LAUSR

TO THE EDITOR: Acute Graft versus Host Disease (aGvHD) after allogeneic hematopoietic stem cell transplant (allo-HSCT) may depend on known factors [1] such as: human leucocyte antigen (HLA) and/or sex mismatch, source of stem cell, disease stage at diagnosis, CMV match-risk and higher regulatory T cells (Tregs) graft content which seems to be correlated with onset and severity of aGvHD through protective effects both in related [2] and in unrelated donor [3] peripheral blood stem cell transplantation (PBSCT). On the contrary, the Tregs reduction seems to be associated with lower chronic GvHD (cGvHD) incidence [4, 5] in allogeneic PBSCT (allo-PBSCT). Of note, this correlation was demonstrated in PB at the time of cGvHD onset, while no evaluation has been performed on allo-graft content at the time of infusion so far. In our previous report focusing on aGvHD [6], we had already shown, by univariate analysis, how cumulative incidence (CI) rate of cGvHD does not correlate with higher graft Tregs (gTregs) content; however, a trend toward a positive association between gTregs and cGvHD incidence has emerged, thus suggesting a possible different impact on aGvHD and cGvHD (protective effect on grade II-IV aGvHD while favoring cGvHD). The aim of this sub-analysis of our previous report [6] was to define the possible correlation between gTregs content (as graft CD3/Tregs ratio) and the cGvHD occurrence in allo-PBSCT. Therefore, we analyzed the cGvHD incidence rate together with patient, donor, and allo-transplantation-related factors in addition to gTregs by extending the observational time (24 months from the last enrolled patient) and adopting a multivariate model. Patients and methods have already been reported [6], Tregs flow-based evaluation has been further specified in supplementary (S) materials (paragraph flow cytometry and Fig. S1). Briefly, we prospectively enrolled 94 patients (18–65yy) affected by acute myeloid leukemia (AML) or acute lymphoblastic leukemia (ALL) in complete remission (CR) who underwent matched related donor (MRD) or matched unrelated donor (MUD) PBSCT, with a myeloablative conditioning regimen. The CD3/Tregs ratio (gCD3/TregsR) was calculated for each PB-allograft obtaining its receiving operating curve-threshold value (≥70) associated with grade II-IV aGvHD onset. The same gCD3/TregsR value was also used in the current cGvHD evaluation to subdivide the study population in two groups, low ratio (LR, gCD3/TregsR <70) and high ratio (gCD3/TregsR≥70), thus allowing us to analyze the conclusive effect of gCD3/TregsR on both aGvHD and cGvHD. All patients received PBSC grafts from an HLA-identical sibling donor (n 35, 37%) or an unrelated HLA-identical donor (n 59, 63%). DNA-based HLA typing of donor and recipient was done using high resolution (four digits) for HLA-A, -B, -C, -DRB1, and -DQB1. In unrelated transplantations (n 59, 63%), an antigenic (9 out of 10) and allelic mismatch was documented in 8 (16%) and 4 (7%) donor-patient pairs, respectively. 65% of the patients were transplanted in CR1, and 35% in CR ≥ 2. The GvHD prophylaxis strategy was mostly anti-thymocyte globulin (ATG) -based (80%) and cyclosporine+methotrexate was used in 95% of the whole study population. Patients belonging to LR group did not differ from the high ratio one in terms of patient age, donor-recipient CMV serostatus and gender combination, donor type, disease type and stage at transplantation and HLA-match [6]. The overall global severity score and target involved organs of the cGvHD case-series are reported in Table S1. Cellular allo-graft composition in patients with and without cGvHD was also evaluated (Table S2). The 3-year CI rate of cGvHD according to any (18 events), mild (1 event), moderate (10 events) and severe (7 events) global severity score was 40 (Fig. S2a), 2%, 29% and 16% (Fig. S2b), respectively. A two-year statistically significant difference of cGvHDCI rate was found between LR and high ratio (37 vs 15%, p= 0.021; Fig. S3a), with and without grade II–IV aGvHD (54 vs 17%, p= 0.045, Fig. S3b) and HLA-matched and HLA-mismatched pairs (14 vs 81%, p < 0.001, Fig. S3c) patients’ group, respectively. The cGvHD-CI rate was higher for patients with a history of grade III-IV aGvHD if compared with those with a history of grade I–II aGvHD (Hazard Ratio (HR)= 0.16; CI 95%: 0.03–0.91, Fig. S4). The 3-year CI rate of relapse in patients with and without cGvHD was of 15% and 33% (p= 0.11, Fig. S5), respectively. The distribution of infections has been shown in Table S3. The 100-day CI rate of bacterial, viral and fungal infections was of 25%, 15% and 5%, respectively (Fig. S6a). The CI-rate difference of infections between patients belonging to the LR and the high ratio group has been shown in Fig. S6b, c, d. No correlation between type and stage of disease, low CMV risk, female donor/male recipient, ATG usage, type of donor, recipient age and cGvHD (or aGvHD) incidence was documented (Table 1). In multivariate analysis, HLA mismatch (antigenic and/or allelic) and LR group were correlated with both cGvHD (HLA-mismatch: HR= 5.33; 95% CI:1.65–17.18, p= 0.005; LR group: HR= 2.99; 95% CI:1.05–8.29, p= 0.04) and aGvHD (HLA-mismatch: HR= 4.43; 95% CI: 1.63–12.09, p= 0.04; LR group: HR= 0.25; 95% CI: 0.08–0.77, p= 0.016) incidence rate (Table 1). When excluding patients without any grade aGvHD, the positive impact of LR on cGvHD was confirmed