|
Hemorrhagic cystitis (HC) is a common complication after allogenic hematopoietic stem cell transplant- tation (allo-HSCT); it causes significant morbidity and considerable expense as it is associated with prolonged hospitalization. Based on time of appearance, HC can be classified into two types. Early onset HC occurs within 48−72 hours of a conditioning regimen and is thought to be the toxic effect of chemo-irradiation.1 Later onset HC (LOHC) occurs beyond 72 hours from the preparative regimen and its definite pathogenesis has not been well elucidated.2 Overall, most of researchers believe that LOHC is a multifactorial disease.3
To date, several risk factors have been reported to be associated with the development of LOHC. The presence of BK virus reactivation is the most significant factor and has been repeatedly confirmed by several studies in the West.4 Adenovirus infection is also thought to play a key role in pathogenesis of LOHC in Japanese patients.5 As heterogeneous viral etiology and genetic factors exist in different populations, the causative agent may differ in different transplantation centers. Cytomegalovirus (CMV) infection is highly prevalent in Chinese patients, which is also the common infective complication in our protocol.6,7 However, to our knowledge, only sporadic case reports provide information about CMV-induced LOHC, which raised the need to investigate the involvement of CMV in the pathogenesis of LOHC.
Currently, the role of GVHD in the development of HC remains obscure.3 In theory, GVHD can lead to HC by increasing the risk of opportunistic infection and by alloimmune injury localized to the bladder epithelium which may also trigger HC. However, one study failed to confirm this correlation.8 In addition, as prior studies utilized many different conditioning regimens, types of transplants and graft manipulations, the independent risk factors for HC were inconsistent.3,8 Based on these reasons, we conducted a retrospective analysis of the clinical records of 250 patients at Peking University Institute of Hematology. In our protocol, we limited our analysis to allo-HSCT and conditioning with BU/CY myeloablative regimens and investigated the incidence and risk factors for HC. The current study also focused on evaluating the role of CMV reactivation and GVHD in the development of HC.
METHODS
Patient characteristics
A total of 250 allogeneic HSCT were carried out in our hospital from April 2003 to November 2005. All patients were treated on institutional protocols approved by the Institutional Review Board. Patient characteristics are summarized in Table 1.
 view in a new window |
Table 1. Patient characteristics |
Conditioning regimen and graft-versus-host disease prophylaxis
Conditioning therapy was all BU/CY-based (busulfan 4 mg∙kg-1∙d-1 administered orally for 3 days and cyclo- phosphamide 1.8 g∙m-2∙d-1 intravenously for 2 days) with the addition of Me-CCNU 250 mg/m2 orally, one dose for all patients, and a single dose of cytarabine 2 g∙m-2∙d-1 intravenously for patients with matched sibling. HLA mismatched and unrelated matched HCT patients were also given cytarabine 4 g∙m-2∙d-1 for two days and ATG (SangStat, France) 2.5 mg∙kg-1∙d-1 intravenously for 4 consecutive days. Prophylaxis against GVHD consisted of cyclosporine (CSA), a short course of methotrexate (MTX) and MMF.
CMV and adenovirus surveillance
All the patients were screened for CMV serostatus pretransplantation. The plasma PCR assay (Roche, Amplicor) was employed weekly from transplantation to 100 days. Urine samples were sent for CMV-PCR assay when CMV-viremia was diagnosed. Once HC developed, blood and urine samples were sent for adenovirus- specific quality-PCR assay (Anhui, China) weekly. As a biopsy was unavailable in most situations, we cannot confirm by pathology viral cystitis for each instance of HC. In this study, CMV-viruria (PCR) in cases of LOHC was thought to be CMV associated viral cystitis.3 For prophylaxis against CMV disease, pre-emptive ganciclovir therapy (5 mg∙kg-1∙d-1) or foscarnet (90−120 mg/kg) was administered to patients with two consecutive PCR tests positive for CMV and continued until the PCR was negative on two occasions.
HC definitions and prophylaxis
HC was defined as the criteria reported by Leung et al.3 In patients with intermittent or recurrent symptoms, the date of onset of HC was defined as the first day of symptoms or laboratory analysis. As prophylaxis for HC, all patients were given intravenous fluid at 3 L∙m-2∙d-1 from 4 hours before to 24 hours after administration of CY. MESNA was given intravenously at a dose of 15 mg/kg prior to cyclophosphamide, and once every 8 hours thereafter, over 24 hours until the last dose of CY.
Event sequence relationship between GVHD and HC
To analyze the event-sequence relationship between GVHD and HC, the time of onset between GVHD and HC was recoded. The onset time was defined as the first day of presentation according to a laboratory test or clinical manifestation, marrow infusion was considered day 0. The Wilconxon rank sum test was employed to analyze the event-sequence relationship.
HC management
Surveillance of urine cultures for bacteria and fungi were done on a weekly basis when hematuria occurred. The initial treatment for HC consisted of intravenous fluid hydration, alkalization and forced diuresis. Platelet transfusions were given to patients when clinically needed; In patients with grade III−IV HC, bladder irrigation was administered to ameliorate symptoms and prevent obstruction by blood clots. If HC could not be resolved, an empirical anti-viral therapy with ganciclovir or foscarnet was given if clinically indicated. Tapering doses of immunosuppressant were given to those patients with HC coexisting viremia or viruria who were without active GVHD. Patients with persistent bleeding despite these measures were defined as having refractory HC.
Statistical analysis
The Cox regression model was used to evaluate risk factors for the development of HC on univariate and multivariate analysis. For this purpose HC was evaluated as a time-dependent variable. Risk factors were analyzed for their effects on the development of HC in univariate analysis, including the patient's sex, age, underlying diagnosis, disease status, match of blood type, donor type, use of ATG in conditioning regimens, infused CD34+ cell, CMV disease or viremia and acute GVHD. To correct for the effect of confounding variables, multivariable analysis was also performed. Actuarial survival from the day of transplantation was estimated by the method of Kaplan-Meier. Statistical analysis was performed using SPSS13.3, and P value less than 0.05 was considered statistically significant.
RESULTS
Clinical features
HC occurred in 72 of the 250 patients within 180 days after transplantation and with a cumulative incidence of 28.8% (SE 0.29%). HC occurred in 12.6% of patients with HLA-matched related donors (MRD), 34.38% of those with HLA-matched unrelated donors (MUD) and 49.45% of those with mismatched related donors (MMRD). No early-onset HC was developed in our cohort. The median time of onset was 33 days after HSCT (range 14−170 days), lasting from 3 to 186 days, with a median duration of 35 days. HC was scored as grade I in 2/72 cases (2.8%), grade II in 49/72 (68.10%), grade III in 20/72 (27.80%) and grade IV in 1/72 (1.40%).
Univariate analysis and multivariate analysis
Factors associated with a significantly elevated risk of HC from univariate analysis were: age younger than 25, high risk status after undergoing diagnosis, CMV-viremia or disease, use of ATG, grafts from MUD or MMRD, GVHD grade II–IV (Table 1). No association of HC was found with the undergoing disease, the number of CD34 cells or nuclear cells infused or platelet engraftment kinetics (Table 2). Results from the multiple regression analysis found that GVHD grade II–IV (RR=2.75, 95% CI 1.63–4.66, P<0.01) and grafts from MUD or MMD (RR=2.60, 95% CI 1.52–5.20, P<0.01) were the independent risk factors. CMV-viremia achieved a marginal significance (RR=2.20, 95% CI 0.9−4.22, P=0.08), LOHC did not increase the mortality (RR=0.67, 95% CI 0.33−1.36).
 view in a new window |
Table 2. Descriptive statistical results by univariate analysis |
Reactivation of CMV associated with incidence of LOHC
CMV reactivation was associated with a high rate of HC by univariate analysis. CMV reactivation, detected by PCR, occurred in 40.00% (n=100) of patients in the whole cohort. The median time of CMV onset was 37 days (range 6−180 days) after HSCT. Percent forty-four of patients with CMV reactivation developed HC compared to 18.67% (n=28) in patients with no CMV reactivation (RR=2.73, P<0.01) and the development of CMV-viremia was marginally significant in the multivariate analysis (P=0.08, Fig. 1). However, among the 15 of 100 patients with CMV reactivation confirmed to have CMV-viruria by PCR assay, only 9 of these 15 patients actually progressed to HC while 4 patients with CMV-viruria had chronic HC exceeding 40 days. Only 4 of patients among the group with HC had positive adenovirus-viruria by the PCR assay and all of these had coexisting CMV-viruria. Three of the four patients with adenovirus-viruria had long-lasting HC.
 view in a new window |
Fig. 1. Cumulative incidence of LOHC associated with CMV reactivation (Cox regression model estimate; P<0.01). The estimate curve shows a higher incidence of LOHC in patients with CMV reactivation than that with no-CMV reactivation. LOHC: later onset hemorrhagic cystitis, CMV: cytomegalovirus. |
Donor type associated with LOHC
Patients who received grafts from MMRD and MUD had higher rates of HC (Fig. 2); recipients of MMRD had the highest rate of HC (49.45%, RR=5.09, P<0.01), followed by patients with grafts from MUD (34.38%, RR=2.75, P<0.01) and MRD (12.60%).
 view in a new window |
Fig. 2. Cumulative incidence of LOHC associated with donor type. Cumulative incidence of HC was as follows: MRD 12.60%, MUD 34.38%, MMD 49.45%. Cox regression model estimate P<0.01. LOHC: later onset hemorrhagic cystitis, HC: hemorrhagic cystitis, MRD: HLA matched related donor, MUD: matched unrelated donor; MMRD: mismatch related donor. |
GVHD associated with LOHC
Patients who experienced grades 2−4 acute GVHD had a significantly increased rate of HC (56.25%), compared with a rate of 15.88% among patients with grades 0−1 acute GVHD (P<0.01, Spearman's rank test, Fig. 3). Multiple analysis indicated that grades II–IV GVHD (RR=2.75, P<0.01, 95% CI 1.63–4.66) was an independent risk factor for HC.
 view in a new window |
Fig. 3. Cumulative incidence of LOHC associated with GVHD (Cox regression model estimate; P<0.01). The estimate curve shows a higher incidence of LOHC in patients with grades II−IV GVHD than that with grades 0−I GVHD. Cox regression model estimate P<0.01. LOHC: later onset hemorrhagic cystitis, GVHD: graft-versus-host disease. |
Event sequence relationship between GVHD and HC
In the group of patients with both HC and GVHD (n=52), 51.92% (27/52) of HC occurred after GVHD initiation (range 8 to 44 days), 15.38% (8/52) before GVHD (range 7 to 33 days), and in 32.69% (17/52) patients HC began around the time of GVHD initiation (range 0 to 7 days). The interval between onset of HC and GVHD was no more than 2 weeks in 71.15% (37/52) patients; the median interval between GVHD and LOHC was 9 days. Therefore, the majority of HC episodes began around the time of GVHD initiation. The Wilconxon rank sum test confirmed that the distribution of the onset time of GVHD and HC were significantly different (P<0.01, Fig. 4).
 view in a new window |
Fig. 4. The distribution of onset time of HC and GVHD (patients in group with GVHD and HC, n=52). Graph shows GVHD is an earlier event in post-transplantation and a majority of HC episodes are initiated around GVHD initiation. HC: hemorrhagic cystitis, GVHD: graft-versus-host disease. |
Outcomes of patients with HC
All patients received conservative therapy. Of the 72 patients 67 (93.06%) recovered from HC after a median of 35 days (range 3−186). Five patients died with HC; the causes of these deaths were lung infection (2), encephalitis (1) and acute GVHD (2). Treatment with hydration alone led to resolution of HC in 6 patients. Thirty-four patients received anti-virus therapy and a gradual tapering of immunosuppressant drugs, twenty-three patients achieved a complete response and 3 patients achieved a partial response. However, five patients developed GVHD in the process of tapering of the immunosuppressant drugs and their HC worsened. In addition, 5 patients had CMV reactivation, by PCR surveillance, and persistent bleeding continued. Seventeen patients with active GVHD received anti-virus therapy combined with anti-GVHD therapy. Nine of these 17 patients achieved a complete remission.
DISCUSSION
In this report, we demonstrated that the incidence of HC in our protocol seems higher than that reported in previous studies,8,9 possibly due to our study containing a higher proportion of patients receiving MUD or MMRD transplants following myeloablative BU/CY conditioning; factors previously identified as predictive for the development of HC.2,8,9 The association of CMV reactivation and CMV-viruria with a higher incidence of HC after HSCT should be confirmed by a multi-center transplantation study.
LOHC is commonly described as being a virus-associated disease.4,5 However, the specific virus responsible for HC is still not identified.3,10 CMV is a less common pathogen associated with LOHC in previous case reports.11 As CMV infection is highly prevalent in China,7 and up to 95% adult donors and recipients are CMV seropositve, we investigated the influence of CMV on development of HC. In our cohort, 40.00% of patients developed CMV-verimia within the first 180 days; univariate analysis indicated that CMV-viremia was significantly associated with an increased incidence of HC. 44.00% of patients with CMV-verimia progressed as compared to 18.67% without CMV-verimia. Multivariable analyses also showed that CMV-verimia was associated HC with a marginal significance. Furthermore, 32/44 of patients with LOHC who received anti-CMV therapy achieved a remission and symptoms improved. We believe CMV is a significant risk factor of LOHC. However, it should be noted that there was not a parallel incidence of CMV- viremia. Therefore, the correlation between CMV-viremia and LOHC does not have an obvious biological explanation and requires further validation.
Previous studies have shown that BK-viremia did not invariably lead to HC.10 Our data also showed that 6 patients with CMV-viruria did not progress to HC. Five patients with CMV-viruria cleared the virus from the urine following antiviral therapy but had persistent bleeding, Which suggest that CMV infection alone is not sufficient to trigger HC, and some viruses other than CMV or ADV, or other unknown mechanisms, might contribute to the pathogenesis of HC. Previous reports have suggested that CMV can induce BK virus-associated HC.12 As our study was limited by the lack of investigation of BK-virus we cannot draw any conclusions regarding the role between BK and CMV in development of HC. Prospective clinical studies, simultaneously quantifying the CMV and BKV urinary loads, might lead to a better understanding of their role in HC during HSCT.
The role of GVHD in the pathogenesis of HC remains debatable.8,13,14 Recently, in a series of 105 allogeneic transplants, the incidence of HC was significantly higher in MUD or UCB transplants but the development of GVHD was not independently associated with HC.15 As these studies were all based on incidence analysis, the observed differences may have resulted from different HSCT protocols. In comparison with the report of MD Anderson, our data showed a comparable incidence of HC and GVHD, which support an association between HC and GVHD based upon the multivariable analyses that indicated the occurrence of grades II−IV acute GVHD was an independent risk factor. Furthermore, the event sequence analysis further supports an association of GVHD with LOHC, since most of GVHD episodes began following, or were coincident with, the onset of acute GVHD. As LOHC frequently overlaps with GVHD it is difficult to determine whether GVHD per se or GVHD related immunosuppressant therapy caused viral infection and was responsible for the high incidence of HC. Based upon our data we conclude that GVHD is a high risk factor for LOHC.
There was a trend toward a higher incidence of HC among MMD and MRD patients, suggesting that more intense immunosuppression may contribute to the increased risk of HC. Also, the kinetics of immune reconstitution after transplant are likely to further to contribute to the development of HC.
In summary, our results indicate that CMV-viremia is a high risk factor for LOHC and our data also showed a good correlation between GVHD and HC, which suggested that an alloimmune response may be involved in the pathogenesis of HC.
Acknowledgements: We would like to thank Prof. Waller EK for his excellent commentary and assistance in editing this manuscript.
REFERENCES
1. Cox PJ. Cyclophosphamide cystitis – identification of acroleinas the causative agent. Biochem Pharmacol 1979; 28: 2045-2049.
2. Russel SJ, Vowels MR, Vale T. Hemorrhagic cystitis in pediatric bone marrow transplant patients: an association with infective agents, GVHD and prior cyclophosphamide. Bone Marrow Transplant 1994; 13: 533-539.
3. Leung AY, Yuen KY, Kwong YL. Polyoma BK virus and hemorrhagic cystitis in hematopoietic stem cell transplantation: a changing paradigm. Bone Marrow Transplant 2005; 36: 929-937.
4. Bedi A, Miller CB, Hanson L, Goodman S, Ambinder RF, Charache P, et al. Association of BK virus with failure of prophylaxis against hemorrhagic cystitis follow-ing bone marrow transplantation. J Clin Oncol 1995; 13: 1103-1109.
5. Akiyama H, Kurosu T, Sakashita C, Inoue T, Mori SI, Ohashi K, et al. Adenovirus is a key pathogen in hemorrhagic cystitis associated with bone marrow transplantation. Clin Infect Dis 2001; 32: 1325-1330.
6. Liu YT. A serological study on anti-CMV and anti-EBV antibodies in population of Beijing, Changzhi Shanxi and Yichang Hubei. Chin J Epidemiol (Chin) 1989; 10: 277-281.
7. Lu DP. Bone marrow transplantation in the People's Republic of China. Chinese bone marrow transplant registry. Bone Marrow Transplant 1994; 13: 703-704.
8. Sencer SF, Haake RJ, Weisdorf DJ. Hemorrhagic cystitis after bone marrow transplantation: risk factor and complications. Transplantation 1993; 56: 875-879.
9. Leung AY, Mak R, Lie AK, Yuen KY, Cheng VC, Liang R, et al. Clinic pathological features and risk factors of clinically overt hemorrhagic cystitis complicating bone marrow transplantation. Bone Marrow Transplant 2002; 29: 509-513.
10. Bogdanovic G, Ljungman P, Wang F, Dalianis T. Presence of human polyomavirus DNA in the peripheral circulation of bone marrow transplant patients with and without hemorrhagic cystitis. Bone Marrow Transplant 1996; 17: 573-576.
11. Tutuncuoglu SO, Yanovich S, Ozdemirli M. CMV-induced hemorrhagic cystitis as a complication of peripheral blood stem cell transplantation: case report. Bone Marrow Transplant 2005; 36: 265-266.
12. Bielorai B, Shulman LM, Rechavi G. CMV reactivation induced BK virus-associated late onset hemorrhagic cystitis after peripheral blood stem cell transplantation. Bone Marrow Transplant 2001; 28: 613-614.
13. Ost L, Lonnqvist B, Eriksson L, Ljungman P, Ringden O. Hemorrhagic cystitis: a manifestation of graft versus host disease? Bone Marrow Transplant 1987; 2: 19-25.
14. Ruutu T, Ruutu M, Volin L, Leskinen R. Severe cystitis as a manifestation of chronic graft-versus-host disease after bone marrow transplantation. Br J Urol 1988; 62: 612.
15. El-Zimaity M, Saliba, R Chan K, Shahjahan M. Hemorrhagic cystitis after allogeneic hamatopoietic stem cell transplantation: donor type matter. Blood 2004; 103: 4674-4680.
|
PDF(838K)
Rapid Response |