Accepted Manuscript Cancer Risk Following Pernicious Anemia in the US Elderly Population Gwen Murphy, Sanford M. Dawsey, Eric A. Engels, Winnie Ricker, Ruth Parsons, Arash Etemadi, Shih-Wen Lin, Christian C. Abnet, Neal D. Freedman
PII: DOI: Reference:
S1542-3565(15)00790-9 10.1016/j.cgh.2015.05.040 YJCGH 54326
To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 26 May 2015 Please cite this article as: Murphy G, Dawsey SM, Engels EA, Ricker W, Parsons R, Etemadi A, Lin SW, Abnet CC, Freedman ND, Cancer Risk Following Pernicious Anemia in the US Elderly Population, Clinical Gastroenterology and Hepatology (2015), doi: 10.1016/j.cgh.2015.05.040. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. All studies published in Clinical Gastroenterology and Hepatology are embargoed until 3PM ET of the day they are published as corrected proofs on-line. Studies cannot be publicized as accepted manuscripts or uncorrected proofs.
ACCEPTED MANUSCRIPT Pernicious anemia and cancer 1
Cancer Risk Following Pernicious Anemia in the US Elderly Population
2 3
Gwen Murphy1, Sanford M. Dawsey1, Eric A. Engels1, Winnie Ricker2, Ruth Parsons2, Arash Etemadi1,
4
Shih-Wen Lin1, Christian C. Abnet1, Neal D. Freedman1
RI PT
5 6
1
7
Health, Bethesda, MD; 2Information Management Services, Inc., Calverton, MD
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
8
Correspondence to: Gwen Murphy, Ph.D., M.P.H., Nutritional Epidemiology Branch, DCEG,
SC
9
National Cancer Institute, 9609 Medical Center Drive, 6E314, Bethesda, MD 20892
11
Phone: 240-276-7199; [email protected]
M AN U
10
12 13 14 15
Funding: This work was supported by the Intramural Research Program of the National Cancer
17
Institute at the US National Institutes of Health.
TE D
16
18
Acknowledgements: The authors acknowledge the efforts of Brenda Edwards and Lynn Ries of the
20
Surveillance Research Program, National Cancer Institute; the Office of Research, Development, and
21
Information, Centers for Medicare and Medicaid Services; Information Management Services, Inc.; and
22
the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the
23
SEER-Medicare database.
24
Disclosure of potential conflicts of interest: All authors have none to declare.
25
Abbreviations: CI: confidence interval; OR: odds ratio
AC C
26 27 28 29
EP
19
Wordcount: 2820
1
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Abstract (264 words) Background & Aims: Pernicious anemia, a result of autoimmune gastritis, is the most common cause of vitamin B12 deficiency, affecting 2%–5% of the elderly population. Treatment with vitamin B12 cures the anemia, but not the gastritis. Findings from small studies indicated that patients with pernicious anemia could have an increased risk of cancer.
SC
RI PT
Methods: We performed a population-based, case–control study of individuals the SEER-Medicare database, comparing 1,138,390 cancer cases (66–99 y old) to 100,000 matched individuals without cancer (controls). Individuals with pernicious anemia were identified based on their medical claims within the year before selection for the study. Odds ratios (OR) and 95% confidence intervals (CI) were calculated using unconditional logistic regression, and models were adjusted for sex, age, and calendar year of diagnosis and selection.
M AN U
Results: Compared with controls, we found individuals with pernicious anemia to be at increased risk for non-cardia gastric adenocarcinoma (OR, 2.18; 95% CI, 1.94–2.45) and gastric carcinoid tumors (OR, 11.43; 95% CI, 8.90–14.69). In addition, people with pernicious anemia have an increased risk of developing tonsilar cancer (OR, 2.00; 95% CI, 1.40–2.85), hypopharyngeal cancer (OR, 1.92; 95% CI, 1.35–2.73), esophageal squamous cell carcinoma (OR, 2.12; 95% CI, 1.76–2.55), small intestinal cancer (OR, 1.63; 95% CI, 1.32–2.02), liver cancer (OR, 1.49; 95% CI, 1.28– 1.73), myeloma (OR, 1.55; 95% CI, 1.37–1.75), acute myeloid leukemia (OR, 1.68; 95% CI, 1.46–1.93), and myelodysplastic syndrome (OR, 2.87; 95% CI, 2.53–3.26). People with pernicious anemia have a lower risk of rectal cancer than the general population (OR, 0.82; 95% CI, 0.74– 0.92).
TE D
Conclusion: In a population-based, case–control study of individuals the SEER-Medicare database, we found individuals with pernicious anemia to have significantly increased risks of gastric carcinoid tumors, adenocarcinomas, and other cancers located throughout the body.
EP
KEY WORDS: chronic atrophic autoimmune gastritis, acid secretion, parietal cells, stomach cancer
AC C
30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
2
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Author’s Contributions Gwen Murphy: concept, data review, data interpretation, wrote the manuscript
63
Sanford M. Dawsey: data interpretation, review
64
Eric A. Engels: study design, review
65
Winnie Ricker: data analysis
66
Ruth Parsons: data analysis
67
Arash Etemadi: data interpretation, review
68
Shih-Wen Lin: data interpretation, review
69
Christian C. Abnet: concept, data interpretation, review
70
Neal D. Freedman: concept, study design, data interpretation, review
AC C
EP
TE D
SC
M AN U
71
RI PT
60 61 62
3
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Introduction
74
The most common cause of vitamin B12 deficiency is pernicious anemia, a slowly worsening disease
75
that is estimated to affect two to five percent of the elderly (60 years and older) in the UK, US, and
76
other countries1, 2. Pernicious anemia is the final stage of a chronic atrophic autoimmune gastritis
77
where autoantibodies target and destroy the acid producing parietal cells in the stomach3. Gastric
78
intrinsic factor which is necessary for vitamin B12 absorption, is produced in parietal cells and over
79
many years their destruction typically leads to vitamin B12 deficiency3. Treatment with regular
80
injections of vitamin B12 corrects the anemia and has transformed the prognosis of pernicious
81
anemia4, however vitamin B12 treatment does not treat the autoimmune gastritis or restore acid
82
secretion in the stomach.
83
Individuals with pernicious anemia have long been suggested to have excess risk of gastric cancer,
84
and in particular gastric carcinoid tumors in a number of studies4-9. Such an association is plausible,
85
as a less acidic stomach (hypochlorhydria) as well as chronic inflammation are among the
86
mechanisms by which H.pylori is thought to cause gastric cancer10. In addition to gastric cancer,
87
pernicious anemia may also be associated with an increased risk of other cancers8, 9, 11-16. However,
88
previous studies have lacked the large number of cancer cases necessary to precisely estimate the
89
magnitude of associations, to examine associations within the stomach by anatomic site or histology,
90
or to examine associations with less common cancer types.
91
We, therefore, evaluated the association between pernicious anemia and subsequent development of
92
cancer using the very large SEER-Medicare database, which links data on malignancies ascertained
93
through the SEER cancer registries to claims from Medicare, the U.S. government insurance program
94
for people over age 65 years.
SC
M AN U
TE D
EP
AC C
95
RI PT
72 73
4
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Methods
97
SEER-Medicare data sources
98
SEER is a National Cancer Institute (NCI) funded program collecting data on cancer incidence and
99
survival from 17 U.S. cancer registries (http://www.seer.cancer.gov), covering approximately 26% of
RI PT
96
the U.S. population. Medicare provides federally funded health insurance for approximately 97% of
101
persons aged ≥ 65 years in the U.S., as well as for individuals under age 65 years who have a medical
102
disability. Hospital inpatient care is included in Part A coverage, which all beneficiaries are entitled
103
to. Approximately 96% of participants pay to subscribe to Part B coverage, which covers physician
104
and outpatient services.
105
The SEER-Medicare database comprises files created during electronic linkage of SEER and
106
Medicare data (http://healthservices.cancer.gov/seermedicare/). The linkage was performed using a
107
deterministic algorithm based on name, social security number, sex, and date of birth. The match
108
links 94% of SEER cancer cases (which, in this paper, includes all tumors reported in the
109
Surveillance, Epidemiology, and End Results (SEER) cancer registries) aged ≥66 years to Medicare
110
claims data. Cases were required to ≥12 months of Medicare data prior to diagnosis (1 year of prior
111
Part A, Part B Medicare coverage while not enrolled in a health maintenance organization [HMO]) so
112
that there was sufficient time in which to identify pernicious anemia information prior to the cancer
113
diagnosis, meaning that individuals included in our analysis were 66 years and older. We included
114
only first reported cancers. In addition, a 5% random sample of Medicare beneficiaries living in
115
SEER regions is also included.
M AN U
TE D
EP
116
SC
100
Nested case-control design within SEER-Medicare
118
For this analysis, cases were defined as those people aged 66 years and older with a first cancer
119
identified in SEER from 1992 through 2005 (national claims history (NCH) and outpatient (OUTPT)
120
claims data is not available prior to 1992). Cases were also required to have ≥13 months of Part A or
121
Part B, non-health maintenance organization (because health maintenance organizations do not
122
submit claims for individual medical conditions/treatments) Medicare coverage preceding cancer
123
diagnosis (assessed using Medicare claims data). Those cases diagnosed only at autopsy or by death
124
certificate were excluded. Case definitions, by site, are list in Supplementary Table 1.
AC C
117
5
ACCEPTED MANUSCRIPT Pernicious anemia and cancer 100,000 controls were selected from the 5% random sample of Medicare beneficiaries living in
126
SEER regions. These controls were alive and cancer free as of July 1st in the calendar year of
127
selection, with 13 months of Part A or Part B, non-health maintenance organization Medicare
128
coverage. Controls were frequency-matched to cases by age category, sex and calendar year of
129
selection. For the selected cases and controls, we then used linked Medicare claims prior to cancer
130
diagnosis/control selection to identify individuals diagnosed with pernicious anemia (ICD-9 code
131
281.0 only) according to at least 2 billable physician or outpatient records, at least 30 days apart, or a
132
single billed hospitalization and pernicious anemia must have been reported >12 months prior to
133
diagnosis/selection.
SC
RI PT
125
134 Statistical Analysis
136
All statistical analyses were carried out using SAS software (SAS Institute, Inc., Cary, North
137
Carolina) and all statistical tests were two-sided. The prevalence of the pernicious anemia in cases
138
and controls was compared using contingency tables and unconditional logistic regression, to allow
139
greater precision in calculating risk estimates. Odds ratios (OR) and 95% confidence intervals were
140
obtained after adjustment for the matching factors: sex, age and calendar year of diagnosis/selection.
141
Where associations between pernicious anemia and cancer were found to pass Bonferroni correction
142
(assuming 60 comparisons, to include all cancer sites: P72 months; P calculated using proc logistic for interaction term: pernicious
145
anemia*latency). The SEER-Medicare data use agreement prevents presenting results for any cell
146
with 10 cases or fewer to preserve confidentiality, so these have been excluded from this analysis.
TE D
EP
AC C
147
M AN U
135
6
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Results
149
The characteristics of cases and controls are presented in Table 1. Cases and controls were matched
150
on sex, age, and calendar year of case diagnosis/control selection.
151
Pernicious anemia was reported in 1.5% of cancer cases and controls and 66% of those identified
152
with pernicious anemia (both cases and controls) had a record of B12 treatment (data not shown). A
153
slightly elevated risk of all cancer was observed for subjects with pernicious anemia, relative to
154
controls (OR: 1.07; 95% CI: 1.01, 1.14; Table 2). Pernicious anemia was significantly associated
155
(passing a Bonferroni correction of P
PII: DOI: Reference:
S1542-3565(15)00790-9 10.1016/j.cgh.2015.05.040 YJCGH 54326
To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 26 May 2015 Please cite this article as: Murphy G, Dawsey SM, Engels EA, Ricker W, Parsons R, Etemadi A, Lin SW, Abnet CC, Freedman ND, Cancer Risk Following Pernicious Anemia in the US Elderly Population, Clinical Gastroenterology and Hepatology (2015), doi: 10.1016/j.cgh.2015.05.040. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. All studies published in Clinical Gastroenterology and Hepatology are embargoed until 3PM ET of the day they are published as corrected proofs on-line. Studies cannot be publicized as accepted manuscripts or uncorrected proofs.
ACCEPTED MANUSCRIPT Pernicious anemia and cancer 1
Cancer Risk Following Pernicious Anemia in the US Elderly Population
2 3
Gwen Murphy1, Sanford M. Dawsey1, Eric A. Engels1, Winnie Ricker2, Ruth Parsons2, Arash Etemadi1,
4
Shih-Wen Lin1, Christian C. Abnet1, Neal D. Freedman1
RI PT
5 6
1
7
Health, Bethesda, MD; 2Information Management Services, Inc., Calverton, MD
Division of Cancer Epidemiology and Genetics, National Cancer Institute, National Institutes of
8
Correspondence to: Gwen Murphy, Ph.D., M.P.H., Nutritional Epidemiology Branch, DCEG,
SC
9
National Cancer Institute, 9609 Medical Center Drive, 6E314, Bethesda, MD 20892
11
Phone: 240-276-7199; [email protected]
M AN U
10
12 13 14 15
Funding: This work was supported by the Intramural Research Program of the National Cancer
17
Institute at the US National Institutes of Health.
TE D
16
18
Acknowledgements: The authors acknowledge the efforts of Brenda Edwards and Lynn Ries of the
20
Surveillance Research Program, National Cancer Institute; the Office of Research, Development, and
21
Information, Centers for Medicare and Medicaid Services; Information Management Services, Inc.; and
22
the Surveillance, Epidemiology, and End Results (SEER) Program tumor registries in the creation of the
23
SEER-Medicare database.
24
Disclosure of potential conflicts of interest: All authors have none to declare.
25
Abbreviations: CI: confidence interval; OR: odds ratio
AC C
26 27 28 29
EP
19
Wordcount: 2820
1
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Abstract (264 words) Background & Aims: Pernicious anemia, a result of autoimmune gastritis, is the most common cause of vitamin B12 deficiency, affecting 2%–5% of the elderly population. Treatment with vitamin B12 cures the anemia, but not the gastritis. Findings from small studies indicated that patients with pernicious anemia could have an increased risk of cancer.
SC
RI PT
Methods: We performed a population-based, case–control study of individuals the SEER-Medicare database, comparing 1,138,390 cancer cases (66–99 y old) to 100,000 matched individuals without cancer (controls). Individuals with pernicious anemia were identified based on their medical claims within the year before selection for the study. Odds ratios (OR) and 95% confidence intervals (CI) were calculated using unconditional logistic regression, and models were adjusted for sex, age, and calendar year of diagnosis and selection.
M AN U
Results: Compared with controls, we found individuals with pernicious anemia to be at increased risk for non-cardia gastric adenocarcinoma (OR, 2.18; 95% CI, 1.94–2.45) and gastric carcinoid tumors (OR, 11.43; 95% CI, 8.90–14.69). In addition, people with pernicious anemia have an increased risk of developing tonsilar cancer (OR, 2.00; 95% CI, 1.40–2.85), hypopharyngeal cancer (OR, 1.92; 95% CI, 1.35–2.73), esophageal squamous cell carcinoma (OR, 2.12; 95% CI, 1.76–2.55), small intestinal cancer (OR, 1.63; 95% CI, 1.32–2.02), liver cancer (OR, 1.49; 95% CI, 1.28– 1.73), myeloma (OR, 1.55; 95% CI, 1.37–1.75), acute myeloid leukemia (OR, 1.68; 95% CI, 1.46–1.93), and myelodysplastic syndrome (OR, 2.87; 95% CI, 2.53–3.26). People with pernicious anemia have a lower risk of rectal cancer than the general population (OR, 0.82; 95% CI, 0.74– 0.92).
TE D
Conclusion: In a population-based, case–control study of individuals the SEER-Medicare database, we found individuals with pernicious anemia to have significantly increased risks of gastric carcinoid tumors, adenocarcinomas, and other cancers located throughout the body.
EP
KEY WORDS: chronic atrophic autoimmune gastritis, acid secretion, parietal cells, stomach cancer
AC C
30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
2
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Author’s Contributions Gwen Murphy: concept, data review, data interpretation, wrote the manuscript
63
Sanford M. Dawsey: data interpretation, review
64
Eric A. Engels: study design, review
65
Winnie Ricker: data analysis
66
Ruth Parsons: data analysis
67
Arash Etemadi: data interpretation, review
68
Shih-Wen Lin: data interpretation, review
69
Christian C. Abnet: concept, data interpretation, review
70
Neal D. Freedman: concept, study design, data interpretation, review
AC C
EP
TE D
SC
M AN U
71
RI PT
60 61 62
3
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Introduction
74
The most common cause of vitamin B12 deficiency is pernicious anemia, a slowly worsening disease
75
that is estimated to affect two to five percent of the elderly (60 years and older) in the UK, US, and
76
other countries1, 2. Pernicious anemia is the final stage of a chronic atrophic autoimmune gastritis
77
where autoantibodies target and destroy the acid producing parietal cells in the stomach3. Gastric
78
intrinsic factor which is necessary for vitamin B12 absorption, is produced in parietal cells and over
79
many years their destruction typically leads to vitamin B12 deficiency3. Treatment with regular
80
injections of vitamin B12 corrects the anemia and has transformed the prognosis of pernicious
81
anemia4, however vitamin B12 treatment does not treat the autoimmune gastritis or restore acid
82
secretion in the stomach.
83
Individuals with pernicious anemia have long been suggested to have excess risk of gastric cancer,
84
and in particular gastric carcinoid tumors in a number of studies4-9. Such an association is plausible,
85
as a less acidic stomach (hypochlorhydria) as well as chronic inflammation are among the
86
mechanisms by which H.pylori is thought to cause gastric cancer10. In addition to gastric cancer,
87
pernicious anemia may also be associated with an increased risk of other cancers8, 9, 11-16. However,
88
previous studies have lacked the large number of cancer cases necessary to precisely estimate the
89
magnitude of associations, to examine associations within the stomach by anatomic site or histology,
90
or to examine associations with less common cancer types.
91
We, therefore, evaluated the association between pernicious anemia and subsequent development of
92
cancer using the very large SEER-Medicare database, which links data on malignancies ascertained
93
through the SEER cancer registries to claims from Medicare, the U.S. government insurance program
94
for people over age 65 years.
SC
M AN U
TE D
EP
AC C
95
RI PT
72 73
4
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Methods
97
SEER-Medicare data sources
98
SEER is a National Cancer Institute (NCI) funded program collecting data on cancer incidence and
99
survival from 17 U.S. cancer registries (http://www.seer.cancer.gov), covering approximately 26% of
RI PT
96
the U.S. population. Medicare provides federally funded health insurance for approximately 97% of
101
persons aged ≥ 65 years in the U.S., as well as for individuals under age 65 years who have a medical
102
disability. Hospital inpatient care is included in Part A coverage, which all beneficiaries are entitled
103
to. Approximately 96% of participants pay to subscribe to Part B coverage, which covers physician
104
and outpatient services.
105
The SEER-Medicare database comprises files created during electronic linkage of SEER and
106
Medicare data (http://healthservices.cancer.gov/seermedicare/). The linkage was performed using a
107
deterministic algorithm based on name, social security number, sex, and date of birth. The match
108
links 94% of SEER cancer cases (which, in this paper, includes all tumors reported in the
109
Surveillance, Epidemiology, and End Results (SEER) cancer registries) aged ≥66 years to Medicare
110
claims data. Cases were required to ≥12 months of Medicare data prior to diagnosis (1 year of prior
111
Part A, Part B Medicare coverage while not enrolled in a health maintenance organization [HMO]) so
112
that there was sufficient time in which to identify pernicious anemia information prior to the cancer
113
diagnosis, meaning that individuals included in our analysis were 66 years and older. We included
114
only first reported cancers. In addition, a 5% random sample of Medicare beneficiaries living in
115
SEER regions is also included.
M AN U
TE D
EP
116
SC
100
Nested case-control design within SEER-Medicare
118
For this analysis, cases were defined as those people aged 66 years and older with a first cancer
119
identified in SEER from 1992 through 2005 (national claims history (NCH) and outpatient (OUTPT)
120
claims data is not available prior to 1992). Cases were also required to have ≥13 months of Part A or
121
Part B, non-health maintenance organization (because health maintenance organizations do not
122
submit claims for individual medical conditions/treatments) Medicare coverage preceding cancer
123
diagnosis (assessed using Medicare claims data). Those cases diagnosed only at autopsy or by death
124
certificate were excluded. Case definitions, by site, are list in Supplementary Table 1.
AC C
117
5
ACCEPTED MANUSCRIPT Pernicious anemia and cancer 100,000 controls were selected from the 5% random sample of Medicare beneficiaries living in
126
SEER regions. These controls were alive and cancer free as of July 1st in the calendar year of
127
selection, with 13 months of Part A or Part B, non-health maintenance organization Medicare
128
coverage. Controls were frequency-matched to cases by age category, sex and calendar year of
129
selection. For the selected cases and controls, we then used linked Medicare claims prior to cancer
130
diagnosis/control selection to identify individuals diagnosed with pernicious anemia (ICD-9 code
131
281.0 only) according to at least 2 billable physician or outpatient records, at least 30 days apart, or a
132
single billed hospitalization and pernicious anemia must have been reported >12 months prior to
133
diagnosis/selection.
SC
RI PT
125
134 Statistical Analysis
136
All statistical analyses were carried out using SAS software (SAS Institute, Inc., Cary, North
137
Carolina) and all statistical tests were two-sided. The prevalence of the pernicious anemia in cases
138
and controls was compared using contingency tables and unconditional logistic regression, to allow
139
greater precision in calculating risk estimates. Odds ratios (OR) and 95% confidence intervals were
140
obtained after adjustment for the matching factors: sex, age and calendar year of diagnosis/selection.
141
Where associations between pernicious anemia and cancer were found to pass Bonferroni correction
142
(assuming 60 comparisons, to include all cancer sites: P72 months; P calculated using proc logistic for interaction term: pernicious
145
anemia*latency). The SEER-Medicare data use agreement prevents presenting results for any cell
146
with 10 cases or fewer to preserve confidentiality, so these have been excluded from this analysis.
TE D
EP
AC C
147
M AN U
135
6
ACCEPTED MANUSCRIPT Pernicious anemia and cancer
Results
149
The characteristics of cases and controls are presented in Table 1. Cases and controls were matched
150
on sex, age, and calendar year of case diagnosis/control selection.
151
Pernicious anemia was reported in 1.5% of cancer cases and controls and 66% of those identified
152
with pernicious anemia (both cases and controls) had a record of B12 treatment (data not shown). A
153
slightly elevated risk of all cancer was observed for subjects with pernicious anemia, relative to
154
controls (OR: 1.07; 95% CI: 1.01, 1.14; Table 2). Pernicious anemia was significantly associated
155
(passing a Bonferroni correction of P
