Impact of in vivo T cell depletion in HLA-identical allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission conditioned with a fludarabine iv-busulfan myeloablative regimen: a report from the EBMT Acute Leukemia Working Party

The main challenge of allo-SCT in AML and other hematological malignancies remains to limit organ life-threatening toxicity while preserving the GVL effect and patients’ quality of life by avoiding severe chronic GVHD. The GVHD and relapse-free survival composite end point is becoming an important end point to improve in allo-SCT [21, 23]. In this objective, while improvements in terms of tolerability of the transplant process have been shown in the last two decades by the development of reduced intensity and toxicity conditioning regimens and by improvement of supportive care including management of infections [24], the increased use of PBSC grafts, reaching 70% of stem cell grafts used in Europe nowadays [25], is also associated with higher incidence of severe cGVHD even with HLA-identical sibling donors [11, 12, 26], thus potentially impairing GRFS. The Flu-ivBu4 RTC associated with PBSC graft has been reported as effective than conventional Bu-cyclophosphamide MAC regimen but with reduced short- and long-term non-relapse mortality in AML patients transplanted in CR1 with an HLA 10/10-matched related or unrelated donor [8‐10]. However, the incidence of chronic GVHD with PBSC (80%) grafts from HLA-identical sibling donors, following such RTC in the absence of ATG, remains high with 68% overall cGVHD and 42% extensive cGVHD incidences at 4 years post-transplant reported recently by the Spanish Cooperative Transplant Group [7]. The current study was limited to AML in CR1 transplanted with HLA-identical donors to reduce bias due to donor type and disease status on transplant outcomes. As reported in other contexts of allo-SCT [14, 16, 17, 27‐30], we confirm in the present study that the addition of intermediate dose of thymoglobulin (median, 5 mg/kg) significantly reduces, after adjustment to other factors, the risk of developing cGVHD (Cox HR = 0.46, p = 0.0001). Compared to patients not receiving ATG, those transplanted with ATG, despite having received more frequently PBSC, had a reduction of 2 years of cumulative incidence of the overall cGVHD from 52 to 31% (p = 0.00026) and that of extensive cGVHD from 26 to 8% (p < 10⁻⁴). Such reduction of cGVHD incidence was not associated with reduced anti-leukemic control since use of ATG did not impact relapse incidence in this series of AML transplanted in CR1 in both univariate and multivariate analyses. Most patients had received a thymoglobulin dose of <6 mg/kg, so we could not analyze the impact of the ATG dose on outcomes. However, these results are in line with preserved GVL effect despite the addition of low or intermediate doses of ATG in the context of allo-SCT for AML performed with MRD and MUD following conventional MAC [31, 32] and RIC [29, 30, 33, 34], in contrast with increased risk of relapse with doses of thymoglobulin >10 mg/kg [28].

By contrast, as reported by others [15, 17, 30], we did not observe protective effect of such doses of ATG against acute GVHD. Actually, in the context of allo-SCT performed with PBSC from matched related or unrelated donors, use of low doses of ATG (2.5 mg/kg of thymoglobulin) was associated with an increased risk of aGVHD [34, 35]; a reduction of the incidence of aGVHD has been observed with thymoglobulin doses starting at 5 mg/kg, although higher doses (≥7.5 mg/kg of thymoglobulin) were associated with increased risk of mortality from infections and relapse in both MAC and RIC settings [28, 36]. The optimal dose of thymoglobulin in both RIC and MAC seems therefore to be about 5 mg/kg. At such intermediate dose of ATG, as described by others [14, 15, 30, 37], we did not observe difference of infection-related mortality between the ATG and no-ATG groups, despite potential increased frequency of viral EBV reactivation manageable by viral load monitoring and preemptive use of rituximab [30, 37].

Altogether, our results show that the addition of intermediate dose of ATG to the Flu-ivBu4 RTC represents an independent factor associated with improved GRFS, as defined by Ruggeri et al. for registry-based studies [21], allowing a probability of being alive without disease and without significant cGVHD at 2 years after allo-SCT in 60% of the patients transplanted for AML in CR1 with a sibling donor, compared to 40% of those not receiving ATG. Although GRFS is not routinely analyzed up to now, a 2-year 60% GRFS compares favorably to 40% GRFS at 3 years reported after allo-SCT performed for AML transplanted in CR1 (79%) or CR2 (21%) with MAC (61%) or RIC and PBSC (82%) from HLA-matched related (55%) or unrelated (45%) donors within the EBMT registry [21], and to 25% at 1 year reported with PBSC HLA-sibling donor allo-SCT by the Minnesota Group [23].

Use of ATG in our series appears also as an independent factor associated with improved LFS and OS, mainly due to reduced incidence of overall and extensive cGVHD leading to a trend towards increased late NRM in the absence of ATG. Although a center effect was observed in the incidence of cGVHD and NRM, possibly due to preferential used of ATG and of prophylactic donor lymphocyte infusion in some centers, we did not detect a center effect on LFS, OS, and GRFS. The only other factor associated with worse survivals was recipient age above 50 years, due to higher NRM, as reported by others with such conditioning regimen [7].