Outcomes of reduced-intensity transplantation for chronic myeloid leukemia: an analysis of prognostic factors from the Chronic Leukemia Working Party of the EBMT

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) has arguably been applied most successfully in the treatment of chronic myeloid leukemia and remains, despite recent advances in therapy, the most effective strategy for inducing durable molecular remission.^1,2  The application of conventional myeloablative allo-HSCT has, however, been limited by the age of the recipient and the availability of matched sibling or unrelated donors.

The original view that the allogeneic cells simply provided a source of tumor-free stem cells has been superseded by the realization of the importance of allogeneic T cells in mediating graft-versus-leukemia effects. This immunologic activity is now thought to be a primary mechanism by which allo-HSCT achieves its antileukemic effect. The clinical responses seen with donor lymphocyte infusions are the clearest demonstration of the potency of this immune action,³  and this effect is particularly pronounced in chronic myeloid leukemia (CML).^4,5 

Reduced-intensity conditioning (RIC) regimens, in contrast to conventional myeloablative strategies, are less myelosuppressive but remain profoundly immunosuppressive.^6-10  Phase 2 studies have demonstrated that reliable and durable engraftment is possible in the majority of patients and the 1-year transplantation-related toxicity is in the range of 10% to 15%.^8,11-14  The long-term antitumor effect of this approach is less well established, with the majority of studies describing patients with heterogeneous diseases and disease phases and limited disease-specific outcome data.

Two studies have reported favorable outcomes using fludarabine (30 mg/m²/d × 5), busulfan 8 mg/kg, and antithymocyte globulin (ATG). Or et al¹⁵  reported the outcome of 24 patients treated in first chronic phase (CP1). Overall survival and progression-free survival were estimated at 85% (confidence interval [CI], 70%-100%) at 5 years. The incidences of acute (grade II-IV) and chronic graft-versus-host disease (GvHD) were 75% and 55%, respectively, but accounted for the only 3 deaths in this series.All patients were reported to be reverse transcriptase–polymerase chain reaction (RT-PCR) negative for BCR-ABL.¹⁵  The second study reported similar outcomes for 15 patients in CP1 using an identical conditioning regimen.¹⁶  However, good results are not universally seen with this regimen, and transplant-related complications remain an issue.^17,18  There are also concerns that the risk of disease relapse may be higher, especially with more advanced disease.¹⁹ 

The purpose of the present study was to estimate the efficacy of RIC allografts in CML in a nonselected patient population. This was performed by a retrospective analysis of patients reported to the Chronic Leukaemia Working Party registry of the European Group for Blood and Marrow Transplantation (EBMT), which collects data from more than 500 transplantation centers in and beyond Europe.

This study was conducted on behalf of the Chronic Leukemia Working Party of the EBMT and was performed in conjunction with a study of RIC in myeloma. All EBMT centers report a minimal essential data set into a central database. This database was interrogated, and all EBMT centers that had reported more than 5 RIC allografts for CML and/or myeloma were contacted for additional data. This analysis is restricted to patients with CML. Informed consent was obtained locally in accordance with the principles laid out in the Declaration of Helsinki and according to the local regulations applicable at the time of transplantation. Since 1 January 2003 the EBMT has required centers to confirm that written informed consent has been obtained prior to data acceptance for new patients and prior to accepting follow-up data from patients who previously received a transplant. The definition of RIC allografts was determined by contributing centers and was based on current EBMT guidance. Patients who received regimens of equivalent or lower intensity to busulfan 8 mg/kg were included in the study. Patients who had received total-body radiotherapy were included, provided the fractionated delivered dose was no more than 6 Gy. The minimal data required for inclusion of a patient in the study were age at transplantation, diagnosis, number of prior transplantations, disease status at transplantation, HLA match of donor, conditioning regimen, remission status after transplantation, disease status at follow-up, date of progression or death, date of follow-up, and cause of death.

Chimerism studies were performed mostly by molecular methodologies (73%), by cytogenetics in 25%, and in 1 patient by both techniques. Chimerism analysis was reported in 1 patient on the basis of ABO group.

Response was documented on or after day 100 after transplantation. Response was subclassified as cytogenetic remission according to established criteria²⁰  or molecular remission (BCR-ABL PCR negativity) when possible. This subclassification of complete response was missing in 43 (23%) patients. Survival was measured in months and defined as the time from the date of transplantation until date of death from any cause or last follow-up. Progression-free survival was defined as the date of transplantation until date of progression or death from any cause, or last follow-up. Transplantation-related mortality (TRM) was defined as death as a result of any cause other than disease progression or relapse. The disease phase at transplantation was defined locally. GvHD was classified according to established criteria.^21,22 

Probabilities of survival and progression-free survival were calculated using the Kaplan-Meier method, whereas TRM and relapse were estimated using the cumulative incidence procedure. Univariate comparisons were made using the log-rank test, and variables found to be significant at the P value less than .1 level were entered into a proportional hazards regression analysis using a backward stepping procedure. Comparisons between groups were made using the chi-square test for categoric data, and the Mann-Whitney test for continuous data. P values are from 2-sided tests with those less than .05 considered significant. Quoted confidence intervals refer to 95% boundaries.

Between January 1994 and October 2002 data were collected on 221 transplantations from 38 centers within the EBMT group. The minimum essential data set was available for 186 patients from 33 centers, and these patients form the basis for this study.

The patient characteristics for all patients and the subset in CP1 are shown in Table 1. Thirty-one different chemotherapy and antibody regimens were used and are summarized in Table 2. The majority were fludarabine based (84%), often combined with busulfan (56%). Cyclophosphamide and melphalan were used less frequently (23% and 12%, respectively). The combination of fludarabine, busulfan, and ATG (Fd/Bu/ATG) was used in 40% of patients, whereas a low-dose TBI regimen was only used in 16 patients, 14 of whom received fludarabine and TBI. The median TBI dose was 2.0 Gy (range, 1.94-3.2 Gy). Ex vivo graft manipulation was uncommon (8 patients), but in vivo T-cell depletion was used in 69%, predominately with ATG (106 patients) or alemtuzumab (Campath 1H; 21 patients). The majority of donors were matched siblings (61%) with a smaller number of mismatched or other family donors (11%). Unrelated volunteers were the donors for 28% of the recipients. The median mononuclear and CD34⁺ cell doses were 3.4 × 10⁸/kg and 4.4 × 10⁶/kg for BM and 9.1 × 10⁸/kg and 5.7 × 10⁶/kg for peripheral blood. Stratification according to EBMT transplantation risk score is given in Table 1.

Engraftment occurred in 94% of the patients irrespective of the stem-cell source. Secondary graft failure occurred in 3%. The median times to neutrophil (> 0.5 × 10⁹/L) and platelet (> 50 × 10⁹/L) recoveries were 16 days (range, 0-380 days) and 17 days (range, 0-384 days), respectively. There were significant differences in the times to neutrophil recovery (15.5 versus 17 days; P = .05) and platelet recovery (16 versus 22 days; P < .001) in patients receiving peripheral-blood stem cells (PBSCs) or marrow. Only one patient who received marrow did not become neutropenic or thrombocytopenic, whereas 12% of those receiving PBSCs had neutrophils greater than 0.5 × 10⁹/L at all times and 11% did not become thrombocytopenic (> 50 × 10⁹/L). Chimerism data were available for 163 patients: 73% achieved 100% donor chimerism and 77% achieved greater than 95% donor chimerism at some time after transplantation. On subsequent analyses these figures had fallen to 65% with 100% donor and 68% greater than 95% donor.

Acute GvHD (aGvHD) occurred in 92 patients (49%) and of these 35 (19%) developed grade I, 40 (21%) grade II, and 17 (9%) grade III to IV aGvHD. The use of alemtuzumab, ATG, the type GvHD prophylaxis, or the use of peripheral blood versus bone marrow did not influence the incidence or severity of aGvHD. Chronic GVHD (cGvHD) was evaluable in 137 patients of whom 26 (19%) developed limited and 32 (23%) extensive cGvHD. In 3 patients the grade of cGvHD was not specified, and 15 patients developed cGvHD, which subsequently resolved. The use of ATG significantly reduced the incidence of cGvHD (44% versus 72%; P = .001). The use of alemtuzumab was of borderline significance (39% versus 59%; P = .06). No effect of the use of PBSCs or the GvHD prophylaxis regimen on the incidence or severity of cGvHD could be demonstrated.

The 100-day and 1- and 2-year TRMs were 3.8% (CI, 1.9%-8%), 13.3% (CI, 9%-19.2%), and 18.9% (CI, 14%-26%). Factors associated with reduction in TRM were disease phase at time of transplantation, 11.6% versus 20.3% for chronic phase and advanced phase at 1 year, respectively, (P = .005; Figure 1) and the use of Fd/Bu/ATG in the conditioning regimen (yes 7.0% versus no 18.0%; P = .004). The use of busulfan alone was of borderline significance (P = .076). In multivariate analysis advanced disease (relative risk [RR], 2.04; CI, 1.05-3.98) and not receiving Fd/Bu/ATG (RR, 2.46; CI, 1.15-5.26) were associated with increased risks of TRM. Development of grade III to IV aGvHD or cGvHD was also associated with an increased TRM regardless of disease phase.

Of the 186 patients, 10 were not evaluable for response because of early death. Overall, 87% (152 of 172) of patients were in remission at the time of first report (on or around day 100 after transplantation). Data on the use of molecular and cytogenetics to define remission are incomplete; however, at least 40% of patients had attained a molecular complete response (CR) and at least 62% a cytogenetic CR (Table 3). When patients in CP1 were analyzed separately, then the overall response rate was 96% with molecular and cytogenetic CRs being achieved in at least 48% and 72%, respectively. The overall response rates in second chronic phase (CP2), accelerated phase (AP), and blast crises (BC) were 92%, 67%, and 22% respectively. The median time to documentation of remission was 92 days.

Overall, 72 patients relapsed with a cumulative probability of relapse of 47% at 3 years. The median time to progression has not been reached. Patients with more advanced disease were at higher risk of relapse at 68.6% and 68.7% for AP and BC, respectively, compared with 34.6% for CP1 and 45.4% for CP2. Other factors associated with an increased risk of relapse were the use of fludarabine + TBI and/or cyclophosphamide, whereas busulfan was associated with a decreased relapse risk. With respect to the use of lymphocyte-depleting antibodies, alemtuzumab was associated with a higher relapse incidence (63%) at 2 years compared with ATG or no antibody (36.0% and 50.5%, respectively; P = .02). The factors that retained significance on multivariate analysis were transplantation for advanced phase disease (RR, 3.73; CI, 2.2-6.3; P < .001; Figure 2) and the use of alemtuzumab (RR, 2.82; CI, 1.5-5.5; P = .002).

The median projected progression-free survival (PFS) was 15.6 months with a 3-year PFS of 33% (CI, 26%-41%; Figure 3). Chronic-phase disease was associated with superior PFS, being 43% at 3 years (CI, 34%-52%). Other important factors were prior transplantation and the use of Fd/Bu/ATG for the conditioning. The effect, if any, of the use of alemtuzumab was difficult to evaluate because the number of patients receiving this drug was small. There was a trend toward an adverse effect (P = .06; Table 4). On multivariate analysis, transplantation in advanced-phase disease retained significance for adverse outcome. The use of Fd/Bu/ATG was associated with a trend to improved outcome (Table 5). The benefit of Fd/BU/ATG was more marked in patients in CP1 (RR, 0.6; CI, 0.3-0.9; P = .03; Figure 4).

Donor lymphocyte infusions (DLIs) were used in 50 patients (27%) with a median time to DLI of 6 months (range, 30 days to 2 years). In 36 patients (74%) the indication for DLI was disease relapse. DLI was used preemptively in 6 patients (12%) and to increase the degree of donor chimerism in 7 (12%). The median number of infusions was 2 (range, 1-10), and the median cell dose for the final infusion was 3.5 × 10⁷ CD3/kg (range, 1 × 10⁶-1.8 × 10⁸ CD3/kg). When DLIs were given for disease relapse, responses were seen in 69% of 35 evaluable patients; of these, 9 (25%) were molecular remissions and 3 (8%) were cytogenetic responses. Data on the use of imatinib were available for 130 patients, including 10 patients who received imatinib before transplantation (8%). An additional 25 patients received imatinib after transplantation, 20 for relapsed disease and 5 preemptively. Of the patients who relapsed, 9 had also received DLI (7 before and 2 after imatinib). The median starting dose was 400 mg (range, 100-600 mg). Responses were seen in 13 of 14 patients, including 6 molecular remissions.

The median follow-up was 35 months (range, 0.3-77 months) and the median survival was 52 months. The probability of survival at 3 years was 54% (CI, 46%-61%). The median survival for patients in CP1 or CP2 has not been reached with 3-year survival probabilities of 69% (CI, 60%-77%) and 57% (CI, 46%-67%), respectively. Transplantation in chronic phase, first transplantation, and the use of Fd/Bu/ATG were associated with improved overall survivals. Conversely, the use of fludarabine and TBI, alemtuzumab, and prior transplantation were associated with an adverse survival (Table 4). Of these factors, transplantation in advanced disease phase retained significance on multivariate analysis (Figure 5) with a trend to adverse outcome with prior transplantation (Table 5).

Overall survival was also predicted by EBMT score (P = .006). Patients with a score of 1 to 2, 3 to 4, and greater than 4 had an overall survival (OS) of 69% compared with 57% and 33%, respectively (Figure 6). EBMT scores of 1 to 2, 3 to 4, and greater than 4 were also significant for PFS at 49% compared with 35% and 19%, respectively.

Chronic myeloid leukemia is theoretically a difficult environment in which to achieve durable hematopoietic engraftment with reduced-intensity conditioning, given the hypercellular nature of the pretransplantation marrow and the lack of any intensive prior chemotherapy. The sustained engraftment rate (91%, including secondary graft failure) is somewhat lower than obtained with conventional conditioning but remains acceptable given that autologous reconstitution is the expected consequence of graft failure. Previous smaller studies^15,16  have demonstrated similar engraftment rates, but this is the largest study to date that shows that engraftment can be achieved using a variety of different conditioning regimens. The median times to neutrophil and platelet engraftment and the lack of severe neutropenia and thrombocytopenia in patients receiving PBSC grafts are consistent with previous reports.^24,25  The attainment of full donor chimerism is important both in assuring stable and durable engraftment and also in providing a basis for the graft-versus-leukemia effect and subsequent immunotherapy. In this study full donor chimerism was achieved in the majority of patients. The results of chimerism at later time points after transplantation were only available in smaller proportions of patients, so the observed fall in the level of donor chimerism may not be representative of the whole patient population.

Overall, the TRM was low, particularly in light of the median age and a median risk assessment score of 3. The day 100 TRM however substantially underestimates the risk of the procedure as the TRM continues to rise for at least the first 2 years. The association between increased TRM and disease phase is expected. The reason for the reduced TRM with the use Fd/Bu/ATG is less clear. In vivo T-cell depletion with ATG may be relevant because the low TRM may be partially explained by the lower incidence of severe aGvHD (8% grade III-IV). Chronic GvHD was more prevalent, possibly reflecting the delay in onset of GvHD that is apparent after RIC.²⁶  As expected, the development of extensive cGvHD was associated with an increase in TRM. In contrast to aGvHD, the incidence of cGvHD was affected both by the use of ATG and alemtuzumab, although the latter was of borderline significance.

For patients in CP1 or CP2 the overall survival was 69% and 57%, respectively, at 3 years. The outcome however for patients in AP or BC is inferior with 2-year OS of 24% and 8%, respectively. At 3 years the overall survivals of patients with scores of 1 and 2 at 69% are similar to those obtained by conventional transplantation at 80% and 70%.²³  Because there is uncertainty about the relapse risk with RIC, these data do not provide any evidence of advantage to justify choosing a reduced intensity rather than a conventional approach in these groups. In contrast, the 3-year survivals for patients with EBMT scores of 3, 4, and 5 were 67%, 42%, and 35% compared with 55%, 45%, and 27% (for score 5-7) for conventional transplantations. Although long-term data with respect to disease-free survivals are clearly not yet available, these results suggest that in the short term RIC transplantation may be the preferred option for patients with a higher EBMT transplantation score. This is of particular relevance at the present time because the encouraging early results of imatinib has led to a reevaluation of the role of transplantation and a substantial reduction in the numbers of patients who receive a transplant.²⁷  The majority of patients are now only offered transplantation if imatinib has failed. These individuals will inevitably come to transplantation with higher risk scores because of their age, the delay from diagnosis to transplantation, and possible advanced-phase disease.

The overall remission rate was 87%. Data relating to more specific remission rates, that is, molecular and/or cytogenetic remissions, are difficult to obtain because of the different forms of assessment of remission in different centers. Relapse remains a significant problem in all disease phases. The type of conditioning does seem to affect this risk with the highest relapse rates being seen after the use of alemtuzumab (76% at 2 years). The association between T-cell depletion and disease relapse is well known,²⁸  and the results with alemtuzumab are consistent with this. Of note, however, lymphocyte depletion with ATG did not have the same adverse outcome, and it is unclear from this study whether this is due to the relative potency/dose or specificity of the drug. The use of fludarabine and TBI was also associated with poorer outcomes. Fludarabine with low-dose TBI is one of the least intensive of the regimens used, and this may be a factor in the poor outcomes seen. However, this observation needs to be interpreted with great caution because it is based on very small patient numbers. This group, for example, contained only 4 patients in CP1. For this reason Fd + TBI was not included in the multivariate analysis. Fludarabine + TBI has been associated in one study with increased rate of graft failure, GvHD, and TRM²⁹ ; however, other groups report better results using this conditioning.^10,30  The importance of the relative conditioning intensity remains an area of considerable uncertainty, and, although some of the more intensive of the RIC regimens report good outcomes in terms of disease control, the benefit with respect to reduction in TRM is less clear.³¹ 

Although relapse is unwelcome, the use of DLI and/or imatinib can restore remissions. The limited data available here are consistent with published data on the efficacy of DLI for relapse³²  after allogeneic transplantation and with the limited data on the use of imatinib in this setting.³³  Given these options, disease relapse may be a more acceptable risk in CML compared with other hematologic malignancies.

In most series, patients with unrelated donors have an unfavorable outcome. In this study survival of patients with unrelated BM/PBSC donors was not significantly inferior to that of patients with matched related donors. This is unlikely to be related to the use of RIC transplantation alone. First, improvements in HLA typing and better supportive care have led to an improvement in unrelated transplantation outcomes.³⁴  Second, T-cell depletion, which is associated with a reduction in early transplant-related mortality, was used in the majority of unrelated transplantations. Finally, the relatively small number of unrelated transplantations might be insufficient to demonstrate any small differences.

No conditioning regimen demonstrated clear superiority. There was a trend toward improved outcomes with respect to PFS in this series in patients receiving Fd/Bu/ATG, although this did not retain significance on multivariate analysis for overall survival. Although it would be inappropriate to conclude that this is the optimal regimen, the use of this combination might form the basis of future prospective studies.

This is a retrospective registry-based study, which documents outcomes being achieved with RIC allografts in CML in a representative cross-section of European transplantation centers. It confirms the feasibility of RIC transplantation in CML and demonstrates efficacy particularly in CP1 or CP2. It also identifies good and poor risk factors and as such provides data for the rational design of the prospective studies that are now required.

The following centers and investigators contributed patients to this study. Austria: H. Greinix, AKH und Universitaetskliniken Wien, Vienna. Belgium: M. Boogaerts, University Hospital Gasthuisberg, Leuven. Czech Republic: A. Vitek, Institute of Hematology and Blood Transfusion, Prague; J. Mayer, University Hospital Brno, Brno. France: J. J. Cahn, Hopital Jean Minjoz, Besancon; J. J. Sotto and F. Garban, Hopital A. Michallon, Grenoble; M. Michallet, Hopital E. Herriot, Lyon; D. Blaise and D. Redortier, Institut Paoli Calmettes, Marseille; G. Socie and L. Ades, Hopital St Louis, Paris; J. M. Boiron, Centre Hospitalier et Universitaire (CHU) Bordeaux Hôpital, Pessac; M. Attal, Hopital de Purpan, Toulouse. Germany: A. Zander and H. Mundhenk, University Hospital Eppendorf, Hamburg. Israel: A. Nagler and A. Shimoni, Chaim Sheba Medical Center, Tel-Hashomer. Italy: A. Rambaldi, Ospedale Bergamo, Bergamo; A. Bacigalupo, Ospedale San Martino, Genova; P. Iacopino, Azienda Ospedaliera, Reggio Calabria; A. Carella, Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS), Casa Sollievo della Sofferenza, San Giovanni Rotondo; F. Locatelli, Azienda Ospedaliera S. Giovanni, Torino. Poland: A. Lange, Lower Silesian Centre for Cellular Transplantation, Wroclaw; M. Komarnicki, K. Marcinkowski University, Poznan. Spain: A. Urbano-Ispizua, Institute of Hematology and Oncology, Barcelona; J. Sierra, Hospital Santa Creu I Sant Pau, Barcelona; D. Caballero, Hospital Clínico, Salamanca; A. Iriondo, Hospital U. Marqués de Valdecilla, Santander. Sweden: M. Brune, Sahlgrenska University Hospital, Goeteborg; O. Ringden and K. Niemelä, Huddinge University Hospital, Huddinge; G. Juliusson, University Hospital, Linkoping. Switzerland: A. Gratwohl, J. Passweg, and A. Genitsch, University Hospital, Basel. United Kingdom: D. Milligan, Birmingham Heartlands Hospital, Birmingham; G. Mufti and M. Kenyon, Guy's, King's, and St Thomas' (GKT) School of Medicine, London; C. Crawley and J. F. Apperley, Hammersmith Hospital, London; R. Chopra and B. Foulkes, Christie Hospital National Health Service (NHS) Trust, Manchester; M. Barnett and J. Matthews, St Bartholomew's and The Royal London Hospital, London.