SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

To assess hemorrhagic cystitis and urinary tract cancer incidence and predictors in cyclophosphamide (CYC)–treated patients with systemic necrotizing vasculitis (SNV).

Methods

The French Vasculitis Study Group database, which contains longitudinal data on SNV patients, was searched for urinary tract cancer and/or hemorrhagic cystitis occurrences in patients diagnosed as having Wegener's granulomatosis (WG), microscopic polyangiitis, Churg-Strauss syndrome, or polyarteritis nodosa. The observed incidence of urinary tract cancer was compared to the expected incidence in the general population by calculating standardized incidence ratios (SIRs). Relationships between urinary tract cancer and/or hemorrhagic cystitis and 10 variables, including CYC dosage and administration route, were investigated by survival analyses for a nested subgroup of patients for whom detailed information on CYC exposure was available.

Results

Among the 805 patients observed over 4,230 patient-years (mean followup 5.3 years), 22 cases of hemorrhagic cystitis and 7 of urinary tract cancer were identified in 27 patients. The SIRs for urinary tract cancer were 5.00 for all patients with SNV (P = 0.001) and 5.96 for patients with WG (P = 0.03). Based on 467 patients with detailed CYC information, cumulative CYC dose (hazard ratio [HR] for 10-gm increments 1.09; P = 0.03), ever-oral CYC administration (HR 5.50; P = 0.001), and WG (HR 2.96; P = 0.01) independently predicted urinary tract cancer and/or hemorrhagic cystitis. According to univariate analyses, smoking (ever) (HR 8.20; P = 0.02) and a prior hemorrhagic cystitis episode (HR 5.20; P = 0.046) significantly predicted urinary tract cancer.

Conclusion

Our findings indicate that CYC treatment of SNV is associated with a 5-fold higher risk of developing urinary tract cancer. Urotoxicity risk in SNV is associated with the cumulative CYC dose and its oral administration, and might be higher in WG.

The alkylating agent cyclophosphamide (CYC) is a cytotoxic drug widely given, by either intravenous (IV) or oral routes of administration, to treat solid and hematologic malignancies (1) and rheumatic diseases, e.g., systemic lupus erythematosus, systemic sclerosis, or systemic necrotizing vasculitis (SNV) (2). It has long been recognized that CYC has substantial adverse effects, such as infections, secondary hematologic cancers, and urotoxicity. CYC-related urotoxicity includes hemorrhagic cystitis (3–5) and urinary tract cancers of the bladder (6, 7) and, less commonly, the ureter and renal pelvis (8–10) that may occur long after the first CYC exposure (2). When CYC was used to treat non-Hodgkin's lymphoma, cumulative 5-year hemorrhagic cystitis incidence rates of 12% and 12-year bladder cancer incidence rates of 11% (11) were reported, and the close link between cumulative CYC dose and secondary bladder cancer risk was established (12).

In the setting of SNV, CYC-related urotoxicity has also raised concerns (13, 14). In Wegener's granulomatosis (WG), an SNV frequently characterized by a chronic relapsing course requiring repeat treatment and for which CYC is a mainstay of therapy (15, 16), elevated urinary tract cancer and/or hemorrhagic cystitis rates have been demonstrated in previous studies (13, 14, 17–22). However, those studies showed wide variations in the frequency of urinary tract cancer and/or hemorrhagic cystitis, left unclear whether there is a cutoff dose below which cumulative CYC has no urotoxic effect, and did not provide information on CYC-related urotoxicity for SNV other than WG. While the results of previous studies tended to support a more favorable safety profile of intermittent IV CYC, as opposed to the originally described daily oral regimen (15, 16, 23), whether the administration route also affects CYC urothelial toxicity has not been evaluated (2).

In this retrospective study, we assessed the incidence of, and risk factors for, urinary tract cancer and/or hemorrhagic cystitis in CYC-treated patients with WG or a related small-vessel or medium-sized–vessel SNV, namely, microscopic polyangiitis (MPA), Churg-Strauss syndrome (CSS), and polyarteritis nodosa (PAN).

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Study setting and study population.

The French Vasculitis Study Group (FVSG) is a national collaborative network dedicated to vasculitis clinical care and research. In 2002, a database was created in Microsoft Access in which information was entered for all patients with any vasculitis who had participated in multicenter clinical trials organized by the FVSG since 1980 (24) or who were followed up at the network's center (Hospital Cochin, Paris, from 2002 to the present and Hospital Avicenne, Bobigny, before 2002). The database contains comprehensive information on demographics, medical histories, employment and lifestyle habits, baseline data at vasculitis diagnoses, and followup data in terms of treatments, vasculitis outcomes, and adverse events. Data on new patients are entered and the database is updated regularly with followup information for patients with existing entries who are still seen at the network's referral center or other primary sites.

When this study was launched (April 2008), the FVSG database was searched to select all patients diagnosed as having WG, MPA, CSS, or PAN. Only those patients whose disease fulfilled the Chapel Hill Consensus Conference Nomenclature (25) definitions of any of these 4 SNVs and had been treated with CYC were selected. Exclusion criteria consisted of followup periods of <3 months, missing treatment information, and participation in as yet unpublished clinical trials. Patients who tested positive for cryoglobulinemia and/or cancer-associated vasculitides were also excluded.

As a general rule, IV pulses and daily oral CYC for SNV were prescribed at doses of 0.6–0.7 gm/m2 of body surface, administered in 2–4–week intervals, and 2 mg/kg/day, respectively. The decision to administer CYC orally or by IV was either dictated by treatment protocols or based on the discretion of the treating physician.

Data extraction.

We searched the database for urinary tract cancer and hemorrhagic cystitis occurrences in the selected cases using the following search terms: cystitis, hematuria, hemorrhagic, cystoscopy, cancer, bladder, and urothelial. Any identified hemorrhagic cystitis episode was included in the analysis if there was documentation of nonglomerular hematuria, defined as microscopic or gross hematuria occurring in the absence of urinary red blood cell casts, declining renal function, and/or urinary tract infection. Cystoscopic confirmation of hemorrhagic cystitis was not required. Cases of urinary tract cancer were included upon biopsy confirmation; noninfiltrative bladder cancers, i.e., in situ intraepithelial (pTis) and superficial noninvasive (pTa, pT1) tumors according to the Tumor Nodes Metastasis (TNM) staging system (26), were also included.

In addition, we extracted from the FVSG database the principal demographic, clinical, therapeutic, and followup characteristics of the patients and information on smoking habits. For the patients identified as having experienced a urotoxic event, supplementary data, e.g., concomitant use of the uroprotective antidote sodium 2-mercaptoethanesulfonate (mesna), were gathered through a review of their medical records.

Cumulative urinary tract cancer and/or hemorrhagic cystitis incidence rates.

Cumulative urinary tract cancer and/or hemorrhagic cystitis incidence rates were assessed with life-table analyses using the Kaplan-Meier method. Because the exact date of starting CYC therapy was not available for all the subjects, time to urinary tract cancer or hemorrhagic cystitis was calculated from the date of SNV diagnosis to the censoring date or the date the corresponding urotoxic event occurred. For patients who experienced both events, the date of hemorrhagic cystitis was used to calculate the time to the combined outcome of urinary tract cancer and/or hemorrhagic cystitis.

Calculation of the urinary tract cancer standardized incidence ratio (SIR).

Age-, sex- and period-standardized incidence ratios for urinary tract cancer were calculated to estimate urinary tract cancer incidence in the study population compared to that expected in the general population of France. Expected numbers of urinary tract cancer cases were computed by multiplying person-years under observation for the SNV study population by the appropriate incidence rates for men and women separately in 1-year age groups and 1-year calendar periods. National cancer incidence rates were obtained from the French Institute for Public Health Surveillance, which includes data from 1980 to 2005 (27). Because this registry provides longitudinal data only for bladder cancer among all urinary tract cancer cases for the period considered, we used national infiltrative bladder cancer rates to approximate the expected numbers of urinary tract cancer cases in our study population. For patient-observation periods before 1980 or after 2005, we applied the general population bladder cancer incidence rates for 1980 and 2005, respectively.

The SIR, calculated as the observed number of cancers divided by the expected number of cancers, was determined, and exact 95% confidence intervals (95% CIs) were computed assuming a Poisson distribution (28). SIRs were calculated for the entire study population and within subgroups for sex, WG and non-WG diagnoses (i.e., MPA, CSS, or PAN), route of CYC administration (ever oral versus IV only), and calendar periods of SNV diagnoses (stratified by the median value).

Analysis of urinary tract cancer and hemorrhagic cystitis predictors.

Survival analyses were conducted to identify variables associated with each of the following 3 outcomes: urinary tract cancer, hemorrhagic cystitis, and a combined urotoxic outcome, i.e., urinary tract cancer and/or hemorrhagic cystitis (including patients with urinary tract cancer only, patients with hemorrhagic cystitis only, and patients with both urinary tract cancer and hemorrhagic cystitis). Because we wanted to specifically analyze the risk of urinary tract cancer and/or hemorrhagic cystitis associated with CYC use, these analyses were restricted to a nested subgroup of patients who participated in prospective trials and for whom detailed information on the cumulative CYC dose over time was available. Time to urinary tract cancer or hemorrhagic cystitis was calculated starting from the date of the first CYC administration.

Based on this population, univariate and multivariate Cox proportional hazards models were fitted using urinary tract cancer and/or hemorrhagic cystitis occurrences as the dependent variables and the following 9 predefined parameters as potential explanatory covariates: age (analyzed as a continuous and a median-dichotomized variable), sex, smoking habits (dichotomized as ever smoked versus never smoked), SNV type (dichotomized as WG diagnosis versus non-WG diagnoses), serum creatinine level at diagnosis (analyzed as a continuous and a dichotomous variable for ≥140 versus <140 μmoles/liter), period of diagnosis (analyzed as a continuous and median-dichotomized variable), antineutrophil cytoplasmic antibody serology, CYC administration route (stratified as ever-oral versus exclusively IV), and cumulative CYC dose over time. In addition, the effect of a prior hemorrhagic cystitis episode on developing urinary tract cancer was analyzed. The values used were those assessed at diagnosis except for the variables “prior hemorrhagic cystitis episode” and “cumulative CYC dose,” which were analyzed as time-varying covariates; for the latter, CYC doses over time were calculated as assessed at 12-month increments of followup. To look for a potential threshold effect of cumulative CYC doses on the risk of urotoxicity, this variable was also analyzed as a categorical covariate by stratification according to quartiles or the median values of their distributions. All parameters that achieved P ≤ 0.20 in univariate analyses were entered into the multivariate regression model, and we then applied a backward stepwise selection algorithm. For variables not assessed as time-dependent variables, the proportionality assumption was checked for all Cox regression models by assessing the statistical significance of an interaction between treatment and the natural logarithm of time.

Statistical analysis.

Continuous variables are expressed as the mean ± SD or, for small sample sizes (n ≤ 30), as the median (range) and compared using Student's t-test or Kruskal-Wallis test, respectively. P values less than or equal to 0.05 (2-tailed) were considered significant. The 95% CIs were calculated. Computations were performed with the SAS Statistical Package, version 9.1.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Study sample selection and characteristics.

Among the 1,793 vasculitis patients registered in the FVSG database, we excluded 522 because of other vasculitides and/or cancer-associated vasculitides (n = 139), cryoglobulinemia (n = 35), inclusion in as yet unpublished clinical trials (n = 59), missing data on treatment (n = 231), and followup periods of <3 months (n = 58). Among the remaining 1,271 patients, 805 patients were identified as having received CYC.

The characteristics of the included 805 SNV patients at diagnosis are summarized in Table 1. Among them, 521 (64.7%) had been included at some time during followup in prospective multicenter clinical trials (24, 29–42). Among the 157 patients with PAN (19.5%), 25 (15.9%) had associated hepatitis B virus infection. The median year of SNV diagnosis was 1998; 19 diagnoses (2.4%) were made before 1980. Over the mean ± SD duration of followup of 5.3 ± 4.6 years, 158 deaths (19.6%) occurred.

Table 1. Characteristics of the patients with SNV treated with CYC*
 All patients (n = 805)Nested subgroup (n = 467)
  • *

    The entire group of patients consisted of 805 patients with systemic necrotizing vasculitis (SNV) who had received cyclophosphamide (CYC). The subgroup consisted of the 467 patients for whom detailed information on CYC treatment was available and who were included in the analysis of risk determinants for the development of urotoxic events. MPA = microscopic polyangiitis; CSS = Churg-Strauss syndrome; WG = Wegener's granulomatosis; PAN = polyarteritis nodosa; NA = not applicable; IV = intravenous.

  • Data on smoking status were missing for 73 patients in the entire group and 23 patients in the subgroup.

  • Data on antineutrophil cytoplasmic antibody (ANCA) serology were missing for 110 patients in the entire group and 35 patients in the subgroup.

  • §

    Data on serum creatinine level were missing for 126 patients in the entire group and 22 patients in the subgroup.

Followup duration, mean ± SD years5.3 ± 4.65.2 ± 4.1
Male sex, no. (%)453 (56.3)256 (54.8)
Age, mean ± SD years52.5 ± 16.254.2 ± 15.1
Ever smoked, no. (%)198 (27.0)118 (26.6)
SNV diagnosis, no. (%)  
 MPA182 (22.6)123 (26.3)
 CSS138 (17.1)86 (18.4)
 WG328 (40.7)173 (37.0)
 PAN157 (19.5)85 (18.2)
ANCA-positive serology, no. (%)446 (65.2)270 (62.5)
Serum creatinine, mean ± SD μmoles/liter§166.3 ± 215.5159.0 ± 180.6
Year of diagnosis, no. (%)  
 Before 1998392 (48.7)244 (52.2)
 1998 or later413 (51.3)223 (47.8)
CYC treatment  
 Cumulative dose, mean ± SD gmNA25.1 (38.0)
 Duration, mean ± SD monthsNA10.4 (13.9)
Route of CYC administration, no. (%)  
 Oral104 (12.9)61 (13.1)
 Oral and IV157 (19.5)84 (18.0)
 IV544 (67.5)322 (69.0)

Urotoxic events.

We identified 22 patients who were diagnosed as having hemorrhagic cystitis and 7 who were diagnosed as having urinary tract cancer. Two patients developed hemorrhagic cystitis before urinary tract cancer diagnoses; consequently, 27 patients had urinary tract cancer and/or hemorrhagic cystitis. One patient experienced 2 consecutive hemorrhagic cystitis episodes. For the 466 (of 1,271) SNV patients excluded because of no CYC exposure, no additional urinary tract cancer cases were noted in the database. The 7 urinary tract cancer diagnoses included 6 bladder cancers, 3 of which were staged as noninfiltrative (1 pTis and 2 pT1) and 1 ureter cancer; their principal characteristics are shown in Table 2. Urinary tract cancer was fatal in 5 patients.

Table 2. Characteristics of the 7 SNV patients treated with CYC and diagnosed as having urinary tract cancer*
Patient/ age/ sexSmoker (pack- years)DiagnosisCumulative CYC dose, gmCumulative CYC duration, monthsCYC routeLatency (years)Tumor site (TNM staging)Prior hemorrhagic cystitisUrinary tract cancer treatmentOutcome
  • *

    TNM = Tumor Nodes Metastasis; HBV-associated = hepatitis B virus–associated; TURBT = transurethral resection of bladder tumors; ND = not determined (see Table 1 for other definitions).

  • Reflects cumulative tobacco-smoking exposure at SNV diagnosis.

  • Interval from CYC start to date of urinary tract cancer diagnosis.

  • §

    The interval from the date of hemorrhagic cystitis diagnosis to urinary tract cancer diagnosis was 9.2 years for both cases.

1/60/FNoHBV- associated PAN3913Oral18Ureter (pT3)NoNephroureterectomy, radiation therapy, chemotherapyDied
2/64/MYes (50)CSS1210IV2.9Bladder (pTis)NoTURBT, mitomycinAlive
3/68/MNoMPA8.17IV3.4Bladder (pT1G3)NoTURBT, mitomycinDied
4/61/MYes (10)WG5035.7IV and oral13.5Bladder (ND)Yes§TURBT, mitomycinAlive
5/51/MYes (50)WG140.640.9IV and oral12.4Bladder (pT4)Yes§Radiation therapy, chemotherapyDied
6/51/MYes (40)PAN213.437.7IV and oral6.5Bladder (pT3)NoPercutaneous nephrostomyDied
7/47/MNoWG70.119.8IV and oral8Bladder (pT1G3)NoCystectomy, chemotherapyDied

The 22 hemorrhagic cystitis diagnoses were confirmed by cystoscopy in 15 patients (68.2%) and led to peremptory CYC discontinuation in 13 patients (59.1%). Hemorrhagic cystitis occurred during or shortly after CYC treatment in all but 1 patient, whose hemorrhagic cystitis was diagnosed 2 years after the last CYC administration. Fourteen patients developed hemorrhagic cystitis during the initial CYC treatment period, whereas the remaining 8 patients had hemorrhagic cystitis during CYC retreatment periods. All 5 patients who developed hemorrhagic cystitis during or after IV CYC administration had received mesna. Although all patients recovered from hemorrhagic cystitis, 1 required massive blood transfusions.

The median times from SNV diagnosis to urinary tract cancer, hemorrhagic cystitis, or urinary tract cancer and/or hemorrhagic cystitis were 8.0 years (range 2.9–29.2 years), 1.4 years (range 0.03–15.5 years), and 2.5 years (range 0.03–29.2 years), respectively. The respective 10-year and 20-year cumulative incidence rates were as follows: for urinary tract cancer, 1.3% (95% CI 0–2.7) and 4.6% (95% CI 0–9.3); for hemorrhagic cystitis, 3.2% (95% CI 1.7–4.6) and 9.8% (95% CI 1.8–17.8); and for urinary tract cancer and/or hemorrhagic cystitis, 4.5% (95% CI 2.4–6.5) and 11.1% (95% CI 3.1–19.1) (Figure 1A).

thumbnail image

Figure 1. A, Cumulative incidences of urinary tract cancer (UTC), hemorrhagic cystitis (HC), and urinary tract cancer and/or hemorrhagic cystitis (including patients with urinary tract cancer only, patients with hemorrhagic cystitis only, and patients with both urinary tract cancer and hemorrhagic cystitis) in 805 patients with systemic necrotizing vasculitis (SNV) treated with cyclophosphamide (CYC). B–F, Cumulative incidence of urinary tract cancer and/or hemorrhagic cystitis in a nested subgroup of 467 patients stratified by age at diagnosis (B), SNV diagnosis (C), CYC administration route (D), period of diagnosis (E), and serum creatinine concentration at diagnosis (F). P values were determined by Cox regression analyses. MPA = microscopic polyangiitis; CSS = Churg-Strauss syndrome; PAN = polyarteritis nodosa; IV = intravenous.

Download figure to PowerPoint

SIR for urinary tract cancer.

The duration of followup for the 805 patients totaled 4,230 patient-years. During this period, 1.40 cases of urinary tract cancer were expected. Table 3 shows the estimated SIR for the entire population and for the predefined subgroups. Exclusion from the numerator of the 1 ureter cancer still yielded a significantly elevated SIR of 4.29 (95% CI 1.57–9.34) (P = 0.006).

Table 3. Age-, sex-, and period-standardized incidence ratios for urinary tract cancer assessed in a cohort of 805 patients with SNV*
 Person-yearsNo. of casesSIR (95% CI)P
ObservedExpected
  • *

    SIR = standardized incidence ratio; 95% CI = 95% confidence interval (see Table 1 for other definitions).

All patients4,23071.405.00 (2.01–10.31)0.001
Sex     
 Male2,44361.254.81 (1.77–10.47)0.004
 Female1,78710.156.60 (0.17–36.76)0.28
Diagnosis     
 WG1,71430.505.96 (1.23–17.41)0.03
 MPA, CSS, or PAN2,51740.904.47 (1.22–11.44)0.03
CYC administration route     
 Ever oral1,76950.539.40 (3.05–21.94)<0.001
 IV only2,46120.872.31 (0.28–8.33)0.43
Period of diagnosis     
 Before 19982,78950.925.45 (1.77–12.72)0.005
 1998 or later1,44120.484.15 (0.50–15.00)0.17

Analysis of predictors of urotoxicity.

All subsequent analyses were based on those subjects for whom detailed information on their CYC treatment modalities was available. Of the 521 patients treated within prospective protocols, we excluded 54 subjects because of incomplete data on CYC therapy. The main baseline characteristics of the remaining 467 patients, which included all 27 patients previously identified as having urinary tract cancer or hemorrhagic cystitis, are shown in Table 1. Their mean ± SD cumulative CYC dose was 25.1 ± 38.0 gm (range 0.72–463.0 gm) and duration of CYC treatment was 10.4 ± 13.9 months (range 0.03–176.8 months). CYC was administered exclusively orally in 61 patients (13.1%), exclusively by IV in 322 patients (69.0%) and via both routes in 84 patients (18.0%). The mean CYC dose was significantly higher for ever-oral versus IV only routes (P < 0.0001) and for patients with WG versus patients with non-WG diagnoses (P = 0.01). Exclusive IV CYC therapy was significantly less common among subjects diagnosed with WG than non-WG vasculitides (P = 0.01).

Figure 2 shows the total cumulative CYC dose distributions according to the occurrence of urotoxic events. For all 3 outcomes, CYC doses were higher for patients diagnosed with a urotoxic event. For the patients with urinary tract cancer versus those without urinary tract cancer, the median cumulative CYC doses were 50.0 gm (range 8.1–213.4 gm) versus 12.0 gm (range 0.7–463.0 gm) (P = 0.02). Stratification of these values according to the presence or absence of urinary tract cancer and/or hemorrhagic cystitis yielded median values of 30.1 gm (range 1.1–435.8 gm) versus 11.9 gm (range 0.7–273.4 gm) (P = 0.002).

thumbnail image

Figure 2. Dot plots of the cumulative CYC doses given to the nested subgroup of 467 SNV patients stratified by whether or not they developed urinary tract cancer, hemorrhagic cystitis, or urinary tract cancer and/or hemorrhagic cystitis. For individuals who experienced a urotoxic event, the values shown refer to the cumulative CYC dose given up to the time of the corresponding event. Horizontal bars indicate the median. P values were determined by Kruskal-Wallis test. See Figure 1 for definitions.

Download figure to PowerPoint

Univariate survival analyses for predictors of urinary tract cancer, hemorrhagic cystitis, or urinary tract cancer and/or hemorrhagic cystitis development are summarized in Table 4, and Kaplan-Meier graphs are shown in Figures 1B–F. In the 2 analyses using the effect of cumulative CYC doses as a categorical variable, the 4 subgroups were defined by quartiles of cumulative CYC doses observed either for the 27 study subjects who developed urinary tract cancer and/or hemorrhagic cystitis or for the 7 patients who developed urinary tract cancer.

Table 4. Cox proportional hazards regression analysis of the influence of selected variables on urinary tract cancer and hemorrhagic cystitis occurrence in a cohort of 467 patients with SNV treated with CYC*
 Univariate analysesMultivariate analysis, urinary tract cancer and/or hemorrhagic cystitis
Urinary tract cancerHemorrhagic cystitisUrinary tract cancer and/or hemorrhagic cystitis
HR (95% CI)PHR (95% CI)PHR (95% CI)PHR (95% CI)P
  • *

    HR = hazard ratio; 95% CI = confidence interval; NS = not significant; NC = not calculable (see Table 1 for other definitions).

  • Treated as a time-varying covariate.

Age (by 10-year increment)1.55 (0.86–2.80)0.140.86 (0.65–1.13)0.290.93 (0.72–1.20)0.57
Age ≥57 years versus <57 years1.61 (0.34–7.52)0.550.59 (0.24–1.47)0.260.66 (0.29–1.49)0.32
Male sex5.31 (0.62–45.21)0.131.16 (0.50–2.72)0.731.42 (0.64–3.12)0.39
Ever smoked8.20 (1.47–45.64)0.020.85 (0.31–2.32)0.751.10 (0.46–2.64)0.83
WG diagnosis1.40 (0.31–6.41)0.674.82 (1.87–12.44)0.0013.46 (1.53–7.82)0.0032.96 (1.28–6.85)0.01
ANCA positivity0.15 (0.02–1.22)0.081.80 (0.70–4.64)0.221.22 (0.55–2.69)0.63
Serum creatinine (by 100- μmole/liter increment)1.10 (0.68–1.77)0.710.89 (0.61–1.30)0.540.95 (0.71–1.27)0.74
Serum creatinine ≥140 versus <140 μmoles/liter2.63 (0.46–15.15)0.280.57 (0.17–1.96)0.370.87 (0.32–2.35)0.78
Year of diagnosis (by 1-year increment)1.15 (0.94–1.39)0.170.99 (0.92–1.06)0.771.00 (0.94–1.07)0.95
Year of diagnosis before 1998 versus 1998 or later0.41 (0.04–3.79)0.432.42 (0.92–6.36)0.071.88 (0.79–4.48)0.16NS
Ever-oral route of CYC administration1.71 (0.29–10.27)0.5612.47 (3.66–42.51)<0.00018.00 (2.99–21.42)<0.00015.50 (1.96–15.44)0.001
Cumulative CYC dose (by 10-gm increment)1.02 (0.95–1.10)0.541.17 (1.07–1.28)0.00061.15 (1.07–1.24)0.00011.09 (1.01–1.17)0.03
Cumulative CYC dose (categorical)        
 ≤13 gm1 (reference) 1 (reference) 1 (reference) 
 ≤35 gmNC1.03.11 (0.98–9.92)0.062.15 (0.74–6.22)0.16
 ≤75 gm1.12 (0.14–9.08)0.923.80 (1.12–12.87)0.033.00 (1.03–8.78)0.04
 >75 gm1.52 (0.17–13.77)0.7113.94 (3.15–61.68)0.00058.54 (2.35–30.97)0.001
 ≤12 gm1 (reference) 1 (reference) 1 (reference) 
 ≤50 gm0.43 (0.05–3.61)0.432.55 (0.79–8.20)0.121.73 (0.63–4.77)0.29
 ≤141 gm0.69 (0.08–5.99)0.747.13 (1.91–26.69)0.0044.37 (1.38–13.86)0.01
 >141 gm1.54 (0.10–23.27)0.7545.32 (6.24–329.23)0.000226.10 (5.17–131.90)<0.0001
Prior hemorrhagic cystitis episode5.20 (1.03–26.17)0.046NA NA 

Because of the small numbers of cases of urinary tract cancer and hemorrhagic cystitis, the multivariate analyses were restricted to the combined urinary tract cancer and/or hemorrhagic cystitis outcome. The following variables were included in the primary regression model: cumulative CYC (as a continuous covariate), CYC administration route, WG versus non-WG diagnoses, and period of diagnosis (as a dichotomous variable); the latter variable did not reach statistical significance and was eliminated from the model. The final model included the following 3 variables: cumulative CYC dose (for 10-gm increments) (hazard ratio [HR] 1.09 [95% CI 1.01–1.17], P = 0.03), ever-oral CYC administration (HR 5.50 [95% CI 1.96–15.44], P = 0.001), and WG diagnosis (HR 2.96 [95% CI 1.28–6.85], P = 0.01), as independent predictors of urinary tract cancer and/or hemorrhagic cystitis. These results remained unchanged when adjusted for age, sex, or the addition of interaction terms between cumulative CYC and administration route or WG diagnosis. In addition, we ran the final multivariate model after removing the 97 subjects who died of causes not related to urinary tract cancer during the observation period; the results of this sensitivity analysis were basically unchanged, compared to those obtained in the entire population. For all models, the assumptions of proportional hazards were met for all tested explanatory variables.

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Among the 805 CYC-treated SNV patients followed up for an average of 5.3 years, 7 developed urinary tract cancer (i.e., 6 had bladder cancer and 1 had ureter cancer), and 22 experienced hemorrhagic cystitis. The observed number of urinary tract cancer cases indicated a 5-fold higher incidence compared to the general population. Our analyses of risk factors, which were based on a nested sample, suggested that cumulative CYC dose, ever-oral CYC administration, and WG diagnosis were independent risk factors for urinary tract cancer and/or hemorrhagic cystitis. According to our univariate analyses, current or past tobacco smoking and a previous hemorrhagic cystitis episode also predicted urinary tract cancer.

Acrolein, a CYC metabolite excreted in the urine, is known to be responsible for CYC urotoxicity. Studies in animals have demonstrated that acrolein initiates bladder cancer (43) and hemorrhagic cystitis in a concentration- and time-dependent manner (44, 45). Further support for the notion of acrolein urotoxicity came from observations indicating a protective effect of adjunctive mesna, which neutralizes acrolein, against hemorrhagic cystitis in patients treated with high-dose IV CYC for hematologic malignancies (46–48) and against bladder cancer in rodents (49). Notably, our findings and a previously calculated summary estimate based on studies of WG, rheumatoid arthritis, and lymphoma patients (2) showed that prior hemorrhagic cystitis significantly multiplied the risk of subsequent urinary tract cancer by a factor of 5–7. This clustering of hemorrhagic cystitis with urinary tract cancer requires further investigation but may support the notion that these 2 events are pathogenetically linked.

The 2.4% estimated 10-year cumulative incidence and 5.0 SIR for urinary tract cancer are consistent with previously published bladder cancer incidence rates in patients with WG (21, 22) and a mixed population of patients with WG or MPA (19). As emphasized previously (19, 22), these data contrast with the higher 5% 10-year incidence rate and 30–31 SIR for bladder cancer observed in another WG cohort (17, 18). It is likely that these discrepancies can be explained, in large part, by the higher cumulative CYC dose (median 75 gm) received in the latter cohort. For our nested subgroup with available detailed CYC exposure data, the mean CYC dose was 25 gm. In addition, our findings confirmed the wide range of time intervals from first CYC exposure to urinary tract cancer development and further underscored the recommendation for prolonged and perhaps lifelong surveillance of CYC-treated SNV patients with periodic urinary cytology screening (50).

We could not reproduce the previously proposed values of 25 gm (21), 33 gm (22), or 100 gm (18) as thresholds associated with increased bladder cancer risk in WG. The wide overlapping of total cumulative CYC doses given to patients who experienced urotoxic events versus their counterparts (Figure 2) rather supports the hypothesis that urotoxicity cannot be predicted based on this parameter alone. That 2 cases of urinary tract cancer occurred in patients with seemingly modest cumulative CYC doses of 8 gm and 12 gm (Figure 2) highlights the likely influence of other risk factors. Whether tobacco smoking, which is a major contributor to bladder cancer (51), acts synergistically with CYC to induce bladder cancer is unknown but should be considered in the urotoxic risk assessment of CYC therapy. While a urotoxic threshold may not exist, our data are consistent with a sublinear dose-response profile with a less steep risk increase at lower, as opposed to higher, cumulative CYC exposures.

Perhaps our most original observation is that the route of CYC administration affected the risk of a urotoxic event. Even when adjusted for cumulative CYC dose, our results indicated that ever-oral CYC recipients had a shorter time to urotoxic events than patients treated exclusively with IV CYC. This observation, which could be made due to the wide use of IV CYC for SNV patients, has biologic plausibility in that the greater urotoxicity of daily oral CYC exposure might reflect prolonged urothelium contact with acrolein. However, this finding could also be partly explained by the more common practice of administering mesna and/or hyperhydration in combination with IV CYC. Oncology study results have shown that hyperhydration has protective efficacy against hemorrhagic cystitis comparable to that of mesna (46–48), but those evaluations were carried out in the setting of IV CYC and leave unresolved whether their results can be extrapolated to oral CYC.

Our study also provides new insight into the question of CYC-related urotoxicity in SNVs other than WG. Our SIR estimates indicated a slightly higher incidence of urinary tract cancer in WG patients than in those with non-WG diagnoses (i.e., MPA, CSS, or PAN). Moreover, our multivariate analyses retained WG diagnosis as an independent predictor of urinary tract cancer and/or hemorrhagic cystitis. In light of the higher cumulative CYC doses and its more common oral administration in WG patients, this finding may have resulted from residual confounding, with the variable WG carrying supplemental information on the burden of high, orally administered CYC exposure. However, we cannot exclude the possibility that the risk of urotoxic adverse events is higher in WG due to other, CYC-independent mechanisms.

Limitations of our study include that the identification of hemorrhagic cystitis and urinary tract cancer cases relied on a spontaneous reporting system. Therefore, our study might have underrated their true incidences. In contrast, the SIR for urinary tract cancer could also represent an overestimation because this calculation used general population incidence data on bladder cancer that did not include noninfiltrative tumors and because 1 of the diagnosed urinary tract cancer cases concerned the ureter. However, our inclusion of noninfiltrative bladder cancers in the numerator should have influenced only the time for these tumors to become infiltrative, whereas, due to their uncommon occurrence when compared to that of bladder cancer (52), the numbers of expected cancers of the ureter or the renal pelvis in our study population were only minimal. Interpretation of our SIR estimates must also consider the fact that they do not account for potential deviations in exposures to smoking or other etiologic urinary tract cancer risk factors between the SNV patients and the general population. Use of SNV patients who were not exposed to CYC as the reference group could have partly circumvented the latter shortcoming, but accurate urinary tract cancer incidence rates are unrealistic to ascertain for this population.

Moreover, our discussed findings on risk factors, i.e., CYC dose and administration route, reached statistical significance for the combined hemorrhagic cystitis and/or urinary tract cancer end point but not for urinary tract cancer alone. Thus, in light of the accumulating evidence that hemorrhagic cystitis might predict urinary tract cancer and considering that the point estimates of these variables consistently pointed in the same direction for either outcome, we think that the combined urotoxic end point has both clinical and methodologic relevance. Due to the small number of urotoxic events observed in this cohort and in previous studies (2), our study also highlights the general complexity of exploring this topic in this particular setting of vasculitis because of limited statistical power and precision, and the constrained possibilities to perform multivariate analyses. Finally, although we cannot assert that our findings are generalizable to SNV populations at large, we have no reason to suspect that our study inclusion criteria selected a subgroup with a distinct urinary tract cancer risk pattern.

In conclusion, the results of this study on urotoxic adverse events in CYC-treated SNV patients strengthen the link between urotoxic adverse events and cumulative CYC exposure and highlight oral CYC and perhaps also WG as additional independent risk factors. Our findings imply that hemorrhagic cystitis may have to be recognized as a marker of increased risk of urinary tract cancer, indicate the need for sustained heightened awareness when prescribing this drug to tobacco smokers, and support the suggested use of uroprotective measures, i.e., mesna and/or hyperhydration, during oral CYC treatment (2, 20). These results contribute to defining high-risk populations who should benefit from close and prolonged screening for urinary tract cancer and to preventing this rare but severe adverse event for patients receiving newly initiated CYC treatment for SNV.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Mahr had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Le Guenno, Mahr, Pagnoux, Dhote, Guillevin.

Acquisition of data. Le Guenno, Mahr, Pagnoux, Dhote, Guillevin.

Analysis and interpretation of data. Le Guenno, Mahr, Pagnoux, Dhote, Guillevin.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We are grateful to Ms Janet Jacobson for editorial assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  • 1
    Emadi A, Jones RJ, Brodsky RA. Cyclophosphamide and cancer: golden anniversary. Nat Rev Clin Oncol 2009; 6: 63847.
  • 2
    Monach PA, Arnold LM, Merkel PA. Incidence and prevention of bladder toxicity from cyclophosphamide in the treatment of rheumatic diseases: a data-driven review [review] [published erratum appears in Arthritis Rheum 2010;62:2555]. Arthritis Rheum 2010; 62: 921.
  • 3
    Coggins PR, Ravdin RG, Eisman SH. Clinical pharmacology and preliminary evaluation of Cytoxan (cyclophosphamide). Cancer Chemother Rep 1959; 3: 911.
  • 4
    Goldman RL, Warner NE. Hemorrhagic cystitis and cytomegalic inclusions in the bladder associated with cyclophosphamide therapy. Cancer 1970; 25: 711.
  • 5
    Rubin JS, Rubin RT. Cyclophosphamide hemorrhagic cystitis. J Urol 1966; 96: 3136.
  • 6
    Cyclophosphamide and the bladder. Br Med J 1971; 2: 7267.
  • 7
    Worth PH. Cyclophosphamide and the bladder [letter]. Br Med J 1971; 3: 182.
  • 8
    Fairchild WV, Spence CR, Solomon HD, Gangai MP. The incidence of bladder cancer after cyclophosphamide therapy. J Urol 1979; 122: 1634.
  • 9
    Levine E. Transitional cell carcinoma of the renal pelvis associated with cyclophosphamide therapy. AJR Am J Roentgenol 1992; 159: 10278.
  • 10
    McDougal WS, Cramer SF, Miller R. Invasive carcinoma of the renal pelvis following cyclophosphamide therapy for nonmalignant disease. Cancer 1981; 48: 6915.
  • 11
    Pedersen-Bjergaard J, Ersboll J, Hansen VL, Sorensen BL, Christoffersen K, Hou-Jensen K, et al. Carcinoma of the urinary bladder after treatment with cyclophosphamide for non-Hodgkin's lymphoma. N Engl J Med 1988; 318: 102832.
  • 12
    Travis LB, Curtis RE, Glimelius B, Holowaty EJ, Van Leeuwen FE, Lynch CF, et al. Bladder and kidney cancer following cyclophosphamide therapy for non-Hodgkin's lymphoma. J Natl Cancer Inst 1995; 87: 52430.
  • 13
    Fauci AS, Haynes BF, Katz P, Wolff SM. Wegener's granulomatosis: prospective clinical and therapeutic experience with 85 patients for 21 years. Ann Intern Med 1983; 98: 7685.
  • 14
    Stillwell TJ, Benson RC Jr, DeRemee RA, McDonald TJ, Weiland LH. Cyclophosphamide-induced bladder toxicity in Wegener's granulomatosis. Arthritis Rheum 1988; 31: 46570.
  • 15
    Fauci AS, Wolff SM, Johnson JS. Effect of cyclophosphamide upon the immune response in Wegener's granulomatosis. N Engl J Med 1971; 285: 14936.
  • 16
    Novack SN, Pearson CM. Cyclophosphamide therapy in Wegener's granulomatosis. N Engl J Med 1971; 284: 93842.
  • 17
    Hoffman GS, Kerr GS, Leavitt RY, Hallahan CW, Lebovics RS, Travis WD, et al. Wegener granulomatosis: an analysis of 158 patients. Ann Intern Med 1992; 116: 48898.
  • 18
    Talar-Williams C, Hijazi YM, Walther MM, Linehan WM, Hallahan CW, Lubensky I, et al. Cyclophosphamide-induced cystitis and bladder cancer in patients with Wegener granulomatosis. Ann Intern Med 1996; 124: 47784.
  • 19
    Westman KW, Bygren PG, Olsson H, Ranstam J, Wieslander J. Relapse rate, renal survival, and cancer morbidity in patients with Wegener's granulomatosis or microscopic polyangiitis with renal involvement. J Am Soc Nephrol 1998; 9: 84252.
  • 20
    Reinhold-Keller E, Beuge N, Latza U, de Groot K, Rudert H, Nolle B, et al. An interdisciplinary approach to the care of patients with Wegener's granulomatosis: long-term outcome in 155 patients. Arthritis Rheum 2000; 43: 102132.
  • 21
    Knight A, Askling J, Granath F, Sparen P, Ekbom A. Urinary bladder cancer in Wegener's granulomatosis: risks and relation to cyclophosphamide. Ann Rheum Dis 2004; 63: 130711.
  • 22
    Faurschou M, Sorensen IJ, Mellemkjaer L, Loft AG, Thomsen BS, Tvede N, et al. Malignancies in Wegener's granulomatosis: incidence and relation to cyclophosphamide therapy in a cohort of 293 patients. J Rheumatol 2008; 35: 1005.
  • 23
    De Groot K, Adu D, Savage CO. The value of pulse cyclophosphamide in ANCA-associated vasculitis: meta-analysis and critical review. Nephrol Dial Transplant 2001; 16: 201827.
  • 24
    Guillevin L, Jarrousse B, Lok C, Lhote F, Jais JP, Le Thi Huong Du D, et al, the Cooperative Study Group for Polyarteritis Nodosa. Longterm followup after treatment of polyarteritis nodosa and Churg-Strauss angiitis with comparison of steroids, plasma exchange and cyclophosphamide to steroids and plasma exchange: a prospective randomized trial of 71 patients. J Rheumatol 1991; 18: 56774.
  • 25
    Jennette JC, Falk RJ, Andrassy K, Bacon PA, Churg J, Gross WL, et al. Nomenclature of systemic vasculitides: proposal of an international consensus conference. Arthritis Rheum 1994; 37: 18792.
  • 26
    Kirkali Z, Chan T, Manoharan M, Algaba F, Busch C, Cheng L, et al. Bladder cancer: epidemiology, staging and grading, and diagnosis. Urology 2005; 66: 434.
  • 27
    French Institute for Public Health Surveillance. Estimation de l'incidence et de la mortalité par cancer en France de 1978 à 2000. URL: http://www.invs.sante.fr/surveillance/cancers/estimations_cancers/donnees_localisation/vessie/default.htm.
  • 28
    Sun J, Ono Y, Takeuchi Y. A simple method for calculating the exact confidence interval of the SMR with a SAS function. J Occup Health 1996; 38: 19697.
  • 29
    Guillevin L, Fain O, Lhote F, Jarrousse B, Le Thi Huong D, Bussel A, et al. Lack of superiority of steroids plus plasma exchange to steroids alone in the treatment of polyarteritis nodosa and Churg-Strauss syndrome: a prospective, randomized trial in 78 patients. Arthritis Rheum 1992; 35: 20815.
  • 30
    Guillevin L, Lhote F, Leon A, Fauvelle F, Vivitski L, Trepo C. Treatment of polyarteritis nodosa related to hepatitis B virus with short term steroid therapy associated with antiviral agents and plasma exchanges: a prospective trial in 33 patients. J Rheumatol 1993; 20: 28998.
  • 31
    Guillevin L, Lhote F, Cohen P, Jarrousse B, Lortholary O, Genereau T, et al. Corticosteroids plus pulse cyclophosphamide and plasma exchanges versus corticosteroids plus pulse cyclophosphamide alone in the treatment of polyarteritis nodosa and Churg-Strauss syndrome patients with factors predicting poor prognosis: a prospective, randomized trial in sixty-two patients. Arthritis Rheum 1995; 38: 163845.
  • 32
    Gayraud M, Guillevin L, Cohen P, Lhote F, Cacoub P, Deblois P, et al, the French Cooperative Study Group for Vasculitides. Treatment of good-prognosis polyarteritis nodosa and Churg-Strauss syndrome: comparison of steroids and oral or pulse cyclophosphamide in 25 patients. Br J Rheumatol 1997; 36: 12907.
  • 33
    Guillevin L, Cordier JF, Lhote F, Cohen P, Jarrousse B, Royer I, et al. A prospective, multicenter, randomized trial comparing steroids and pulse cyclophosphamide versus steroids and oral cyclophosphamide in the treatment of generalized Wegener's granulomatosis. Arthritis Rheum 1997; 40: 218798.
  • 34
    Guillevin L, Cohen P, Mahr A, Arene JP, Mouthon L, Puechal X, et al, and the French Vasculitis Study Group. Treatment of polyarteritis nodosa and microscopic polyangiitis with poor prognosis factors: a prospective trial comparing glucocorticoids and six or twelve cyclophosphamide pulses in sixty-five patients. Arthritis Rheum 2003; 49: 93100.
  • 35
    Guillevin L, Mahr A, Cohen P, Larroche C, Queyrel V, Loustaud-Ratti V, et al, for the French Vasculitis Study Group. Short-term corticosteroids then lamivudine and plasma exchanges to treat hepatitis B virus–related polyarteritis nodosa. Arthritis Rheum 2004; 51: 4827.
  • 36
    Cohen P, Pagnoux C, Mahr A, Arene JP, Mouthon L, Le Guern V, et al, for the French Vasculitis Study Group. Churg-Strauss syndrome with poor-prognosis factors: a prospective multicenter trial comparing glucocorticoids and six or twelve cyclophosphamide pulses in forty-eight patients. Arthritis Rheum 2007; 57: 68693.
  • 37
    Martinez V, Cohen P, Pagnoux C, Vinzio S, Mahr A, Mouthon L, et al, for the French Vasculitis Study Group. Intravenous immunoglobulins for relapses of systemic vasculitides associated with antineutrophil cytoplasmic autoantibodies: results of a multicenter, prospective, open-label study of twenty-two patients. Arthritis Rheum 2008; 58: 30817.
  • 38
    Ribi C, Cohen P, Pagnoux C, Mahr A, Arene JP, Lauque D, et al, for the French Vasculitis Study Group. Treatment of Churg-Strauss syndrome without poor-prognosis factors: a multicenter, prospective, randomized, open-label study of seventy-two patients. Arthritis Rheum 2008; 58: 58694.
  • 39
    Pagnoux C, Mahr A, Hamidou MA, Boffa JJ, Ruivard M, Ducroix JP, et al. Azathioprine or methotrexate maintenance for ANCA-associated vasculitis. N Engl J Med 2008; 359: 2790803.
  • 40
    De Menthon M, Cohen P, Pagnoux C, Sibilia J, Jaussaud R, Delaval P, et al. Infliximab versus rituximab for refractory Wegener's granulomatosis: a prospective, randomized, multicenter study on 21 patients [abstract]. APMIS 2009; 117 Suppl: S72.
  • 41
    Ribi C, Cohen P, Pagnoux C, Mahr A, Arene JP, Puechal X, et al, for the French Vasculitis Study Group. Treatment of polyarteritis nodosa and microscopic polyangiitis without poor-prognosis factors: a prospective randomized study of one hundred twenty-four patients. Arthritis Rheum 2010; 62: 118697.
  • 42
    Hiemstra TF, Walsh M, Mahr A, Savage CO, de Groot K, Harper L, et al, for the European Vasculitis Study Group (EUVAS). Mycophenolate mofetil vs azathioprine for remission maintenance in antineutrophil cytoplasmic antibody–associated vasculitis: a randomized controlled trial. JAMA 2010; 304: 23818.
  • 43
    Cohen SM, Garland EM, St John M, Okamura T, Smith RA. Acrolein initiates rat urinary bladder carcinogenesis. Cancer Res 1992; 52: 357781.
  • 44
    Batista CK, Brito GA, Souza ML, Leitao BT, Cunha FQ, Ribeiro RA. A model of hemorrhagic cystitis induced with acrolein in mice. Braz J Med Biol Res 2006; 39: 147581.
  • 45
    Cox PJ. Cyclophosphamide cystitis—identification of acrolein as the causative agent. Biochem Pharmacol 1979; 28: 20459.
  • 46
    Hows JM, Mehta A, Ward L, Woods K, Perez R, Gordon MY, et al. Comparison of mesna with forced diuresis to prevent cyclophosphamide induced haemorrhagic cystitis in marrow transplantation: a prospective randomised study. Br J Cancer 1984; 50: 7536.
  • 47
    Shepherd JD, Pringle LE, Barnett MJ, Klingemann HG, Reece DE, Phillips GL. Mesna versus hyperhydration for the prevention of cyclophosphamide-induced hemorrhagic cystitis in bone marrow transplantation. J Clin Oncol 1991; 9: 201620.
  • 48
    West NJ. Prevention and treatment of hemorrhagic cystitis. Pharmacotherapy 1997; 17: 696706.
  • 49
    Habs MR, Schmahl D. Prevention of urinary bladder tumors in cyclophosphamide-treated rats by additional medication with the uroprotectors sodium 2-mercaptoethane sulfonate (mesna) and disodium 2,2′-dithio-bis-ethane sulfonate (dimesna). Cancer 1983; 51: 6069.
  • 50
    Bernatsky S, Ramsey-Goldman R, Clarke AE. Malignancies and cyclophosphamide exposure in Wegener's granulomatosis. J Rheumatol 2008; 35: 113.
  • 51
    Zeegers MP, Tan FE, Dorant E, van den Brandt PA. The impact of characteristics of cigarette smoking on urinary tract cancer risk: a meta-analysis of epidemiologic studies. Cancer 2000; 89: 6309.
  • 52
    ParkinDM WhelanSL FerlayJ TeppoL ThomasDB, editors. Cancer incidence in five continents, vol. VIII. IARC scientific publication 155. Lyon (France): International Agency for Research on Cancer (WHO); 2002. URL: http://www.iarc.fr/en/publications/pdfs-online/epi/sp155/index.php.