Introduction

Allogeneic stem cell transplantation (allo-HSCT) is a potentially curative treatment for patients diagnosed with high-risk hematological disorders [1, 2]. However, it remains a challenging treatment with high morbidity and mortality [3].

Posttransplant cyclophosphamide (PTCy) decreases the risk of GVHD and graft rejection by inducing apoptosis of rapidly proliferating early alloreactive T cells and sparing regulatory T cells [4,5,6]. The efficacy of PTCy in combination with other immunosuppressive drugs preventing GVHD has been reported by several studies [7,8,9,10]. PTCy effectiveness using different graft sources was explored in a retrospective registry-based analysis and PTCy in combination with bone marrow sources resulted in lower rates of GVHD but higher cumulative incidence of relapse [11].

Anti-thymocyte globulin (ATG) is a commonly used agent for GVHD prophylaxis [12]. ATG depletes donor and recipient T cells reducing the risk of graft rejection and GVHD [12,13,14,15,16,17,18]. The efficacy of ATG in controlling cGVHD has been demonstrated on six randomized clinical trials [12,13,14,15,16,17,18]. In five out of six studies [12,13,14,15,16,17], the use of ATG did not affect survival or relapse rates. However, there remains a concern that the use of ATG may reduce graft-versus-leukemia effect, and impair passive transfer of memory T cells that reconstitute early immunity [18,19,20].

In October 2015, RIC regimen combined with rabbit-ATG, PTCy and cyclosporine (CsA) for the prevention of GVHD was established as the institutional standard of care in the Princess Margaret Cancer Centre, using peripheral blood stem cell (PBSC) grafts. The hypothesis for this protocol was that combining ATG and PTCy in T-cell replete-PBSC transplants would minimize GVHD rates without affecting engraftment and relapse rates. Peripheral blood was the selected donor source in order to reduce relapse rate, reduce time to engraftment, and thereby reduce infection rates.

The aim of the study is to report the outcome for all patients who underwent allo-HSCT with this protocol from October 2015 to May 2018 in the Center, and to explore the impact of donor type on GVHD and survival rates. There have been published reports documenting the safety of this protocol in MUD and haplo-HSCT [21,22,23,24,25,26]. However, this study includes a larger and more heterogeneous cohort of patients with an updated follow-up.

Patients and methods

Patient selection

Between October 2015 and May 2018, 270 adults diagnosed with hematological malignancies underwent allo-HSCT in the Center. All patients received reduced conditioning regimen and ATG–PTCY–CsA for GVHD propylaxis and were included in the analysis. General eligibility criteria for allo-HSCT are summarized in the Supplementary material. All patients provided informed consent prior allo-HSCT.

The present study was conducted in accordance with the Declaration of Helsinki, and was approved by the Ethics Committee at the Princess Margaret Cancer Centre, Toronto, Canada.

Donor selection and source of graft

High resolution molecular typing for HLA classes I (A, B, C) and II (DR, DQ) was performed for recipients and donors. Donor-specific antibodies (DSA) were assessed once the donor was identified, and repeated prior to transplant in all recipients. A MRD was preferred as first choice. In the absence of a MRD, a 10/10 MUD was selected as the second choice. For recipients who did not have a suitable MRD or MUD, alternative stem cell sources such as 9/10 HLA mismatched unrelated donors, and HLA haploidentical family members were considered.

All recipients received T-cell replete-PBSC grafts. The median CD34 + cell dose requested for infusion was 4 × 10^6 CD34+/kg for MRD, 10/10MUD and 9/10MUD, and 6 × 10^6 CD34+/kg for haploidentical donors.

Conditioning regimen, posttransplant immunosuppression, and supportive care

The present conditioning regimen and GVHD prophylaxis are summarized in Fig. 1.

Fig. 1
figure 1

Conditioning regimen and GVHD prophylaxis diagram.

Full size image

RIC regimen comprised fludarabine (Flu) (30 mg/m2/day × 4 days from days-5 to -2), busulfan (Bu) (3.2 mg/kg/day × 2 days from days-3 to -2), and 200 cGy of TBI administered on day-1. GVHD prophylaxis consisted of rabbit- ATG (Thymoglobulin; Genzyme-Sanofi, Lyon, France), administered in doses of 0.5 mg/kg on day -3, 2 mg/kg on day-2 and 2 mg/kg on day-1 (total dose of 4.5 mg/kg), PTCy (50 mg/kg/day on days +3 to +4), and CsA, started at a dose of 2.5 mg/kg q12h IV on day +5 adjusted to achieve a therapeutic level of 200-400 ng/L, d. CsA tapering was started around day +45 to +60 for all patients without GVHD. Pharmacokinetics levels of ATG and cyclophosphamide were not routinely measured.

Definitions and disease monitoring are reported in the Supplementary material

Statistical method

Data was collected through retrospective chart review and updated in January 2019. Categorical variables were presented as counts and percentages. Continuous variables were presented with medians with ranges. Overall survival (OS), relapse-free survival (RFS) and GVHD-free/RFS survival (GRFS) were calculated using the Kaplan–Meier product-limit method and the impact of variables was assessed using Log-rank test. Non-relapse mortality (NRM) was estimated using the cumulative incidence method considering relapse as a competing risk. Cumulative incidence of relapse (CIR) was estimated considering death without relapse as competing event. The impact of variables on competing event analysis was compared using Gray’s test. Cumulative incidence of GHVD was estimated considering death and relapse as competing events. A multivariable Cox proportional hazards regression model was used to determine the effect of factors that were found to be statistically significant on the univariate analysis, and other factors considered clinically important. The multivariable analysis for OS and RFS included the following independent variables as control variables: Karnofsky performance status (70–80% vs 90–100%); disease risk index (low moderate risk vs high very-high risk); and donor type (all four donor types were included).

All p-values were two-sided and for the statistical analyses p < 0.05 was considered to indicate a statistically significant result. Statistical analysis was performed using the version 9.4 of the SAS system for Windows (Copyright© 2002–2012 SAS Institute, Inc., Cary, NC).

Results

Patient characteristics

Patient characteristics are detailed in Table 1. Median age was 58 years (range: 18–74.5 years), and 154 (57%) of patients were males. Ethnicity and race were not recorded. Acute myeloid leukemia (AML) was the most prevalent diagnosis representing 53% of the entire cohort, followed by myelodysplastic syndrome (18%) and myelofibrosis (10%). Among patients diagnosed with AML, 26 (18%) had complex/high risk cytogenetic diagnosis. One hundred and twenty-eight (88.5%) recipients were in first complete remission (CR), 12 (8.3%) in second CR, and one in third CR. Three (0.2%) recipients were not in remission prior allo-HSCT. The hematopoietic cell transplantation-specific comorbidity index (HCT-CI) and Karnofsly performance status (KPS) were assessed routinely at the time of the pre-transplant assessment. Median HCT-CI score was 1 (range: 0–7). Seventy-one (26%) of patients had a HCT-CI score ≥3, and 44 (16%) of recipients had a KPS ranged between 70 and 80%. Median follow-up in all 270 patients was 12.7 months (range 0.6–39).

Table 1 Patient and donor characteristics.
Full size table

Disease risk index (DRI) was the prognosis tool used to stratify disease-risk prior transplantation. Fifty-five (20%) patients had high (52) and very high [3] DRI prior allo-HSCT. Among these 55 patients, the median age was 61 (range: 20–74) and the baseline diagnoses were as follows: AML in 25 (45.4%) cases (22 patients had AML with complex cytogenetics in CR prior allo-HSCT and three recipients were not in CR), myelodysplastic syndrome with complex karyotype in 25 (45.5%) cases, acute lymphoblastic leukemia in second CR in five (9%) patients. Seven (12.7%) patients had an HCT-CI score ≥ 3, and 12 (21.8%) a KPS between 70 and 80%. Nine (16.4%) patients received grafts from MRD, 25 (45.5%) from 10/10 MUD, 7 (12.7%) from 9/10 MUD and 14 (25.5%) from haploidentical sibling donors.

Transplant characteristics

Transplant details are shown in Table 2. Cytokine release syndrome (CRS) was observed in five cases and appeared between day 0 and +1. It presented with fever in all cases, only grade 1 and 2 were documented, and were managed supportively in all cases. Engraftment syndrome occurred in 13 (5%) recipients, and all of them were successfully treated with steroids. Day 60 chimerism was assessed in 233 (86%) recipients; and in 198 (85%) cases was higher than 95% donor cells.

Table 2 Main post-transplant information and outcomes.
Full size table

Graft failure (GF) was documented in 23 (9%) patients. Four (2%) recipients had primary GF and 19 (7%) secondary GF. Patients who underwent 9/10 MUD (21%) and haploidentical donor (17%) allo-HSCT, had a higher incidence of GF than those recipients who received MRD (2%) and MUD (3%) grafts (p < 0.0001). DSA were reassessed and in all cases were negative. Nine recipients underwent second salvage allo-HSCT and one recipient received donor lymphocyte infusions. Seventeen (74%) patients with GF died, and in most of them infection was the immediate cause of death.

Median length of transplant hospitalization was 29 days (range 20–180). Two hundred and forty-seven (92%) recipients had mucositis; 15 (8%) of them were grade 3–4, and 11 required a short course of total parenteral nutrition. Neutropenic fever occurred in 221 (82%) patients and 108 (40%) of them showed positive blood cultures. Forty-four (16%) recipients developed veno-occlusive syndrome (VOD). Moderate VOD was documented in 18 (7%) recipients and severe VOD in one. All cases resolved with fluid restriction, diuretics, dopamine intravenous infusion, and defibrotide (in one case).

Infectious complications

Cytomegalovirus (CMV) reactivation was observed in 156 (58%) patients. CMV titers were monitored by quantitative PCR in plasma samples during the first six months of follow-up or longer if on immunosuppression. The cut-off value for test positivity was >200 copies of CMV DNA per milliliter of plasma. Any positive result above that level was considered a CMV reactivation and preemptive treatment was initiated. First line of treatment was oral valgancyclovir or intravenous ganciclovir until viremia clearance. Eleven (4%) recipients developed CMV disease. First line of treatment was intravenous ganciclovir and intravenous immunoglobulin (Ig) replacement in those recipients with CMV pneumonitis, and in those with low IgG levels. Eight cases resolved with intravenous ganciclovir therapy.

Epstein–Barr virus (EBV) reactivation was documented in 171 (63.6%) recipients. The cut-off value for test positivity was >600 copies of EBV DNA per milliliter of plasma. Any positive result over that level was considered EBV reactivation. Preemptive treatment with rituximab for high EBV-DNA-emia was not routinely done. Twenty-three (8.5%) patients developed probable (6 patients) or biopsy proven (17 patients) posttransplant lymphoproliferative disease (PTLD). PTLD was managed by tapering immunosuppression and weekly rituximab for a maximum of four doses. One recipient required second line of treatment with R-CHOP chemotherapy. Twenty-one (91%) recipients achieved clinical responses with regression of affected nodal tissue (if any) and clearance of EBV-viremia. Two patients died secondary to PTLD.

Fifty-seven (21%) recipients developed BK hemorrhagic cystitis. Among these 57 patients, only four recipients had grade 3 and 4 BK cystitis, presenting significant urinary retention caused by macroscopic hematuria with clots. These four recipients required prolonged hospitalization, intravesical cidofovir, and one of them IV cidofovir. No patient developed acute kidney injury. Three of these four recipients had clinical response to treatment while the fourth developed secondary graft failure concomitantly with multiple infections.

Other viral infections were diagnosed in 80 (30%) patients. Proven or probable invasive fungal infection was identified in 21 (8%) recipients during the posttransplant course. Nine of these 21 patients died with an ongoing fungal infection. Three of these nine patients had antecedent graft failure.

Graft versus host disease

Detailed results are summarized in Table 3 and Supplementary Fig. 1. Overall, aGVHD was identified in 94 (35%) patients. However, grade III and IV aGVHD was diagnosed in 12 (4%) and 1 (1%) recipient, respectively. Nine (3%) cases had steroid refractory aGVHD. Fifty-three (20%) recipients were diagnosed with cGVHD. Moderate cGVHD was documented in 29 (11%) recipients and severe cGVHD in 7 (3%). Five patients with moderate/severe cGVHD had steroid refractory cGVHD. Four patients died secondary to severe GVHD: one due to acute gut GVHD, one secondary to steroid refractory acute gut and liver GVHD, one due to idiopathic pneumonia syndrome, and one secondary to chronic severe lung GVHD.

Table 3 Graft versus host disease information.
Full size table

Figure 2 shows the cumulative incidence of GVHD according to donor type. The difference between the cumulative incidence of grade II–IV aGVHD (p = 0.440) and moderate–severe cGVHD according to donor type was not statistically significant (p = 0.738). Grade III–IV aGVHD was not analyzed according to donor type because of the low number of events.

Fig. 2
figure 2

Overall Survival, Progression Free Survival in all patients and stratified as per donor type.

Full size image

Outcome

Main outcome information is summarized on Table 2 and on Table 4, and Figs. 2, 3 and 4. With a median follow-up of 12.7 months (range 0.6–39), 105 (39%) patients died and 62 (23%) relapsed. Main causes of death were infection (15%) and relapse (14%).

Table 4 Risk factors for overall survival and relapse-free survival.
Full size table
Fig. 3
figure 3

Cumulative incidence of NRM and GVHD stratified as per donor type.

Full size image
Fig. 4
figure 4

GVHD free and Relapse Free Survival for all patients and stratified as per donor type.

Full size image

Results from the univariate and multivariate analysis are provided in the Table 4. Donor type had a significant impact on OS, RFS, NRM, CIR, and GRFS in the univariate analysis. Karnofsky performance status between 70 and 80% (HR 1.92 (95% 1.25–2.9); p = 0.003) and high/very high DRI prior allo-HSCT (HR 1.79 (95% CI 1.21–2.66); p = 0.003) were significant predictors for worse RFS in the multivariate analysis. KPS between 70 and 80% (HR 1.99 (95% CI 1.3–3.04); p = 0.008) and higher DRI (HR 1.99% (95% CI 1.3–3.04); p = 0.001) were also significant risk factors for worse OS. The use of 10/10 matched unrelated donors provided a protective effect in comparison with other donor sources for OS and RFS.

A sub-analysis among the 55 patients with higher DRI was done to explore this cohort in detail. Among these 55 patients, 33 (60%) died and 20 (36.4%) relapsed. Among these 20 recipients who relapsed, 80% died during the follow-up. Two years OS and RFS were respectively 31.2% (95% CI 17.8–45.6) and 29–8% (95% CI 17.1–43.5). NRM and CIR at 2 years were respectively 31% (95% CI 18.1–45) and 39.2% (95% CI 25.8–52.5).

Discussion

Allo-HSCT using un-manipulated-T-cell replete-PBSC grafts with ATG, PTCy, and cyclosporine, results in low rates of GVHD. A pilot report with data from patients undergoing unrelated and haploidentical allo-HSCT was reported [21,22,23, 26], and the efficacy of this protocol in AML and myelofibrosis patients has been explored [24, 25]. This study provides additional information from a more heterogeneous cohort and with an updated follow-up.

The combination of fludarabine, low-dose of busulfan, and TBI (200 cGy) is a well-tolerated regimen and induces sufficient immunosuppression to allow consistent engraftment. Transplant age was not a significant risk factor for survival. These results support that age may have less of an impact on defining eligibility criterion for allo-HSCT, and they underline the importance of pre-transplant performance status [27]. Relapse rates documented in the study were 23%. The feasibility of RIC allo-HSCT for high-risk hematological malignancies is well documented [28,29,30,31]. Those patients with high/very-high DRI had a significant higher risk of mortality in our study. DRI is a valuable prognostic tool [32] that can be used to identify a very high-risk population. Posttransplant disease control strategies could be considered for these patients, such as prophylactic donor lymphocyte infusions or maintenance chemotherapy.

Survival rates differed among donor type. Those recipients who received 10/10 MUD grafts had higher survival rates than those who received grafts from MRDs. This results contrast with findings from other studies [33], and may be partly explained by the smaller number of patients who received MRD grafts. However, with the evidence reported here, we hypothesize that MRD transplants would require less immunosuppression to control GVHD. The optimal dose of immunosuppression for each donor source is not well established. Further investigations are needed to determine the best dose of ATG, and to determine if it should be used in MRD transplant. Survival rates were comparable between patients who received 9/10 MUD and haploidentical donor grafts. Taking into account these results, and easier availability and lower costs, haploidentical donors may be preferred to 9/10 MUD for allo-HSCT, using the present conditioning regimen.

The combination of ATG, PTCy, and CsA reduces GVHD rates when PBSC grafts are used. PTCy may also potentiate the antineoplastic activity of the conditioning regimen decreasing relapse rates in patients with higher risk diseases [34]. PTCy decreases the risk of GVHD and graft rejection by inducing cell apoptosis of rapidly proliferating early alloreactive T-cells sparing regulatory T-cells and preserving nondividing hematopoietic stem cells [4,5,6]. The addition of ATG results in faster achievement of donor chimerism and increases control of GVHD [12,13,14,15,16,17,18]. A remarkable finding of the study is the absence of statistically significant differences in clinically relevant GVHD between donor types, suggesting that the new combination of ATG, PTCy, and CsA overcomes the HLA-barrier and its effect on GVHD.

Primary GF rates were comparable with published data; however secondary GF rates were higher [35, 36]. Observed rates of secondary GF, especially in mismatched allo-HSCT, could be secondary to the protracted T-cell depletion induced by this GVHD prophylaxis, leading to delayed donor T-cell engraftment. Unfortunately, T-cell reconstitution data are not available.

Infection was the most frequent cause of death. Despite the use of T-cell replete-grafts, observed rates CMV were high. The use of novel CMV prophylactic agents such as letermovir could be explored [37]. We attribute the high proportion of EBV-R and PTLD observed in the study to the use of ATG for T-cell depletion [38,39,40]. The selection of EBV matched donors and the use of pre-emptive rituximab could be considered when this GVHD prophylaxis is used [41, 42]. To reduce infectious toxicity and with the hypothesis of preserving the GVHD-preventing properties of this regimen, we have decided to reduce the dose of ATG to 2 mg/kg from current dose of 4.5 mg/kg for our patients. This modification may stimulate the recovery of donor T-cells and thereby reduce secondary GF rates.

The main limitations of this study are the retrospective design and the heterogeneity of the sample size. Future research will focus on differences in the effectiveness of the new treatment across different hematological disorders. Prospective multicenter clinical trials could be considered to validate the results in different transplant centers.

In conclusion, the evidence reported in this paper indicates that the combination of ATG, PTCy, and CsA is effective in reducing both acute and chronic GVHD in all donor types. Further optimization of the present protocol is needed to reduce NRM and disease relapse, for example using posttransplant disease control strategies for the very-high risk group of patients.