Relapse to Prior Autograft and Chronic Graft-versus- Host Disease Are ...

1 downloads 2682 Views 289KB Size Report
... D-20246 Hamburg, Germany (e-mail: [email protected]). ... also significantly diminished by chronic graft-versus-host disease (GVHD) in a ...
Biology of Blood and Marrow Transplantation 10:698-708 (2004) 䊚 2004 American Society for Blood and Marrow Transplantation 1083-8791/04/1010-0005$30.00/0 doi:10.1016/j.bbmt.2004.06.002

Relapse to Prior Autograft and Chronic Graft-versusHost Disease Are the Strongest Prognostic Factors for Outcome of Melphalan/Fludarabine-Based DoseReduced Allogeneic Stem Cell Transplantation in Patients with Multiple Myeloma Nicolaus Kro¨ger,1 Jose A. Perez-Simon,2 Han Myint,3 Hans Klingemann,3 Avichai Shimoni,4 Arnon Nagler,4 Rodrigo Martino,5 Adrian Alegre,6 Jose F. Tomas,7 Rainer Schwerdtfeger,8 Michael Kiehl,9 Axel Fauser,9 Herbert Gottfried Sayer,10 Angel Leon,11 Jo¨rg Beyer,12 Tatjana Zabelina,1 Francis Ayuk,1 Jesus F. San Miguel,2 Ronald Brand,13 Axel Rolf Zander1 1

Bone Marrow Transplantation, University Hospital Hamburg, Hamburg, Germany; 2Hospital Clinico Universitario, Salamanca, Spain; 3Department of Bone Marrow Transplantation, RUSH University Medical Center, Chicago, Illinois; 4Department of Bone Marrow Transplantation, Chaim Sheba Medical Center, Tel Hashomer, Israel; 5Hospital de la Santa Creu I Sant Pau, Barcelona, Spain; 6Hospital de la Princesa, Madrid, Spain; 7 Fundacion Jimenez Diaz, Madrid, Spain; 8Department of Bone Marrow Transplantation, Wiesbaden, Germany; 9 Department of Oncology and Hematology and BMT, Clinic for BHT, Idar-Oberstein, Germany; 10Department of Oncology and Hematology, University Jena, Jena, Germany; 11Hospital de la Seguridad Social, Jerez, Spain; 12 Department of Hematology and Oncology, University Hospital Marburg, Marburg, Germany; 13Department of Medical Statistics, Leiden University Medical Center, Leiden, The Netherlands Correspondence and reprint requests: Nicolaus Kro¨ger, MD, Bone Marrow Transplantation, University Hospital Hamburg-Eppendorf, Martinistrasse 52, D-20246 Hamburg, Germany (e-mail: [email protected]). Received February 10, 2004; accepted May 29, 2004

ABSTRACT We evaluated prognostic factors of melphalan/fludarabine-based dose-reduced allografts in patients with multiple myeloma. From 1998 to 2002, 120 patients with multiple myeloma were treated with melphalan/ fludarabine followed by allogeneic stem cell transplantation. The cumulative risk at 1 year for treatmentrelated mortality (TRM) was 18% (95% confidence interval [CI], 12%-28%). In a multivariate analysis, relapse after prior high-dose chemotherapy was the most significant risk factor for TRM (hazard ratio [HR], 2.80; 95% CI, 1.16-6.74; P ⴝ .02), relapse (HR, 4.14; 95% CI, 2.04-8.38; P < .001), event-free survival (HR, 3.11; 95% CI, 1.77-5.46; P < .001), and overall survival (HR, 2.69; 95% CI, 1.35-5.35; P ⴝ .005). In addition, relapse was also significantly diminished by chronic graft-versus-host disease (GVHD) in a time-dependent Cox model (HR, 0.37; 95% CI, 0.16-0.87; P ⴝ .02). At transplantation, 8% of the patients were in complete remission, whereas 27% had progressive disease. After allografting, 49% achieved complete remission, and 38% achieved partial remission. In a subgroup of patients with chemosensitivity at transplantation and no relapse after prior high-dose chemotherapy who underwent transplantation with peripheral blood stem cells (n ⴝ 46), the cumulative risk of TRM at 1 year was only 8% (95% CI, 1%-54%). The 2-year estimated event-free and overall survival was 60% (95% CI, 42%-78%) and 75% (95% CI, 59%-91%), respectively, for related donors (n ⴝ 34) and was 81% (95% CI, 59%-100%) and 92% (95% CI, 76%-100%), respectively, for unrelated donors (n ⴝ 12). © 2004 American Society for Blood and Marrow Transplantation

KEY WORDS Allogeneic stem cell transplantation ● Dose-reduced conditioning ● Nonmyeloablative transplantation ● Graft-versus-host disease ● Autologous stem cell transplantation ● Multiple myeloma 698

Dose-Reduced Conditioning and Allografting for Multiple Myeloma

INTRODUCTION Allogeneic stem cell transplantation may cure multiple myeloma (MM). Despite high treatment-related mortality (TRM) of 17% to 40% [1-5], lower relapse rates in comparison to autologous transplantation have been reported. These are probably due to a graft-versus-myeloma effect mediated by immunocompetent donor lymphocytes [6]. After allogeneic stem cell transplantation, a portion of patients achieved well-documented molecular remission, resulting in long-term freedom from disease [6,7]. Despite the reduced TRM after standard conditioning over recent years [6], new strategies are still needed to decrease the high treatment-related morbidity and mortality. Recently, reduced-intensity conditioning or nonmyeloablative regimens followed by allogeneic stem cell transplantation have been successfully undertaken in a variety of hematologic malignancies, including myeloma [8-15]. In myeloma patients, the TRM was lower than after standard conditioning, and this allowed us to investigate this approach successfully in unrelated donors [16]. Despite the high efficacy with high complete remission (CR) rates even in treatment-refractory patients, the curative potential and optimal timing of this approach still have to be determined. We here report a multi-institutional study on factors influencing the outcome of patients with MM after dose-reduced conditioning consisting of melphalan and fludarabine and allogeneic stem cell transplantation. This study mainly focused on the optimal timing and effect of graft-versus-host disease (GVHD) on the outcome of this new approach in the treatment of myeloma patients and the identification of patients who may most benefit from this strategy. One hundred twenty patients with advanced MM were enrolled between July 1998 and December 2002 in 3 international studies investigating dose-reduced conditioning consisting of melphalan/fludarabine followed by allogeneic stem cell transplantation in MM. Thirty-eight patients underwent dose-reduced allogeneic stem cell transplantation as part of an autologous/ allogeneic tandem approach with melphalan (70-140 mg/m2) and fludarabine (150-180 mg/m2) followed by allogeneic stem cell transplantation from related or unrelated donors [12,16]. The remaining 82 patients were treated with dose-reduced allogeneic transplantation as salvage therapy after a relapse to a previous autograft (n ⫽ 58) or after a response to chemotherapy alone (n ⫽ 24) [17]. These patients received melphalan (140-160 mg/m2) and fludarabine (90-180 mg/m2) To be included, patients were required to have sufficient cardiac function (ejection fraction ⬎30%), a creatinine clearance ⬎30 mL/min, a lung diffusion capacity of at least 50%, and liver transaminases ⬍3 times the upper limit of normal. Patient characteristics are shown in detail in Table 1. The median age of the

BB&MT

patients was 52 years (range, 31-65 years). There were 73 male and 47 female patients. Seventy-two donors were male, and 48 were female. The median ␤2-microglobulin level at diagnosis was 2.9 mL/dL (range, 1.0-7.0 mL/dL). Cytogenetic fluorescent in situ hybridization (FISH) analysis was available in 45 patients and showed deletion 13 in 15 patients. Because of this low number, deletion 13 could not be included in the multivariate analysis. Thirty-four patients had received prior radiation therapy, and the median number of previous chemotherapy cycles was 6 (range, 4-24). Ninety-six patients had received at least 1 previous cycle of high-dose chemotherapy followed by autologous (n ⫽ 95) or allogeneic (n ⫽ 1) stem cell transplantation. Twelve patients had received 2 cycles of previous high-dose chemotherapy. The disease status before allogeneic transplantation was as follows: CR, n ⫽ 8; partial remission (PR), n ⫽ 63; minor response, n ⫽ 3; no change, n ⫽ 13; and progressive disease, n ⫽ 33. Eighty-five patients received transplants from related and 35 from unrelated donors. Two patients received stem cell transplants from HLA-B–mismatched related donors, and 2 patients received transplants from mismatched unrelated donors (HLA-A and HLA-DRB1 and -DQB1, respectively). The median time from diagnosis to transplantation was 23 months (range, 13-105 months). The cytomegalovirus serostatus was positive in 86 patients and negative in 34 patients. The source of stem cells was peripheral blood in 112 patients and bone marrow in 8 patients. No manipulation of the graft was performed. The median follow-up of the surviving patients was 16 months (range, 3-46 months).

METHODS Engraftment was defined as a leukocyte count ⬎1 ⫻ 109/L for 3 consecutive days and a nontransfused platelet count ⬎20 ⫻ 109/L. HLA-A and -B antigens were typed by serologic methods, and HLADRB1 alleles were typed with sequence-specific oligonucleotide probes. Standard criteria were used for the grading of acute and chronic GVHD [18,19]. Acute GVHD was treated with high-dose steroids, and extensive chronic GVHD was treated with cyclosporin A and steroids. Chronic GVHD was evaluated in patients who survived at least 100 days with sustained engraftment. Chimerism analysis was performed with an allelespecific multiplex polymerase chain reaction (PCR) technique or with FISH in case of sex mismatch. Antibiotic, antiviral, and antifungal prophylaxis was performed according to the center’s policy. Response to treatment was defined according to the European Group for Blood and Marrow Transplantation (EBMT)/International Bone Marrow Transplant 699

N. Kro¨ger et al.

Table 1. Patient Characteristics Variable No. patients Age, y, median (range) Sex Male Female Stage I II III Plasma cell leukemia Bence-Jones proteinuria (light chain) IgA IgG Nonsecretory IgM Deletion 13 (FISH positive) Deletion 13 (FISH negative) Deletion 13 (unknown) ␤2-Microglobulin, mg/dL, median (range) (n ⴝ 65) Prior radiotherapy Yes No Unknown No. prior chemotherapies, median (range) No. prior high-dose (HD) chemotherapies 0 1 2 Patients with relapse after HD chemotherapy Patients with autograft as part of an autologous allorganic approach Patients with relapse to at least 1 chemotherapy (including HD chemotherapy) Status before allografting CR PR NC MR PD Donor Related Unrelated Related mismatch Unrelated mismatch Donor sex Female Male Donor age, y, median (range) ABO Match Mismatch Unknown Patient CMV status Positive Negative Stem cell source Peripheral blood stem cells Bone marrow Conditioning Fludarabine 90 mg/m2 125 mg/m2 150-180 mg/m2

700

Data 120 52 (31-65) 73 47 2 38 80 2 21 18 69 7 5 15 30 75 2.9 (1-7) 34 54 32 4 (1-26) 24 84 (including 1 prior allograft) 12 58 38 65 8 63 13 3 33 85 35 2 (2 ⴛ HLA-B) 2 (1 ⴛ HLA-A, 1 ⴛ HLA-DRB1, -DQB1) 48 72 49 (10-68) 57 32 31 86 34 112 8

47 18 55

Dose-Reduced Conditioning and Allografting for Multiple Myeloma

Table 1. Patient Characteristics (Continued) Variable

Data

Melphalan 70 mg/m2 100 mg/m2 140-160 mg/m2 ATG none (Fresenius) 30 mg/kg BW (Fresenius) 60 mg/kg BW Horse ATG: 15 mg/kg BW (SangStat) 6.0 mg/kg BW Median CD34ⴙ cell dose/kg BW GYHD prophylaxis CsA/MTX CsA/MMF CsA alone None FK506/MMF

5 34 81 62 41 15 1 1 5.0 ⴛ 106 (range, 0.4-23) 111 3 3 1 2

Ig indicates immunoglobulin; FISH, fluorescent in situ hybridization; CR, complete remission; PR, partial response; NC, no change; MR, minor response; PD, progressive disease; CMV, cytomegalovirus; BW, body weight; ATG, antithymocyte globulin; CsA, cyclosporin A; MTX, methotrexate; MMF, mycophenolate mofetil; GVHD, graft-versus-host disease.

Registry criteria [20]. Briefly, CR required a disappearance of monoclonal gammopathy in serum and urine as determined by immunofixation analysis for at least 6 weeks and ⬍5% plasma cells in a bone marrow aspirate. A PR was defined as a ⬎50% reduction, and a minor response was defined as a ⬎25% reduction of the paraprotein level. No change was defined as a ⬍25% decrease or an increase of the paraprotein. Relapse was defined as recurrence of the monoclonal protein or bone marrow plasmocytosis in case of prior CR. Progression of non-CR patients required at least a 25% increase of paraprotein or development of new bone lesions. Written, informed consent was obtained

from each patient, and the local ethics committee approved the studies (Figure 1). GVHD Prophylaxis

GVHD prophylaxis consisted of cyclosporin A plus short-course methotrexate in 111 patients, cyclosporin A plus mycophenolate mofetil in 3 patients, cyclosporin A alone in 3 patients, and tacrolimus plus mycophenolate mofetil in 1 patient. The dose of cyclosporine A was adjusted to serum levels. In general, cyclosporine A was tapered between day 60 and day 100 if no signs of GVHD were observed. All patients

Figure 1. Treatment protocols and number of patients included.

BB&MT

701

N. Kro¨ger et al.

with an unrelated donor and 23 patients with a related donor received pretransplantation antithymocyte globulin (ATG; rabbit Fresenius, 30-60 mg/kg body weight, n ⫽ 56; rabbit SangStat, 6.0 mg/kg body weight, n ⫽ 1; and horse, 15 mg/kg body weight, n ⫽ 1). Statistics

For the analysis, the data of all 3 studies were pooled. Survival curves were estimated initially by the Kaplan-Meier method for descriptive purposes. The log-rank test was used for comparison of subgroups; if the number of subgroups exceeded 2 and the subgroups had an ordering, the trend version of the logrank test was used. Cox proportional hazards models were used for multivariate analyses and to produce overall (average) hazard ratio (HR) estimates in univariate comparisons of subgroups. The HR is assumed to be a valid estimate within the framework of the Cox model, even in the presence of competing risks. All computations were performed in SPSS version 10/11 (SPSS Inc., Chicago, IL). In each instance, the proportionality assumption of the Cox model was verified both by graphical methods (comparison with the Kaplan-Meier curves) and by computing a time-dependent Cox model, in which a dummy variable indicating whether an event occurred before or after the median time to event for all uncensored observations was introduced as an interaction term with the main risk factor. This formally tests whether the HR of that risk factor is significantly different at the first and second half of the survival curve. The effect of GVHD in a multivariate setting was analyzed in a time-dependent Cox model in which the risk factor GVHD was a dichotomous variable that switched from 0 to 1 when the event occurred after transplantation. Although the HR associated with such a time-dependent covariate can be interpreted clinically, survival curves themselves can not be drawn in a meaningful way to reflect the effect of GVHD. Chronic GVHD analysis was performed in a landmark setting, ie, by selecting all patients alive at day 100 (as required by the definition of GVHD). Cumulative incidences themselves for TRM and relapse (the heights of the survival curves) were estimated in the framework of a competing risk model after transfer of the data to NCSS version 2001 (NCSS, Kaysville, UT), which has a module for cumulative incidence estimation. The probability of TRM and relapse was estimated in 1 model, where they were considered competing risks for each other. Note that the sum of the cumulative incidence estimates of TRM and relapse thus obtained is equal to (1 ⫺ the estimated event-free survival probability), which is not the case with the Kaplan-Meier estimates. 702

RESULTS Engraftment and Chimerism

Three patients died of treatment-related causes before engraftment; all other patients engrafted. The median time until leukocyte (⬎1/nL) and platelet (⬎20/nL) engraftment was 15 days (range, 6-47 days) and 16 days (range, 4-65 days), respectively. Leukocyte and platelet engraftment was faster after related than after unrelated transplantation (14 versus 15 days, P ⫽ .009; and 14 versus 20 days, P ⫽ .005, respectively). Faster engraftment of leukocytes and platelets was also observed after peripheral blood stem cell (PBSC) as compared with bone marrow transplantation (14 versus 18 days, P ⫽ .002; and 15 versus 26 days, P ⫽ .002, respectively). Chimerism data are available in 95 patients, and 98% showed complete chimerism in peripheral blood at day 100, whereas in 2% a mixed chimerism was described. T-cell chimerism analysis was performed in 42 patients, and 93% showed complete and 7% mixed chimerism. Graft-versus-Host Disease

Acute GVHD grade II to IV was noted in 55 (46%) patients. Forty-five (38%) did not experience any acute GVHD, whereas 30 (25%) experienced mild acute GVHD grade I. Severe grade III to IV acute GVHD was observed in 21%: grade III in 21 patients (18%) and grade IV in 4 patients (3%). There was no difference in grades II to IV and grades III to IV acute GVHD between related and unrelated donors (51% versus 50% and 18% versus 22%, respectively). No significant difference in terms of grades II to IV acute GVHD was observed between PBSC and bone marrow transplantation (49% versus 75%; P ⫽ .10). Chronic GVHD was observed in 47% of the patients: 25% experienced limited and 22% extensive chronic GVHD. The incidence of chronic GVHD was similar among patients receiving PBSC or bone marrow transplants (49% versus 42%; P ⫽ .30), although the number of patients receiving bone marrow was too small (n ⫽ 8) to draw definite conclusions. It is interesting to note that the incidence of chronic GVHD was lower among patients who received transplants from unrelated as compared with related donors (29% versus 56%; P ⫽ .006); this was probably due to the incorporation of ATG in all unrelated transplants. Although ATG had no effect on acute GVHD, the incidence of chronic GVHD was significantly influenced by the use of ATG (23% versus 83%; P ⫽ .006). In a multivariate analysis, patient treated with ATG had a less than 2 times lower incidence of chronic GVHD compared with those who did not receive ATG (HR, 0.34; 95% confidence interval [CI], 0.16-0.71; P ⫽ .004).

Dose-Reduced Conditioning and Allografting for Multiple Myeloma

Table 2. Treatment-Related Mortality Variable

n

Overall Sepsis, pneumonia, other infections Fungal infection GVHD Cardiac failure Toxicity (multiorgan failure, alveolar hemorrhage)

24 8 5 5 3 3

only 8 patients (8%) were in CR, whereas 33 (27%) had progressive disease. After allogeneic transplantation, 53 patients achieved a CR (49%), 41 patients a PR (38%), and 4 patients (4%) a minor response according to the EBMT criteria, resulting in an overall response rate of 91%. Five patients (5%) achieved no change, and 5 patients (5%) had progressive disease after allogeneic transplantation.

Treatment-Related Mortality

Relapse/Progression

The cumulative risk at 1 year for TRM was 18% (95% CI, 12%-28%). The causes of TRM are listed in Table 2. In a univariate analysis, the TRM was lower among patients who had not relapsed with a prior high-dose chemotherapy regimen (13% versus 25%; P ⫽ .01), patients with chemosensitive disease at transplantation (15% versus 23%; P ⫽ .05), patients with a related donor (15% versus 26%; P ⫽ .03), patients who received PBSCs as the graft source (17% versus 37%; P ⫽ .02), and patients who underwent transplantation within 1 year after diagnosis (3% versus 28%; P ⫽ .002). A trend toward a lower TRM was noted among patients who did not receive prior radiotherapy (14% versus 35%; P ⫽ .08) and those with a male donor (16% versus 31%; P ⫽ .10). No effect on TRM was observed for acute GVHD, CD34⫹ cell dose, dose of melphalan, age, use of ATG, patient’s sex, or ABO mismatch. In a multivariate analysis using the Cox regression model, relapse after prior highdose chemotherapy was the only factor that significantly influenced TRM, with an HR of 2.80 (95% CI, 1.16%-6.74%) and a P value of .02 (Figure 2). In a subgroup of patients with chemosensitivity and no relapse to prior high-dose chemotherapy who underwent transplantation with PBSCs (n ⫽ 46), the cumulative risk of nonrelapse mortality at 1 year was only 6% (95% CI, 1%-22%) for related (n ⫽ 34) and 8% (95% CI, 1%-54%) for unrelated (n ⫽ 12) donors.

The cumulative incidence of relapse/progression at 2 years for all patients was 43% (95% CI, 33%57%). In the univariate analysis, relapse after prior high-dose chemotherapy (66% versus 33%; P ⬍ .001), chemorefractory disease at transplantation (55% versus 38%; P ⫽ .02), and late transplantation ⬎2 years after diagnosis (43% versus 41%; P ⫽ .05) were the most significant factors for a higher relapse rate. Limited data on deletion 13 were available, but despite the low number of included patients, the cumulative incidence of relapse was higher in patients with (n ⫽ 15) than without (n ⫽ 30) deletion 13 (56% versus 14%; P ⫽ .03). Because of the low number of patients, this factor could not be included in the multivariate analysis. The use of ATG was not associated with a higher incidence of relapse or progression (P ⫽ .60). In a multivariate analysis, relapse after a prior high-dose chemotherapy regimen was the only factor that significantly influenced relapse (HR, 4.14; 95% CI, 2.048.38; P ⬎ .001; Tables 3 and 4 and Figure 2). For results concerning the effect of GVHD on relapse, see below.

Response

One hundred eight patients were evaluable for response. At the time of allogeneic transplantation,

Overall Survival

The estimated 2-year overall survival for all patients was 59% (95% CI, 29%-89%). In the univariate analysis, relapse after prior high-dose chemotherapy (43% versus 73%; P ⫽ .001), no chemosensitivity at the time of transplantation (51% versus 66%; P ⫽ .03), bone marrow as the stem cell source (31% versus 72%; P ⫽ .003), a matched unrelated donor (38% versus 66%; P ⫽ .01), the use of ATG (47% versus

Figure 2. Cumulative incidence of relapse and treatment-related mortality in patients with (right) or without (left) relapse to a prior autograft.

BB&MT

703

N. Kro¨ger et al.

Table 3. Univariate Analyses of Overall and Event-Free Survival after Melphalan/Fludarabine Variable Age 50 y Patient sex Female Male Deletion 13 Yes (n ⴝ 15) No (n ⴝ 30) ␤2-Microglobulin* 2 mg/dL Prior radiotherapy Yes No Relapse after prior HD chemotherapy Yes No Status at transplantation CR/PR NC/PD Transplantation Unrelated donor Related donor Donor sex Female Male Donor age >50 y 100 mg/m2