Aplastic Anemia Preconditioned with Fludarabine, Cyclophosphamide, and Anti-Thymocyte Globulin

Background

Acquired severe aplastic anemia (SAA) is a life-threatening hematologic disease associated with significant morbidity and mortality. Immunosuppressive therapy and allogeneic hematopoietic cell transplantation (HCT) are effective treatments for SAA [1,2]. It was recently suggested that allogeneic transplantation using bone marrow as a stem cell source from an HLA matched-related donor should be considered first-line therapy for patients with SAA, especially for those patients under the age of 20 years, whereas immunosuppressive therapy is typically used as first-line therapy for older patients (over 40 years of age) with SAA or for younger patients who lack an HLA identical bone marrow donor [3–5]. The outcomes of HCT for treating SAA have improved dramatically over the past 2 decades due to the modification of treatment regimens, improved selection of donors by high resolution HLA matching, and better supportive care [6–9]. Recently, the number of allogeneic peripheral blood stem cell transplantations (allo-PBSCTs) has increased in the majority of centers, due to the convenient collection of grafts and faster engraftment. However, as yet, there is not sufficient data to identify the outcomes of allo-PBSCT in SAA patients, especially for adult SAA patients without matched-related donors.

The traditional preconditioning regimen for SAA patients in HCT is cyclophosphamide (Cy) in combination with anti-thymocyte globulin (ATG), which is named the CA regimen [6]. However, higher incidence of graft rejection or GvHD have been reported with CA regimens, especially among older patients and patients with an unrelated donor [6,7]. Recent data suggest that survival of SAA patients treated with HCT can be significantly improved by using a fludarabine (Flu)-based conditioning regimen [3,10–23]. However, the outcome of PBSCT with Flu-based regimens in SAA patients has not yet been defined, especially for older patients and patients with unrelated donors.

In this retrospective study, we report our experience with 46 adult SAA patients who underwent allo-PBSCT from 2010 to 2017 with Flu in combination with the CA regimen (FCA regimen), and 30 SAA patients who received only the CA regimen from 2006 to 2011 were used as historical controls. Our study demonstrated that the FCA regimen for allo-PBSCT in adult SAA patients significantly improved event-free survival (EFS) and reduced graft rejection, especially for patients with unrelated donor PBSCT.

Material and Methods

PATIENTS AND DONOR SELECTION:

This study cohort included patients with acquired SAA who underwent their first allo-PBSCT from an HLA-identical siblings or from unrelated donors at our department from January 2006 to September 2017. Eligibility criteria were 1) age older than 18 years; 2) diagnosis of acquired SAA; and 3) peripheral blood as a stem cell source. Patients who had significant functional problems in major organs, or congenital disorders were excluded, leaving 76 eligible patients in the cohort. All patients received either FCA regimen (n=46 patients) or CA regimen (n=30 patients).

Clinical information from all patients was collected. Informed consent forms were obtained from all patients before starting the conditioning therapy in accordance with the transplantation procedure. For donor selection, HLA-A, -B, -C, -DQ, and -DRB1 matching was confirmed by a high-resolution molecular method for all patients and donors. Patient eligibility criteria of the FCA group was similar to the historical control group (CA group). Patient characteristics are summarized in Table 1.

CONDITIONING REGIMEN AND GVHD PROPHYLAXIS:

The FCA regimen consisted of Flu (30 mg/m² on days −6 to −3), Cy (50 mg/kg/day on days −6 to −3), and ATG (Fresenius, Germany, 5 mg/kg/day on days −4 to −1). Forty-six patients (61%) received the FCA regimen, whereas the other 30 patients (39%) received a CA regimen, which included Cy (50 mg/kg/day on days −6 to −3) and ATG (Fresenius, Germany, 5 mg/kg/day on days −4 to −1).

GvHD prophylaxis consisted of cyclosporine A and methotrexate. Methotrexate was administered on day +1 after transplantation at a dose of 15 mg and on day +3 and day +6 after transplantation at a dose of 10 mg. Intravenous cyclosporine A (2–3 mg/kg daily) was administered from day −5; then switched to oral drug administration (4–5 mg/kg daily). Cyclosporine A was tapered from month +9 and discontinued by 1 year in patients with no GvHD after PBSCT.

STEM CELL COLLECTION:

All donors were treated with granulocyte colony-stimulating factor (G-CSF, 10 μg/kg daily) for 4–5 days before apheresis (peripheral blood stem cells were collected). The median number of mononuclear cells infused was 7.36×108/kg (range, 3.5–13.47×108/kg) and 6.21×108/kg (range, 2.23–13.57×108/kg) in the FCA group and the CA group, respectively. The median number of CD34+ cells infused was 3.12×106/kg (range, 1.789–9.96×106/kg) and 3.00×106/kg (range, 1.77–6.52×106/kg) in the FCA group and the CA group, respectively. The difference in the number of cells between the 2 groups was not statistically significant (Table 1).

STUDY ENDPOINTS AND DEFINITIONS:

Study endpoints included hematopoietic recovery, graft rejection, GvHD, infection, and mortality. The single nucleotide polymorphism (SNP)-polymerase chain reaction (PCR) of bone marrow aspirates at 15, 30, 60, 90, 180, and 360 days post-PBSCT was used for chimerism monitoring [24].

Primary graft rejection, which we called graft failure, was defined as no engraftment followed by severe neutropenia (absolute neutrophil count [ANC] <0.5×109/L) or the absence of donor cells in the bone marrow or peripheral blood up to day 30 after transplantation. Secondary graft rejection was defined as subsequent loss of ANC to <0.5×109/L and < 5% donor chimerism after neutrophil and/or platelet recovery. Acute GvHD (aGVHD) and chronic GvHD (cGvHD) were graded according to consensus criteria [25,26]. Regimen-related toxicity after transplantation was graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0 [27,28].

Only serious infections were analyzed in this study. Serious infections were defined as infections associated with death or severe clinical compromise, and included shock or organ failure, cytomegalovirus (CMV) end-organ disease, lower respiratory tract infection with respiratory viruses, invasive molds in sinus or lung, or disseminated Aspergillus species infection [29].

STATISTICAL ANALYSIS:

All data were measured from the date of PBSCT to the date of disease progression or death from any event (EFS), the date of death from any cause (overall survival, OS), the date of rejection according to the SNP-PCR of bone marrow (rejection rate [RR] including primary rejection and secondary rejection), and the date of serious infection (infection rate, [IR]). Data on patients were evaluated in May 2018. Actuarial probabilities of survival were calculated using the Kaplan-Meier method, and the log-rank test (univariate) was used to evaluate differences between groups. A Cox proportional hazard regression model was used for the multivariate analysis of prognostic factors affecting survival. SPSS 20 was used to perform the statistical analyses. All P values <0.05 indicated statistical significance. The 95% confidence interval (CI) was calculated for standard error computation.

Results

PATIENTS:

A total of 46 patients underwent allo-PBSCT with the FCA regimen, and 30 patients underwent allo-PBSCT with the CA regimen. The median age of patients was 29 years (range, 18–47 years) and 32 years (range 18–49 years) in the FCA group and the CA group, respectively. Fifty-one donors were a sibling of the patient, and 25 donors were unrelated to the patient. The median duration between diagnosis and PBSCT was 6.5 months (range 1–120 months). Thirty-four patients (43%) received cyclosporine A as a single immunosuppressive therapy agent before PBSCT, and only 1 patient received cyclosporine A in combination with ATG for 2 cycles. None of the patients had undergone a prior HCT. The FCA group and the CA group were comparable in terms of age, sex, time from diagnosis to HCT, donor type, ABO blood type, HLA matching, history of immunosuppressive therapy, ferritin, and transfusions before PBSCT (P>0.05, Table 1).

ENGRAFTMENT AND REJECTION:

The median number of days with an ANC ≥0.5×10⁹/L was 14 days (range, 11–27 days) in the FCA group and 13 days (range, 8–30 days) in the CA group. Platelet recovery (≥20×10⁹/L) without transfusion required was a median of 15 days (range, 9–31 days) in the FCA group and 15 days (range, 13–30 days) in the CA group. The difference in the engraftment time between the 2 groups was not statistically significant (P=0.57 [ANC]; P=0.56 [PLT]). Forty-one patients (41 out of 46, 89.1%) in the FCA group and 18 patients (18 out of 30 patients, 60%) in the CA group achieved >95% donor-type chimerism by day 30 post-PBSCT (P=0.004). Moreover, the donor cell percentage in bone marrow on day 30, detected by SNP-PCR, in the FCA group was higher than that of the CA group (98.8±1.1% versus 92.7±3.2%, respectively, P=0.007, 95% CI: 1.678–17.814).

Of all 76 patients, only 1 patient in the CA group experienced primary graft rejection and that patient achieved reconstitution after immunosuppressive therapy (cyclosporine A) 3 months later. No patient in the FCA group underwent primary graft rejection. Five patients (5 out of 30 patients, 16.67%) in the CA group underwent secondary graft rejection (median 2.9 months after PBSCT, range, 1.8–4.4 months). One patient self-recovered without treatment and 1 patient received a second HCT from the same donor; both of these patients had long-term survival. The other 3 of the 5 patients were given a donor lymphocyte infusion. Of these 3 patients, 1 patient died of serious infection of the lung 2 months later and the other 2 patients survived with disease (aplastic anemia). In the FCA group, secondary graft rejection occurred in 2 patients (2 out of 46 patients, 4.35%). One of the 2 patients received a successful second HCT from a second matched-unrelated donor and ultimately achieved hematologic recovery. The other patient underwent donor lymphocyte infusion but died 1 year later because of disease that was not controlled. The 5-year RR in the FCA group was 4.6±3.1%, which was significantly lower than that in the CA group (23.6±7.8%, P=0.013; Figure 1A). Multivariate analysis revealed the FCA regimen as an independent factor affecting RR (95% CI: 0.016–0.766, P=0.026, Table 2).

OS AND EFS:

The median follow-up was 54 months (range, 9–120 months). At the time of last follow-up, 45 patients (97.83%) were alive in the FCA group. The probability of 5-year OS in the FCA group was 97.4±2.5%. Twenty-seven patients (90%) in the CA group were alive, and the probability of 5-year OS was 90.0±5.5% (P=0.157; Figure 1B). In addition, the probability of 5-year EFS in the FCA group was 95.4±3.1%, which was significantly higher than that in the CA group (73.3±8.1%, P=0.006; Figure 1C). Multivariate analysis revealed the FCA regimen as an independent factor affecting EFS (95% CI: 0.012–0.667, P=0.019, Table 2).

TOXICITY AND INFECTION:

Regimen-related toxicity after PBSCT occurred in 18 patients (39%) in the FCA group and in 11 patients (37%) in the CA group. There was no significant difference between the 2 groups (P=0.691). The main toxicities were due to cytomegalovirus (CMV) infection (23.9% in the FCA group, 20% in the CA group), hemorrhagic cystitis (10.9% in the FCA group, 13.3% in the CA group), and pulmonary infection (10.9% in the FCA group, 6.7% in the CA group), which were successfully treated and did not cause directly mortality (Table 3).

CMV infection occurred in 11 patients in the FCA group and 5 patients in the CA group (23.9% versus 16.7%, respectively, P=0.199), and the median time to reactivation was 31 days after transplantation (range, 20–64 days). CMV disease occurred in only 1 patient in the CA group. More information regarding infection and toxicity rates is provided in Table 3. Two patients in the FCA group had Epstein-Barr virus (EBV) infection; post-transplant lymphoproliferative disorder (PTLD) developed in 1 patient. The EBV infection was controlled in both patients after rituximab treatment (375 mg/m2 per week). The 5-year serious infection rate after PBSCT was 28.3±6.6% and 23.3±7.7% in the FCA group and the CA group, respectively (P=0.157; Figure 1D).

GVHD:

In the FCA group, acute GvHD (aGvHD) was scored in 4 patients (4 out of 46 patients, 8.7%) as grade II to IV at a median time of 38 days (range, 32–49 days) after PBSCT. Compared with the FCA group, grades II to IV aGvHD appeared in 4 patients (4 out of 30 patients, 13.3%) of the CA group at a median time of 24 days (range, 19–39 days), (8.7±4.2% for the FCA group versus 13.3±6.3% for the CA group, P=0.51; Figure 2A). Except for 1 patient who died of transplant-related disease, 75 SAA patients survived longer than 3 months after transplantation and were eligible for evaluation for cGvHD. In the FCA group, 9 patients (19.5%) developed cGvHD at a median time of 7 months (range, 4–11 months). Compared with the FCA group, cGvHD appeared in 10 patients (34.5%) in the CA group at a median time of 5 months (range, 4–9 months) (19.5±6.2% in the FCA group versus 34.5±8.8% in the CA group, P=0.213; Figure 2B). All cases were limited cGvHD; no extensive or severe cGvHD occurred.

UNRELATED DONOR PBSCT IN THE FCA GROUP AND CA GROUPS:

Overall, there were 25 patients out of the 76 patients included in the study who received unrelated donor PBSCT. Sixteen patients (16 out of 46 patients, 35%) were from the FCA group and 9 patients (9 out of 30 patients, 30%) were from the unrelated donor CA subgroup. In the unrelated donor FCA subgroup, all patients (16 out of 16 patients, 100%) were alive at the last follow-up; whereas in the unrelated donor CA subgroup, 7 out of 9 patients (77.8%) were alive at the last follow-up. The probability of 5-year OS for the unrelated donor subgroups was 100% for FCA and 77.8±13.9% for CA (P=0.052, Figure 3A). However, the probability of 5-year EFS in the FCA subgroup was 93.8±6.1%, which was significantly higher than that in the CA subgroup (44.4±16.6%, P=0.013; Figure 3B). Only 1 patient (6.2%) in the FCA subgroup experienced graft rejection after PBSCT, whereas 4 patients (44.4%) in the CA subgroup experienced graft rejection. Furthermore, the FCA subgroup had a significantly lower 5-year RR than the CA subgroup (6.2±6.1% versus 44.4±17.3%, P=0.016; Figure 3C).

Seven patients in the FCA subgroup and 2 patients in the CA subgroup experienced serious infection after PBSCT. The serious infection rates in the FCA subgroup and CA subgroup were 43.7±12.4% and 22.2±13.9%, respectively (P=0.371, Figure 3D).

Discussion

Recently, Flu-based regimens have been introduced with promising results in SAA patients with bone marrow transplantation (BMT) [10,13]. However, the efficacy of this regimen is less well demonstrated in PBSCT, which is used more frequently in many transplantation centers, especially those patients with matched-unrelated donor in China, because the China Marrow Donor Program only collects peripheral blood stem cells from donors [3,6,9,30–34]. We reviewed the literature and summarized reports using Flu-based regimens in both BMT and PBSCT for SAA patients published in recent years (Table 4) [18–23]. As shown in Table 4, most of these studies reported data for younger patients with BMT or PBSCT, mixed together, with 4-year or 5-year OS varying from 67.9% to 96.7%, and few studies reported the long-term outcome in large cohorts [18,19,21–23].

Since aplastic anemia is a non-malignant hematologic disease, the effect of graft versus host reaction has no benefits for patients and might lead to serious complications and even death [35]. It is well known that, compared to PBSCT, BMT results in less incidence and severity of GvHD [36,37]. As shown in Table 4, the incidences of GvHD varies greatly, ranging from 11.7% to 46.4% for aGvHD and from 16.2% to 37.9% for cGvHD. In the current study of PBSCT, however, the rates of grade II to IV aGvHD (8.7%) and cGvHD (19.5%, all limited) in the FCA group were more acceptable. This favorable result might be related to the enhancement of immunosuppression with the FCA regimen in this study.

On the other hand, graft rejection is another key complication after allo-HCT, which happens more frequently in SAA cases, ranging from 0% to 25%, than other hematological malignancies [38–41]. Multiple transfusion is known to result in increased rates of graft rejection in SAA patients after HCT because of alloimmunization, which adversely affects survival rates [42,43]. Iron overload related to transfusion is also known to increase transplant-related complications, including graft rejection, and transplant-related mortality after HCT [44,45]. In the current study with 46 peripheral blood recipients older than 18 years of age who were preconditioned with FCA regimen, however, only 2 patients (4.35%) underwent a second graft rejection, and the estimated 5 year RR was much lower in the FCA group compared to the CA group (P=0.013), while the amount of transfusion and the serum level of ferritin were similar between the 2 groups. Moreover, the day 30 donor type chimerism detected by SNP-PCR was also higher in the FCA group than that in the CA group. These results indicated that the higher engraftment property of the PB graft might play an important role in the lower RR in this cohort of adult SAA patients.

BMT with matched-unrelated donor is considered a first-line treatment approach for children with SAA who lack a matched-related donor. Lee et al. reported that matched-unrelated donor BMT conditioned with total body irradiation (TBI) and Cy regimen resulted in a 5-year OS of 88% in adult patients with SAA [46]. A recent retrospective study showed that the survival outcome of matched-unrelated BMT was not statistically inferior to that of matched-related donor BMT [47]. However, the outcome of matched-unrelated donor PBSCT for adult SAA is still unclear. Three previous studies using Flu-based regimen for matched-unrelated donor HCT (including BMT and PBSCT) with SAA patients showed that the 4-year or 5-year OS ranged from 67.9% to 96.7% (Table 4) [18,19,23]. In the current study, one-third of SAA patients (25 out of 76 patients) underwent a matched-unrelated donor PBSCT. Of these patients, the estimated 5-year OS of patients who received the FCA regimen and the CA regimen was 100% and 77.8%, respectively (P=0.052). The 5-year EFS for patients treated with the FCA regimen was superior to that of patients who received the CA regimen (P=0.013). These results can majorly be attributed to a lower RR rate in the FCA group compared to the CA group (P=0.016). The RR rate of the FCA group shown in this study was similar to that reported by Bacigalupo et al., with an RR of 5% among young SAA patients who received matched-unrelated donors BMT [12]. Therefore, Flu can improve long-term survival mainly due to stronger immunosuppressive activity with lower RR rate, especially in adult patients with matched-unrelated donor PBSCT. More importantly, in our study, the stronger immunosuppressive activity of the FCA regimen was not associated with higher regimen-related toxicities or complications such as serious infections and organ toxicities (Table 3).

Conclusions

Our findings indicated that the FCA regimen for allo-PBSCT is safe and feasible for adult patients with SAA, especially for those with a matched-unrelated donor. Compared to the CA regimen, the FCA regimen preconditioning for SAA resulted in higher rates of engraftment and EFS, fewer graft rejections, and acceptable incidence of GvHD and organ toxicity. The nature of this retrospective study which used historical controls and had a limited number of patients was a major limitation for this study. Further multicenter studies are needed to confirm this conclusion.