Annals of Oncology
original article doi:10.1093/annonc/mdr333
Autoimmune disease and subsequent digestive tract
cancer by histology
K. Hemminki
1,2
*, X. Liu
2
, J. Ji
2
, J. Sundquist
1,3
& K. Sundquist
2
1
Division of Molecular Genetic Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany;
2
Center for Primary Health Care Research, Lund
University, Malmo¨, Sweden;
3
Stanford Prevention Research Center, Stanford University School of Medicine, Stanford, USA
Received 21 March 2011; accepted 1 June 2011
Background: Dysregulation of the immune function in autoimmune diseases could potentially lead to cancer
development and there is definite evidence linking some autoimmune mechanisms with cancer. We analyzed
systematically the occurrence of histology-specific digestive tract cancers in patients diagnosed with 33 different
autoimmune diseases in order to address the question of shared susceptibility.
Patients and methods: Standardized incidence ratios (SIRs) were calculated for subsequent digestive tract
cancers up to the year 2008 and in patients hospitalized for autoimmune disease after the year 1964.
Results: Myasthenia gravis associated with five different cancers with SIRs ranging from 1.35 to 2.78. Pernicious
anemia, Crohn disease, ulcerative colitis, systemic lupus erythematosis and psoriasis were also associated with
cancers at multiple sites. Rheumatoid arthritis associated with no cancer and the standardized incidence ratio was
decreased for colon adenocarcinoma, also in ankylosing spondylitis patients.
Conclusions: Increased risks of cancer were observed in patients with several autoimmune diseases. Myasthenia
gravis and pernicious anemia were associated with many cancers; this is possibly related to immunosuppressant
medication in myasthenia gravis. The decreased risks in colon and rectal adenocarcinomas in rheumatoid arthritis and
ankylosing spondylitis suggest underlying inflammatory mechanisms as the risks may have been suppressed by the
use of anti-inflammatory medication.
Key words: cancer, digestive tract, autoimmune disease
introduction
Cancer risk is vastly increased in immune deficiency conditions
and immunosuppressed patients [1]. A normal immune system
can act against tumor formation at multiple levels. It can eliminate
and suppress microbial infections and it can resolve inflammation
before it becomes chronic and tumor promoting. Transformed
cells express tumor-specific antigens, which are recognized by
tumor immune surveillance and subsequently eliminated. Thus,
dysregulation of the immune function in conditions such as
autoimmune diseases could potentially lead to cancer
development. Autoimmune diseases, which affect 5% to 10% of
the population in the developed countries, include some common
diseases, such as rheumatoid arthritis and autoimmune
thyroiditis, and many rare ones [2]. For unknown reasons, many
autoimmune diseases are more common in women than in men.
There are also large ethnic differences in their prevalence.
Autoimmune diseases are often defined as clinical syndromes
caused by the activation of T cells or B cells, or both, in the absence
of an ongoing infection or other discernible cause [3]. These cells
recognize specific foreign antigens but even in healthy individuals,
they possess a low level of autoreactivity toward own (self or auto)
antigens. In autoimmune diseases, autoreactivity is increased
either systemically (as in systemic lupus erythematosus) or organ
specifically (type 1 diabetes). The diseases are diagnosed on the
basis of various criteria, such as clinical symptoms or signs and by
specific types of autoantibodies. Genetic epidemiological studies
have demonstrated that genetic factors are crucial determinants of
susceptibility to autoimmune disease [3].
Many previous studies have shown associations between
autoimmune diseases and cancer, particularly lymphomas and
myeloma [4, 5]. However, many of the previous studies have
focused on a single or a few autoimmune diseases and cancers, and
most have covered small populations. As an exception, a recent
large study reported alimentary tract cancers after hospitalization
for autoimmune diseases in 4.5 million USA male veterans [6]. In
the present follow-up study, we examined digestive tract cancer
risks in those who had been hospitalized for an autoimmune
disease, covering the total Swedish population of 9 million. The
major differences to the USA study are that the present population
included also women and children. Moreover, 27 autoimmune
diseases were covered by both studies but we included 6 additional
diseases. The present study included tumor histologies to examine
whether the many autoimmune diseases affect a specific histology.
materials and methods
All the registers used in this study were nationwide, covering the whole
population of Sweden over a defined period of time (9.0 million in 2005).
o r i g i n a l
a r t i c l e
*Correspondence to: K. Hemminki, Division of Molecular Genetic Epidemiology, German
Cancer Research Center (DKFZ), Im Neuenheimer Feld 580, D-69120 Heidelberg,
Germany. Tel: +49-6221-421800; Fax: +49-6221-421810; E-mail: k.hemminki@dkfz.de
ª The Author 2011. Published by Oxford University Press on behalf of the European Society for Medical Oncology.
All rights reserved. For permissions, please email: journals.permissions@oup.com
Annals of Oncology Advance Access published August 2, 2011
at Gesellschaft f?r Biotechnologische Forschung on August 17, 2011 annonc.oxfordjournals.org Downloaded from The research database used for this study is a subset of the national MigMed
2 datasets at the Center for Primary Health Care Research, Lund University,
Malmo¨, Sweden. Data on autoimmune diseases were obtained from the
Swedish Hospital Discharge Register that records complete data on all
discharges with dates of hospitalization and diagnoses in some regions since
1964 and nationwide since 1986. The International Classification of
Diseases codes used were described earlier [7, 8]. We did not want to be
restrictive in including diseases into the present study. Rather, we wanted to
include diseases such as multiple sclerosis and amyotrophic lateral sclerosis
even though some consider these neurodegenerative rather than
autoimmune diseases. A total of 33 diseases were covered. However, five of
them had less than five cases for any cancer and data are not shown for
these diseases that were autoimmune hemolytic anemia, chorea minor,
discoid lupus erythematosus, lupoid hepatitis and Reiter disease; these
diseases were not significantly associated with any cancer. The linkages were
carried out by the use of an individual national identification number that
is assigned to each person in Sweden for their lifetime. This number was
replaced by a serial number for each person in order to provide anonymity.
Cancers were obtained from the nationwide Swedish Cancer Registry.
Person-years of follow-up were calculated from date of discharge with the
first main diagnosis of autoimmune disease until diagnosis of cancer, death,
emigration, or closing date, 31 December 2008. In control studies to
check for the possibility of surveillance bias in patients who underwent
treatment, the follow-up was started from the last hospitalization.
Standardized incidence ratios (SIRs) were calculated as the ratio of observed
(O) to expected (E) number of cases. Expected numbers were calculated for
anyone not hospitalized for autoimmune disease. The expected numbers
were calculated as age (5-year groups), sex, period (5-year groups), region
and socioeconomic status-specific standard incidence rates. An additional
adjustment, because of potential confounding, was made for hospitalization
for obesity using codes ICD-7 = 287.00, 287.09; ICD-8 = 277.99; ICD-9 =
278A; ICD-10= E65-E68 (26). A total of 30 020 individuals had been
hospitalized for obesity. Similar adjustments were made for smoking using
hospitalization for chronic obstructive pulmonary disease as a surrogate
with 260 243 individuals affected (codes: ICD-7 = 500–502; ICD-8 = 490–
493; ICD-9 = 490–496; ICD-10= J40–J49), and alcohol using alcoholisms as
a surrogate (codes: ICD-9 = 303; ICD-10 = F10.1–F10.9) with 181 862
individuals affected, respectively. Assuming a Poisson distribution, 95%
confidence intervals (95% CI) were calculated.
The study was approved by the regional ethical review board at Lund.
results
Table 1 shows the risks of upper digestive tract, esophageal and
stomach cancers after 28 autoimmune diseases; five
autoimmune diseases had less than five cases even for the most
common cancers and they were excluded but none of the
results were significant. The first column shows the personyears at risk for autoimmune disease patients. The risks of
upper digestive tract squamous cell carcinoma (SCC) were
increased after hospitalization for seven autoimmune diseases
with an overall standardized incidence ratio (SIR) of 1.37; the
highest risks were seen for systemic lupus erythematosus (2.86)
and pernicious anemia (2.17). Esophageal adenocarcinoma was
only increased in localized scleroderma (5.77) and
polymyosistis/dermatomyositis patients (16.58), both
diagnosed in three patients. Esophageal SCC was more
common in the present population than was adenocarcinoma
and it was increased in patients diagnosed with pernicious
anemia (5.62), Addison disease (5.54) and myasthenia gravis
(2.78); the overall SIR was 1.67. Stomach adenocarcinoma was
increased in seven patient groups, with the highest risks in
patients with pernicious anemia (4.09) and immune
thrombocytopenic purpura (3.04). Pernicious anemia and
myasthenia gravis were associated with increases in three of the
studied cancers in Table 1.
Small intestinal adenocarcinoma is a rare cancer but the
overall SIR was 2.33 (Table 2). Very high risks were noted for
Crohn disease (11.05) and pernicious anemia (7.64); the SIR for
ulcerative colitis was 4.10. Colon and rectal adenocarcinomas
were increased in Crohn disease (2.76 and 1.83) and ulcerative
colitis patients (3.13 and 1.94) and also in chronic rheumatic
heart disease and myasthenia gravis patients. Both cancers were
decreased in ankylosing spondylitis (0.55 and 0.35) patients
and colon cancer was also decreased in rheumatoid arthritis
patients (0.70). Anal cancer is a rare cancer with SCC histology;
it was increased in systemic lupus erythematosus (7.18),
sarcoidosis (3.51) and psoriasis patients (3.18).
Gastrointestinal carcinoids occur mainly in the stomach,
small intestine and colorectum (Table 3). Particularly high risks
for stomach carcinoids were observed in patients with primary
biliary cirrhosis (78.29), pernicious anemia (46.67), type 1
diabetes (39.21) and rheumatoid arthritis (5.38); the overall SIR
was 4.53. Crohn disease and ulcerative colitis patients were at
a risk of small intestinal (7.33 and 3.31) and colorectal
carcinoids (8.18 and 7.32). Small intestinal carcinoids were also
increased in sarcoidosis patients (3.67). In order to assess the
effect of lead time bias, the risks for carcinoid tumors were
analyzed in two follow-up periods after the last hospitalization:
£2 years and >2 years. Small intestinal and colorectal carcinoids
were highly increase within 2 years of hospitalization and the
only SIR that was significant in the >2 period was that for
colorectal carcinoids in Crohn disease patients (N = 15, SIR =
1.95, 95% CI 1.09–3.22).
The present data indicated cancer risks after the first
hospitalization for an autoimmune disease. The analyses were
repeated considering cancer after last hospitalization for an
autoimmune disease; the case numbers were fewer but the SIRs
were essentially unchanged.
discussion
A follow-up study of the present design may raise a number of
technical queries. Diagnostics should be of high quality because
most of the patients would have been treated in specialist
departments and the indicated diagnosis was the one on which
the patients were discharged. Coverage is another issue that
depends, e.g. on the severity of the condition. We have
discussed diagnostic accuracy and coverage in many previous
papers on familial clustering of autoimmune diseases in Sweden
[9–12]. Severe debilitating conditions, such as Graves disease,
type 1 diabetes mellitus, Crohn disease, ulcerative colitis,
rheumatoid arthritis and celiac disease, lead to hospitalization
at one point or another in Sweden. In contrast, Hashimoto
disease/hypothyroidism is an example of a disease with a low
degree of hospitalization.
Another issue relating to the interpretation of the results is
the lead time bias, i.e. the possibility of earlier diagnosis of any
other disease, such as cancer, in patients with chronic disease
original article Annals of Oncology
2 | Hemminki et al.
at Gesellschaft f?r Biotechnologische Forschung on August 17, 2011 annonc.oxfordjournals.org Downloaded from and frequent medical contacts. The Swedish Cancer Registry
records all new cases of cancer and close to 100% of the cases
are histologically or cytologically confirmed. Thus, the lead
time bias would only shift the diagnoses earlier as the
diagnostic accuracy is not compromised. Benign tumors,
particularly carcinoids in the present study, are of particular
concern because they are usually indolent tumors diagnosed
incidentally at medical examination conducted for other
purposes [13, 14]. As all the observed risks were much higher
immediately after hospitalization for autoimmune diseases, it is
Table 1. Risk of cancers of the upper digestive tract, esophagus and stomach after a specified autoimmune disease
Autoimmune disease (case
number)
Upper digestive tract Esophageal
adenocarcinoma
Esophageal SCC Stomach
adenocarcinoma
Pyrs O SIR 95% CI O SIR 95% CI O SIR 95% CI O SIR 95% CI
Digestive tract involvement
Ankylosing spondylitis
(5173)
92 881 14 1.05 0.57–1.77 5 2.99 0.94–7.03 2 0.73 0.07–2.70 13 0.92 0.49–1.57
Celiac disease (4124) 67 572 1 0.75 0.00–4.32 0 1 3.38 0.00–19.37 0
Crohn disease (28 349) 488 461 58 1.32 1.00–1.71 3 0.60 0.11–1.77 9 1.00 0.45–1.90 43 0.87 0.63–1.17
Immune
thrombocytopenic
purpura (1709)
28 310 2 1.36 0.13–5.01 0 1 3.13 0.00–17.92 6 3.04 1.09–6.66
Localized scleroderma
(3128)
63 257 7 1.38 0.55–2.86 3 5.77 1.09–17.09 1 0.87 0.00–5.01 11 1.56 0.70–2.55
Pernicious anemia
(11 839)
68 625 34 2.17 1.50–3.03 3 1.45 0.27–4.28 23 5.62 3.56–8.44 108 4.09 3.36–4.94
Polyarteritis nodosa
(12 046)
94 093 23 1.19 0.75–1.78 4 1.59 0.41–4.12 8 1.51 0.64–2.99 35 1.02 0.71–1.42
Primary biliary cirrhosis
(835)
7110 1 0.85 0.00–4.86 0 0 2 1.29 0.12–4.75
Sarcoidosis (9053) 136 747 16 0.95 0.54–1.54 0 2 0.54 0.05–1.98 31 1.45 0.98–2.06
Sjo¨gren syndrome (3769) 46 309 7 1.30 0.51–2.69 1 1.69 0.00–9.67 3 2.26 0.43–6.68 12 1.42 0.73–2.48
Systemic lupus
erythematosus (5318)
62 007 16 2.86 1.63–4.65 0 2 1.52 0.14–5.60 10 1.20 0.57–2.21
Systemic sclerosis (1195) 8199 1 0.94 0.00–5.40 0 0 2 1.32 0.12–4.87
Ulcerative colitis (16 363) 15 883 17 1.14 0.66–1.83 2 1.08 0.10–3.97 1 0.32 0.00–1.81 15 0.88 0.49–1.45
No digestive tract involvement
Addison disease (1594) 21 314 2 0.84 0.08–3.08 0 3 5.54 1.05–16.41 9 2.74 1.24–5.23
Amyotrophic lateral
sclerosis (4262)
19 135 2 0.74 0.07–2.70 0 0 4 0.96 0.25–2.49
Behcet disease (2860) 51 636 9 1.40 0.64–2.67 1 1.34 0.00–7.67 0 11 1.66 0.83–2.99
Chronic rheumatic heart
disease (16 770)
133 597 30 1.10 0.74–1.58 5 1.42 0.45–3.33 5 0.74 0.23–1.73 59 1.40 1.07–1.81
Diabetes mellitus type I
(20 554)
323 086 5 1.70 0.54–4.00 0 1 4.59 0.00–26.33 5 2.64 0.83–6.21
Grave/hyperthyroidism
(36 240)
468 161 66 1.36 1.05–1.73 2 0.40 0.04–1.48 9 0.75 0.34–1.42 103 1.31 1.07–1.59
Hashimoto/hypothroidism
(10 682)
100 612 20 1.70 1.04–2.63 2 1.56 0.15–5.75 6 1.95 0.70–4.28 26 1.34 0.87–1.96
Multiple sclerosis (12 553) 157 915 19 1.13 0.68–1.77 2 1.08 0.10–3.97 2 0.55 0.55–2.03 12 0.55 0.28–0.97
Myasthenia gravis (17 974) 286 399 104 1.66 1.36–2.02 11 1.29 0.64–2.31 41 2.78 2.00–3.78 117 1.38 1.14–1.65
Polymyalgia rheumatica
(14 745)
172 528 23 0.79 0.50–1.18 4 1.08 0.28–2.80 6 0.81 0.29–1.79 63 1.45 1.11–1.85
Polymyositis/
dermatomyositis (1256)
11 188 2 1.32 0.12–4.86 3 16.58 3.13–49.07 2 5.52 0.52–20.30 6 2.74 0.99–6.01
Psoriasis (15 592) 222 027 59 1.78 1.36–2.30 6 1.62 0.58–3.54 25 3.36 2.17–4.96 49 1.28 0.94–1.69
Rheumatic fever (3458) 65 040 14 1.48 0.81–2.49 0 3 1.48 0.28–4.39 16 1.50 0.86–2.44
Rheumatoid arthritis
(26 937)
23 5491 38 1.07 0.76–1.47 9 2.12 0.96–4.03 13 1.44 0.76–2.46 60 1.07 0.82–1.38
Wegener granulomatosis
(945)
9398 4 2.84 0.74–7.34 0 0 1 0.45 0.00–2.59
All (290 665) 3 606 353 599 1.37 1.26–1.48 65 1.24 0.96–1.58 172 1.67 1.43–1.94 833 1.36 1.27–1.46
CI, confidence interval; O, observed number of cases; SIR, standardized incidence ratio; Pyrs, person years.
Annals of Oncology
original article
doi:10.1093/annonc/mdr333 | 3
at Gesellschaft f?r Biotechnologische Forschung on August 17, 2011 annonc.oxfordjournals.org Downloaded from likely that the findings were due to lead time bias or
overdiagnosis [15]. Carcinoids are rare tumors at most
gastrointestinal sites and the results would be only marginally
affected if histology was not considered. However, the
exceptionally high risks of stomach carcinoids after pernicious
anemia (46.67) and primary biliary cirrhoses (78.29) would have
influenced stomach cancer risks; after pernicious anemia all
stomach cancer SIRs would have been 4.53, instead of 4.09 for
adenocarcinoma and after primary biliary cirrhoses, 2.35 for all
stomach cancer and 1.29 for adenocarcinoma. The small
intestine was an exceptional site because there carcinoids
outnumbered adenocarcinomas. Pernicious anemia was strongly
associated with small intestinal adenocarcinoma only (7.64).
Previous Swedish studies have also noted the high risk of
carcinoid tumors in patients with pernicious anemia [16, 17].
The above example on carcinoids shows the importance of
considering histology. Another example is provided by
esophageal tumors. Albeit the case numbers were small, none of
the six autoimmune diseases appeared to be associated with
both histological types of esophageal cancer. None of these six
increases were found in the USA study that did not consider
histology [6]. Similarly, rectal and anal cancers differ in
histology. Again, none of the seven autoimmune diseases with
an increased risk of cancer at these sites showed a concordant
increase. For the four autoimmune diseases that associated with
rectal adenocarcinoma in our study, only ulcerative colitis was
also associated in the USA study that considered rectum and
anus together [6]; the USA study found an association with
sarcoidosis, which in our study associated with anal cancer
only.
Table 2. Risk of cancers of the small intestine, colon, rectum and anus after a specified autoimmune disease
Autoimmune disease Small intestinal adenocarcinoma Colon adenocarcinoma Rectal adenocarcinoma Anus SCC
Pyrs O SIR 95% CI O SIR 95% CI O SIR 95% CI O SIR 95% CI
Digestive tract involvement
Ankylosing spondylitis 92 881 0 17 0.55 0.32–0.88 7 0.35 0.14–0.73 2 2.01 0.19–7.41
Celiac disease 67 572 1 10.70 0.00–61.31 9 2.05 0.93–3.90 1 0.42 0.00–2.39 0
Crohn disease 488 461 32 11.05 7.55–15.62 340 2.76 2.48–3.07 132 1.83 1.53–2.17 8 1.66 0.71–3.29
Immune
thrombocytopenic
purpura
28 310 0 13 2.61 1.39–4.48 3 1.11 0.21–3.28 0
Localized scleroderma 63 257 0 29 1.35 0.90–1.94 9 0.83 0.38–1.58 2 2.11 0.20–7.77
Pernicious anemia 68 625 9 7.64 3.46–14.56 72 1.16 0.91–1.46 35 1.04 0.72–1.44 3 1.86 0.35–5.50
Polyarteritis nodosa 94 093 2 1.15 0.11–4.23 106 1.13 0.92–1.36 55 1.15 0.87–1.50 4 1.50 0.39–3.89
Primary biliary cirrhosis 7110 0 8 1.64 0.70–3.24 3 1.21 0.23–3.59 1 5.42 0.00–31.06
Sarcoidosis 136 747 2 1.65 0.16–6.05 73 1.30 1.02–1.63 34 1.09 0.75–1.52 7 3.51 1.39–7.27
Sjo¨gren syndrome 46 309 0 29 1.21 0.81–1.73 11 0.90 0.45–1.62 0
Systemic lupus
erythematosus
62 007 0 38 1.59 1.13–2.19 10 0.82 0.39–1.51 7 7.18 2.85–14.88
Systemic sclerosis 8199 0 7 1.55 0.61–3.20 0 0
Ulcerative colitis 15 883 4 4.10 1.07–10.61 133 3.13 2.62–3.71 49 1.94 1.43–2.56 4 2.72 0.71–7.03
No digestive tract involvement
Addison disease 21 314 0 5 0.57 0.18–1.34 7 1.49 0.59–3.08 0
Amyotrophic lateral
sclerosis
19 135 1 5.84 0.00–33.46 13 1.50 0.80–2.58 4 0.79 0.21–2.05 0
Behcet disease 51 636 0 14 0.92 0.50–1.55 10 1.09 0.52–2.01 1 2.07 0.00–11.88
Chronic rheumatic heart
disease
133 597 3 1.54 0.29–4.56 126 1.26 1.05–1.50 71 1.29 1.01–1.63 5 1.83 0.58–4.31
Diabetes mellitus type I 323 086 0 4 0.96 0.25–2.49 5 2.06 0.65–4.84 0
Grave/hyperthyroidism 468 161 5 1.08 0.34–2.54 248 1.06 0.93–1.20 130 1.12 0.94–1.33 12 1.33 0.68–2.32
Hashimoto/hypothroidism 100 612 2 1.77 0.17–6.51 63 1.07 0.82–1.37 15 0.52 0.29–0.87 1 0.48 0.00–2.76
Multiple sclerosis 157 915 2 1.62 0.15–5.96 62 1.09 0.84–1.40 23 0.73 0.46–1.10 2 0.88 0.08–3.23
Myasthenia gravis 286 399 1 0.28 0.00–1.61 233 1.35 1.18–1.53 159 1.49 1.26–1.74 5 1.25 0.39–2.94
Polymyalgia rheumatica 172 528 4 1.69 0.44–4.36 114 0.94 0.78–1.13 55 0.86 0.65–1.12 3 0.79 0.15–2.34
Polymyositis/
dermatomyositis
11 188 1 9.02 0.00–51.70 10 1.83 0.87–3.38 4 1.34 0.35–3.47 0
Psoriasis 222 027 1 0.49 0.00–2.81 112 2.19 0.98–1.44 65 1.23 0.95–1.56 10 3.18 1.52–5.88
Rheumatic fever 65 040 1 1.88 0.00–10.77 28 1.19 0.79–1.73 10 0.69 0.33–1.27 0
Rheumatoid arthritis 235 491 4 1.33 0.34–3.43 109 0.70 0.58–0.85 64 0.80 0.62–1.02 3 0.59 0.11–1.76
Wegener granulomatosis 9398 0 7 1.35 0.54–2.80 3 1.06 0.20–3.14 0
All 3 606 353 75 2.33 1.83–2.92 2026 1.30 1.24–1.35 979 1.15 1.08–1.22 80 1.55 1.23–1.93
CI, confidence interval; O, observed number of cases; SIR, standardized incidence ratio; Pyrs, person years.
original article Annals of Oncology
4 | Hemminki et al.
at Gesellschaft f?r Biotechnologische Forschung on August 17, 2011 annonc.oxfordjournals.org Downloaded from Some autoimmune diseases were associated with many
cancers, excluding carcinoid tumors. Myasthenia gravis
associated with five cancers with SIRs ranging from 1.35 to
2.78 (upper digestive tract, esophageal SCC, stomach
adenocarcinoma and colon and rectal adenocarcinomas);
pernicious anemia associated with four cancers with SIRs
ranging from 2.17 to 7.64 (upper digestive tract, esophageal
SCC, stomach adenocarcinoma and small intestinal
adenocarcinoma); Crohn disease associated with four sites and
ulcerative colitis, systemic lupus erythematosis and psoriasis
associated each with three sites. A common disease, rheumatoid
arthritis associated with no cancer. However, rheumatoid
arthritis is an old age disease with a relatively short follow-up
time; consequently, the numbers of person-years at risk for
rheumatoid arthritis were only less than half compared with
Crohn and Graves diseases.
The finding on pernicious anemia being associated with
upper gastrointestinal cancers ranging from oral cavity to small
intestine has been reported by others [6, 16, 17]. Pernicious
anemia is caused by a gradual loss of gastric parietal cells
through autoimmune destruction and subsequent inability to
absorb vitamin B12. The increased risk of stomach cancer has
been ascribed to a chronic inflammation in the course of an
atrophic gastritis but the effects appear to be widespread as the
present risks were higher for small intestinal adenocarcinoma
and esophageal SCC than for gastric adenocarcinoma.
Increased risks of cancer have been reported also at sites outside
the gastrointestinal tract [6, 16].
The multiple cancers diagnosed after myasthenia gravis are
novel findings based on a cohort study. However, it is well
known that myasthenia gravis patients have frequently benign
and malignant thymomas [18, 19]. There are many reports on
Table 3. Risk of carcinoid tumors in the stomach, small intestine and colorectum after a specified autoimmune disease
Autoimmune disease Stomach Small intestine Colorectum
Pyrs O SIR 95% CI O SIR 95% CI O SIR 95% CI
Digestive tract involvement
Ankylosing spondylitis 92 881 0 1 0.82 0.00–4.72 2 2.35 0.22–8.63
Celiac disease 67 572 0 0 0
Crohn disease 488 461 2 2.46 0.23–9.04 33 7.33 5.04–10.31 39 8.18 5.82–11.19
Immune
thrombocytopenic
purpura
28 310 0 0 0
Localized scleroderma 63 257 0 3 4.32 0.81–12.79 1 1.38 0.00–7.91
Pernicious anemia 68 625 12 46.67 24.00–81.78 4 2.16 0.56–5.59 0
Polyarteritis nodosa 94 093 0 3 1.15 0.22–3.40 1 0.73 0.00–4.21
Primary biliary cirrhosis 7110 2 78.29 7.38–287.91 0 0
Sarcoidosis 136 747 1 3.11 0.00–17.80 7 3.67 1.46–7.61 3 2.11 0.40–6.25
Sjo¨gren syndrome 46 309 1 8.30 0.00–47.57 0 1 1.81 0.00–10.36
Systemic lupus
erythematosus
62 007 0 1 1.35 0.00–7.73 0
Systemic sclerosis 8199 0 1 7.23 0.00–41.44 0
Ulcerative colitis 15 883 1 3.80 0.00–21.80 5 3.31 1.04–7.79 11 7.32 3.63–13.14
No digestive tract involvement
Addison disease 21 314 0 0 1 4.38 0.00–25.10
Amyotrophic lateral
sclerosis
19 135 0 1 3.53 0.00–20.24 0
Behcet disease 51 636 0 2 3.51 0.33–12.90 0
Chronic rheumatic heart
disease
133 597 1 2.27 0.00–13.01 3 0.97 0.18–2.87 0
Diabetes mellitus type I 323 086 3 39.21 7.39–116.06 0 3 1.25 0.24–3.70
Grave/hyperthyroidism 468 161 2 1.65 0.16–6.05 10 1.44 0.69–2.66 5 0.85 0.27–2.00
Hashimoto/hypothroidism 100 612 2 6.93 0.65–25.48 2 1.20 0.11–4.40 3 2.39 0.45–7.09
Multiple sclerosis 157 915 0 2 1.03 0.10–3.78 2 1.15 0.11–4.22
Myasthenia gravis 286 399 3 3.68 0.69–10.90 4 0.68 0.18–1.75 3 0.96 0.18–2.83
Polymyalgia rheumatica 172 528 2 3.43 0.32–12.60 8 2.23 0.95–4.41 6 2.69 0.97–5.90
Polymyositis/
dermatomyositis
11 188 0 1 5.75 0.00–32.97 0
Psoriasis 222 027 1 1.90 0.00–10.88 6 1.88 0.68–4.11 2 0.83 0.08–3.04
Rheumatic fever 65 040 0 1 1.16 0.00–6.64 0
Rheumatoid arthritis 235 491 4 5.38 1.40–13.91 7 1.52 0.60–3.15 4 1.33 0.35–3.43
Wegener granulomatosis 9398 0 0 0
All 3 606 353 37 4.53 3.19–6.26 105 2.10 1.72–2.54 87 2.22 1.78–2.74
CI, confidence interval; O, observed number of cases; SIR, standardized incidence ratio; Pyrs, person years.
Annals of Oncology
original article
doi:10.1093/annonc/mdr333 | 5
at Gesellschaft f?r Biotechnologische Forschung on August 17, 2011 annonc.oxfordjournals.org Downloaded from patient series on extrathymic tumors in myasthenia gravis
patients with or without thymoma. In the latter patient group,
an excess of squamous cell skin cancer has been reported [19].
Myasthenia gravis is diagnosed earlier than many other
autoimmune diseases, the mean ages have been 34.9 years for
women and 48.5 years for men in Sweden [18]. In Sweden, as
elsewhere, a large proportion of patients receive an
immunosuppressive therapy with either azathioprine or
ciclosporin, two known human carcinogens [18, 19]. These
drugs are used in some other autoimmune diseases but rarely
for equally long periods. To test for the possible effect of
medication, a case-only study compared myasthenia gravis
patients with and without extrathymic malignancies [20]. The
therapies by these drugs showed a borderline association with
extrathymic neoplasms in univariate analysis but no longer in
multivariate analysis. Whether the present associations at five
sites are related to autoimmunity or medication remain to be
established.
To our knowledge, only a Danish study has previously
analyzed anal cancers in autoimmune diseases patients, with
somewhat lower case numbers than in the present study [21].
Both studies found a significant and equal risk (3.1) for
psoriasis but the present associations with sarcoidosis and
systemic lupus erythematosus were not significant in the
Danish study, albeit over 1.00. Addison disease was associated
with an increased risk of esophageal SCC and stomach
adenocarcinoma. The risks were above unity in the USA study
but not significant [6].
In studies involving multiple comparisons, chance findings
are an issue. It is, however, noteworthy that only five
significant decreases were observed and three of them are
probably true findings. The decreased risks of colon and rectal
adenocarcinomas in ankylosing spondylitis and rheumatoid
arthritis may be related to the prolonged use of
anti-inflammatory medication. According to a recent joint
analysis of randomized trials on a daily use of aspirin, a >5
year use reduced colorectal cancer mortality to about one half
[22]. An equal reduction of esophageal cancer mortality was
observed but no reduction was apparent in the present study.
As anti-inflammatory medication is used in many
autoimmune diseases, and some effects, at least for colorectal
cancer, may have been masked by the beneficial effects of the
treatment.
In conclusion, increased risks of digestive tract
adenocarcinoma and SCC were observed in patients with
hospitalization for some autoimmune diseases. Myasthenia
gravis and pernicious anemia were associated with many
cancers, in myasthenia gravis probably related to
immunosuppressant medication. The decreased risks in colon
and rectal adenocarcinomas in ankylosing spondylitis and
rheumatoid arthritis suggest inflammatory mechanisms to the
observed risks as these were alleviated by the assumed use of
anti-inflammatory medication.
funding
Supported by the Swedish Council for Working Life and Social
Research, the Swedish Cancer Society and Deutsche Krebshilfe.
The funding sources had no influence on the study
disclosure
The authors declare no conflicts of interest.
references
1. Vajdic CM, van Leeuwen MT. Cancer incidence and risk factors after solid organ
transplantation. Int J Cancer 2009; 125: 1747–1754.
2. Forabosco P, Bouzigon E, Ng MY et al. Meta-analysis of genome-wide linkage
studies across autoimmune diseases. Eur J Hum Genet 2009; 17: 236–243.
3. Zhernakova A, van Diemen CC, Wijmenga C. Detecting shared pathogenesis
from the shared genetics of immune-related diseases. Nat Rev Genet 2009; 10:
43–55.
4. Goldin LR, Landgren O. Autoimmunity and lymphomagenesis. Int J Cancer 2009;
124: 1497–1502.
5. Brown LM, Gridley G, Check D, Landgren O. Risk of multiple myeloma and
monoclonal gammopathy of undetermined significance among white and black
male United States veterans with prior autoimmune, infectious, inflammatory,
and allergic disorders. Blood 2008; 111: 3388–3394.
6. Landgren AM, Landgren O, Gridley G et al. Autoimmune disease and subsequent
risk of developing alimentary tract cancers among 4.5 million US male veterans.
Cancer 2011; 117: 1163–1171.
7. Hemminki K, Li X, Sundquist K, Sundquist J. Familial risks for chronic obstructive
pulmonary disease among siblings based on hospitalizations in Sweden.
J Epidemiol Community Health 2008; 62: 398–401.
8. Hemminki K, Li X, Sundquist K, Sundquist J. Familial risks for asthma among
twins and other siblings based on hospitalizations in Sweden. Clin Exp Allergy
2007; 37: 1320–1325.
9. Hemminki K, Li X, Sundquist J, Sundquist K. Familial associations of rheumatoid
arthritis with autoimmune disorders and related conditions. Arthritis Rheum
2009; 60: 661–668.
10. Hemminki K, Li X, Sundquist J, Sundquist K. Familial association between type 1
diabetes and other autoimmune and related diseases. Diabetologia 2009; 52:
1820–1828.
11. Hemminki K, Li X, Sundquist J, Sundquist K. The epidemiology of Graves’
disease: evidence of a genetic and an environmental contribution. J Autoimmun
2010; 34: J307–J313.
12. Hemminki K, Li X, Sundquist K, Sundquist J. Familial association of inflammatory
bowel diseases with other autoimmune and related diseases. Am J Gastroenterol
2010; 105: 139–147.
13. Hemminki K, Li X. Incidence trends and risk factors of carcinoid tumors. Cancer
2001; 92: 2204–2210.
14. Hemminki K, Li X. Familial carcinoid tumors and subsequent cancers: a nationwide epidemiological study from Sweden. Int J Cancer 2001; 94: 444–448.
15. Welch HG, Black WC. Overdiagnosis in Cancer. J Natl Cancer Inst 2010; 102:
605–613.
16. Hsing AW, Hansson LE, McLaughlin JK et al. Pernicious anemia and subsequent
cancer. A population-based cohort study. Cancer 1993; 71: 745–750.
17. Ye W, Nyren O. Risk of cancers of the oesophagus and stomach by histology or
subsite in patients hospitalised for pernicious anaemia. Gut 2003; 52:
938–941.
18. Kalb B, Matell G, Pirskanen R, Lambe M. Epidemiology of myasthenia gravis:
a population-based study in Stockholm, Sweden. Neuroepidemiology 2002; 21:
221–225.
19. Tsinzerling N, Lefvert AK, Matell G, Pirskanen-Matell R. Myasthenia gravis:
a long term follow-up study of Swedish patients with specific reference to thymic
histology. J Neurol Neurosurg Psychiatry 2007; 78: 1109–1112.
20. Citterio A, Beghi E, Millul A et al. Risk factors for tumor occurrence in patients
with myasthenia gravis. J Neurol 2009; 256: 1221–1227.
21. Sunesen KG, Norgaard M, Thorlacius-Ussing O, Laurberg S. Immunosuppressive
disorders and risk of anal squamous cell carcinoma: a nationwide cohort study in
Denmark, 1978-2005. Int J Cancer 2010; 127: 675–684.
22. Rothwell PM, Fowkes FG, Belch JF et al. Effect of daily aspirin on long-term risk
of death due to cancer: analysis of individual patient data from randomised trials.
Lancet 2011; 377: 31–41.
original article Annals of Oncology
6 | Hemminki et al.
at Gesellschaft f?r Biotechnologische Forschung on August 17, 2011 annonc.oxfordjournals.org Downloaded from