Rheumatoid arthritis and the risk of malignancy.

ARTHRITIS & RHEUMATISM
Vol. 40, No. 9, Septembcr 1997, pp 1580-1586
0 1997, American College of Rheumatology
1580
RHEUMATOID ARTHRITIS AND THE RISK OF MALIGNANCY
JOLANDA CIBERE, JOHN SIBLEY, and MAY HAGA
Objective. To determine the relative risks of malignancy and of site-specific malignancies in patients
with rheumatoid arthritis (RA).
Methods. In a prospective cohort study, 862 patients with RA (96% white) were enrolled from 1966 to
1974 and were followed up for up to 35 years (mean 17.4
years) at the University of Saskatchewan Rheumatic
Disease Unit. All diagnoses of cancer were crossreferenced with the Provincial Cancer Registry. Expected cancer incidence rates were determined based on
province of Saskatchewan population statistics matched
to each study patient for age, sex, and calendar year.
Standardized incidence ratios (SIRS) of the observedto-expected cancer incidence and 95% confidence intervals (95% CI) were then calculated.
Results. A total of 136 cases of cancer were
observed compared with 168 expected (SIR 0.80, P =
0.011 [95%CI 0.67-0.951). The relative risk of colorectal
malignancy was significantly reduced in the RA study
population (SIR 0.52, P = 0.037 [95%CI 0.25-0.961). A
significant excess of leukemia was found (SIR 2.47, P =
0.026 [95% CI 1.12-4.69]), whereas the incidence rates
for Hodgkin’s disease and non-Hodgkin’s lymphoma
and all other site-specific malignancies were not found
to be significantly different from general population
rates.
Conclusion. In our cohort of RA patients, colorectal cancer was detected in only half the expected number
of patients. This risk reduction may be related to
long-term nonsteroidal antiinflammatory d ru g
(NSAID) use in RA, as has been suggested in several
other studies of long-term NSAID use. An increased risk
of leukemia was confirmed. This may be due to the
persistent immune stimulation associated with RA itSupported by NIH grant AM-21393 to the Arthritis, Rheumatism and Aging Medical Information System.
Jolanda Cibere, MD, John Sibley, MD, May Haga, B A
University of Saskatchewan, Saskatoon, Saskatchewan, Canada.
Address reprint requests to John Sibley, MD, Royal University Hospital, Saskatoon, SK S7N OW& Canada.
Submitted for publication December 16, 1996; accepted in
revised form April 11, 1997.
self, since other potential explanatory factors for increased leukemia were not apparent. Despite the excess
of hemopoietic malignancy and despite treatment of RA
with potentially oncogenic agents, the overall risk of
malignancy was reduced in this RA cohort.
Several authors have investigated an association
between rheumatoid arthritis (RA) and malignancy (16). This information is summarized in Table 1. In 1978,
a large cohort study by Isomaki et a1 (1) demonstrated
an increased risk of leukemia, lymphoma, and multiple
myeloma as well as a significantly increased risk of
cancer of the respiratory organs in men, and a decreased
risk of stomach and rectal cancer in women. Subsequent
studies (2-5), with the exception of one (6), have
generally confirmed the higher incidence of lymphoproliferative and myeloproliferative malignancies in RA.
However, there is disagreement regarding what subtypes
of hemopoietic cancers are increased in RA. An increased incidence of non-Hodgkin’s lymphoma was confirmed in a study by Prior et a1 (2) and in a large Swedish
study by Gridley et a1 ( 5 ) , although, in the latter study,
non-Hodgkin’s lymphoma was significantly increased in
women only. Other studies have found an increased risk
of multiple myeloma (3,4). An excess of leukemia has
not been reported since the study by Isomaki et a1 (1) in
1978. Similarly, differences in the rates of solid tumors in
RA have not been widely confirmed in the literature
with the exception of the study by Gridley et a1 ( 5 ) ,which
found a reduced risk of colon cancer in women.
Also germane to any examination of RA and
malignancy is the cancer-sparing or cancer-promoting
potential of RA therapies. Animal studies have shown
that various nonsteroidal antiinflammatory drugs
(NSAIDs) can prevent colon carcinogenesis in rats
(7-9). Recently, several large studies of human cohorts
(10-13) have shown a reduced risk of colorectal cancer
in association with regular, long-term aspirin use. In
addition, Schreinemachers and Everson ( 3 3) reported
reduced rates of lung and breast cancer in women who
took aspirin regularly. This is of particular interest in
RISK OF MALIGNANCY IN RA
1581
Table 1. Published studies of rheumatoid arthritis and the risk ratios for malignancy
0bserved:expected risk of malignancy
Study (ref. no.)
Isomaki et al, 1978 (1)
No. of
patients
11,483 men,
34,618 women
Prior et al, 1984 (2)
448
Katusic et al, 1985 (3)
521
Gridley et al, 1993 (5)
3,750 men,
7,933 women
Moritomo et al, 1995 (6) 665
Hemopoietic
Non-llodgkin’s Hodgkin’s Multiple Colorectal Colon Rectal Lung
cancer
Leukemia
lymphoma
disease myeloma
cancer
cancer cancer cancer
2.7*
2.78
23.0$
3.2
1.8
1.9t
2.2
3.18
7.1
2.1.12.2$
-
-
1.2
-
4.6t
0.9
5.01
1.8
0.8
-
-
2.5*
1.4
4.9
1.9
1.9
0.8
1.3
-
-
-
-
1.7
-
8.0$
-
-
-
1.1
0.8
~
1.2
0.7
O.ht
~
0.6
0.6t
-
0.5
0.7
0.7
-
1.3*
1.3
0.8
1.4
1.2
1.5
0
* P < 0.01.
t P < 0.05.
$ P < 0.001.
RA, since most R A patients are long-term users of
aspirin or other NSAIDs. Indeed, it is somewhat surprising that RA studies have generally not reported a
lower-than-expected incidence of colorectal cancer. In
1985, Fries et a1 (14) assessed the effects of gold or
prednisone therapy on mortality and found a trend of
reduced deaths due to malignancy in R A patients who
were treated with either of these 2 drugs.
Conversely, several studies have investigated the
possibility of an increased risk of malignancy in RA
patients treated with disease-modifying antirheumatic
drugs (DMARDs) (15-19). Azathioprine has been reported to be associated with an 8-10-fold increased risk
of lymphoproliferative malignancies in RA patients
(16,17). Cyclophosphamide therapy is well known to
increase the risk of lymphoproliferative malignancy
(20,21), and this trend has been found in patients with
RA as well (15,18). Methotrexate may also be associated
with hemopoietic malignancies, although the literature
consists predominantly of case reports (22-26). In a
recent case-control study, Williams et a1 (19) found that
R A patients with leukemia or lymphoma showed a trend
toward increased prior methotrexate or azathioprine use
compared with matched RA controls without leukemia
or lymphoma.
In our institution, we have a large R A cohort that
has been followed up prospectively since 1966. This has
given us the opportunity to further evaluate the longterm risk of malignancy in RA.
PATIENTS AND METHODS
Patients. During the study enrollment period, our
university-based Rheumatic Disease Unit (RDU) was the sole
referral center for the northern half of the province of
Saskatchewan, serving a population of -450,000 with a racial
composition of -96% white and 4% North American Indian.
All consecutive new patients with a clinical diagnosis of RA in
whom symptoms started after age 16 and who were seen
between 1966 and 1974 (n = 1,191) were assessed by a
rheumatologist and enrolled into a long-term prospective study
of outcome in RA (27). More than 90% were first seen as
outpatients. Prospective data collection began in 1966. Patients
who had been seen in the RDU between 3955 and 1965 and
who were still being followed up after 1966 were also entered
into the study (n = 285). For these patients, the 195.5-1965
data were retrospectively abstracted from all available clinic
and hospital records.
Our study population does not represent an inception
cohort. We could not account for those patients with RA for
whom cancer may have preempted their eventual referral to
our RDU. Thus, we excluded from our analysis all patients
who had a diagnosis of cancer made prior to their first visit to
the RDU (n = 26). In addition, patients who developed cancer
within the first year of followup (n = 5) were excluded in order
to reduce the possibility of recruitment bias due to inadvertent
inclusion of patients who had rheumatologic manifestations
due to an underlying malignancy (28,29). Study closure was
January 1 , 1995.
Evaluation. After the initial comprehensive examination, patients were seen once yearly or more frequently as
clinically indicated. Standard clinical data (30) were gathered
throughout the study course. Since 1981, followup also consisted of a semi-annual mailing of the Stanford Health Assessment Questionnaire (HAQ) (31). The HAQ inquires into a
broad spectrum of health concerns, including comorbid
diagnoses.
All diagnoses of cancer were confirmed and crossreferenced with the Provincial Cancer Registry. This registry is
a record of all premorbid diagnoses of cancer in our province
since 195.5. Prior to 1988, diagnoses of cancer that were
established only on the death certificate or at autopsy were not
recorded in the Cancer Registry. Therefore, in order to ensure
like comparisons, cancers within our study cohort that were
diagnosed only at death prior to 1988 were excluded.
Prevalence rates in relation to the risk factors of
smoking and obesity were determined from HAQs completed
between 1984 and 1986. Obesity was defined as a body mass
index >27.8 for men and >27.3 for women (32).
Sjogren’s syndrome was diagnosed clinically if there
CIBERE ET AL
1582
was believed to be substantial dryness of the eyes or mouth for
no other apparent reason. Many, but not all, patients diagnosed with Sjogren's syndrome were referred to ophthalmologists for diagnostic confirmation and therapy. Vasculitis was
also diagnosed clinically. Biopsies for Sjogren's syndrome and
vasculitis were done if hclieved to be clinically warranted, but
were not required as part of this study. RA functional class was
defined in accordance with the classification system of Steinbrocker et a1 (33).
Statistical analysis. Provincial cancer statistics were
used to calculate the expected cancer incidence rates for each
study patient, corrected for age, sex, and calendar year. The
relative risks of any cancer and of each site-specific cancer for
the study cohort were then calculated as standardized incidence ratios (SIRS), that is, the ratio of observed-to-expectcd
cancers, along with the corresponding 95% confidence intervals (95% CI) (34). The first year of followup was censored for
each patient. The last year of followup was censored for each
patient, unless a malignancy of interest was diagnosed in that
year, or unless followup was known to extend beyond study
closure. All data after the diagnosis of cancer were excluded.
For those patients lost t o followup at study closure, all data up
to the loss of followup were included.
Rates for smoking and obesity in the general population were derived from Canadian epidemiology studies conducted during the 1980s (35,36), and were compared with the
rates found in the study group, using the chi-square test.
Statistical significance was defined as P < 0.05.
RESULTS
Of the original 1,191 patients enrolled with a
clinical diagnosis of RA, 893 were found to have definite
or classic RA (37). Of these, 26 patients had cancer
diagnosed prior to study enrollment (i.e., prior to first
contact at our RDU) and 5 developed cancer within the
first year of followup. These 31 patients were excluded
from further analysis. The 5 malignancies excluded
during the first year of followup were 1 lymphosarcoma,
1 colorectal, 1 lung, 1 prostate, and 1 skin cancer.
Inclusion of these patients in the analysis did not change
the significance of our results. The study population of
interest thus consisted of 862 patients. At study closure,
their status was as follows: 330 (38%) were still under
observation, 466 (54%) had died, 26 (3%) had refused to
participate, and 40 (5%) were lost to followup.
Our findings are based on 14,998 patient-years of
followup. A summary of the demographic and clinical
features, as well as the medication use, of the study
cohort is shown in Table 2. As expected, almost all the
patients had been receiving NSAID therapy, most of
them on a long-term basis. Azathioprine, methotrexate,
and cyclophosphamide were used by 7 % , 7 % , and 2% of
patients, respectively. Any 1 of these 3 cytotoxic agents
was used in 12% of the patients.
Table 2. Characteristics of the rheumatoid arthritis (RA) study
patients"
RA patients
(n = 862)
Feature
Demographic
Sex, 5% female
Race, % white
Mean followup (range), years
Mean age (range) at onset of RA, years
Clinical
Rheumatoid factor positive
Subcutaneous nodules
Joint erosions
RA functional class 2-3
Vasculitis
Sjogren's syndrome
Drug use
All NSAIDs
Aspirin
Prednisone
Parenteral gold
Hydroxychloroquine
Penicillamine
Azathioprine
Methotrexate
Cyclophosphamide
67
96
17.4 (0.6-35)
43 (16-82)
91
57
89
13
10
32
97
95
63
50
29
23
7
7
2
Except where otherwise indicated, values are the percentage of
patients. NSAIDs = nonsteroidal antiinflammatory drugs.
Potential lifestyle risk factors for cancer in our
patients and in the general population are compared in
Table 3. The proportion of current smokers was greater
in the study patients than in the general population
(36% versus 29%; P > 0.05), particularly in women,
where the difference reached statistical significance
(44% versus 28%; P < 0.001). The prevalence of obesity
in the study cohort was similar to that in the general
population (35% versus 31%), with a slight predominance in both male (42% versus 35%) and female (32%
versus 27%) RA patients, although these differences
Table 3. Incidence rates of cancer risk factors in rheumatoid arthritis
(RA) patients and the general population"
Prevalence, %
Risk factor
Smoker
Total
Male
Female
Obesity
Total
Male
Female
" NS
=
RA patients
(1984-1986)
General population
(1986-1990)
P
36
28
44
29
30
28
NS
NS
<:0.001
35
42
32
31
35
27
NS
NS
NS
not signficant (P> 0.05).
RISK OF MALIGNANCY IN RA
1583
Table 4. lncidence of site-specific cancers in rheumatoid arthritis
patients*
Cancer site
All
Bladder
Breast
Colorectal
Gynecologic
Ovary
Uterus
Cervix
Kidney
Liverigallbladder
Lung
Female
Male
Hemopoie tic
Leukemia
Non-Hodgkin’s lymphoma
Hodgkin’s disease
Oropharynx
Pancreas
Prostate
Skin
Stomachiesophagus
Observed:
expected
incidence
136:167.5
2:5.4
18:22.8
10:19.0
10:12.71
5:5.6
4:6.4
112.0
5r3.1
512.6
16:14.8
516.0
11:8.8
12:9.6
913.6
3:5.5
0:0.43
3:1.00
5:3.9
12:12.0
45151.6
415.6
SIR
95% CI
P
0.80
0.36
0.90
0.52
0.78
0.89
0.62
0.49
1.62
1.93
1.08
0.84
1.24
1.25
2.47
0.55
0.00
3.00
1.27
1.00
0.83
0.71
0.67-0.95
0.04-1.33
0.46-1.24
0.25-0.96
0.36-1.44
0.28-2.09
0.16-1.60
0.06-2.76
0.52-3.78
0.62-4.50
0.61-1.75
0.27-1.96
0.62-2.22
0.64-2.18
1.12-4.69
0.11-1.60
0.00-8.53
0.60-8.76
0.41-2.96
0.51-1.74
0.63-1.16
0.19-1.84
0.01 1
NS
NS
0.037
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
0.026
NS
NS
NS
NS
NS
NS
NS
*SIR = standardized incidence ratio; 95% CI
interval; NS = not signficant (P > 0.05).
=
95% confidence
were not statistically significant. Data on other risk
factors for malignancy, such as hormone therapy, dietary
factors, and family history, were not available.
By 35 years of followup, 136 of the 862 RA
patients were identified as having malignancies, whereas
168 were expected (SIR 0.80, P = 0.011 [95% CI
0.67-0.951) (Table 4). Colorectal cancer risk was reduced by nearly half, with 10 cases observed compared
with 19 expected (SIR 0.52, P = 0.037 [95% CI 0.250.961). There was a trend toward a reduced risk for
breast cancer in women, although the apparent reduction was not statistically significant. The overall risk of
hemopoietic malignancy was only slightly increased (SIR
1.25, P > 0.05 [95% CI 0.64-2.181). However, subanalysis showed that leukemia was significantly overrepresented (SIR 2.47, P = 0.026 [95% CI 1.12-4.691). No
cases of Hodgkin’s disease were identified, and only 3
patients developed non-Hodgkin’s lymphoma (SIR 0.55,
P > 0.05 [95% CT 0.11-1.601). Lung cancer was more
frequent than expected, but not significantly so. Because
of the significant differences in smoking rates observed
in our cohort, lung cancer was analyzed separately for
men and women, but the results showed no difference
from expected rates for either subgroup. There was no
difference in observed and expected rates of prostate
cancer. No other statistically significant differences were
found, although the number of cases of each of the other
site-specific malignancies was small.
DISCUSSION
In our cohort of RA patients, we found a reduced
risk of colorectal cancer. This has not been a universal
finding in other studies. Isomaki et a1 (1) and Gridley et
al ( 5 ) did report reduced rates of colorectal cancer in
their series of RA patients. In contrast, several studies
(2,3,6) found trends toward an increased risk for colorectal cancer, although these series involved small numbers of malignancies and the excess occurrences were
not statistically significant. Our results are consistent
with reports of reduced colorectal malignancy among
patients who are regular users of aspirin (10-13). In
those studies, it was noted that long-term aspirin use was
associated with a progressive decrease in risk. However,
this risk reduction does not become apparent until after
at least 10 years of aspirin use on a regular basis. Most
of our patients had been on a long-term regimen of
NSAIDs, and our followup was sufficient to be able to
confirm a reduced risk of colorectal cancer.
Factors other than NSAID use that may have
contributed to the reduced occurrence of colorectal
cancer observed in our cohort are not apparent. Corticosteroids also inhibit prostaglandin synthesis, and 65%
of our study cohort received corticosteroids at some
time. Although we had an insufficient number of patients to assess for any cancer-sparing effect of corticosteroids independent of NSAIDs, we did not observe any
enhanced reduction of colorectal cancer in those patients who received corticosteroids (unpublished observations). It is conceivable that increased frequency of
gastrointestinal (GI) symptoms or GI bleeding could
lead to increased investigations and therefore early
detection of precancerous lesions in NSAID users. However, this is unlikely to account for the large risk
reduction found in our cohort. Generally, the concern
about GI blood loss in NSAID users is focused on
bleeding in the upper GI area. We did not have a
screening program for occult GI bleeding. We do have
coding for bowel polypectomy, but we had no information that any of our patients underwent benign polypectomy despite our followup with a semi-annual HAQ.
Furthermore, our study patients were not subjected to
additional surveillance with fecal occult blood testing or
sigmoidoscopy other than the routine surveillance that
applies to the general population.
The risk reduction in colorectal cancer could
1584
have been conceivably due, in part, to RA itself, although to date no such pathogenetic mechanism has
been established or suggested. Furthermore, the degree
of risk reduction for colorectal cancer that we observed
(52%) is similar to that reported in non-RA NSAID
users (10-13). Premature mortality is a feature in patients with RA, including our cohort (38). However, that
should not have been a confounding factor in our results,
since our controls were matched for age, as well as sex
and calendar years of followup.
Dietary factors appear to play a role in colorectal
cancer, with obesity being associated with an increased
risk of cancer (32). The prevalence of obesity in our
study patients was slightly higher than in the general
population. Thus, if anything, one would have expected
more, not less, colorectal cancer in our study cohort,
thereby adding to the significance of our results.
Our data showed a trend toward an increased risk
of lymphoproliferative malignancies. A similar trend has
been observed in some studies (3,14), while other studies
found not only a trend, but a significant excess of
hemopoietic malignancies (1,2,5). In our subanalysis,
leukemia was significantly overrepresented. This has not
been consistently observed in the literature. In studies
where subanalyses were done, only some (1,2,5) showed
an excess of leukemia. In 2 of those studies (1,5), results
were significant only for men, while in a third study (2),
only women had an increased risk of leukemia. Similar
inconsistencies have been noted for lymphoma, and are
likely due to small numbers of observed malignancies.
Nevertheless, there is a persistent overrepresentation of
hematologic malignancies in previously published studies. This excess risk of hemopoietic malignancy was
confirmed in our study.
Although several factors have been proposed to
account for the increased risk of lymphoproliferative
malignancy in RA, those factors are not evident in our
cohort. Cytotoxic medications, in particular azathioprine
and cyclophosphamide, have been associated with an
increased risk of hemopoietic malignancy (15-19). However, of our 12 patients with hemopoietic malignancy,
only 2 had received cytotoxic agents. One had been
receiving cyclophosphamide, while the other had received both cyclophosphamide and azathioprine. Thus,
cytotoxic agents do not account for the increased leukemia seen in our cohort.
Sjogren’s syndrome is also associated with an
increased risk of lymphoproliferative malignancy, but
this is an unlikely factor in this study. We diagnosed
Sjogren’s syndrome on clinical grounds, not requiring
biopsy or ophthalmology confirmation. Yet, despite
CIBERE ET AL
these less stringent diagnostic criteria, only 3 of our 12
patients with hemopoietic malignancies had Sjogren’s
syndrome. Two of these 3 patients with Sjogren’s syndrome had received cytotoxic agents.
Viruses, in particular Epstein-Barr virus, and
other infectious agents have been investigated as causative agents in the pathogenesis of both lymphoma and
RA (39,40), although to date no clear link has been
established. Thus, it may well be that increased lymphoproliferative malignancy is a consequence of RA itself.
RA is characterized by persistent immune stimulation.
This chronic immune stimulation and polyclonal lymphocyte proliferation could conceivably increase the
potential for malignant transformation, predisposing to
an increased risk of hemopoietic malignancy.
A trend toward increased lung cancer in RA has
been reported (1,3,5), and in a large Finnish study (l),
the difference reached statistical significance in men.
The incidence of lung cancer in our study cohort was
increased, although not significantly so. Whether the
increase we observed is related to RA is not clear.
Matteson et a1 reported increased lung cancer in RA
patients treated with azathioprine (16), but the number
of patients who were taking azathioprine or other cytotoxic agents was insufficient to make any inferences.
Recent evidence has suggested that smokers are at
greater risk of developing RA (41-43). Unfortunately,
we had information on smoking history for only 20% of
our RA patients (174 of 862), since the importance of
smoking was not appreciated during the study enrollment period of 1966-1974, nor during the early followup
years. Among those study patients for whom the smoking history was known, smoking rates were higher in RA
patients compared with the general population (36%
versus 29%; P > 0.05), and this difference was significant for female smokers (44% versus 28%; P < 0.001).
This higher-than-expected prevalence of smoking may
have contributed to the increased rate of lung cancer in
our study population. Of course, RA itself and rheumatoid lung disease may predispose to lung cancer, as
suggested by case studies in which malignant cells were
found in rheumatoid lung nodules (44).
There are several strengths to this study, including the large cohort and, in particular, the communitybased nature of the cohort. This should allow for the
generalizability of our results to most white patients with
RA. All cancer diagnoses in our province are recorded
in the Provincial Cancer Registry, thus ensuring that
ascertainment of cancer diagnoses was complete and
accurate. Followup rates of the study patients have been
good. In addition, the period of followup was sufficiently
RISK OF MALIGNANCY IN RA
long to allow for observation of differences in malignancy rates.
Limitations of this study include the fact that
there was an insufficient number of patients taking
DMARDs and cytotoxic agents to assess for an association with malignancy. Only 50% of patients received
p a r e n t e d gold and only 29% received hydroxychloroquine. Azathioprine, methotrexate, and cyclophosphamide were used in only 7%, 7%, and 2% of
patients, respectively, and any 1 of these agents was used
in only 12% of the patients. These rates of DMARD and
cytotoxic usage are low by today’s standards, but they
simply reflect the less aggressive approach taken in the
management of RA that was the norm during the era of
enrollment and in the early followup period of this study.
An association of cytotoxic agents with lymphoproliferative malignancy would have been of particular interest,
although in an earlier case-control study, we were
unable to confirm an increased risk of lymphoproliferative malignancy in RA patients taking azathioprine,
cyclophosphamide, or methotrexate (19). In addition,
information on smoking and obesity was not complete,
and may have been biased toward collecting such information from individuals at risk for those variables.
By study conclusion, 54% of our cohort had died.
While this is certainly a sizable portion, it is likely that
the rate and distribution of malignancies will differ with
continued followup as the survivors age. However, in this
large cohort of predominantly white patients with RA,
comprising 15,000 patient-years of followup, we found
the risk of colorectal malignancy was significantly decreased. This risk reduction may be due, at least in part,
to a cancer-protective effect of NSAID therapy.
Whether other factors such as RA itself contribute to a
cancer-protective effect will need to be determined in
future investigations. In addition, a significant excess of
leukemia was found, and a trend toward increased
hemopoietic malignancy waf confirmed. Our data suggest that this increased risk of leukemia, rather than
being related to external factors, may be secondary to
persistent immune stimulation from R A itself. For many
site-specific cancers, the numbers of cases were too
small to assess. Nonetheless, despite an excess of hemopoietic malignancy and despite therapy with potentially
oncogenic cytotoxic agents in 12% of our patients, the
overall risk of malignancy was reduced in this R A
cohort.
-
ACKNOWLEDGMENTS
The authors gratefully recognize the work of Dr. D. M.
Mitchell who initiated a long-term observational study of
patients with RA in 1966, and followed up this large cohort,
from which this study is drawn, until his death in 1988. The
authors also thank the Saskatoon Cancer Centre for providing
access to the Cancer Registry files, and for supplying provincial
cancer statistics.
REFERENCES
1. Isomaki HA, Hakulinen T, Joutsenlahti U: Excess risk of lymphomas, leukemia and myeloma in patients with rheumatoid arthritis.
J Chronic Dis 31591-696, 1978
2. Prior P, Symmons DPM, Hawkins CF, Scott DL, Brown R: Cancer
morbidity in rheumatoid arthritis. Ann Rheum Dis 43:128-131,
1984
3. Katusic S, Beard CM, Kurland LT, Weis JW, Bergstralh E:
Occurrence of malignant neoplasms in the Rochester, Minnesota,
rheumatoid arthritis cohort. Am J Med 7850-55, 1985
4. Eriksson M: Rheumatoid arthritis as a risk factor for multiple
myeloma: a case-control study. Eur J Cancer 29A:259-263, 1993
5. Gridley G, McLaughlin JK, Ekbom A, Klareskog L, Adami H,
Hacker DG, Hoover R, Fraumeni JF: Incidence of cancer among
patients with rheumatoid arthritis. J Natl Cancer Inst 85307-311,
1993
6. Moritomo H, Ueda T, Hiyama T, Hosono N, Mori S, Komatsubara Y: The risk of cancer in rheumatoid patients in Japan. Scand
J Rheumatol 24:157-159, 1995
7. Narisawa T, Satoh M, Sano M, Takahashi T: Inhibition of initiation and promotion by N-methylnitrosoruea-induced colon carcinogenesis in rats by non-steroid anti-inflammatory agent indomethacin. Carcinogenesis 4:1223-1227, 1983
8. Metzgeer U, Meier J, Uhlschmid G, Weihe H: Influence of various
prostaglandin synthesis inhibitors on DMH-induced rat colon
cancer. Dis Colon Rectum 27:366-369, 1084
9. Reddy BS, Nayini J, Tokumo K, Rigotty J, Zang E, Kelloff G:
Chemoprevention of colon carcinogenesis by concurrent administration of piroxicam, a nonsteroidal antiinflammatory drug with
D,L,-a-difluoromethylornithine, an ornithine decarboxylase inhibitor, in diet. Cancer Res 50:2562-2568, 1990
10. Thun MJ, Namboodiri MM, Heath CW Jr: Aspirin use and
reduced risk of fatal colon cancer. N Engl J Med 121:241-246,
1991
11. Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Ascherio A,
Willett WC: Aspirin use and the risk for colorectal cancer and
adenoma in male health professionals. Ann Intern Med 121:241246, 1994
12. Giovannucci E, Egan KM, Hunter DJ, Stampfer MJ, Colditz GA,
Willett WC, Speizer FE: Aspirin and the risk of colorectal cancer
in women. N Engl J Med 333:609-614, 1995
13. Schreinemachers DM, Everson RB: Aspirin use and lung, colon,
and breast cancer incidence in a prospective study. Epidemiology
5138-146, 1994
14. Fries JF, Bloch D, Spitz P, Mitchell DM: Cancer in rheumatoid
arthritis: a prospective long-term study of mortality. Am J Med
78:56-59, 19x5
15. Baker GL, Kahl LE, Zee BC, Stolzer BL, Aganval AK, Mcdsgcr
TA: Malignancy following treatment of rheumatoid arthritis with
cyclophosphamide. Am J Med 83:l-9, 1987
16. Matteson EL, Hickey AR, Maguire L, Tilson HH, Urowitz MB:
Occurrence of neoplasia in patients with rheumatoid arthritis
enrolled in a DMARD registry. J Rheumatol 18:809-814, 1991
17. Silman AJ, Petrie J, Hazleman B, Evans SJW: Lymphoproliferative cancer and other malignancy in patients with rheumatoid
arthritis treated with azathioprine: a 20 year follow up study. Ann
Rheum Dis 47:988-992, 1988
18. Radis CD, Kahl LE, Baker GL, Wasko MCM, Cash JM, Gallatin
A, Stolzer BL, Aganval AK, Medsger TA Jr, Kwoh C K E€fects of
1586
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
cyclophosphamide on thc development of malignancy and on
long-term survival of patients with rheumatoid arthritis: a 20-year
followup study. Arthritis Rheum 38:1120-1127, 1995
Williams CA, Bloch DA, Sibley J, Haga M, Wolfe F, Raynauld J-P,
Singh G, Hickey AR, Fries JF: Lymphoma and leukemia in
rheumatoid arthritis: are they associated with azathioprine, cyclophosphamide, or methotrexate? A matched case-control study in
the Arthritis, Rheumatism and Aging Medical Information System
(ARAMIS) population. J Clin Rheumatol 2:64-72; 1996
Penn I: Depressed immunity and the development of cancer. Clin
Exp Immunol46:459-474, 1981
Kaldor JM, Day NE, Pcttersson F, Clarke EA, Pedersen D,
Mehnert W, Bell J, Host H, Prior P, Karjalaine S, Neal F, Koch M,
Band P, Choi W, Kirn VP, Arslan A, ZarCn B, Belch AR, Storm
H, Kittelmann B, Fraser P, Stovall M: Leukemia following chemotherapy for ovarian canccr. N Engl J Med 322: 1-6, 1990
Ellman MH, Hurwitz H, Thomas C, Kozloff M: Lymphoma
developing in a patient with rheumatoid arthritis taking low dose
weekly methotrexate. J Kheumatol 18:1741-1743, 1991
Shiroky JB, Frost A, Skelton JD, Haegert DG, Newkirk MM,
Neville C: Complications of immunosuppression associated with
weekly low dose methotrcxate. J Rheumatol 18:1172-1175, 1991
Taillan B, Garniert G, Castanet J, Ferrari E, Pesce A, Dujardin P:
Lymphoma developing in a patient with rheumatoid arthritis
taking methotrexate. Clin Rheumatol 12:93-94, 1993
Pointud P, Prudat M, Pcron JM: Acute leukemia after low dose
methotrexate therapy in a patient with rheumatoid arthritis.
J Rheumatol 20:1215-1216, 1993
Zimmer-Galler I, Lie JT: Choroidal infiltrates as the initial
manifestation of lymphoma in rheumatoid arthritis after treatment
with low-dose methotrexate. Mayo Clin Proc 69:258-261, 1994
Mitchell DM, Fries JF: An analysis of the American Rheumatism
Association criteria for rheumatoid arthritis. Arthritis Rheum
25:481-487, 1982
Sheon RAP, Kirsner AB, Tangsintanapas P, Samad F; Garg ML,
Finkel RI: Malignancy in rheumatic disease: interrelationships.
J Am Geriatr Soc 25:20-27, 1977
Dorfman HD, Siege1 HL, Perry MC, Oxenhandler R: NonHodgkin’s lymphoma of the synovium simulating rheumatoid
arthritis. Arthritis Rheum 30:155-161, 1987
The Computer Committee of the American Rheumatism Association: A uniform database for rheumatic diseases. Arthritis
Rheum 22:1029-1033, 1979
Fries JF, Spitz P, Kraincs RG, Holman HR: Measurements of
patient outcome in arthritis. Arthritis Rheum 23:136-145, 1980
CIBERE ET AL
32. National Institutes of Health Consensus Development Panel on
the Health Implications of Obesity: Health implications of obesity:
National Institutes of Health consensus development conference
statement. Ann Intern Med 103:147-151, 1985
33. Steinbrocker 0, Traeger CH, Batterman RC: Therapeutic criteria
in rheumatoid arthritis. JAMA 140:659-662, 1949
34. Breslow NE, Day NE: Rates and rate standardization. In, Statistical Methods in Cancer Research. Vol. 2. Edited by NE Breslow,
NE Day. New York, Oxford University Press, 1987
35. Stachenko SJ, Keeder BA, Lindsay E, Donovan C, Lessard R,
Balram C: Smoking prevalence and associated risk factors in
Canadian adults: Canadian Heart Health Surveys Research
Group. Can Med Assoc J 146:1989-1996, 1992
36. Reeder BA, Angel A, Ledoux M, Rdbkin SO, Young TK, Sweet
LE: Obesity and its relation to cardiovascular disease risk factors
in Canadian adults: Canadian Heart Health Surveys Research
Group. Can Med Assoc J 146:2009-2019, 1992
37. Ropes MW, Bennett GA, Cobhs, Jacox R, Jessar RA: 1958
revision of diagnostic criteria for rheumatoid arthritis. Bull Rheum
Dis 9: 175-176, 1958
38. Wolfe F, Mitchell DM, Sibley JT, Fries JF, Bloch DA, Williams
CA, Spitz PW, Haga M, Kleinheksel SM, Cathey M A The
mortality of rheumatoid arthritis. Arthritis Rheum 37:481-494,
1994
39. Henle W, Henlc G: Epstein-Barr virus and human malignancies.
Adv Viral Oncol 5:201-238, 1985
40. Depper JM, Zvaifler NJ: Epstein-Barr vi
its relationship to the
pathogenesis of rheumatoid arthritis. Art
s Rheum 24755-761,
1981
41. Hernandez-Avila M, Liang MH, Willett WC, Stampfer MJ, Colditz GA, Rosner B, Roberts WN, Hennekens CH, Speizer FE:
Reproductive factors, smoking, and the risk for rheumatoid arthritis. Epidemiology 1:285-291. 1990
42. Voigt LF, Kocpsell TD, Nelson JL; Dugowson CE, Daling JR:
Smoking, obesity, alcohol consumption, and the risk of rheumdtoid arthritis. Epidemiology 5:525-532, 1994
43. Silman AJ, Newman J, MacGregor AJ: Cigarette smoking increases the risk of rheumatoid arthritis: results from a nationwide
study of diseasc-discordant twins. Arthritis Rheum 39:732-735,
1996
44. Shcnberger KN, Schned AR, Taylor TH: Rheumatoid disease and
bronchogenic carcinoma: case report and review of the literaturc.
J Rheumatol 11:226-228, 1984