Outcomes and risk factors of hemorrhagic cystitis in pediatric allogeneic hematopoietic stem cell transplantation recipients using different graft source and condition with severe aplastic anemia

ABSTRACT
 Background Hemorrhagic cystitis (HC) is a severe complication of allo-HSCT, characterized by irritative symptoms of the urinary tract and a higher morbidity rate. The risk factors and prognosis of HC are still unclear. Objective The objective of this study is to identify risk factors and outcomes to improve treatment in pediatric SAA patients undergoing HSCTs in the Children's Hospital of Soochow University. Methods A total of 97 SAA patients as a cohort were enrolled from 2010 to 2019 in the Children's Hospital of Soochow University and a number of factors related to HC and outcomes were analysed. In all transplants (except UCBT), patients received a combination of G-CSF stimulated bone marrow (BM) and peripheral blood stem cell (PBSC). The minimum number of CD34 + cells is 5 × 106 cells/kg. Results Mononuclear cells dose (MNC, cut off: 8.53 × 108/kg) and grade II–IV acute graft versus host disease (aGVHD) were identified as independent risk factors for HC. Patients without HC had better overall survival (OS) than with HC (No HC: 98.6%±1.4% vs HC: 87.4% ± 6.8%, p = 0.03). Conclusion We concluded that aGVHD and MNC dose in graft might play an important role in the development of HC in pediatric SAA patients undergoing allo-HSCT. HC is also a key complication affecting the prognosis of children with SAA after allo-HSCT.


Introduction
Aplastic anemia (AA) is a rare but potentially fatal disorder characterized by hypocellular bone marrow and pancytopenia [1]. Severe aplastic anemia (SAA) is a lifethreatening type which allogeneic hematopoietic stem cell transplantation (allo-HSCT) and immunosuppressive therapy(IST) are the available curative approaches [2,3]. However, IST has obvious shortcomings, approximately one-third of patients did not respond to IST and one-third of responders relapsed after initial therapy [4,5]. HSCT from an HLA-matched related donor (MRD) is the first-line treatment for children with SAA [6]. But due to the limitation of the source of MRD, unrelated cord blood transplantation (UCBT), haploidentical donor (HID) and matched unrelated donor (MUD) of allo-HSCT are gradually applied in the treatment of SAA [7][8][9][10]. The long-term disease-free survival rate of allo-HSCT from different donor types for the treatment of SAA over 85%. Although allo-HSCT for SAA has achieved a high disease-free survival rate, complications related to allo-HSCT are still important factors affecting the quality of life of children. Among them, hemorrhagic cystitis (HC) is a potential life-threatening complication after allo-HSCT [11].
HC is an inflammatory process of the bladder that occurs after allo-HSCT and is a challenging condition to treat [12]. The potential for life-long sequelae from fibrosis and contracture of the bladder is of great consequence in pediatric patients. Clinically, HC is characterized by hematuria, dysuria, irritative symptoms of the urinary tract and even renal failure due to hydronephrosis [13,14]. Early onset HC, often caused by cyclophosphamide (CY), busulfan, and irradiation [15]. And Late onset HC is often associated with a consequence of virus reactivation, most commonly involving BK virus (BKV), but also adenovirus or cytomegalovirus (CMV) [16,17]. Many studies have explored the risk factors for HC after HSCT. Older age, acute Graft versus host disease (aGVHD), HLA matching, T-cell depletion, and severe immunosuppression have been implicated as risk factors to late onset HC [12].
However, to the best of our knowledge, the outcomes and risk factors of HC following allo-HSCT in a pediatric cohort with SAA have not yet been established. Therefore, in the present study, we addressed this question by retrospectively analyzing the possible risk factors (such as age, sex, disease, disease course, donor type, HLA matching, ABO match, CMV seropositive, EBV seropositive, neutrophil engraftment time, platelet engraftment time, MNC, CD34, engraftment syndrome, aGVHD, aGVHD localization) and outcomes (such as cGVHD and overall survival) of HC in 97 pediatric SAA patients after allo-HSCT. We hope that through our research we could identify the most potentially close factors and help in dealing with HC to improve the quality of life of children with SAA after allo-HSCT.

Methods and patients
Patients This study included 97 pediatric patients (aged 1-18 years), who were diagnosed with AA received allo-HSCT in Children's Hospital of Soochow University, Suzhou, China from 2010 to 2019. Due to the retrospective nature of the study, it was not necessary to request informed consent from the patients, but patient confidentiality was strictly maintained. All patients and stem cell donors signed an informed consent before beginning the transplant process. The criteria for selection of patients were as follows: (i) diagnosed with SAA. Patients with congenital bone marrow disorders (Fanconi anemia, Diamond-Blackfan anemia, and dyskeratosis congenital (DKC)) were excluded. (ii) aged 0-18 years at the time of transplantation; (iii) no prior HSCT.

Immunophenotypic analysis
PB samples were collected using a heparin tube. PB mononuclear cells (PBMCs) were separated within 2 h using Ficoll-Paque (Sigma, San Francisco, CA, USA). T cells were tested using FITC-CD3, PE-CD4, and APC-CD8. B cells were tested using PE-CD19. NK cells were tested using FITC-CD16 and APC-CD56. All antibodies used for immunofluorescence staining were purchased from BD Biosciences (San Jose, CA, USA). Abs staining were performed according to the manufacturer's instructions. Cells were then washed twice in staining buffer (Phosphate Buffered Saline + 2.5% Fatal Bovine Serum + 0.05% NaN3). Flow cytometry analyses were performed using Flowjo software (BD Biosciences, CA) and Gallios cytometers(Beckman, Los Angeles, CA, USA).

Definitions
Hemorrhagic cystitis was defined as microscopic or macroscopic hematuria and dysuria with a negative bacterial culture in the urine and no other hemorrhagic conditions. The symptoms of HC are classified as follows: grade 1 shows microscopic hematuria; grade 2, macroscopic hematuria; grade 3, macroscopic hematuria with clots, and grade 4, macroscopic hematuria with clots, urinary obstruction, kidney damage, and bladder damage [18]. Overall survival (OS) was considered as time from allo-HSCT to death. Neutrophil engraftment was defined as achieving an absolute neutrophil count ≥0.5 × 10 9 /L for 3 consecutive days. Platelet engraftment is defined as achieving a platelet count ≥20 × 10 9 /L without transfusion support for 7 consecutive days.
Myeloablative conditioning regimen, defined as busulfan (BU) >8 mg/kg. The diagnosis and grading of acute GVHD (aGVHD) and chronic GVHD (cGVHD) were assigned by the transplant center using standard criteria [19,20]. Engraftment syndrome (ES) was defined as fever in absence of infection, skin rash, diarrhea, capillary leak syndrome, and pulmonary injury which occur before or during neutrophil engraftment [21].

Statistical analysis
To compare differences between groups, the chisquare test was used for categorical variables and the Mann-Whitney U test for continuous variables. The univariate and multivariate analysis of prognostic factors was performed with a logistic regression model to estimate the odds ratio(OR) effect of variables on HC. Variables with a p < 0.10 in univariate analyses were included in the multivariate model. Receiver operating characteristic (ROC) analyses were performed and areas under the curve (AUC) were calculated with HC as the actual state variable. Analyses of OS, CI of ES, aGVHD, cGVHD were performed using the Kaplan-Meier method, with differences compared by log-rank tests. All statistical analyses were performed using SPSS version 22.0(IBM, Armonk, NY).

Characteristics of patients
A total of 97 patients were eligible for this study. Divide the patients into two groups (HC group vs No HC group) according to the presence or absence of hemorrhagic cystitis. The characteristics of patients and transplantation are presented in Tables 1 and 2. In No HC group, 38 males and 33 females, with a median age of 7.75 years (range 1-15.75 years). Forty-seven patients (66%) were SAA, and the remaining 24 patients (34%) were VSAA. Forty patients (56%) underwent HSCT with a HID, 19 (27%) with a MRD, 1 (1%) with a UCB, and 11(16%) with a MUD. The median time interval from SAA diagnosis to SCT was 4.3 months (range 1-114.4 months). The median MNC and CD34 +⍰⍰cell doses were 6.04 × 10 8 /kg (range 1.32-14.89 × 10 8 /kg) and 4.97 × 10 6 /kg (range 0.16-12.50 × 10 6 /kg), respectively.
There are 26 patients in the HC group (14 males and 12 females), with a median age of 5.71 years (range 2.42-11.33 years). The overall cumulative incidence of HC (grade 1-4) was 27% (26 of 97 patients). The median time to HC onset was 21 days (range 1-95 days). Median time to HC duration was 14 days (range 3-95 days). There were 5 patients with grade I HC, 15 patients with grade II HC, and 6 patients with grade III HC, there were no patients with grade IV HC. Eleven patients (42%) were SAA, and 15 patients (58%) were VSAA. Seventeen patients (65%) underwent HSCT with a HID, 3(12%) with a MRD, 4(15%) with a UCB, and 2 (8%) with a MUD. The median  Figure 1(A), platelet: p = 0.03, Figure 1(B)) in the HC and no HC groups. We evaluated whether the frequency and cell count of CD3 + T cells, CD4 + T cells, CD8 + T cells, CD19 + B, CD16 + CD56 + NK cells and total lymphocyte count at D30 after HSCT impact the risk of HC. There was no statistically significant difference in lymphocyte subsets in patients with and without HC ( Figure 2).

Risk factors for HC
Further, we evaluated a number of other parameters for a possible association with increased risk of HC using univariate analysis, including age, sex, disease, disease course, donor type, HLA matching, ABO match, CMV seropositive, EBV seropositive, neutrophil engraftment time, platelet engraftment time, MNC, CD34, engraftment syndrome, aGVHD, and aGVHD localization (Table 3) (Table 5). We performed ROC analysis to test the specificity and sensitivity of MNC. MNC had a great ability to predict HC ( Figure 3, AUC = 0.65, p = 0.026). When the MNC is above the threshold (8.53×10 8 /kg, sensitivity: 75%, specificity: 58%), the probability of HC occurrence is significantly increased. We further analyzed CI of engraftment and aGVHD. The CI of engraftment in the HC group was higher than that in the no HC group (Figure 4, 65.4% ± 9.3% vs 46.6% ± 5.9%, p = 0.08). But this difference was not statistically significant (p > 0.05). The CI of aGVHD between HC group and no HC group was significantly different (p < 0.05). Regardless of the CI of grade I-IV aGVHD ( Figure 5(A), HC: 65.4% ± 9.3% vs no HC: 36.6% ± 5.7%, p = 0.018) or  grade II-IV aGVHD ( Figure 5(B), HC: 39.5% ± 9.5%% vs no HC: 8.5% ± 3.3%, p < 0.001), the HC group was higher than the no HC group.

Outcomes and HC
Finally, we explored the relationship between the long-term prognosis after HSCT and HC. All patients did not have disease relapse and late transplantation failure. We compared the difference in the CI of cGVHD ( Figure 5(C), HC: 31.0% ± 9.9% vs No HC 40.4% ± 8.7%, p = 0.99) and CI of moderate and severe cGVHD ( Figure 5(D), HC: 22.5% ± 9.0% vs No HC 27.0% ± 9.2%, p = 0.65) between the HC group and the no HC group. However, HC was not associated with cGVHD. We further compared the survival of the two groups. Overall, 4 patients died, including 3 with HC. In 3/4 cases the death was caused by infections (bacterial, fungal) and aGVHD. In 1/4 cases, the death was caused by encephalitis and bronchiolitis obliterans. Patients without HC had better OS ( Figure 6, No HC: 98.6% ± 1.4% vs HC: 87.4% ± 6.8%, p = 0.03) than with HC.

Discussion
Herein, we reported the incidence of HC after allo-HSCT in children with SAA, and explored the potential risk factors for the development of HC and the prognosis of children with HC. In recent years, more and more children with SAA choose allo-HSCT as a treatment option. Although a good disease-free survival rate has been achieved, multiple complications after allo-HSCT have become an important factor restricting the quality of life. Among these complications, HC is a noteworthy complication particularly in this pediatric population. Therefore, a complete understanding of the risk and outcomes of HC is important. To our knowledge, there are no reports that have analyzed the risk factors of HC after HSCT in children with SAA. In our study, the incidence of clinical signs of HC was 26.8%, which was similar to the incidence of HC as described in the literature [18].
Adequate immune reconstitution post-HSCT is crucial for the success of transplantation, and is associated with many post-HSCT complications [22,23]. Caroline Jepsen et al analyzed 96 patients who underwent haplo-HCT, the medium time to myeloerythroid reconstitution in the whole cohort was 11 days for neutrophils and 14 days for platelets. Among the HCpositive patients, this was 15 (neutrophils) and 14 (platelet) days. The study did not perform a statistical comparison of engraftment time [24]. In our study, we found that the time of neutrophil engraftment and platelet engraftment in children with HC were longer than those without HC (Table 1, Figure 1). The immune disorder caused by HC may be an important reason for the delay of engraftment. Recovery of NK-, B-and T-lymphocytes is usually slow following HSCT     and increases the risk of viral infections [25][26][27]. And virus infection is an important factor in causing HC [28]. We compared frequency and cell counts of lymphocyte subsets on D30 after HSCT in patients who did and did not develop HC. Somewhat surprisingly, we found no significant difference between these two groups ( Figure 2). In our study, we only analyzed the lymphocyte subsets on D30 after HSCT in the two groups, and the subsequent immune reconstitution was not analyzed. In our cohort, we do not find lymphocyte subsets on D30 playing a major role in developing HC.
Many HC risk factors such as male gender, older age, clinical conditions, and GVHD have been described [29,30]. Hassan et al analyzed 834 patients who underwent HSCT, busulfan-based conditioning, myeloablative conditioning, and CMV infection were found to be risk factors for grades II-IV HC, and bacteremia and severe acute GVHD were the risk factors for severe HC (grades IIIIV) [31]. Barbaros sahin karagun et al. reported patients with a HID, a history of busulfan, and cyclophosphamide treatment or total body irradiation had a higher risk of BK virus-associated HC [32]. In our subject, we identified potential risk factors for HC, and found that VSAA, UCB, MNC, gut aGVHD, and II-IV aGVHD were risk factors in our univariate analysis (Table 3). MNC and II-IV aGVHD were significant risk factors in our multivariable analysis (Table 4; p = 0.002 and p = 0.011, respectively). Since only UCB used MAC conditioning regimen among all donor types, this may be the reason why UCB is a risk factor for HC. To our knowledge time to VSAA and MNC dose have not been previously reported as risk factors for HC. Transplantation with high MNC dose ( Figure 3, cut off: 8.53 × 10 8 /kg) may further increase the risk of HC after transplantation. Higher numbers of MNCs in grafts may be more likely to cause damage to the bladder epithelium. Therefore, a reasonable dose of MNC is essential for the prevention of HC.
Many studies have shown that there may be a correlation between GVHD and HC or immunosuppressive therapy used to treat GVHD increases the chance of subsequent opportunistic infections of HC [33][34][35]. However, it still remains not clear whether aGVHD targeting bladder epithelium is manifested as HC or whether immunosuppressive drugs used to treat acute GVHD which could affect viral reactivation. Our study found that regardless of the cumulative incidence of grade I-IV aGVHD or grade II-IV aGVHD, the HC group was significantly higher than the no HC group ( Figure 5(A, B)). However, no significant correlation between HC and the occurrence and severity of cGVHD was found ( Figure 5(C, D)). Regarding overall survival, we found that HC patients were lower than those without HC ( Figure 6, No HC: 98.6%  ±1.4% vs HC: 87.4%±6.8%, p = 0.03). The patient's cause of death was GVHD or infection. This may be due to the higher CI of aGVHD in the HC group. The use of stronger immunosuppressive strategies further increased the risk of infection in children [36,37].
Research limitations include its retrospective nature, the small number of samples analyzed, and the small number of HC patients. In the risk factor analyses, we analyzed relatively small patient numbers, higher numbers are needed to achieve a more robust risk analysis. And our center has not tested the BK virus. Prospective multicenter studies are required to firmly establish risk factors for HC development.
In conclusion, we found MNC (cut off: 8.53×10 8 /kg) and II-IV aGVHD were independent risk factors and sufficient for developing HC through univariate and multivariate analysis. And the platelet engraftment in children with HC is late. Children with HC have a worse overall survival than children without HC. Further validation of these findings may help clarify our understanding. Knowledge of the natural history and factors associated with HC is important for the day-to-day management of SAA children undergoing HSCT and the design of clinical trials testing novel therapies.

Disclosure statement
No potential conflict of interest was reported by the author(s).