Cyclophosphamide-Induced Cystitis and Bladder Cancer in
Patients with Wegener Granulomatosis
Cheryl Talar-Williams, MPH, PA-C; Yasmine M. Hijazi, MD; McClellan M. Walther, MD;
W. Marston Linehan, MD; Claire W. Hallahan, MS; Irina Lubensky, MD; Gail S. Kerr, MD;
Gary S. Hoffman, MD; Anthony S. Fauci, MD; and Michael C. Sneller, MD
Objective: To describe the incidence of, clinical manifes-
tations off and risk factors for cyclophosphamide-induced
urinary bladder toxicity in patients treated for nonmalig-
W egener granulomatosis is a necrotizing gran-
ulomatous vasculitis that typically involves the
upper and lower respiratory tract and the kidneys.
Standard therapy for Wegener granulomatosis in-
Design: Retrospective analysis of patients followed at the cludes the daily administration of low-dose oral cy-
National Institutes of Allergy and Infectious Diseases from clophosphamide and corticosteroid therapy (1, 2).
1967 to 1993.
This therapeutic regimen has dramatically improved
Setting: The Warren G. Magnuson Clinical Center of the the survival of patients with this otherwise fatal
National Institutes of Health (NIH). disease: More than 90% of patients treated with
Patients: 145 patients who received cyclophosphamide cyclophosphamide and corticosteroid therapy im-
for the treatment of Wegener granulomatosis and were prove markedly, and 75% achieve complete remis-
followed for 0.5 to 27 years (median, 8.5 years), for a total sion of disease (1, 3). However, extended follow-up
of 1333 patient-years. of patients with Wegener granulomatosis indicates
Measurements: Clinical characteristics, cystoscopic find- that relapse of disease is common, and repeated
ings, results of cytologic examination of urine, surgical and prolonged courses of cyclophosphamide can be
pathology, and total dose and duration of cyclophospha- associated with serious long-term toxicities, includ-
mide therapy were recorded and analyzed using a comput- ing bone marrow suppression, infertility, hemor-
er-based information retrieval system.
rhagic cystitis, and the development of cancer (3).
Results: Nonglomerular hematuria occurred in 73 of 145 Hemorrhagic cystitis and bladder cancer are well-
patients treated with cyclophosphamide (50%). Sixty of recognized complications of cyclophosphamide ther-
the 73 patients with nonglomerular hematuria (82%) had apy for both malignant (4) and nonmalignant dis-
cystoscopy at the NIH. Forty-two of the 60 patients (70%)
eases (3, 5-7). However, the relations among total
who had cystoscopy had macroscopic changes consistent
with cyclophosphamide-induced bladder injury. Seven pa- cyclophosphamide dose, the development of cystitis,
tients (5%) developed transitional-cell carcinoma of the and the occurrence of bladder cancer have not been
urinary bladder. In 6 of these 7 patients, the total cumula- well defined. In this report, we describe the inci-
tive cyclophosphamide dose exceeded 100 g, and the cu- dence, clinical manifestations, and natural history of
mulative duration of cyclophosphamide therapy exceeded cyclophosphamide-mediated urotoxicity in a cohort
2.7 years. Before they were given a diagnosis of bladder of patients with Wegener granulomatosis. We iden-
cancer, all 7 patients had had one or more episodes of tify risk factors associated with the development of
microscopic or gross nonglomerular hematuria. In con- cyclophosphamide-induced bladder cancer and dis-
trast, none of the 72 patients who had never had nonglo-
cuss recommendations for surveillance.
merular hematuria developed bladder cancer. Cox propor-
tional hazards regression analysis showed that only
microscopic nonglomerular hematuria was a significant
risk factor for the development of bladder cancer (P < 0.01).
Conclusion: Long-term oral cyclophosphamide therapy is
associated with substantial urotoxicity, including the de-
velopment of transitional-cell carcinoma of the urinary Patients
bladder. In this cohort of patients, the estimated incidence From 1967 to 1993, 145 patients with Wegener
of bladder cancer after the first exposure to cyclophospha- granulomatosis were treated with cyclophosphamide
mide was 5% at 10 years and 16% at 15 years. Nonglo- at the Warren G. Magnuson Clinical Center of the
merular hematuria was a frequent manifestation of cyclo- National Institutes of Health (NIH). The clinical
phosphamide-induced cystitis, and it identified a subgroup
features of the underlying disease in all but 3 of
of patients at high risk for the development of bladder
these patients have been reported previously (3);
the clinical and demographic characteristics of these
Ann Intern Med. 1996;124:477-484. patients are summarized in Table 1. Individual pa-
From the National Institutes of Health, Bethesda, Maryland. For tients were followed for 0.5 to 27 years (median, 8.5
current author addresses, see end of text. years), for a total of 1333 patient-years.
1 March 1996 • Annals of Internal Medicine • Volume 124 • Number 5 477
Table 1. Demographic and Clinical Characteristics of 145 Patients were evaluated at the NIH every 1 to 3
Patients with Wegener Granulomatosis Who
Were Treated with Cyclophosphamide*
months. Those who achieved remission of disease
and maintained it for 1 year were subsequently seen
Characteristic Patients, n(%) every 6 months. For each patient, urinalysis was
done at every visit, and a cytologic examination of
Sex urine was done every 6 to 12 months.
Female 72 (49) All patients received cyclophosphamide as part of
Race clinical research protocols approved by the National
Black 3 (2) Institute of Allergy and Infectious Diseases (NI-
Other 2(1.4) AID) Institutional Review Board, the NIAID Clin-
Smoker 61 (42) ical Director, and the director of the NIH Clinical
Nonsmoker 73 (50) Center. All patients gave written informed consent.
Unknown 11 (8)
History of renal disease
Presentt 116 (80) Urine Cytology
Absent 29 (20)
Cumulative cyclophosphamide dose Cytologic examination of urine was done at each
<50 g 52 (36) evaluation. Sediments obtained from voided urine
50-100 g 42(29)
101-200 g 31(21.4) specimens were fixed in Saccomanno solution (Le-
>200g 20(13.8) rner Laboratories, Pittsburgh, Pennsylvania), immo-
* Median age of patients at time of first cyclophosphamide dose, 40 years (range, 10 to bilized on membrane filters (Millipore, Chicago, Il-
t Glomerulonephritis due to active Wegener granulomatosis was documented at some
linois) or by cytospin, placed in 95% ethanol, and
point in the course of illness. stained with Papanicolaou stain. Urine samples ob-
tained as much as 6 months before each cystoscopic
examination were reviewed retrospectively and were
Treatment Protocol correlated with subsequent bladder biopsy speci-
We used the standard cyclophosphamide treat- mens. Cellular cytologic features were placed in the
ment regimen, which has been described previously following categories according to the following cri-
(1, 3). Therapy consisted of 1) oral cyclophospha- teria (8-10):
mide, 2 mg/kg of body weight per day and 2) pred- 1. Negative: no important epithelial abnormali-
nisone, 1 mg/kg of body weight per day. If patients ties.
improved substantially after the first month of treat- 2. Atypia: some nuclear abnormalities in epithe-
ment, the prednisone dose was gradually tapered to lial cells, but the changes could not be definitely
an alternate-day regimen, and prednisone therapy placed in categories 3, 4, or 5.
was eventually discontinued. Cyclophosphamide 3. Therapeutic or viral: cytologic changes consis-
therapy was continued for at least 1 year after pa- tent with polyomavirus infection or cyclophospha-
tients achieved complete remission. The cyclophos- mide toxicity. In the absence of diagnostic inclusions,
phamide dose was then tapered by 25-mg decre- the two types of changes are indistinguishable, and
ments of therapy every 2 to 3 months until we therefore grouped them together. The nuclear
discontinuation of therapy or until disease recur- enlargement and hyperchromasia associated with
rence required an increase in dose. The cyclophos- polyomavirus or chemotherapeutic effect should be
phamide dose was adjusted as needed to keep the distinguished from high-grade dysplasia or carcinoma.
absolute neutrophil count higher than 1.0 X 106/L. 4. Dysplasia or possible low-grade transitional-
If substantial toxicity required the permanent dis- cell carcinoma: a few cells in a voided urine sample,
continuation of cyclophosphamide therapy and if either singly or in clusters, that have slightly en-
signs of active vasculitis were present, azathioprine, larged, irregular nuclei with increased granularity in
chlorambucil, or (after 1990) low-dose weekly meth- chromatin distribution and small or absent nucleoli.
otrexate therapy was allowed. 5. Transitional-cell carcinoma: high-grade lesions
Eight patients with fulminant disease initially re- meeting unequivocal criteria of malignancy.
ceived intravenous cyclophosphamide at daily doses
of 3 to 5 mg/kg. When their disease stabilized, these Definition of Terms
patients were switched to the standard oral cyclo- Nonglomerular hematuria was defined as micro-
phosphamide regimen. Fourteen patients received scopic or gross hematuria not associated with the
monthly pulses of intravenous cyclophosphamide, 1 presence of erythrocyte casts or declining renal
g/m2 body surface area; 13 of these 14 patients also function. Glomerulonephritis causing hematuria as-
received one or more courses of the standard oral sociated with erythrocyte casts (glomerular hematu-
cyclophosphamide regimen. Thus, 144 of the 145 ria) occurred at least once in 116 of the 145 patients
patients received oral cyclophosphamide therapy for (Table 1). If hematuria persisted after the treatment
some period of time. of glomerulonephritis and the disappearance of
478 1 March 1996 • Annals of Internal Medicine • Volume 124 • Number 5
erythrocyte casts, or if hematuria not associated Table 2. Clinical Characteristics of the 73 Patients Treated
w i t h Cyclophosphamide W h o Developed
with the presence of erythrocyte casts ever devel- Nonglomerular Hematuria
oped, patients were considered to have nonglomeru-
lar hematuria and had cystoscopy (see below). Characteristic Patients, n(%)
Cyclophosphamide-induced cystitis was defined
as nonglomerular hematuria associated with charac- Initial episode of hematuria
Microscopic 41 (56)
teristic cystoscopic bladder changes. These changes Gross 32 (44)
included patchy areas of neovascularity and telangi- Had initial episode of hematuria while receiving
cyclophosphamide 55 (75)
ectasia manifested as an increased number of tor- Had initial episode of hematuria after discontinuation
tuous, thin-walled veins and small areas of hemor- of cyclophosphamide 18 (25)
Had recurrent episodes of hematuria 28 (38)
rhage in or under the bladder epithelium. The Examined by cystoscopy 60 (82)
mucosa between the hypervascular areas may ap- Cystitis documented 42/60 (70)
Developed bladder cancer 7 (10)
pear normal or pale with decreased vascularity.
Cystoscopy was done to evaluate nonglomerular Results
hematuria (microscopic or gross) in patients receiv-
ing cyclophosphamide. Only the results of cystosco- Hematuria
pies done at the NIH Clinical Center by members Seventy-three of 145 patients treated with cyclo-
of the Urologic Oncology Section of the National phosphamide (50%) developed nonglomerular he-
Cancer Institute are included in this report. Most of maturia; the median time to development for all
these cystoscopies were done by two of the authors; patients was 37 months of receipt of cyclophospha-
all cystoscopy results were reviewed by these two mide (95% CI, 32 to 55 months), and the median
authors. Random biopsies were done only if the dose before development was 124 g (CI, 82 to 149
results of cytologic examination of urine suggested g) (Table 2, Figure 1). Forty-one patients (56%)
malignancy. Patients having cystoscopy had intrave- presented with microscopic hematuria; 32 (44%)
nous pyelography or retrograde pyelography, or both, presented with gross hematuria. Twenty-eight pa-
at least once to evaluate the upper urinary tract. tients had more than one recurrent episode of non-
glomerular hematuria (microscopic or gross), some-
Statistical Analysis times years after cyclophosphamide therapy had
The frequencies of clinical findings were com- been discontinued. Eighteen of the 73 patients de-
pared by using the Fisher exact test for association veloped hematuria after cyclophosphamide therapy
with bladder cancer and nonglomerular hematuria; was discontinued. Manifestations of active vasculitis
adjustments were made for multiple comparisons of necessitated that cyclophosphamide therapy be con-
microscopic and gross hematuria with bladder can- tinued in 26 of the 55 patients who developed non-
cer using the modified Bonferroni method (11). The glomerular hematuria while receiving this therapy.
effects of fixed covariates (sex, history of smoking, Three patients (2.1%) had gross hematuria with
age at onset of Wegener granulomatosis disease, clinically significant blood loss. Two men and one
and duration of disease before first cyclophospha- woman were transfused with 35 to 80 units of blood
mide treatment) and time-varying covariates (micro- and required intravesical treatment with formalin or
scopic hematuria, gross hematuria, total cyclophos- silver nitrate to stop the bleeding. Bleeding was attrib-
phamide dose, and duration of cyclophosphamide uted to cyclophosphamide-induced cystitis in two
therapy) on the development of bladder cancer patients and was related to tumor (of colonic origin)
were examined using Cox proportional hazards re- invasion of the bladder in one patient. The total
gression analysis (12, 13). Medians and other per- doses and durations of cyclophosphamide therapy
centiles for variables dependent on follow-up time for these three patients were 55 g administered over
were estimated by using the Kaplan-Meier method 12 months, 128 g administered over 43 months, and
(14). The cumulative distributions determined by 163.7 g administered over 36.5 months.
the Kaplan-Meier method were compared with the Sixty of the 73 patients with nonglomerular he-
log-rank test. Risk estimates for the development of maturia (82%) had a total of 139 cystoscopies done
bladder cancer in patients with Wegener granulo- at the NIH (mean, 2.3 cystoscopies; range, 1 to 13
matosis who were treated with cyclophosphamide cystoscopies). Cystoscopy was done with similar fre-
were determined by comparing observed rates with quency in patients with and those without glomer-
the expected rates for the United States population, ulonephritis. Eleven of 29 patients (38%) who had
which were obtained from the Surveillance, Epide- no episodes of glomerulonephritis during the study
miology, and End Results (SEER) Cancer Statistics period had a total of 20 cystoscopies; 49 of 116
Review, 1973-1991 (15). patients (42%) who had at least one episode of
1 March 1996 • Annals of Internal Medicine • Volume 124 • Number 5 479
years]). The effect of smoking on the development
of hematuria is shown in Figure 2. The proportions
of smokers and nonsmokers who developed hema-
turia (47% and 57%, respectively) were not statis-
tically different, but smokers developed hematuria
after fewer months of cyclophosphamide treatment
than did nonsmokers. The median time to the de-
velopment of hematuria was 34 months for smokers
compared with 56 months for nonsmokers (P =
0.03). Exact data are not available, but, in our ex-
perience, most patients with microscopic nonglo-
merular hematuria have no symptoms related to the
Transitional-Cell Carcinoma of the Bladder
Seven patients (5%) treated with oral cyclophos-
phamide developed transitional-cell carcinoma of
the urinary bladder at a median age of 53 years
(range, 49 to 83 years). The time receiving cyclo-
phosphamide and the cumulative dose at which 5%
of the sample developed bladder cancer were 46
months and 113 g, respectively. In 6 of the 7 patients
who developed bladder cancer, the total cumulative
cyclophosphamide dose exceeded 100 g, and the
Figure 1. Cumulative risk for developing nonglomerular hematu-
cumulative duration of cyclophosphamide treatment
ria. Top. As a function of time receiving cyclophosphamide therapy. Bot- exceeded 2.7 years (Table 3). Before bladder cancer
tom. As a function of total cyclophosphamide dose. was diagnosed, all 7 patients had had one or more
episodes of microscopic or gross nonglomerular he-
glomerulonephritis had a total of 119 cystoscopies. maturia (Tables 2 and 3). In contrast, none of the
Forty-two of the 60 patients (70%) having cysto- 72 patients who did not have a documented episode
scopy each had macroscopic changes consistent with of nonglomerular hematuria developed bladder can-
cyclophosphamide-induced bladder injury on at cer.
least one occasion. Three patients were given a At the time of diagnosis, three patients had non-
diagnosis of bladder cancer on the basis of their invasive papillary lesions, one had invasion into the
initial cystoscopies, which also showed changes con- lamina propria, one had invasion into the deep mus-
sistent with cyclophosphamide-induced bladder in- cle, one had carcinoma in situ, and one had a
jury. In the remaining 15 patients, the bladder mu- noninvasive papillary lesion and carcinoma in situ.
cosa appeared normal or showed nonspecific Five patients had transurethral resection of their
abnormalities. The cystoscopic manifestations of bladder tumors. Five patients had bladder instilla-
cystitis often changed over time in individual pa- tion therapy with thiotepa, mitomycin, or Mycobac-
tients. These abnormalities resolved in some pa- terium bovis Bacillus Calmette-Guerin vaccine. Two
tients; in others, bladder lesions remained stable or patients ultimately required cystectomy because of
progressed. No patient had evidence of Wegener
granulomatosis involving the bladder.
Figure 1 shows the Kaplan-Meier estimates of
the cumulative percentage of patients developing
nonglomerular hematuria with cyclophosphamide
therapy. Although men developed hematuria more
often than women, a sex difference over time was
not significant after adjustment for the total dose
and duration of cyclophosphamide therapy. Al-
though the median ages of men (39 years) and
women (43 years) at the time of initial cyclophos-
phamide treatment were not statistically different,
men tended to develop hematuria earlier than
Figure 2. Cumulative risk for developing nonglomerular hematuria
women (median for men, 56 years [CI, 48 to 62 as a function of time receiving cyclophosphamide therapy for smok-
years]; median for women, 65 years [CI, 55 to 73 ers compared with nonsmokers.
480 1 March 1996 • Annals of Internal Medicine • Volume 124 • Number 5
Table 3. Patients with Bladder Cancer after Cyclophosphamide Treatment for Wegener Granulomatosis*
Patient Age/ Smoker Cumulative Duration Gross Micro- Results of La- Lag Tumor Date Outcome Follow-up
Sex Cyclophos- of Hema- scopic Urine tencyt Time* Grade Diag- Status
phamide Cyclophos- turia Hema- Cytology Stage! nosed
Dose phamide turia
y 9 y y
1 52/M Yes 19 0.6 No Yes Negative 0.6 0 I-III, Ta 5/90 TURBT; no recur- Alive
rence as of
2 70/M Yes 144 2.7 Yes Yes Dysplasia 11.7 9 ll-lll, Ta 12/90 1/91 Thiotepa Alive
3 55/F No 184 8 Yes Yes Therapeutic 10.8 0 l-lll, Ta 4/88 TURBT; 12/89 re- Deceased
effect currence, BCG
4 83/M Yes 111 5.1 No Yes Atypia 15.3 10 ll-lll, T1 7/91 TURBT; 12/91 mit- Deceased
5 72/F Yes 105 4.5 Yes No Atypia 5.0 0.6 ll-lll, 4/83 7/83 stage D; Deceased
T3aN1 8/83 radical cys-
6 51/M Yes 215 4.2 Yes Yes Dysplasia, 8.2 0 Ill-Ill, 11/87 TURBT; 12/90 re- Alive
suspicious Ta and currence, BCG
for TCC Tis then mitomycin;
7 49/M Yes 251 3.6 No Yes Dysplasia, 12.9 3.7 Tis 12/93 Mitomycin Alive
* BCG = Bacillus Calmette-Guerin; TCC = transitional-cell carcinoma; TURBT = transurethral resection of bladder tumor,
t Day 1 of cyclophosphamide therapy to date of diagnosis of bladder cancer.
* Date of last cyclophosphamide dose to date of diagnosis of bladder cancer.
§ TNM (tumor, node, metastasis) classification.
invasive disease; a third patient with invasive blad- seen in our cohort. No neoplasms of the ureter or
der cancer did not have a cystectomy because of renal pelvis were seen in our patient sample.
severe underlying medical problems (Table 3). The rate of occurrence of bladder cancer among
The latency period, defined as the time from the the 145 patients receiving cyclophosphamide who
initiation of cyclophosphamide therapy to the diag- had Wegener granulomatosis was compared with
nosis of bladder cancer, ranged from 7 months to that in the SEER Cancer Statistics Review (1973-
15.3 years (Table 3). The lag time, defined as the 1991) for adult men and women in the general U.S.
time from the date of the last cyclophosphamide population (15). A 31-fold increase in the incidence
dose to the date of the diagnosis of bladder cancer, of bladder cancer (CI, 13-fold to 65-fold) was seen
ranged from 0 to 10 years. The patient who devel- in our patients on the basis of the SEER rate ad-
oped bladder cancer after only 7 months of therapy justed for all ages; those younger than 65 years of
was receiving cyclophosphamide for active Wegener age had a 51-fold increase (CI, 14-fold to 132-fold;
granulomatosis at the time of diagnosis. The patient P < 0.001).
who developed bladder cancer 15 years after his
first dose of cyclophosphamide had been in remis-
sion and had not received cyclophosphamide for 10
years. Figure 3 shows the Kaplan-Meier estimate of
the cumulative risk for developing bladder cancer as
a function of time from the first cyclophosphamide
Six of the seven patients with bladder cancer had
a history of smoking. None had a history of expo-
sure to aromatic amines, including aniline dyes. No
patient had bladder symptoms or cystoscopic find-
ings compatible with decreased bladder capacity.
Other clinical conditions that may enhance cyclo- Figure 3. Cumulative risk for bladder cancer in 145 patients
phosphamide-induced bladder toxicity, such as ob- treated w i t h cyclophosphamide w h o had Wegener granulomatosis;
time from first cyclophosphamide dose t o development of bladder
structive uropathy and neurogenic bladder, were not cancer. Dashed lines represent 95% CIs.
1 March 1996 • Annals of Internal Medicine • Volume 124 • Number 5 481
Predictors of the Development of Bladder Cancer amination of urine showed no epithelial cell abnor-
We analyzed the fixed covariates (sex, history of malities in 10 patients, therapeutic effect in 1 pa-
smoking, age at onset of Wegener granulomatosis, tient, and atypia in 1 patient. In 3 patients, cytologic
and duration of disease before first cyclophospha- examination of urine was not done within 2 months
mide treatment) and the time-varying covariates of cystoscopy. For the detection of bladder cancer
(microscopic hematuria, gross hematuria, total cy- in the 60 patients with nonglomerular hematuria
clophosphamide dose, and duration of cyclophospha- who had cystoscopy, the finding of dysplasia on
mide therapy) using the Cox proportional hazards cytologic examination of urine had a sensitivity of
regression model to determine whether any of these 43% and a specificity of 100%; the finding of atypia
factors were significant predictors of the develop- had a sensitivity of 29% and a specificity of 89%.
ment of bladder cancer. Of these variables, only A total of 34 bladder biopsy specimens were
microscopic nonglomerular hematuria was found to obtained from 25 patients. Biopsy specimens from 9
be statistically significantly associated with the de- patients were interpreted as showing acute and
velopment of bladder cancer (P < 0.01). Bladder chronic hemorrhagic cystitis or chronic inflamma-
cancer developed in 6 of 51 patients (12%) in whom tion. Epithelial atypia was seen in 7 patients, epi-
microscopic nonglomerular hematuria was docu- thelial dysplasia was seen in 1 patient, and no ab-
mented at some point during follow-up and in 1 of normalities were seen in 1 patient. Transitional-cell
94 patients without microscopic nonglomerular he- carcinoma of the bladder was reported in one or
maturia (P = 0.02). more biopsy specimens from 7 patients, of whom 1
A history of smoking tobacco was not found to had no cystoscopically identifiable lesion.
be an independent risk factor for the development
of bladder cancer. However, more patients with a
history of smoking developed bladder cancer at a Discussion
lower total dose of cyclophosphamide (P = 0.03)
and after fewer months of cyclophosphamide ther- Cyclophosphamide is highly effective for the
apy (P < 0.01). treatment of Wegener granulomatosis and other se-
rious autoimmune diseases, but its use can be asso-
ciated with substantial toxicity of the urinary tract.
Urine Cytology The drug is metabolized in the liver to various
Five of the seven patients with bladder cancer chlormethine metabolites and acrolein (16). The
had abnormal results on cytologic examination of chlormethine metabolites are responsible for the
urine before cancer was diagnosed by cystoscopy therapeutic effects; the toxic effects on the urinary
(Table 3). Cellular atypia was seen in two of these tract are thought to be mediated by the acrolein
patients, and dysplasia suspicious for transitional- metabolite (17). Although the entire urinary collect-
cell carcinoma was seen in three patients. In two of ing system is at risk for acrolein-mediated toxicity,
the seven patients, cytologic results did not suggest the bladder is most susceptible because of its pro-
malignancy. Both of these patients had low-grade longed exposure to the drug. Hemorrhagic cystitis
malignant lesions (Table 3). Patient 1, in whom and the induction of bladder cancer are well-recog-
bladder cancer was diagnosed after only 7 months nized complications of oral cyclophosphamide therapy
of cyclophosphamide therapy, had had normal re- in patients with Wegener granulomatosis, systemic lu-
sults on cytologic examination of urine 2 days be- pus erythematosus, and rheumatoid arthritis (3,
fore his bladder cancer was diagnosed by cystos- 5-7). However, the incidence of cyclophosphamide-
copy. The cytologic changes in urine seen in patient induced cystitis and its relation to the development
3 were thought to be due to therapeutic or viral of bladder cancer have not been defined.
effects (Table 3). In our study, nonglomerular hematuria devel-
The results of cytologic examination of urine did oped in 50% of patients treated with oral cyclophos-
not consistently correlate with the degree of bladder phamide. It occurred earlier in patients who smoked
injury seen at cystoscopy in patients with cyclophos- tobacco and was chronic and recurrent in 38% of
phamide-induced cystitis. All 42 patients with evi- patients, even after the cessation of cyclophosphamide
dence of cyclophosphamide-induced bladder changes therapy. Seventy percent of patients with nonglomeru-
had at least one cytologic examination of urine done lar hematuria who had cystoscopy had macroscopic
within 2 months of one or more cystoscopic exam- changes of cyclophosphamide-induced cystitis. Thus,
inations. No epithelial-cell abnormalities were seen nonglomerular hematuria usually indicates the pres-
in 26 patients; atypia was seen in 5 patients; and ence of cyclophosphamide-induced bladder injury.
therapeutic changes were seen in 15 patients. In the Bladder cancer developed in 5% of our patients
15 patients with nonglomerular hematuria and no treated with cyclophosphamide; compared with the
cystoscopic evidence of bladder injury, cytologic ex- general population, this represents a 31-fold in-
482 1 March 1996 • Annals of Internal Medicine • Volume 124 • Number 5
crease in the incidence of bladder cancer and a limited usefulness in monitoring cyclophosphamide-
51-fold increase for persons younger than 65 years treated patients for the development of bladder can-
of age. By Kaplan-Meier analysis (Figure 3), the cer. In three of the seven patients with bladder
estimated incidences of bladder cancer after first cancer, such examination showed changes suggestive
exposure to cyclophosphamide are 2% at 5 years, of transitional-cell carcinoma. Atypia was seen in
5% at 10 years, and 16% at 15 years. The rate of two patients and therapeutic or viral changes were
bladder cancer in our study is higher than that seen in one patient; cytologic examination of urine
previously reported for patients with Wegener gran- in one patient showed no epithelial cell abnormali-
ulomatosis who are treated with oral cyclophospha- ties (Table 3). These results, along with the findings
mide (1, 3, 5, 18), probably because follow-up was of other investigators (5), indicate that cytologic
longer in our study. It should be emphasized that examination of voided urine is a relatively insensi-
this is a minimum estimate of the occurrence of tive method for detecting low-grade bladder cancer
bladder cancer in patients receiving cyclophospha- in patients treated with cyclophosphamide. How-
mide. Given the long latency period between the ever, the sensitivity of cytologic examination of
initiation of cyclophosphamide therapy and the di- urine increases for higher-grade malignant lesions.
agnosis of bladder cancer (Table 3), the cumulative The most sensitive method for detecting cyclo-
incidence of bladder cancer in this patient cohort phosphamide-induced cystitis or early malignant le-
may continue to increase with time. sions is a routine urinalysis with careful examination
We examined several clinical variables for their for nonglomerular microscopic hematuria. Because
association with the development of bladder cancer. nonglomerular hematuria may be the first manifesta-
Neither the total dose of cyclophosphamide nor the tion of bladder cancer or, more likely, of cyclophos-
cumulative time receiving cyclophosphamide ther- phamide-induced cystitis, we strongly recommend that
apy were found to be significant risk factors for the all patients treated with cyclophosphamide have uri-
development of bladder cancer in our patients. nalysis every 3 to 6 months, even after cyclophospha-
However, bladder cancer developed in 6 of the 51 mide therapy is discontinued. Patients who develop
patients (12%) whose cumulative cyclophosphamide nonglomerular hematuria (microscopic or gross)
dose exceeded 100 g. Continued follow-up of these should be evaluated by cystoscopy. If cystitis is doc-
patients may show the cumulative dose of cyclo- umented, cyclophosphamide therapy may be contin-
phosphamide to be an independent risk factor. Al- ued if medically necessary to treat potentially life-
though 6 of the 7 patients had a history of smoking threatening manifestations of Wegener granuloma-
tobacco (Table 3), this history was not found to be tosis. However, these patients are probably at great-
a statistically significant risk factor for the develop- er risk for additional cyclophosphamide-induced blad-
ment of bladder cancer. However, smokers devel- der toxicity, and other immunosuppressive agents,
oped both nonglomerular hematuria and bladder such as methotrexate, should be considered (19).
cancer after fewer months of cyclophosphamide It is still unclear how best to monitor the sub-
treatment than did nonsmokers, suggesting that group of patients with cyclophosphamide-induced
smoking may enhance the urotoxicity of cyclophos- cystitis for the development of bladder cancer. Cy-
phamide and its metabolites. tologic examination of urine is useful for detecting
Of the variables examined, only microscopic non- higher-grade bladder cancer and should be done
glomerular hematuria was found to be a significant every 6 to 12 months. The finding of atypia or
risk factor for the development of bladder cancer. dysplasia on such examination should lead to a cys-
Bladder cancer developed in 12% of patients with toscopic evaluation. However, cytologic examination
and in only 1% of patients without microscopic of urine is relatively insensitive for detecting lower-
nonglomerular hematuria (P = 0.02). Although six grade malignant lesions, and this limits its usefulness
of the seven patients with bladder cancer had at in monitoring these patients. Thus, routine cystos-
least one documented instance of nonglomerular copy every 1 or 2 years should be considered for all
microscopic hematuria, these episodes were not al- patients with cyclophosphamide-induced hematuria.
ways frequent and recurrent. In one patient, bladder This approach would be similar to that taken in
cancer was diagnosed 5 years after the last docu- patients with ulcerative colitis, in whom the risk for
mented episode of microscopic hematuria (patient developing carcinoma of the colon is estimated to
2; Table 3). In combination with the cystoscopy be 7% to 14% after 20 to 25 years of disease
results described above, these data strongly suggest (20-22). For these patients, annual or biennial
that nonglomerular hematuria is a sensitive marker colonoscopic surveillance with random biopsies is
for cyclophosphamide-induced bladder injury and advocated (23, 24).
that it identifies a subgroup of patients at high risk In summary, 5% of our patients treated with
for developing bladder cancer. cyclophosphamide developed transitional-cell carci-
We found cytologic examination of urine to be of noma of the bladder. We found that nonglomerular
1 March 1996 • Annals of Internal Medicine • Volume 124 • Number 5 483
microscopic hematuria was a frequent manifestation References
of cyclophosphamide-induced cystitis and that it
identified a subgroup of patients at high risk for 1. Fauci AS, Haynes BF, Katz P, Wolff SM. Wegener's granulomatosis: pro-
spective clinical and therapeutic experience with 85 patients for 21 years. Ann
developing bladder cancer. The best way to monitor Intern Med. 1983;98:76-85.
this subgroup of patients is unclear, but routine 2. Fauci AS, Wolff SM. Wegener's granulomatosis: studies in eighteen patients
and a review of the literature. Medicine (Baltimore). 1973;52:535-61.
cystoscopy every 1 or 2 years should be considered. 3. Hoffman GS, Kerr GS, Leavitt RY, Hallahan CW, Lebovics RS, Travis
WD, et al. Wegener granulomatosis: an analysis of 158 patients. Ann Intern
Because the risk for bladder cancer persists for Med. 1992;116:488-94.
many years after cyclophosphamide therapy has 4. Pedersen-Bjergaard J, Ersboll J, Hansen VL, Sorensen BL, Christof-
fersen K, Hou-Jensen K, et al. Carcinoma of the urinary bladder after
been discontinued, lifelong surveillance for bladder treatment with cyclophosphamide for non-Hodgkin's lymphoma. N Engl J
cancer may be necessary. Med. 1988;318:1028-32.
5. Stillwell TJ, Benson RC Jr, DeRemee RA, McDonald TJ, Weiland LH.
Note: Since the manuscript for this article was Cyclophosphamicie-induced bladder toxicity in Wegener's granulomatosis. Ar-
submitted, an additional study patient has been thritis Rheum. 1988;31:465-70.
6. Baker GL, Kahl LE, Zee BC, Stolzer BL, Agarwal AK, Medsger Ta Jr.
given a diagnosis of a grade II to III transitional-cell Malignancy following treatment of rheumatoid arthritis with cyclophospha-
carcinoma of the bladder. The tumor developed mide. Long-term case-control follow-up study. Am J Med. 1987;83:1-9.
7. Plotz PH, Klippel JH, Decker JL, Grauman D, Wolff B, Brown BC, et al.
after a lag time of 3 years and a latency period of Bladder complications in patients receiving cyclophosphamide for systemic
17 years. The patient is a 41-year-old nonsmoking lupus erythematosus or rheumatoid arthritis. Ann Intern Med. 1979;91:221-3.
8. Weiss M. Urinary cytology of transitional cell neoplasms of the bladder. In:
white man whose cumulative cyclophosphamide Young RH, ed. Pathology of the Urinary Bladder. New York: Churchill Living-
dose was 264 g administered over 7 years. The stone; 1989:245-84.
9. Forni A, Koss L, Geller W. Cytological study of the effect of cyclophospha-
indication for cystoscopy was routine follow-up for a mide on the epithelium of the urinary bladder in man. Cancer. 1964;17:1348-
history of cyclophosphamide-induced cystitis (pre- 55.
10. Coleman DV. The cytodiagnosis of human polyomavirus infection. Acta Cy-
senting with microscopic nonglomerular hematuria). tol. 1975;19:93-6.
11. Simes RJ. An improved Bonferroni procedure for multiple tests of signifi-
The results of the cytologic examination of voided cance. Biometrika. 1986;73:751-4.
urine obtained the day before cystoscopy showed 12. Cox DR. Regression models and life-tables. Journal of the Royal Statistical
changes consistent with therapeutic or viral effect 13. Crowley J, Hu M. Covariance analysis of heart transplant survival data.
and no evidence of atypia, dysplasia, or malignancy. Journal of the American Statistical Association. 1977;72:27-36.
14. Kaplan EL, Meier P. Non-parametric estimation from incomplete observa-
tions. Journal of the American Statistical Association. 1958;53:457-81.
Acknowledgments: The authors thank the clinical fellows and
15. Ries LA, Miller BA, Hankey BF, Kosary CL, Harras A, Edwards BK, eds.
physician assistants at the National Institute of Allergy and In- SEER Cancer Statistics Review, 1973-1991: Tables and Graphs. Bethesda,
fectious Diseases and the Clinical Center nursing staff for the MD: National Cancer Institute; 1994: NIH Pub. No. 94-2789.
care of our patients; and Drs. David Ailing, Steven Banks, and 16. Fraiser LH, Kanekal S, Kehrer JP. Cyclophosphamide toxicity. Characteris-
Mary Foulkes for their assistance with the statistical analysis. ing and avoiding the problem. Drugs. 1991;42:781-95.
17. Cox PJ. Cyclophosphamide cystitis—identification of acrolein as the causative
Requests for Reprints: Michael C. Sneller, MD, National Institute agent. Biochem Pharmacol. 1979;28:2045-9.
18. Reza MJ, Dornfeld L, Goldberg LS, Bluestone R, Pearson CM. Wegen-
of Allergy and Infectious Diseases, Building 10, Room 11B-13/10,
er's granulomatosis. Long-term followup of patients treated with cyclophos-
Center Drive, MSC 1876, Bethesda, MD 20892-1876. phamide. Arthritis Rheum. 1975;18:501-6.
19. Sneller MC, Hoffman GS, Talar-Williams C, Kerr GS, Hallahan CW,
Current Author Addresses: Ms. Talar-Williams, Ms. Hallahan, and Fauci AS. An analysis of forty-two Wegener's granulomatosis patients
Drs. Sneller and Fauci: Laboratory of Immunoregulation, Na- treated with methotrexate and prednisone. Arthritis Rheum. 1995;38:608-13.
tional Institute of Allergy and Infectious Diseases, National In- 20. Gyde SN, Prior P, Allan RN, Stevens A, Jewell DP, Truelove SC, et al.
stitutes of Health, Building 10, Room 11B-13, 9000 Rockville Colorectal cancer in ulcerative colitis: a cohort study of primary referrals from
three centres. Gut. 1988;29:206-17.
Pike, Bethesda, MD 20892.
21. Lennard-Jones JE, Melville D M , Morson BC, Ritchie JK, Williams CB.
Dr. Hijazi: National Cancer Institute, Building 10, Room 2A 10, Precancer and cancer in extensive ulcerative colitis: findings among 401 pa-
9000 Rockville Pike, Bethesda, MD 20892. tients over 22 years. Gut. 1990;31:800-6.
Drs. Walther and Linehan: National Cancer Institute, Building 22. Lofberg R, Brostrom O, Karlen P, Tribukait B, Ost A. Colonoscopic
10, Room 2B 47, 9000 Rockville Pike, Bethesda, MD 20892. surveillance in long- standing total ulcerative colitis—a 15-year follow-up
Dr. Lubensky: National Cancer Institute, Building 10, Room 2N study. Gastroenterology. 1990;99:1021-31.
212, 9000 Rockville Pike, Bethesda, MD 20892. 23. Jewell DP. Ulcerative colitis. In: Sleisenger MH, Fordtran JS, eds. Gastrointes-
tinal Disease: Pathophysiology, Diagnosis, Management. 5th ed. Philadelphia:
Dr. Kerr: Veterans Affairs Medical Center, 50 Irving Street NW,
WB Saunders; 1993:1305-30.
Washington, DC 20422. 24. Korelitz Bl. Inflammatory bowel disease: cancer. In: Haubrich WS, Schaffner
Dr. Hoffman: Cleveland Clinic Foundation, 9500 Euclid Avenue, F, Berk JE, Bockus HL, eds. Bockus Gastroenterology. 5th ed. Philadelphia: WB
Cleveland, OH 44195. Saunders; 1995:1364-81.
484 1 March 1996 • Annals of Internal Medicine • Volume 124 • Number 5