Anti-thymocyte globulin (ATG) mitigates graft vs host disease (GVHD) risk in patients undergoing allogeneic hematopoietic cell transplantation (HCT). Due to T cell depletion there remains a concern that ATG administration may be associated with an increased risk of malignancy, relapse, and infection in recipients of reduced intensity conditioning (RIC). It was hypothesized that ATG infusion early in the course of conditioning will promote rapid immune reconstitution because of lower levels at the time of graft infusion and will thus help to improve clinical outcomes in RIC.

Rabbit ATG (Thymoglobulin, Sanofi Aventis) was administered from day (d) -9 to -7 to HLA matched-unrelated (MUD; 5 mg/kg in divided doses) and -related (MRD; 3.5 mg/kg) donor transplant recipients conditioned with Fludarabine and Melphalan (ATG -9 cohort; N=36). Immune reconstitution and clinical outcomes were compared with a historical control group of patients who received the same doses of ATG from d-3 to -1 (ATG -3 cohort; N=28). Standard GVHD prophylaxis with calcineurin inhibitor and antimetabolite was administered, with CMV and EBV monitoring. ATG -9 cohort had more, MUD recipients 80% vs. 64%; myeloid malignancy (AML, MDS, MPD) 83% vs. 35%. Age (56 vs. 57), graft type (97 vs. 92% PBSC) and CD34+cell dose infused (4.8 vs 4.7 E6/KG) were similar.

Immune reconstitution was uniformly superior in ATG -9 cohort, with significantly higher absolute monocyte counts (AMC) at d30, 60 and 90 (P<0.001), as well as higher donor derived CD3+ (ddCD3+) and CD3+/8+ cell counts at d60 and 90 (P<0.01), and CD3+/4+ cells at d90 (P=<0.05). T cell - monocyte interactions were modelled as 2 dimensional vectors in the immune phase space, (Figure 1) with a consistently higher vector magnitude observed in ATG-9 cohort (P<0.01). Rate of T cell reconstitution was determined by calculating the derivative of ddCD3+ cell count as a function of time (dT/dt) post-transplant (Figure 2) and was generally higher in the ATG-9 cohort, particularly at d60. With a median follow up of 14.9 months in the ATG-9 cohort, and 47.2 months in the ATG -3 cohort, there is a non-significant trend for improved survival (72.2% vs. 46.4% at 2 years) and relapse (19.4% vs. 35.7% at 2 years) in the ATG -9 cohort. TRM and acute GVHD were similar, with a trend towards greater risk for chronic GVHD in the ATG -9 cohort, albeit of a lower severity.

Given the relatively low number of patients in each cohort, the effect of immune reconstitution on clinical outcomes was evaluated in the pooled population. Survival was improved in those with AMC and ddCD3+ cell counts >200/µL at d60 (P=0.004 & 0.016 respectively), and in patients with T cell - monocyte vector magnitude > median (577.48/µL) at that time (P= 0.008), as well as in those with a calculated dT/dt > median (1.96 cells/µL/day) at d45 (P=0.047). The latter was also associated with a reduction in relapse rate (P=0.04), as was ddCD8+ cell count >145/µL at d60 (P=0.04). Acute GVHD risk was increased when dT/dt was >median (7.60 cells/µL/day) at d15 (P=0.0095), and correspondingly with T cell - monocyte vector magnitude > median (1033.3/µL) at d30 (P=0.017).

In conclusion, this retrospective study demonstrates that equal doses of ATG administered earlier (d -9 to -7 as opposed to d-1 to -3) during conditioning yield more rapid and robust immune reconstitution. Monocyte and ddCD3+ cell recovery kinetics have a favorable impact on survival and relapse risk following HLA matched HCT. Patients at risk for acute GVHD may be identified as early as d15 post HCT by analyzing ddCD3+ cell reconstitution kinetics. Different ATG administration schedules should be studied prospectively with a focus on immune reconstitution kinetics as a determinant of clinical outcomes.


No relevant conflicts of interest to declare.

OffLabel Disclosure:

Rabbit anti-thymocyte globulin for GVHD prophylaxis.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution