT in 47 of 50 patients. This was a relatively heavily pretreated patient group with a median of 3 years from the time of diagnosis. NRM was, as may be expected, higher than in the upfront approach, reaching 26%, which is similar to other studies in the same setting.14-16, 24, 25 With a median follow-up of >6 years, there was a clear plateau in the PFS curve after 3 years, with 34% of patients alive and 26% free of disease progression, 7 years after allogeneic SCT. The EBMT study showed, in a heterogeneous group of 229 patients, similar 3-year OS and PFS rates of 41% and 21%, respectively, with no plateau.15 The advantages of allogeneic SCT may only become apparent after a long follow-up, when a plateau becomes apparent, and this study, as well as others in the salvage setting, may not have had sufficient follow-up.24, 25 To the best of our knowledge, the only other study in the salvage setting with a long follow-up was the Seattle auto-allo study published by Rotta et al.17 This study included both upfront and previously treated patients, but not patients with a prior autologous SCT. With a median follow-up of 6.3 years, the OS and PFS rates were 64% and 36%, respectively, with better results for patients treated early (<10 months from the start of initial therapy). This study demonstrated a continuous risk of disease recurrence with no apparent plateau. It is currently unknown whether the auto-allo approach offers an advantage over standard salvage therapy followed by RIC in the salvage setting. Whether an allogeneic SCT should be offered as part of a first-line treatment or as salvage therapy for recurrent/refractory disease is a matter of debate. As discussed, upfront therapy is associated with a lower NRM, but long-term outcome still requires better confirmation. Kroger et al demonstrated that the outcome after allogeneic SCT of patients who develop disease recurrence after autologous SCT is significantly worse than in other patients.14 Patient after disease recurrence may harbor highly resistant myeloma cells that may also not respond to the GVM effect. However, NRM is still significant when considering the possible long-term survival of newly diagnosed patients with modern therapy.22 If not given early, allogeneic SCT can still salvage effectively a subset of patients. The conventional therapeutic options for patients who develop disease recurrence after an autologous SCT are limited. The 2-year OS rate after disease recurrence in the IFM study was 36%4; and the median OS in 2 other studies was <3 years.6, 23 Notwithstanding that the patients in our study may be a selected group, their long-term outcome compares favorably with these results, and suggests that salvage allogeneic SCT should be considered in this setting.
We identified several prognostic factors for predicting patient outcome. Chemosensitivity at the time of SCT is one of the most important factors in allogeneic SCT for all hematologic malignancies including myeloma.30, 31 Similarly, the EBMT study demonstrated that patients with progressive disease do not benefit from allogeneic SCT with RIC.15 The GVM effect may not be potent enough to overcome bulky or highly resistant disease. SCT earlier in the disease course was associated with better outcome. In our study that included heavily treated patients, the cutoff was 5 years from diagnosis. The majority of other studies demonstrated similar trends but used an earlier cutoff of 1 to 2 years.14, 15, 17 This observation must be taken with caution because the worse outcome of patients with a long disease duration may be related to poor prognostic characteristics such as refractory disease (53% of patients with disease duration of >5 years in the current study) and heavy prior therapy (35% had ≥2 prior autologous transplants, but the total lines of therapy was not different). We could not correlate the number of lines of therapy with outcome in this limited data set. Patients with a long first remission and sensitive disease can expect a favorable outcome with allogeneic SCT; however, there were only a few such patients in our series. SCT from a female donor to a male recipient was also detected in our study as well as the EBMT study15 as a strong independent adverse prognostic factor. This was related to a high incidence of NRM, and in particular due to acute GVHD, which is not well tolerated in myeloma patients. In a large EBMT analysis, female donor to male recipient was associated with both increased risk for GVHD, but decreased risk for disease recurrence, possible due to allogeneic responses mediated against H-Y minor histocompatibility antigens.32 Age was not found to be a significant factor in RIC, and an unrelated donor had a marginal adverse prognostic impact.14, 15 Although we have not used alemtuzumab in the conditioning, its use was found in other studies to be associated with adverse prognosis, due to loss of the GVM effect.15, 16 Cytogenetics has become an important tool in determining prognosis of myeloma. We had only partial cytogenetics data in this analysis, not allowing us to make a definitive conclusion. Earlier studies showed an adverse effect of deletion 1333; however, more recent analysis suggests that allogeneic SCT can overcome the adverse prognostic effect of deletion 13 as a sole fluorescence in situ hybridization finding and even when associated with t(4;14). Deletion 17p remains a poor prognostic factor after allogeneic SCT.17, 34
We have identified 2 post-transplant endpoints assessed after Day 100 post-transplant as being predictive of prognosis: occurrence of chronic GVHD and achievement of a CR. Chronic GVHD was associated with improved outcome in the majority14, 15 but not all studies.17, 18 The association of chronic GVHD with reduced recurrence risk supports the existence of GVM; however, it may also occur in its absence. Achievement of CR after autologous SCT has become an important surrogate marker for prolonged survival.35 Achievement of a better response before autologous SCT by the use of novel agents may also improve the post-transplant outcome.1 As also seen in the current study, allogeneic SCT is associated with a high CR rate. The depth of remission is related to prolonged response. In the current study, the achievement of VGPR was associated with longer responses than PR; however, long-term PFS was possible almost exclusively in patients achieving CR. Corradini et al demonstrated that a molecular remission achieved after myeloablative conditioning, as determined by polymerase chain reaction for immunoglobulin gene rearrangements generating a clone-specific molecular marker, was associated with possible cure, whereas patients with persistent minimal residual disease were destined to develop disease recurrence.36 Similar results were seen in the RIC setting as well.18 Patients who develop disease recurrence after SCT may respond to donor lymphocyte infusion (DLI).37 They may still demonstrate remarkable responses to the novel agents alone or in combination with DLI.38, 39 In particular, thalidomide and lenalidomide may induce T-cell and natural-killer (NK) cell activity against the myeloma, whereas bortezomib may inhibit GVHD while maintaining the GVM effect. These responses explain the wide separation of PFS and OS curves seen in this and other studies, and why the plateau in survival curves is only seen 4 to 5 years after SCT. These observations led Kroger et al to use novel agents and DLI in patients achieving only PR after SCT, not waiting until the inevitable disease progression in these patients, and trying to upgrade remission depth and improve survival.40 In the current study, 5 patients achieved CR with maintenance thalidomide after transplant and only 1 of them developed disease recurrence. These results were better than those of patients achieving CR and not having maintenance therapy (10 of 18 developed disease progression), although not reaching statistical significance due to low numbers. This observation suggests a role for maintenance therapy, although this study was not designed to explore this option. Future studies will continue to explore the integration of novel agents and novel cellular therapies with RIC, trying to reduce the incidence of disease recurrence, which is the major obstacle for cure after RIC.
In conclusion, RIC allogeneic SCT can result in prolonged PFS in a subset of patients with multiple myeloma. Best results are achieved when SCT is given early in disease course and in remission. However, allogeneic SCT can also be successful in the salvage setting.