close

Вход

Забыли?

вход по аккаунту

?

Effects of cyclophosphamide on the development of malignancy and on long-term survival of patients with rheumatoid arthritis a 20-year followup study.

код для вставкиСкачать
ARTHRITIS & RHEUMATISM
Val. 38, No. 8, August 1995, pp 1120-1127
0 1995, American College of Rheumatology
1120
EFFECTS O F CYCLOPHOSPHAMIDE ON THE DEVELOPMENT OF
MALIGNANCY AND ON LONG-TERM SURVIVAL OF
PATIENTS WITH RHEUMATOID ARTHRITIS
A 20-Year Followup Study
CHARLES D. RADIS, LESLIE E. KAHL, GARY L. BAKER, MARY CHESTER MORGAN WASKO,
JOSEPH M. CASH, AILEEN GALLATIN, BERTRAND L. STOLZER, AMRIT K. AGARWAL,
THOMAS A. MEDSGER, JR., and C. KENT KWOH
Objective. To examine the effects of cyclophosphamide (CYC) on the development of malignancies and
on the long-term survival of patients with rheumatoid
arthritis (RA).
Methods. We used a longitudinal cohort design in
which 119 patients (76 women and 43 men) with refractory RA who were treated with oral CYC between 1968
and 1973 were compared with 119 control patients with
RA (matched for age, sex, disease duration, and functional class) who were evaluated during the same time
period but did not receive CYC.
Results. There was increased risk of malignancy
in the CYC-treated group, with 50 cancers found in 37
patients in the CYC group compared with 26 cancers in
25 of the control patients (P< 0.05). The relative risk of
cancer for those treated with CYC was 1.5 (95% confiPresented in part at the 57th National Meeting of the
American College of Rheumatology, San Antonio, TX, November
1993.
Supported in part by a grant from the Saint Margaret
Memorial Hospital Foundation.
Charles D. Radis, DO, Mary Chester Morgan Wasko, MD,
MSc, Thomas A. Medsger, Jr., MD: University of Pittsburgh School
of Medicine, Pittsburgh, Pennsylvania; Leslie E. Kahl, MD: Washington University, St. Louis, Missouri; Gary L. Baker, MD: Rheumatology Associates, Minneapolis, Minnesota; Joseph M. Cash,
MD, Aileen Gallatin, RN, Bertrand L. Stoker, MD, Amrit K.
Agarwal, MD: Saint Margaret Memorial Hospital, Pittsburgh, Pennsylvania; C. Kent Kwoh, MD: University of Pittsburgh School of
Medicine and Graduate School of Public Health, and Saint Margaret
Memorial Hospital. Dr. Radis is currently with Rheumatology
Associates, Portland, Maine; Dr. Cash is currently with the Cleveland Clinic, Cleveland, Ohio; Dr. Agarwal is currently with the
Beaver Medical Center, Beaver, Pennsylvania; Dr. Kwoh is currently with University Hospitals and Case Western Reserve University, Cleveland, Ohio.
Address reprint requests to Charles D. Radis, DO, Rheumatology Associates, 51 Sewell Street, Portland, Maine 04102.
Submitted for publication May 13, 1994; accepted in revised
form February 15, 1995.
dence interval 0.9S5.5). Nine of the malignancies in the
CYC group were bladder cancers and 19 were skin
cancers, compared with no bladder cancers and 6 skin
cancers in the control group. The total dose of CYC was
higher in those who developed cancer, particularly in
those with bladder cancer. Three of the bladder cancers
occurred 14, 16, and 17 years after CYC had been
discontinued
Conclusion. The risk of malignancy, particularly
bladder cancer, in RA patients treated with oral CYC
continues even 17 years after discontinuation of the drug.
.
Cyclophosphamide (CYC), an alkylating agent
first utilized in the treatment of malignancy, was
investigated in the late 1960s as a potential treatment
for refractory rheumatoid arthritis (RA). At that time,
CYC was the most effective drug available for the
treatment of RA (1-3), although its use was frequently
associated with such side effects as bone marrow
suppression, infection, gonadal failure, alopecia, and
hemorrhagic cystitis.
As a result of these risks, many investigators
have abandoned the use of CYC. Because of its
efficacy, however, as recently as 1986 low-dose CYC
was advocated in a combined therapeutic regimen for
RA (4). In addition, oral CYC continues to play an
important role in the treatment of other systemic
rheumatic diseases, such as Wegener’s granulomatosis, polyarteritis nodosa, and other life-threatening
vasculitidies (5-7).
An increased risk of developing a second malignancy has been well-documented in cancer patients
who had been treated with high-dose CYC (8,9). In
nonmalignant conditions such as RA, however, the
evidence regarding the risk of cancer from treatment
1121
CYC AND MALIGNANCY IN RA
with oral CYC is conflicting. Although Parsons et a1
(10) reported the occurrence of 7 myeloproliferative
malignancies in CYC-treated patients with RA, casecontrol studies by Farber et a1 in 1979 (11) and by
Boyle et a1 in 1981 (12)did not detect an increased risk
of malignancy after mean followup periods of 4.7 and
10 years, respectively. Baltus et a1 (13), in 1983,
demonstrated a 4-fold increase in the incidence of
malignancy among RA patients treated with CYC, but
the followup was limited to 6.5 years.
In 1987,we reported our findings in a retrospective cohort study (mean followup > I 1 years) of 119
patients with RA who had been treated with oral CYC
and 119 age- and sex-matched patients with RA who
had not taken CYC (14). An increased number of
malignancies was detected in the CYC-treated patients
compared with the controls. The incidence of bladder
cancer and skin cancer was significantly increased in
the CYC-treated group, and myeloproliferative disorders showed a similar trend. This increased rate of
new malignancies in the CYC-treated group persisted
throughout the followup period.
The purpose of the present study was to determine whether the increased risk of malignancy continued during the second decade after CYC administration and to report the mortality experience in our
original cases and controls.
PATIENTS AND METHODS
One hundred nineteen patients (cases) with refractory RA in 1 private rheumatology practice in Pittsburgh
were treated with oral CYC between 1968 and 1973. A
control group of 119 RA patients from the same practice was
selected from among persons hospitalized during the same
time period who had never received CYC. Details regarding
the indications for CYC, the assembly of the 2 groups, and
the matching process have been previously described (14).
Briefly, patients were treated according to a multicenter investigational drug protocol. The indications for
CYC were RA that had failed to respond to conventional
second-line therapy and current disease flare. Patients requiring CYC for extraarticular disease manifestations (e.g.,
vasculitis) were not included as part of this protocol. At the
time, patients with an RA flare that had failed outpatient
management were routinely admitted to an inpatient rehabilitation unit for a comprehensive program of physical
therapy, occupational therapy, patient education, and social
service counseling. Since the CYC-treated patients were
hospitalized on this same unit for the initiation of CYC
therapy, control patients were selected from among patients admitted to the rehabilitation unit within 6 months of
the initiation of CYC in the index case. The cases and
controls were matched by age, sex, disease duration, and
Steinbrocker’s functional class (15) at the time of study
entry. All of this information was routinely recorded by the
attending physician.
Patients were contacted by telephone in 1985 for the
initial survey and again during 1992-1993 to obtain followup
information for the present report. For each of the patients,
data were collected regarding race, tobacco use, serum
rheumatoid factor test result at the time of study entry, use
of other antirheumatic drugs, dates and total doses of CYC
treatment, adverse effects attributable to CYC, diagnosis
and specific cell type of malignancy, current survival status,
and cause of death. Available office and hospital records,
pathology reports, and death certificates were utilized to
assess survival status and to verify the presence or absence
of malignancy. Malignancies were confirmed by pathology
reports; only malignancies diagnosed before death were
included in the analysis.
Statistical procedures included matched-pair analysis utilizing chi-square tests, paired r-tests and Student’s
t-tests where appropriate. Kaplan-Meier curves were used to
evaluate survival distributions, using the time of diagnosis of
malignancy and the time of death as end points. The log rank
test was used to evaluate between-group differences in
survival rates and time to first malignancy. Ninety-five
percent confidence intervals (95% CI) for the rates of malignancy based on person-years of followup for the two groups
were calculated using the Poisson distribution.
RESULTS
Baseline characteristics of the 2 study groups
are shown in Table 1. There were 76 women and 43
men in each study group. Age at entry, disease duration, and functional class at entry were similar for the
2 groups. Drug therapy was comparable, with corticosteroids prescribed at some time during the disease
course in 95% of CYC patients and 76% of controls,
gold therapy in 92% and 90%, respectively, and Dpenicillamine in 36% and 26%, respectively. Nineteen
percent of the CYC-treated group and 14% of the
controls received another immunosuppressive drug at
some other time in their illness.
As of February 1993,we were able t o ascertain
the followup status of 96% of the patients in both
groups. The mean duration of followup was 13.1 years
for the CYC-treated group and 14.4 years for the
control group. Patients who were still living at that
time had been followed up for 20-25 years. The shorter
mean duration of followup is accounted for by the
deaths of patients in both groups. We found no significant between-group differences in mortality rates (Table 1). The mean age at death in the CYC-treated
patients was 2.4 years earlier than in the control
patients, but this difference was not statistically significant ( P = 0.09). The mean age at death was lower in
1122
RADIS ET AL
Table 1. Demographic and clinical characteristics of the patient
groups
Table 2. Development of malignancy in cyclophosphamidetreated versus control patients with rheumatoid arthritis
Cyclophosphamide Control
group
group
(n = 119)
(n = 119)
At study entry
Fema1es:males
Age, mean 2 SD
Duration of rheumatoid arthritis,
mean 2 SD
Functional class, n
I
I1
76:43
54.7 k 8.8
9.2 f 5.7
111
IV
At followup in 1993
Ascertainment, %
Drug treatment, %
Corticosteroids
Gold
D-penicillamine
Other immunosuppressive
agents
Died
Mean Age at Death
Males
Females
76:43
55.3 8.8
9.0 f 6.8
*
0
24
74
2
0
31
66
3
96
96
95
92
36
19
76
90
26
14
85*
75
69.6
69.3
69.9
67.2*
65.5t
68.6"
=
0.02 versus matched controls.
CYC-treated men (P = 0.02), whereas in women, the
age at death was similar in both groups (P > 0.05).
Figure 1 illustrates the cumulative long-term
survival rates of the CYC-treated and control groups
after study entry. The Kaplan-Meier curves sugge
Percent Survival
100%
80%
60%
40%
20%
0%
0
5
10
15
20
Control
group
(n = 119)
37 (31)
19/43 (44)
18/76 (24)
25 (21)
14/43 (33)
11/76 (14)
51 (43)*
26/43 (60)*
25/76 (33)t
26 (22)
11/43 (26)
15/76 (20)
No. of patients with at least
1 malignancy
Total
Males
Females
No. of malignancies
Total
Males
Females
No. of patients with
multiple malignancies
Total
Males
Females
10 (8)t
4/43 (9)
6/76 (7)
1 (0.8)
0/43 (0)
1/76 (1.3)
* P < 0.01 versus matched controls.
t P < 0.05 versus matched controls.
* P > 0.05 versus matched controls.
tP
Cyclophosphamide
group
(n = 119)
25
Years of Follow-up
Figure 1. Long-term survival in patients with rheumatoid arthritis
treated with cyclophosphamide (solid line) compared with rheumatoid arthritis control subjects (dotted line). P = 0.08, by log rank
test.
that the CYC-treated group had a trend toward reduced survival compared with the control group, but
this difference did not reach statistical significance
(P = 0.08). The curves separate after year 7, but
converge somewhat after year 20.
As shown in Table 2, there were significantly
more individuals with malignancies as well as more
patients with multiple malignancies in the CYC-treated
group. Fifty-one malignancies were identified in 37
CYC-treated patients (3 1%) versus only 26 malignancies in 25 control patients (21%) (P < 0.05). Among the
men, 26 malignancies occurred in the CYC-treated
group and 11 in the controls. Among the women, 25
malignancies occurred in the CYC-treated group and
15 in the controls. Overall, males were more likely
than females to develop malignancies (P = 0.03).
Multiple malignancies were detected in 10 patients, 9
of whom had received CYC. The relative risk for the
development of malignancy in the CYC-treated group
was 1.5 (95% CI 0.93-5.5).
The specific types of tumors that occurred in
the 2 groups are listed in Table 3. All 9 bladder cancers
were found in the CYC-treated patients (P < 0.001).
Since our previous report (14), 3 additional bladder
cancers have occurred at 14, 16, and 17 years after
discontinuing CYC, corresponding to 19.6, 20.5, and
22.8 years after study entry. There were 19 skin
cancers in the CYC-treated group versus 6 in the
control group (P = 0.04). The majority of these cutaneous malignancies were basal cell carcinomas; there
were 3 squamous cell cancers and 1 melanoma in the
1123
CYC AND MALIGNANCY IN RA
Table 3. Types of malignancy in 119 rheumatoid arthritis patients
who were treated with cyclophosphamide and 119 who were not
~~
Malignancy
Bladder
Skin
Basal cell
Squamous cell
Melanoma
Myeloproliferative
Acute leukemia
Non-Hodgkin’s lymphoma
Multiple myeloma
Chronic lymphocytic leukemia
Other
Lung
Gynecologic (cervix, ovary,
endometrium)
Breast
Colon
Oropharynx
Prostate
Brain
Esophagus
Stomach
Undifferentiated
Total
Cyclophosphamide
group
Control group
(n = 119)
(n = 119)
0
9
6
19
5
1
0
15
3
1
5
1
2
2
1
0
0
0
1
0
18
19
6
5
5
3
1
0
3
1
1
1
0
0
4
4
0
1
51
26
0
0
1
1
CYC-treated group, and 1 squamous cell carcinoma in
the control group. (There was no between-group difference, however, in the numbers of skin cancers
reported since.)
Myeloproliferative disorders occurred in 5
CYC-treated patients (2 acute leukemias, 2 nonHodgkin’s lymphomas, and 1 multiple myeloma) and
in 1 control patient (chronic lymphocytic leukemia)
(P = 0.06). The rates of myeloproliferative disorders
based on person-years of followup were 190/100,000
person-years in the CYC-treated group (95% CI 40560/100,000 person- years) and 60/ 100,000person-years
in the control group (95% CI 10-320/100,000 personyears). Of note, no additional myeloproliferative disorders have occurred during the second decade of
followup. None of the patients developed myelodysplasia. A wide variety of other tumors occurred in the
2 treatment groups, but there were too few in any single
category to detect a significant difference in frequency.
The time to diagnosis of the first malignancy
differed between the 2 groups (Figure 2). The KaplanMeier curves showed that patients treated with CYC
developed malignancies at an accelerated rate compared with controls ( P < 0.024 by log rank test). This
difference became apparent approximately 6 years
after treatment with CYC was initiated and continues
even after 20 years. Hematologic malignancies tended
to occur relatively earlier in the followup period (after
4.3, 6.5, 7.2, 7.8, and 8.8 years of followup in the CYC
group and after 10.6 years in the control group). The
occurrences of skin and bladder cancers were more
evenly distributed over the 20-25-year followup period
(data not shown).
Risk factors for the development of malignancy
were evaluated in our 1987 report (14). Using a linear
regression model, only the history of tobacco use was
identified as a significant risk factor. Insufficient data
were available to assess additional malignancy risk
factors in the present study. Patients who had ever
smoked were evenly distributed between the 2 groups,
with 69 (58%) in the CYC-treated group and 64 (54%)
in the control group. As expected, smokers were more
likely to develop a malignancy (33%) than were nonsmokers (16%) (P = 0.02). There was no difference,
however, in the mean number of pack-years between
those who developed cancer (35 pack-years) and those
who did not develop cancer (33.5 pack-years). Cigarette smoking, also recognized as a risk factor for
bladder cancer, may have played an additional role in
our patient population. Six of the 9 patients who
developed bladder cancer were heavy smokers.
As shown in Table 4, the mean total CYC dose
in the 37 patients who developed malignancy was 79.0
gm, compared with 41.2 gm in the 82 patients without
malignancy (P < 0.001). In those patients who received more than 80 gm of CYC, the risk of malignancy was 53%. The mean duration of CYC treatment
Percent without Malignancy
100%
20%
I
-
_
- -
~
-
~
-
RADIS ET AL
1124
Table 4. Association between total dose and duration of cyclophosphamide and later development of malignancy in 119
cyclophosphamide-treated patients with rheumatoid arthritis
All malignancies, n
Mean total dose, gm
Mean duration, months
Bladder malignancies, n
Mean total dose, gm
Mean duration, months
Malignancy
No
malignancy
37
79.0 2 50.4%
45.6 2 30.5*
9
119.8 2 55.6
60.1 -+ 31.9
82
41.2 & 32.4
24.8 ? 18.3
110
47.5 36.4
28.8 2 21.8
*
~~
* P < 0.001 versus no malignancy.
t P < 0.001 versus no bladder malignancy.
was 45.6 months in the patients who developed malignancies versus 24.8 months in those who did not ( P <
0.00 1).
For the 9 patients who developed bladder cancer, the mean total dose of CYC was 119.8 gm compared with 47.5 in those without bladder malignancies
(P< 0.001). Seven of the 9 bladder cancers developed
in patients who had taken more than 80 gm of CYC,
representing 25% of the patients (7 of 28) who had
received such a high total dose. Three of the 9 patients
with bladder cancer had preceding hemorrhagic cystitis, compared with I 1 of 110 patients without bladder
cancer ( P = 0.07). An additional 4 patients had microscopic hematuria during treatment, and 2 subsequently
developed bladder cancer. All patients in whom microscopic hematuria was detected underwent cystoscopy. One patient’s death was attributed to radiation
treatment for bladder cancer. No patient underwent
cystectomy. According to patient self-report, quality
of life was more related to their RA and other comorbidities than to their bladder cancer.
The mean duration of treatment was 60.1
months in those with bladder malignancy versus 28.8
months in those without this complication ( P < 0.001).
The latency period (from the start of CYC therapy to
the development of bladder cancer) ranged from 4.0 to
22.8 years, with a mean of 156 months. Age-adjusted
and sex-specific incidence rates for western Pennsylvania were utilized to calculate the expected number
of cases of bladder cancer over this 20-year interval.
There were 0.4 expected cases compared with 9 observed cases, yielding a relative risk of 22 (95% CI
8.8-55.5).
Atherosclerotic (cardiovascular and cerebrovascular) diseases were the leading causes of death in
both study groups, accounting for 44% of the CYCtreated group and 38% of the control group. As expected, more CYC-treated patients (20%) died of
malignancy than did control patients (13%) ( P <
0.001). Among males, 26% of the CYC-treated group
died of malignancy versus 19% of the control group,
while among females, these frequencies were 8% and
3%, respectively. All 5 hematologic malignancies in
the CYC-treated group were fatal, while in only 1 of
the 9 bladder cancers did the malignancy directly
result in death. Infection as the cause of death occurred in similar proportions in both groups (18% and
17%, respectively). Rheumatoid arthritis or its complications was infrequently reported as a direct cause of
death in either group (3% in the CYC-treated group
versus 8% in the control group).
DISCUSSION
Although our original study demonstrated an
increased risk of malignancy in patients with RA
treated with oral CYC after 10 years of followup (14),
important long-term questions remain. Until an effective, low-toxicity treatment for RA is developed, physicians will have to continue to rely on drugs with
well-described short-term benefits and toxicities but
ill-defined long-term risks in making their treatment
decisions. As patients with RA are treated earlier and
more aggressively in their disease course, often with
combined immunosuppressive therapies, the quantification of delayed treatment toxicity becomes extremely important.
The present study has shown that the increased
risk of malignancy previously described (14) continues
throughout the second decade after the discontinuation of CYC. The occurrence of bladder and skin
cancer was significantly more frequent in the CYCtreated group than in the control group, with a similar
trend for myeloproliferative malignancies. The total
dose and duration of CYC treatment remained significant risk factors for the development of these malignancies.
The estimation of malignancy risk associated
with CYC and other therapies is complicated by the
potentially increased risk of malignancy associated
with RA itself. Various studies have shown either an
overall increased risk (17), an increased risk for myeloproliferative disorders (16,18-22), or no increased risk
of malignancy due to RA (23,24). None of these
studies, however, considered specific therapies that
may be associated with subsequent risk of malignancy,
nor was information reported regarding risk in specific
subsets of patients with RA. Gridley et a1 (16) noted no
overall increased risk for cancer, but examination of
site-specific risks revealed decreased risks for cancers
CYC AND MALIGNANCY IN RA
of the colon, rectum, and stomach and an increased
risk for lymphoma. A recent report noted a 2-fold
increased risk of malignancy for men with Felty’s
syndrome and a 12-fold increased risk for nonHodgkin’s lymphoma (25). Only 1 of the 19 patients
with lymphoma had been treated with a cytotoxic
agent (azathioprine). Despite the methodologic difficulties inherent in clearly defining the baseline risk of
malignancy in RA, the weight of evidence suggests no
overall increased risk of malignancy, but a small
increase in risk for myeloproliferative disorders.
Two of the 4 previous studies did not demonstrate an increased risk of malignancy in RA patients
treated with CYC (11,12). One study was limited by
the relatively brief duration of followup (11). In the
present study, malignancy rates were roughly similar
in cases and controls until after the sixth year of followup, when they markedly diverged. The second negative study was also likely to have missed a significant
dBerence, since only 27 patients were followed over 10
years. In addition, the treated patients had received a
relatively low total dose of CYC (39.5 gm) (12).
In 1983, Baltus et a1 reported a 4-fold increase
in malignancy among 81 CYC-treated patients compared with matched RA controls (13). Although specific malignancies were not analyzed according to the
total dose or duration of CYC therapy, the mean total
dose of 82 gm and the mean duration of therapy of 6.2
years were both higher than were reported in the 2
previous studies (1 1,12). Four myeloproliferative disorders were reported in the CYC-treated group, compared with none in the controls (13). Of interest,
bladder cancer was not reported in any of these 3
studies (1 1,12,13).
Bladder malignancy is a well-recognized complication among patients treated with oral CYC, and
has previously been reported in CYC-treated patients
with Wegener’s granulomatosis (5) and systemic lupus
erythematosus (26). The long delay between CYC
treatment and the development of bladder cancer
strongly demonstrates the need for prolonged urologic
surveillance in this patient population. Present recommendations include semiannual urine cytology, periodic intravenous pyelogram to assess for possible
upper tract abnormalities, and cystoscopy to evaluate
abnormalities found by routine urinalysis or cytology
(27). The presence of nonglomerular hematuria, usually microsopic, has been reported to be the most
sensitive indicator in screening for the development of
bladder toxicity related to CYC therapy, and all patients with this condition should be evaluated by
cystoscopy (28).
1125
Our current report confirms that the risk of
bladder cancer is related to the total CYC dose received. Seven of the 9 patients with bladder cancer
had received >80 gm of CYC. A similar dose-response
was noted by Pedersen-Bjergaard et a1 (8) in their
study of patients with non-Hodgkin’s lymphoma: 8 of
9 patients who developed bladder cancer had received
>80 gm of oral CYC. In our series of CYC-treated RA
patients, bladder cancer has developed in an alarming
proportion (25%) of all patients who received >80 gm
of this drug. A similarly high cumulative risk of
bladder cancer may be likely with current oral regimens of CYC in patients who are being treated for
vasculitis. A report by Hoffman et a1 (5) indicated that
CYC-treated patients with Wegener’s granulomatosis
had a 33-fold increase in bladder malignancies compared with the general population, with a latency
period that ranged from 7 months in a patient taking
daily doses of CYC to 12 years in a patient who had
discontinued CYC 10 years previously (5).
Our results suggest that skin cancers were
found more frequently in CYC-treated patients than in
controls, but this between-group difference decreased
with time. Up to 1985 all 10 skin cancers were found in
CYC-treated patients; between 1986 and 1993, 9 additional skin cancers were reported in CYC-treated
patients and 6 in the control patients. Unfortunately,
potential differences in skin exposure between the 2
groups could not be assessed. It is possible that
patients who had taken CYC were more likely than
control patients to seek medical attention for skin
lesions. Given the potential problems associated with
ascertainment bias, the data regarding the increased
risk of skin cancers should be interpreted cautiously.
Baltus et a1 also reported 3 basal cell carcinomas in CYC-treated patients and 1 in control patients
after 6.5 years of followup (13). Other immunosuppressants may similarly increase the risk of skin cancer.
One previous study reported 3 basal cell cancers in
azathioprine-treated patients and none in controls (19).
In our study, skin cancer developed in 2 CYC-treated
patients who had subsequently also been treated with
azathioprine.
Although myeloproliferative disorders occurred
in 5 CYC-treated patients and in 1 control, the trend
did not reach statistical significance. This lack of clear
association may be the result of a type I1 error, given
the wide 95% confidence intervals. Based on the most
recent Surveillance, Epidemiology, and End Results
(SEER) Program report, among persons living in the
US, there is an annual incidence rate of 28.2/100,000
person-years for all myeloproliferative disorders (i.e.,
1126
9.7 for all leukemias, 15.5 for non-Hodgkin’s lymphomas, and 3.0 for Hodgkin’s lymphomas per 100,000
person-years) (29). The rates in our 2 groups (1901
100,000 person-years for the CYC-treated group and
60/100,000 person-years for the control group) exceed
the rates reported by the SEER program, but the
results are difficult to interpret since 95% CI are not
available from the most recent SEER data set. Gridley
et a1 (16) have reported an increased risk of all myeloproliferative disorders among men with RA, and only
an increased risk of non-Hodgkin’s lymphoma among
women with RA. Previous treatment with cytotoxic
agents was not specifically assessed, however. Those
authors reported a clinical impression that few Swedish patients with RA had been treated with cytotoxic
agents .
Pancytopenia developed after 4 years of CYC
treatment in 1 of our patients, and the drug was
discontinued. Four months later, acute myelocytic
leukemia was diagnosed in this patient. In another
patient, non-Hodgkin’s lymphoma occurred 18 months
after CYC was discontinued; the total cumulative dose
was 149 gm. In the 3 remaining patients who developed hematologic malignancies, these occurred 2, 5,
and 7.5 years after CYC treatment had been stopped.
Further evidence that this risk is related to CYC
therapy is the clustering of these disorders during the
early years of follawup. This timing of myeloproliferative malignancies suggests defective immunologic
surveillance, allowing pathologic clones of cells to
escape normal pathways of suppression or destruction. It has also been found that RA patients previously treated with CYC have an increased frequency
of haploidy and chromosomal breaks (30).
An increased risk of secondary malignancies in
cancer patients treated with alkylating agents may be
an acceptable trade-off, particularly in those with
Hodgkin’s disease, for which chemotherapy is often
curative. Any increased risk of malignancy in the
treatment of chronic rheumatic disorders cannot be so
easily justified, however. In this study, a significant
proportion of the excess cancer risk can be attributed
to the number of basal cell and squamous cell skin
cancers, which account for 18 of the 50 cancers in the
CYC group and 10 of the 26 malignancies in the control
group. As noted above, identification of skin cancers
may be associated with biased ascertainment. If these
less life-threatening malignancies are excluded from
the statistical analysis, the overall difference between
the 2 groups is less clinically impressive but still
significant (P < 0.02). Of the “serious” malignancies
RADIS ET AL
which differed between the 2 groups, 15 were found in
the CYC group and only 1 in the control group. In 6 of
the CYC-treated patients (the 5 with myeloproliferative disorders and 1 with bladder cancer), the malignancy was the primary cause of death.
The results suggest that survival may be decreased in a subset of RA patients treated for prolonged periods with an immunosuppressive agent. In
men, the mean age at death was 65.5 years in the
CYC-treated group versus 69.3 in the control group
(P = 0.02). Since the subset was relatively small (43
men in each group), this finding should be viewed with
caution. In contrast, the mean survival for women was
similar in the 2 groups. Furthermore, since patients
were not randomly assigned to the 2 groups, those in
the CYC-treated group may have had more severe
disease, even though we attempted to match patients
as closely as possible. In addition, reasons for starting
CYC therapy were not recorded. To assess whether
this difference in survival rates was due to the higher
mean cumulative dose of CYC taken by the men (65
gm) than by the women (46 gm), patients were stratified according to the total dose of CYC and survival.
No relationship was seen between life span and total
CYC dose. Thus, although as a group, men treated
with CYC received higher doses than did women, their
reduced rates of survival cannot be attributed to this
difference.
Intravenous (IV) pulse CYC may be the means
of administering CYC with less risk of malignancy.
Although there are no controlled data comparing the
risk of malignancy in patients given oral versus IV
pulse treatment, only 5 cases of malignancy following
IV pulse therapy have been reported (including 1 acute
nonlymphocytic leukemia in a patient with systemic
lupus erythematosus who received a pulse dose every
3 months for 12 months) (31-33). No bladder malignancies have been reported following IV pulse CYC.
This paucity of cases may be partly due to lower total
doses, shorter followup time, the use of mesna in some
protocols as a bladder protectant, and/or a greater
awareness of the need to hydrate patients before and
after therapy, thus reducing the contact time between
the CYC metabolite acrolein and the bladder mucosa.
It is unlikely that IV CYC is as effective a treatment
for R 4 as it is for systemic lupus erythematosus,
however. In one study of monthly IV pulse CYC in
RA, there was no significant benefit after 6 months of
therapy (34).
This study has shown that therapy with oral
CYC carries a long-term increased risk of malignancy,
CYC AND MALIGNANCY IN RA
specifically bladder and skin cancer. The risk of bladder cancer continues for at least 20 years after discontinuation of CYC therapy, is higher in males, and is
highest in patients receiving the highest cumulative
dose of CYC. In all instances, careful assessment of
the risks and benefits of CYC treatment should be
made before initiating this form of therapy.
ACKNOWLEDGMENTS
The authors thank Sherry L. Swatchick and the
physicians and staff of H. M. Margolis, MD, and Associates,
Ltd. for their technical assistance, and Deborah M. Whitley
for her secretarial assistance.
REFERENCES
1. Currey HLF: Comparison of azathioprine, cyclophosphamide
and gold in treatment of RA. Br Med J 3:763-766, 1974
2. Cooperating Clinics Committee of the American Rheumatism
Association: A controlled trial of cyclophosphamide in rheumatoid arthritis. N Engl J Med 283:883-889, 1970
3. Ianuzzi L, Dawson N, Zein N, Kushman I: Does drug therapy
slow radiographic progression in rheumatoid arthritis? N Engl J
Med 309: 1023-1028, 1983
4. Csuka M, Carrera G, McCarty D: Treatment of intractable
rheumatoid arthritis with combined cyclophosphamide, azathioprine, and hydroxychloroquine: a follow-up study. JAMA 255:
2315-2319, 1986
5 . Hoffman GS, Kerr GS, Leavitt RY, Hallahan CW, Lebovics
RS, Travis WD, Rottem M, Fauci AS: Wegener’s granulomatosis: an analysis of 158 patients. Ann Intern Med 116:488498,
1992
6. Fauci AS, Katz P, Haynes B, Wolff S: Cyclophosphamide
therapy of severe systemic necrotizing vasculiiis. N Engl J Med
301:235-240, 1979
7. Fauci AS, Haynes B, Katz P, Wolff S: Wegener’s granulomatosis: prospective clinical and therapeutic experience with 85
patients for 21 years. Ann Intern Med 9k76-85, 1983
8. Pedersen-Bjergaard J, Ersoboll J, Hansen V, Sorensen BL,
Christofferson K, Hou-Jensen K, Nissen NI, Knudson JB,
Hansen MM: Carcinoma of the urinary bladder after treatment
with cyclophosphamide for non-Hodgkin’s lymphoma. N Engl J
Med 318:1028-1032, 1988
9. Tucker M, Meadows A, Boice J: Leukemia after therapy with
alkylating agents for childhood cancer. J Natl Cancer Inst
78:455-464, 1987
10. Parsons J, Strong J, Fosdick W: The causes of death in patients
with rheumatoid arthritis treatment with cytotoxic agents (abstract). J Rheumatol 1 (suppl):75, 1974
11. Farber SJ, Sheon RP, Kirsner AB, Finkel RI: Incidence of
malignant disease in patients receiving cytotoxic therapy for
rheumatoid arthritis (abstract). Arthritis Rheum 22 (suppl 6):
608, 1979
12. Boyle DJ, Day JF, Kassan SS, Thomas MR, Robinson WA,
Steigenvald JC: Incidence of malignancy 10 years following
cyclophosphamide use for rheumatoid arthritis (abstract). Arthritis Rheum 27 (suppl 4):S71, 1981
13. Baltus 3, Boersma W, Hartman A, Vandenbroucke J: The
occurrence of malignancies in patients with rheumatoid arthritis
treated with cyclophosphamide: a controlled retrospective follow-up. Ann Rheum Dis 42:36%373, 1983
14. Baker G, Kahl L, Zee B, Stoker B, Agarwal A, Medsger T:
1127
Malignancy following treatment of rheumatoid arthritis with
cyclophosphamide. Am J Med 83:l-9, 1987
15. Steinbrocker 0, Traeger C, Batterman R: Therapeutic criteria in
rheumatoid arthritis. JAMA 140:659-652, 1949
16. Gridley G, McLaughlin J, Ekbom E, Klareskog L , Adami HO,
Hacker DG, Hoover R, Fraumen: J F Jr.: Incidence of cancer
among patients with rheumatoid arthritis. J Natl Cancer Inst
85:307-311, 1993
17. Isomaki H, Hakulinen T, Joutsenlahti U: Excess risk of lymphomas, leukemia and myeloma in patients with rheumatoid
arthritis. J Chron Dis 31:691-696, 1978
18. Prior P, Symmons D, Hawkins C, Scott D, Brown R: Cancer
morbidity in rheumatoid arthritis. Ann Rheum Dis 43: 128-131,
1984
19. Silman A, Petrie J, Hazleman B, Evans S: Lymphoproliferative
cancer and other malignancy in patients with rheumatoid arthritis treated with azathioprine: a 20-year follow up study. Ann
Rheum Dis 47:98&992, 1988
20. Porter D, Madhok R, Capell H: Non-Hodgkin’s lymphoma in
rheumatoid arthritis. Ann Rheum Dis 50:275-276, 1991
21. Katusic S, Beard M, Kurland L, Weis J, Bergstralh E: Occurrence of malignant neoplasms in the Rochester, Minnesota,
rheumatoid arthritis cohort. Am J Med 78:5&55, 1985
22. Laakso M, Mutru 0, Isomaki, Koota K: Cancer mortality in
patients with rheumatoid arthritis. J Rheumatol 13522-526, 1986
23. Fries J, Bloch D, Spitz P, Mitchell D: Cancer in rheumatoid
arthritis: a prospective long-term study of mortality. Am J Med
7856-59, 1985
24. Lewis P, Hazleman B, Hanka B, Roberts S: Cause of death in
patients with rheumatoid arthritis with particular reference to
azathioprine. Ann Rheum Dis 39:457-461, 1980
25. Gridley G, Klipple J, Hoover R, Fraumeni JJF: Incidence of
cancer among men with the Felty Syndrome. Ann Intern Med
120:35-39, 1994
26. Plotz P, Klippel J, Decker J, Grauman D, Wolff B, Brown BC,
Rutt G: Bladder complications in patients receiving cyclophosphamide for systemic lupus erythematosus or rheumatoid arthritis. Ann Intern Med 91:221-223, 1979
27. Stillwell T, Benson R: Cyclophosphamide-induced hemorrhagic
cystitis: a review of 100 patients. Cancer 61:451-457, 1988
28. Stillwell TJ, Benson RC Jr, DeRemee RA, McDonald TJ,
Weiland LH: Cyclophosphamide-induced bladder toxicity in
Wegener’s granulomatosis. Arthritis Rheum 3 1:465-470, 1988
29. Reis L, Miller B, Hankey B, Kosary C, Harras A, Edwards B,
eds: SEER Cancer Statistics Review, 1973-1991: Tables and
Graphs. NIH publication no. 94-2789. Bethesda, MD: National
Cancer Institute, 1994
30. Tolchin SF, Winkelstein A, Rodnan GP, Pan SF, Nankin HR:
Chromosome abnormalities from cyclophosphamide therapy in
rheumatoid arthritis and progressive systemic sclerosis (scleroderma). Arthritis Rheum 17:375-382, 1974
31. Sessons S, Kovarsky J: Monthly intravenous cyclophosphamide in the treatment of severe systemic lupus erythematosus.
Clin Exp Rheumatol 2:247-251, 1984
32. Grananger W, Benke-Studnicka A, Smolen J: Side effects of high
dose bolus cyclophosphamide therapy of systemic lupus erythematosus (abstract). ILAR Conference of Rheumatology, Rio de
Janeiro, Brazil, September, 1989
33. Gibbons RB, Westerman E: Acute nonlymphocytic leukemia
following short-term, intermittent, intravenous cyclophosphamide treatment of lupus nephritis. Arthritis Rheum 31:15521554, 1988
34. Arnold MH, Janssen B, Schrieber L, Brooks PM: Prospective
pilot study of intravenous pulse cyclophosphamide therapy for
refractory rheumatoid arthritis. Arthritis Rheum 32:933-934,
1989
Документ
Категория
Без категории
Просмотров
2
Размер файла
831 Кб
Теги
development, years, patients, long, cyclophosphamide, followup, effect, terms, survival, stud, arthritis, malignant, rheumatoid
1/--страниц
Пожаловаться на содержимое документа