The optimal combination of fludarabine, busulfan, and antithymocyte globulin (ATG) for reduced-intensity conditioning (RIC) transplantation has not been established. ATG plays a pivotal role in the prevention of graft-versus-host disease (GvHD), but it is associated with a higher relapse rate and an elevated incidence of infections when high doses are used.
The authors retrospectively compared 2 different doses of ATG combined with fludarabine and busulfan in 229 adult patients who underwent transplantation at their institution. ATG was administered over 1 day (FBA1) or over 2 days (FBA2) at a daily dose of 2.5 mg/kg.
There were 124 patients in the FBA2 cohort and 105 patients in the FBA2 cohorts. Patients in the FBA2 cohort were older and more frequently underwent transplantation from an unrelated donor; 93% of patients in the FBA2 cohort received intravenous busulfan versus only 5% in the FBA1 cohort. The incidence of grade 2 through 4 acute GvHD was 23% in the FBA2 cohort versus 42% in the FBA1 cohort (P = .002); the incidence of grade 3 through 4 acute GvHD was 10% versus 23%, respectively (P = .006); and the incidence of chronic GvHD was 35% versus 69%, respectively (P < .0001). The 2-year rates of overall survival, nonrelapse mortality, and relapse/progression for the FBA1 and FBA2 cohorts were 65% versus 67%, respectively (P = .99), 20% versus 19%, respectively (P = .61), and 30% versus 19%, respectively (P = .09). The results were confirmed in multivariate analysis.
Allogeneic hematopoietic stem cell transplantation performed after reduced-intensity conditioning (RIC-AlloSCT) is an alternative option for patients who are not fit for a myeloablative (MA) preparation.1-3 In this setting, the combination of fludarabine, busulfan, and antithymocyte globulin (ATG) is widely used.1, 4-6 It is known that ATG is effective in preventing graft-versus-host disease (GvHD)7-9; however, because of in vivo T-cell depletion, ATG is associated with a potential risk of increasing the relapse probability10-12; therefore, titration of the dose is crucial.13, 14 Whereas an ATG dose in the range of 10 mg/kg appears to be associated with loss of disease control,4 we recently reported that a 2.5 mg/kg total dose did not preclude patients presenting with high rates of severe acute GvHD (aGvHD) and chronic GvHD (cGvHD).15 An intermediate dose of approximately 4.5 mg/kg reportedly delivers satisfactory GvHD prevention and antitumor effects.16, 17
Given these data, we decided to increase the dose of ATG in our recently reported conditioning regimen15 with the objective of optimizing the prevention of potential complications and maintaining disease control. In the current report, we retrospectively compare the outcome of 229 consecutive patients who received treatment at our center with 2 different doses of ATG, 2.5 mg/kg (n = 124 [the FBA1 cohort]) and 5 mg/kg (n = 105 [the FBA2 cohort]), in combination with 5 to 6 days of fludarabine and 2 days of busulfan.
MATERIALS AND METHODS
Conditioning for the FBA1 cohort consisted of oral busulfan (1 mg/kg every 6 hours on days −4 and −3; Myleran, GlaxoSmithKline, Marly-le-Roi, France), rabbit ATG (2.5 mg/kg on day −3; Thymoglobuline; Genzyme, St. Germain-en-Lay, France), and fludarabine (30 mg/m2 daily from day −6 or −5 to day −1; Fludara; Bayer Sante, Puteaux, France). For patients in the FBA2 cohort, busulfan was mainly administered as an intravenous formulation (0.8 mg/kg every 6 hours on days −4 and −3; Busilvex; Pierre Fabre, Boulogne-Billancourt, France) in combination with 2 days of ATG (2.5 mg/kg daily on days −2 and −1) and fludarabine as specified above. GvHD prophylaxis was mainly realized by cyclosporine alone, and supportive care was performed in accordance with previously reported procedures.15 All patients with hematologic malignancies who received the FBA1 or FBA2 regimen between January 1, 2000 and January 30, 2010 were included in the study. Diseases were classified according to the risk of relapse.18 Patients underwent RIC-AlloSCT according to our institutional policy. All donors were human leukemic antigen (HLA)-matched (related or unrelated) or mismatched (unrelated with 1 HLA mismatch). The graft source was peripheral blood stem cells harvested after the administration of granulocyte–colony-stimulating factor. This study was approved by the Paoli-Calmettes Institute review board.
The primary objective of the study was to assess the incidence of aGvHD and cGvHD with the newer FBA2 regimen. Other endpoints were overall survival (OS), nonrelapse mortality (NRM), relapse, and progression.
Categorical variables are expressed as proportions, and continuous variables are expressed as the median with the respective range. Comparisons between groups were performed with the chi-square and Mann-Whitney tests for categorical and continuous variables, respectively. Survival outcomes were calculated from the date of transplantation. NRM was defined as death without evidence of disease progression or relapse, and death after disease progression was treated as a competing event in the NRM analyses. The incidence of aGvHD and cGvHD was estimated considering disease progression or death as competing events. Patients without an event were censored at their last contact date, except with respect to aGvHD incidence (day 100). The Kaplan-Meier method, log-rank tests, and Cox proportional hazards models19, 20 were used for the OS analyses. Hazard ratios were estimated with their respective 95% confidence intervals (CIs). Prentice estimation, the Gray test, and Fine and Gray models21, 22 were used for the NRM, relapse/progression, and aGvHD and cGvHD analyses. When investigating its impact on survival outcomes, aGvHD was treated as a time-dependent covariate. Estimates are given at 2 years after transplantation, unless otherwise specified. aGvHD was graded according to the criteria published by Glucksberg et al,23 and cGvHD was classified as either limited or extensive.24 A multivariate analysis was performed to determine the impact of the conditioning regimen relative to the main pretransplantation and peritransplantation variables (patient and donor age and sex, diagnosis, disease status at transplantation, comorbidities, year of AlloSCT, donor type, total cell surface glycoprotein [cluster of differentiation 34 (CD34)]-positive and T-cell coreceptor [CD3]-positive counts infused with the graft). Models were built when the ATG dose was associated with the outcome of interest with a significance of P ≤ .10 in univariate analysis. P values ≤ .05 were considered significant. The statistical software packages SPSS version 16.0 (SPSS Inc., Chicago, Ill) and R version 2.12.0 (R Foundation for Statistical Computing, Vienna, Austria) were used for statistical analyses.
Data from 229 consecutive adult patients were entered into the analysis of the 2 study cohorts: FBA1 and FBA2. The median follow-up was 38 months (range, 3-108 months) for the patients who remained alive. Their primary characteristics are provided in Table 1. The 2 cohorts differed significantly in the length of observation time, donor sex, and busulfan formulation. Moreover, patients in the FBA2 cohort were significantly older, received a graft with higher CD34-positive and lower CD3-positive counts, and more frequently underwent transplantation from an unrelated donor than patients in the FBA1 cohort. Diagnoses were distributed equally except for chronic myeloid leukemia, which was more frequent in the FBA1 cohort (P = .02). No significant differences in disease risk were observed between the 2 cohorts: low-risk, standard-risk, and high-risk diseases were distributed equally between the 2 groups (see Table 1).
Table 1. Patient and Transplantation Characteristics
The cumulative incidence of grade 2 through 4 aGvHD differed significantly between the 2 groups and was 42% (95% CI, 33%-51%) in the FBA2 cohort and 23% (95% CI, 15%-31%) in the FBA2 cohort (P = .002) (Fig. 1, top). This difference persisted when considering grade 3 through 4 aGvHD, for which the cumulative incidence was 23% (95% CI, 15%-30%) and 10% (95% CI, 4%-15%) in the FBA1 and FBA2 cohorts, respectively (P = .006) (Fig. 1, bottom).
The cumulative incidence of cGvHD was 69% (95% CI, 61%-78%) in the FBA1 cohort and 35% (95% CI, 25%-45%) in the FBA2 cohort (P < .0001) (Fig. 2). The extensive form of cGvHD occurred more frequently in the FBA1 cohort compared with the FBA2 cohort (56% vs 23%; P < .0001). cGvHD occurred a median of 121 days after transplantation among all patients, but it occurred significantly later in the FBA2 cohort (FBA1 cohort, 111 days; FBA2 cohort, 148 days; P < .0001).
The OS rate at 3 years was 63% (95% CI, 56%-70%) for the whole patient series; no difference in survival probability was observed between the 2 cohorts (P = .99). Sixty-five patients (52%) in the FBA1 cohort and 77 patients (73%) in the FBA2 cohort were alive at the last follow-up (Fig. 3).
Thirty-two patients died of NRM in the FBA1 cohort compared with 16 patients in the FBA2 cohort (P = .07; chi-square test). The cumulative incidence of NRM was 20% (95% CI, 11%-29%) in the FBA2 cohort and 19% (95% CI, 12%-26%) in the FBA1 cohort (P = .61). The causes of NRM are detailed in Table 2. In total, 29 events of relapse or progression were observed in the FBA1 cohort compared with 27 events in the FBA2 cohort, and the cumulative incidence was 19% (95% CI, 12%-26%) and 30% (95% CI, 20%-40%), respectively (P = .09). We observed a correlation between the patient's disease risk group and relapse/progression incidence in the multivariate analysis, as expected: the hazard ratio was 3.00 (95% CI, 1.06-8.47) for high-risk versus low risk disease (P = .04), and the hazard ratio was 1.56 (95% CI, 0.58-4.15) for standard-risk versus low-risk disease (P = .37).
Table 2. Causes of Nonrelapse Mortality According to Treatment With 1 Day or 2 Days of Combined Fludarabine, Busulfan, and Antithymocyte Globulin
Table 3 provides the adjusted hazard ratios for the FBA2 versus FBA1 regimens for each outcome of interest, confirming that the ATG dose was an independent predictor of the occurrence and severity of both aGvHD and cGvHD, but it did not have any significant impact on OS, NRM, or relapse/progression.
Table 3. Multivariate Hazard Ratios for Graft-Versus-Host Disease and Relapse/Progression According to Treatment With 1 Day or 2 Days of Combined Fludarabine, Busulfan, and Antithymocyte Globulina
In this retrospective analysis of 2 sequential cohorts of patients who received treatment over a 10-year period at a single institution, we observed a significant reduction in the incidence of aGvHD and cGvHD without loss of disease control after the introduction of 1 more day of ATG in the conditioning regimen for RIC transplantation. We recently reported our previous experience with an ATG dose of 2.5 mg/kg15 in the context of RIC transplantation, in which patient outcomes still were impaired by the occurrence of long-lasting, extensive cGvHD; for this reason, we decided to modify the conditioning regimen by adding 1 more day of ATG in an attempt to lower the GvHD incidence. The results obtained from the current study confirmed our hypothesis, because the incidence of both aGvHD and cGvHD was significantly lower in the FBA2 cohort. This is particularly notable because the FBA2 cohort was both older and had a higher proportion of unrelated donors.
Our results are in line with those reported in other studies that aimed to identify the optimal ATG dose in fludarabine/busulfan-based conditioning regimens.4, 25-27 A series of 75 patients who received an “FBA” regimen using a total ATG dose of 6 mg/kg was reported by Hamadani et al,27 who observed a subsequent incidence of grade 2 through 4 aGvHD, grade 3 and 4 aGvHD, and cGvHD of 34%, 12%, and 48%, respectively. Mohty et al4 analyzed the incidence of GvHD among 101 patients who received different ATG doses (2.5-10 mg/kg) and reported that aGvHD was correlated significantly with ATG dose. A recent study25 reported incidences of grade 2 through 4 aGvHD, grade 3 and 4 aGvHD, and cGvHD of 23%, 10%, and 45%, respectively, among 207 patients who received a myeloablative conditioning regimen using fludarabine plus busulfan and a total ATG dose of 4.5 mg/kg. Recently, Bashir et al28 compared a total ATG dose of 4.5 mg/kg with a total dose of 7.5 mg/kg among 20 patients who underwent RIC-AlloSCT from unrelated donors after a fludarabine-melphalan conditioning regimen. The treatment-related mortality in that study was 20% in the 4.5 mg/kg arm and 60% in the 7.5 mg/kg arm, thus favoring the use of 4.5 mg/kg in patients undergoing RIC-AlloSCT from unrelated donors.
A large retrospective analysis recently questioned the beneficial impact of ATG,29 indicating inferior outcomes with T-cell depletion (anti-CD52 or ATG) compared with non–T-cell depleted grafts; our results appear to contradict this finding, but it is likely that differences in ATG dose and brand are responsible. In fact, in the Center for International Blood and Marrow Transplant Research (CIBMTR) study,29 the median ATG dose was higher than ours (7 mg/kg), and 27% of patients in that study received horse ATG. Inferior disease-free survival in the ATG group was mainly because of the high relapse incidence (49% vs 23% in our study), which may be explained by the higher ATG dose used, although other factors, such as busulfan dose (≤8 mg/kg), use of other alkylating agents during conditioning, and pretransplantation therapies, also may be evoked.
Caution is needed when interpreting NRM and relapse results uniquely with regard to ATG dose changes; indeed, we cannot exclude the contribution of disease control and/or toxicity by the intravenous busulfan formulation, which was used mostly with the FBA2 regimen. Unfortunately, it has not been possible to analyze the ATG dose and busulfan formulation separately. It is reasonable to speculate that less variability in the pharmacokinetics associated with intravenous busulfan may allow better disease control and less organ toxicity, and this latter effect possibly may influence relapse/progression and NRM incidence; however, heterogeneity of diseases makes any further interpretation difficult, especially with respect to relapse or progression. A single-center, matched-pair analysis16 and, more recently, a long-term analysis of a prospective randomized trial30 and a retrospective registry-based analysis31 have confirmed that the addition of ATG does not increase the incidence of relapse and, thus, does not affect OS or progression-free survival.
In conclusion, our results demonstrate a reduced incidence of both acute and chronic GvHD with the FBA2 regimen compared with the FBA1 regimen, and this reduction was not accompanied by a significantly higher relapse incidence or by any significant differences in other transplantation outcomes. The current data indicate that 5 mg/kg of ATG provides satisfactory T-cell depletion in this type of RIC transplantation; the FBA2 regimen may be a platform for the development of strategies aimed at improving disease control, such as intensification of the busulfan dose in the conditioning regimen32 and/or the addition of post-transplantation therapeutics in high-risk patients.33-35
We thank all personnel working at the Hematology Department and Cellular Therapy Unit the Paoli-Calmettes Institute for their remarkable contribution to patient care and the assistance provided to the patients' families.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
Dr. Blaise has contributed to educational events sponsored by Genzyme.