The incidences of inflammatory bowel disease (IBD) 1 3

Transcription

1 ORIGINAL ARTICLE Survival in Danish Patients with Breast Cancer and Inflammatory Bowel Disease: A Nationwide Cohort Study Kirstine Kobberøe Søgaard, BA,* Deirdre P. Cronin-Fenton, PhD,* Lars Pedersen, MSc,* Henrik T. Sørensen, DMSc,* and Timothy L. Lash, DSc* Background: Incidences of inflammatory bowel disease (IBD) and of breast cancer have increased over the last decades. The influence of IBD on breast cancer prognosis, however, is unknown. We therefore examined the impact of IBD on treatment receipt and survival in breast cancer patients. Methods: Information on breast cancer patients (stage and treatment) diagnosed between 1980 and 2004 was sourced from the Danish Cancer Registry. Data on IBD and potential confounders were extracted from the Danish National Registry of Patients covering all Danish hospitals. Cox regression was used to compute mortality rate ratios (MRRs) among breast cancer patients with IBD, compared to their non-ibd counterparts, adjusting for age, stage, comorbidity measured by the Charlson Index, and calendar year. Results: We identified 71,148 breast cancer cases; 67 also had Crohn s disease (CD) and 216 had ulcerative colitis (UC). Patients with CD had more advanced stage and received radiotherapy less, and chemotherapy more, frequently than patients without IBD. In the adjusted analyses there was no substantial survival difference in breast cancer patients with and without IBD (MRR CD 1.22; 95% confidence interval [CI] ; MRR UC 1.09; 95% CI ). In a stratified analysis, chemotherapy was associated with poorer survival in patients with CD (MRR CD 1.93; 95% CI ). Conclusions: Breast cancer patients with UC receive the same treatment and have similar survival to breast cancer without IBD. In contrast, breast cancer patients with CD are treated with radiotherapy less often. Survival of breast cancer in patients with CD treated Received for publication August 24, 2007; Accepted October 19, From the *Department of Clinical Epidemiology, Aarhus University Hospital, DK-8000 Aarhus C, Denmark, Department of Epidemiology, Boston University School of Public Health, Boston, Massachusetts, USA, Department of Medicine V (Hepatology and Gastroenterology), Aarhus University Hospital, 8000 Aarhus C, Denmark. Supported by the Danish Medical Research Council, Danish Agency of Science, Technology and Innovation, and the Karen Elise Jensen Foundation. There is no conflict of interest. Reprints: Kirstine Kobberøe Søgaard, Department of Clinical Epidemiology, Aarhus University Hospital, Ole Worms Allé 150, DK-8000 Aarhus C, Denmark ( kks@dce.au.dk). Copyright 2007 Crohn s & Colitis Foundation of America, Inc. DOI /ibd Published online 10 December 2007 in Wiley InterScience (www. interscience.wiley.com). with chemotherapy is poorer compared to survival in patients without IBD. (Inflamm Bowel Dis 2008;14: ) Key Words: Crohn s disease; ulcerative colitis; breast cancer; breast cancer treatment; prognosis The incidences of inflammatory bowel disease (IBD) 1 3 and of breast cancer 4,5 have increased over the last decades, although the breast cancer incidence currently seems to be declining. 6 In Denmark the incidence of these 2 diseases ranks among the highest in the world. 1,4,7 Although the etiology of IBD is largely unknown, research has suggested that use of oral contraceptives and hormone replacement therapy (HRT) are risk factors, 8,9 both of which also impact breast cancer risk. A recent case-control study found an increased risk of breast cancer in female first-degree relatives of Crohn s disease (CD) patients, 10 which may suggest an association between IBD and breast cancer 7 that could be genetic or environmental. However, other studies have reported no increased risk of breast cancer among patients with IBD In addition to any potential common etiology, the impact of IBD on breast cancer prognosis is unknown and unpredictable. Breast cancer prognosis depends on several factors, including the biological nature of the disease, i.e., tumor stage and grade (estrogen receptor status [ER / ]), diagnostic utility, treatment effectiveness, and clinical performance. The contraindications of anticancer therapy may limit the extent of such treatment in patients with coexistent IBD. Research also indicates that cancer patients with comorbid conditions are less likely to receive guideline treatment 14 and have poorer survival than those without comorbidities. 15 Furthermore, patients with IBD are treated with immunosuppressive drugs, which may have a negative impact on tumor progression. 16,17 Cancer patients with IBD have decreased tolerance to radiotherapy 18 and the use of chemotherapy in these patients may increase the risk of severe diarrhea and injuries in the intestinal mucosa. 19 IBD may therefore impact not only the underlying breast cancer prognosis, but the choice and effectiveness of breast cancer treatment. For clinicians taking care of IBD and breast cancer patients it is Inflamm Bowel Dis Volume 14, Number 4, April

2 Søgaard et al Inflamm Bowel Dis Volume 14, Number 4, April 2008 important to know whether IBD modifies the prognosis of breast cancer. Previous research is limited to 1 study with short follow-up and imprecise estimates of the standardized mortality ratio. 20 We therefore conducted a large population-based cohort study to examine the impact of IBD on breast cancer prognosis. We also aimed to determine the impact of IBD on the receipt of anticancer therapy in breast cancer patients. MATERIALS AND METHODS Study Population and Design We conducted this cohort study within the entire Danish female population (2.5 million women) in the period from January to December The Danish Civil Registration System (CRS) is a national registry of all Danish residents; since April 1968 a unique personal identifier number (CPR number) has been assigned to every Danish citizen at birth or immigration. Information on changes in vital status, such as emigration and death, are kept electronically in the CRS. 21 The CPR number can be used to link individuals in different Danish registries. The Danish National Health Service provides tax-supported healthcare for all residents, guaranteeing unrestricted and cost-free access to all general practitioners and hospitals. Data on breast cancer patients and treatment were sourced from the Danish Cancer Registry (DCR), which records incident cases of cancer on a nationwide basis since 1943 with accurate and virtually complete ascertainment. 22 The registry receives notifications from hospital departments, institutes of forensic medicine, general practitioners, and practicing specialists. Cancer diagnoses are classified according to a modified version of the 7th revision of the International Classification of Diseases (ICD-7) 23 ; cases of breast cancer are identified by code 170. Information on treatment included no treatment, radical surgery (mastectomy), nonradical surgery (breast conserving), chemotherapy, radiotherapy, and antihormonal therapy. Data are complete through Patients who did not receive any of the 5 treatments (24 non-ibd patients) were excluded. Data on breast cancer patients were linked to information on IBD occurrence and other comorbid diseases reported in the Danish National Registry of Patients (NPR). The registry, established in 1977, records 99.4% of all discharges from Danish hospitals. For each hospital admission, and since 1995, for all outpatient clinic visits the hospital discharge registry records the civil registration number of the patient, dates of admission and discharge, surgical procedure(s) performed, and up to 20 discharge diagnoses, coded by physicians on the date of discharge according to the ICD 8th revision (ICD-8) until the end of 1993 and the 10th revision (ICD-10) thereafter. 24 Of the registered discharge diagnoses, 1 is registered as primary and the others as secondary. 24 We identified 71,148 women with an incident diagnosis of breast cancer between January and June recorded in the DCR. Inflammatory Bowel Disease We searched the NPR for the years for all primary and secondary discharge diagnoses of IBD (codes , , or , in the ICD-8 and codes K50.0, K50.1, K50.8, K50.9, or K in the ICD-10), and linked data using the CPR number to the DCR. For all included patients we obtained all discharge diagnoses from inpatient hospitalizations from 1977 and inpatients and outpatients from 1995, until the patient was given a discharge diagnosis of breast cancer. We found 283 breast cancer patients with IBD, 67 with CD, and 216 with ulcerative colitis (UC). Comorbidity at Breast Cancer Diagnosis We included information on comorbidity before breast cancer diagnosis. Comorbidities were identified using the NPR and were classified according to the Deyo adaptation of the Charlson Comorbidity Index, 25,26 which was computed for each patient. This index reflects the number and severity of 19 comorbid diseases, originally validated as predictors of 1-year mortality in medical patients. 25 The index relies on the ICD categories and was used to adjust for the severity of illness. Breast cancer was excluded from the index. Three levels of comorbidity were defined: 0 ( none ), corresponding to patients with no recorded underlying diseases implemented in the Charlson Index; 1 2 ( medium ); and 2 ( high ). 27 Stage at Diagnosis Information on breast cancer stage was obtained from the Danish Cancer Registry. Stage is coded at three levels: local, regional, and distant, 28 although for 8.6% of the patients stage was reported as unknown. Mortality The outcome was time to death. We obtained information on the vital status of cohort members from the Danish Civil Registration System. All patients entered the study at the admission date of breast cancer diagnosis and were followed until death or censored at emigration or 1 January 2007, whichever came first. Statistical Analysis We calculated mortality rates using person time and number of deaths for each variable in patients with and without IBD. To detect statistically significant differences in mortality rates we calculated confidence intervals (CIs). We analyzed the treatment patterns (mastectomy, breast-conserving surgery, radiotherapy, chemotherapy, and antihormonal therapy) in breast cancer patients with and without IBD by 520

3 Inflamm Bowel Dis Volume 14, Number 4, April 2008 Survival in Danish Patients TABLE 1. Descriptive Data on Breast Cancer Patients with and without Inflammatory Bowel Disease (Mortality Rate per 1000 Person-Years [Number of Deaths/Person-Years]) Comparison Cohort N 70,841 Crohn s Disease N 67 Ulcerative Colitis N 216 Age mean SD Overall mortality rate 69 [38,723/563,995] 105 [39/370] 62 [100/1,616] Age category [9,612/253,082] 73 [10/136] 29 [21/737] [8,344/141,41] 74 [8/10] 57 [25/438] [10,059/110,034] 137 [10/73] 93 [26/281] [10,708/59,466] 212 [11/52] 174 [28/161] Stage Local 47 [15,224/325,566] 95 [16/169] 44 [41/935] Regional 83 [15,509/186,777] 90 [17/188] 84 [41/487] Distant 404 [4,009/9,927] 534 [5/9] 521 [6/12] Unknown 95 [3,981/41,725] 281 [1/4] 66 [12/183] Charlson Comorbidity Score None (0) 61 [30,815/501,189] 86 [25/292] 50 [62/1,239] Medium (1-2) 117 [6,511/55,515] 179 [14/78] 78 [24/306] High ( 2) 192 [1,397/7,291] [14/71] Calendar year [9,453/125,124] 207 [4/19] 87 [10/115] [9,516/131,185] 182 [6/33] 32 [11/346] [9,071/133,102] 70 [7/100] 67 [31/459] [7,059/112,092] 108 [14/129] 66 [29/436] [3,624/62,492] 90 [8/88] 73 [19/260] We calculated the mortality rates by dividing number of deaths by person years. creating contingency tables for the main treatment groups and used chi-square tests for homogeneity of proportions. We summarized the survival of breast cancer patients using Kaplan Meier survival curves. The proportions of patients surviving at 5 years, 10 years, and 15 years with and without IBD were calculated along with associated 95% CIs using the life table method. We also tested for homogeneity of survivor functions of CD, UC, and the comparison group. We used Cox proportional hazards regression to compute the hazard ratio (HR) as a measure of mortality rate ratio (MRR) among breast cancer patients with IBD compared to breast cancer patients without IBD. The diagnostic plot indicated that the mortality hazards were proportional over the entire time period. We calculated the MRR both for the first 5 years and for the whole follow-up period. We adjusted for age categories ( 55 years, years, years, and 76 years), stage (local, regional, distant, and unknown), comorbidity (Charlson Comorbidity Index: 0, 1 2, and 3 ), and calendar year of breast cancer diagnosis ( , , , , ). Because patients with CD were less often treated with radiotherapy and more often with chemotherapy, we stratified our analysis by radiotherapy and chemotherapy receipt. We analyzed whether results changed including and excluding a category for missing stage. As the estimates were unchanged, we included women with missing stage. We also analyzed the data without adjusting for stage. All analyses were conducted using STATA v. 9.2 (College Station, TX). Ethical Considerations Our study was not considered for publication elsewhere and all authors contributed to the study and agreed to its submission. RESULTS Table 1 shows descriptive data for patients with and without IBD. Crude mortality rates in the whole follow-up period were statistically higher in patients with CD (105 per 1000 person years [PY]; 95% CI ), but similar in patients with UC (62 per 1000 PY; 95% CI 51 75) and patients without IBD (69 per 1000 PY; 95% CI 68 69). The same relation was found within each age category for CD, whereas patients with UC had lower mortality rates. Patients with CD and local, regional, distant, or unknown stage breast cancer had higher mortality rates than their non-ibd coun- 521

4 Søgaard et al Inflamm Bowel Dis Volume 14, Number 4, April 2008 TABLE 2. Clinical and Treatment Characteristics Comparison Cohort N 70,841 Crohn s Disease N 67 Ulcerative Colitis N 216 N (%) N (%) N (%) Breast cancer stage P Local 34,151 (48%) 30 (45%) 110 (51%) Regional 26,341 (37%) 31 (46%) 75 (35%) Distant 4,294 (6.1%) 5 (7.5%) 7 (3.2%) Missing 6,055 (8.6%) 1 (1.5%) 24 (11%) Surgery P Mastectomy 46,123 (65%) 50 (75%) 147 (68%) Breast-conserving surgery 19,066 (27%) 14 (21%) 61 (28%) None 5,652 (8.0%) 3 (4.5%) 8 (3.7%) Radiotherapy P Yes 18,556 (26%) 9 (13%) 58 (27%) No 52,285 (74%) 58 (87%) 158 (73%) Chemotherapy P Yes 12,536 (18%) 19 (28%) 34 (16%) No 58,305 (82%) 48 (72%) 182 (84%) Antihormonal therapy P Yes 14,139 (20%) 15 (22%) 39 (18%) No 56,702 (80%) 52 (78%) 177 (82%) P-values of the homogeneity of proportions, calculated by chi-square tests. terparts, as had patients with UC and distant stage. However, patients with UC and unknown or local stage had lower mortality rates. Mortality rates in patients with Charlson scores of 0 and 1 2 were higher in patients with CD and lower in patients with UC compared to patients without IBD. Table 2 shows descriptive data on stage distribution and treatment in breast cancer patients with and without IBD. Although not statistically significant, patients with CD had more advanced stage than patients without IBD, whereas patients with UC more often were diagnosed with local or missing stage. While patients with UC received the same treatment as patients without IBD, patients with CD were treated with radiation therapy less frequently (P 0.018) and chemotherapy more frequently (P 0.022) than patients without IBD. Slightly more CD patients had mastectomies than non-ibd patients, consistent with the more advanced stage distribution of CD breast cancer patients. Figure 1 shows survival for patients with breast cancer and IBD compared to patients without IBD. Survival at 5 years and 10 years was lower in patients with CD compared to patients without IBD: 52% versus 67%; and 38% versus 50%, respectively. At 15 years, survival was poorer in patients with CD (27%) but improved for those with UC (45%) compared to patients without IBD (38%). The P-value for FIGURE 1. Survival in patients with breast cancer with and without inflammatory bowel disease. 522

5 Inflamm Bowel Dis Volume 14, Number 4, April 2008 Survival in Danish Patients TABLE 3. Adjusted Mortality Rate Ratios at 5 Years After Breast Cancer Diagnosis and Over All Years of Follow-up for Patients with Breast Cancer and Inflammatory Bowel Disease, Adjusted for Age Category, Stage, Comorbidity, and Calendar Year Adjusted (95% CI) 5 Years Adjusted (95% CI) All Years No IBD Crohn s disease 1.22 ( ) 1.19 ( ) Ulcerative colitis 1.09 ( ) 0.97 ( ) Age categories (reference) ( ) 1.52 ( ) ( ) 2.26 ( ) ( ) 4.29 ( ) Stage Local (reference) Regional 2.34 ( ) 1.89 ( ) Distant ( ) 7.49 ( ) Unknown 2.17 ( ) 1.68 ( ) Charlson Comorbidity Score None (0) (reference) Medium (1-2) 1.52 ( ) 1.49 ( ) High ( 2) 2.27 ( ) 2.10 ( ) Calendar year (reference) ( ) 0.91 ( ) ( ) 0.82 ( ) ( ) 0.68 ( ) ( ) 0.54 ( ) difference in survivor functions of CD, UC, and the comparison group was Table 3 shows MRRs for patients with IBD. Although the cumulative survival rates and the crude estimates indicated that survival was improved in patients with UC and poorer in patients with CD, there was no substantial survival difference after adjusting for other factors in the Cox proportional hazard regression models. Older age, more advanced stage, and higher comorbidity score were, as expected, associated with a poorer prognosis. In the adjusted analysis stratified by receipt of radiotherapy and chemotherapy, receipt of chemotherapy was associated with a poorer prognosis in patients with CD compared to patients without IBD (MRR 1.93; 95% CI ) (data not presented). Furthermore, in the analysis without adjustment for stage the MRR in patients with CD was 1.52 ( ) after 5 years and 1.47 ( ) for the whole follow-up period (data not presented). DISCUSSION Findings from this large population-based study suggest that breast cancer patients with IBD overall had a similar prognosis to breast cancer patients without IBD. Patients with UC had the same stage distribution and received the same treatment as their non-ibd counterparts. Among patients with CD there was a tendency toward more advanced stage at breast cancer diagnosis and patients with CD received radiotherapy less and chemotherapy more often than their non-ibd counterparts. CD patients treated with chemotherapy had a poorer prognosis compared to patients without IBD; this finding is possibly due to confounding by indication arising from only CD patients with poor breast cancer prognosis receiving chemotherapy. Since IBD often causes significant morbidity to a patient s health and quality of life, this morbidity may influence the planning of various cancer treatments. Several studies have indicated that patients with comorbidities are less likely to receive appropriate treatment, 14,29,30 which may be due to physician or patient preferences to refrain from the potential toxicities of cancer-directed therapy. We did note decreased radiotherapy and, nonsignificantly, decreased breast-conserving surgery and increased mastectomy in patients with CD over time. However, the change in surgery over time did not account for the whole decrease in use of radiotherapy; restricting to patients treated with breast-conserving treatment, slightly fewer patients with CD received radiotherapy compared to their non-ibd counterparts. This treatment pattern may indicate a preference for more invasive primary surgical therapy for patients with CD and a preference to minimize any subsequent anticancer therapy that may exacerbate IBD. In population-based studies, patients with CD have an increased mortality at 10 years compared to the general population, 31,32 whereas patients with UC are found to have the same overall survival as non-ibd patients, 33 or even a reduced mortality rate. 34 Crude results from the current study indicated the same tendency, but we did not see a difference in prognosis in the adjusted analysis. The tendency of patients with CD to be diagnosed at a more advanced stage may be a cause of their treatment with immunosuppressive drugs, 16,17 although a recent meta-analysis found that the risk of malignancy is not increased in IBD patients receiving immunosuppressive drugs. 35 Furthermore, treatment with chemotherapy in patients with CD was predictive of a poorer prognosis. This may be explained by serious side effects like extensive ulcerations of the intestines; the disruption of mucosal integrity together with neutropenia can result in sepsis. 19 Another possible explanation is confounding by indication; the poorer survival is a consequence of general differences in the baseline health or tumor characteristics of patients who receive chemotherapy and those who do not. In the adjusted analysis we included stage as a covariate. However, if instead we thought that IBD treatment in- 523

6 Søgaard et al Inflamm Bowel Dis Volume 14, Number 4, April 2008 fluenced the progression of breast cancer, and therefore would shift the stage distribution toward more advanced stage, then stage would not be a confounder but an intermediate, and should in this case not be adjusted. We therefore carried out a subanalysis omitting stage as an adjustment variable and found that CD was associated with a poorer prognosis after 5 years and for the whole follow-up period, in accordance with previous studies showing excess mortality in CD patients. Our survival estimates are derived from a populationbased follow-up study in a setting with a national health service that largely removed referral and diagnostic biases. Selection bias was minimized because all patients who were diagnosed with breast cancer and IBD in the study period were included in the study, and they were compared to all other patients with breast cancer in the same period. It is likely that patients with very mild CD or UC, never seen as in- or outpatients in a Danish hospital, are included in the non-ibd breast cancer group. Such misclassification would tend to bias a potential effect of IBD on the breast cancer survival toward the null. 36 Ongoing validation of DCR data and collection of supplemental information ensure that the Registry s data are of high quality. 37 The completeness and validity of the data in the DCR is estimated to be 95% 98%. 22 A validation study documented that 94% of all patients with a pathologically confirmed CD or UC diagnosis were included in the NPR and that the positive predictive value of a CD diagnosis is 97% and of UC is 90%. 38 These misclassifications of 3% and 10% might have diminished any potential effect of IBD on the prognosis of breast cancer, but the bias would be small. Among the limitations, some factors must be considered. The validity of our estimates depends on the ability to control for confounding, such as by comorbidity and stage. The Charlson Index is widely used but may not control for confounding from comorbidity as effectively as more comprehensive clinical data 39 or other strategies. 40 In addition, since IBD patients received more health care, there may be differential misclassification in the Charlson Index. Due to lack of clinical details, we used a relatively broad classification of stage according to the extent of spread of disease, and up to 10% of our patients had missing stage. Handling missing data can be complex, since both including and excluding missing data can create bias. 41 Since only 8.6% of the patients included in our study were lacking information on stage, and including them in the analysis did not change our main results (compared to an analysis restricted to patients with known stage), we chose to include them. We used all-cause mortality rather than breast cancer-specific mortality due to incomplete data registration of the specific death cause. With information on breast cancer-specific mortality, we would have been able to estimate a more exact impact of IBD on breast cancer survival. We had no information on hormone receptor (estrogen receptor, progesterone receptor expression) or other biomarkers (e.g., HER-2/neu) that influence breast cancer therapy and prognosis. However, patients who received any antihormonal therapy should have been those with estrogen receptor-positive disease. In conclusion, breast cancer patients with UC receive the same treatment as breast cancer patients without IBD and survival of breast cancer in these patients does not differ from patients without IBD. In contrast, however, breast cancer patients with CD are treated with radiotherapy less and chemotherapy more frequently than their non-ibd counterparts and survival of breast cancer in patients with CD treated with chemotherapy is poorer compared to survival in patients without IBD. REFERENCES 1. Vind I, Riis L, Jess T, et al. Increasing incidences of inflammatory bowel disease and decreasing surgery rates in Copenhagen City and County, : a population-based study from the Danish Crohn colitis database. Am J Gastroenterol. 2006;101: Fonager K, Sorensen HT, Olsen J. Change in incidence of Crohn s disease and ulcerative colitis in Denmark. A study based on the National Registry of Patients, Int J Epidemiol. 1997;26: Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease: incidence, prevalence, and environmental influences. Gastroenterology. 2004;126: Althuis MD, Dozier JM, Anderson WF, et al. Global trends in breast cancer incidence and mortality Int J Epidemiol. 2005;34: Smigal C, Jemal A, Ward E, et al. Trends in breast cancer by race and ethnicity: update CA Cancer J Clin. 2006;56: Decline in breast cancer incidence United States, MMWR Morb Mortal Wkly Rep. 2007;56: Sant M, Francisci S, Capocaccia R, et al. Time trends of breast cancer survival in Europe in relation to incidence and mortality. Int J Cancer. 2006;119: Boyko EJ, Theis MK, Vaughan TL, et al. Increased risk of inflammatory bowel disease associated with oral contraceptive use. Am J Epidemiol. 1994;140: Garcia Rodriguez LA, Gonzalez-Perez A, Johansson S, et al. Risk factors for inflammatory bowel disease in the general population. Aliment Pharmacol Ther. 2005;22: Riegler G, Caserta L, Castiglione F, et al. Increased risk of breast cancer in first-degree relatives of Crohn s disease patients. An IG-IBD study. Dig Liver Dis. 2006;38: Bernstein CN, Blanchard JF, Kliewer E, et al. Cancer risk in patients with inflammatory bowel disease: a population-based study. Cancer. 2001;91: Jess T, Winther KV, Munkholm P, et al. Intestinal and extra-intestinal cancer in Crohn s disease: follow-up of a population-based cohort in Copenhagen County, Denmark. Aliment Pharmacol Ther. 2004;19: Winther KV, Jess T, Langholz E, et al. Long-term risk of cancer in ulcerative colitis: a population-based cohort study from Copenhagen County. Clin Gastroenterol Hepatol. 2004;2: Louwman WJ, Janssen-Heijnen ML, Houterman S, et al. Less extensive treatment and inferior prognosis for breast cancer patient with comorbidity: a population-based study. Eur J Cancer. 2005;41: Cronin-Fenton DP, Norgaard M, Jacobsen J, et al. Comorbidity and survival of Danish breast cancer patients from 1995 to Br J Cancer. 2007;96: Kandiel A, Fraser AG, Korelitz BI, et al. Increased risk of lymphoma among inflammatory bowel disease patients treated with azathioprine and 6-mercaptopurine. Gut. 2005;54: Ljung T, Karlen P, Schmidt D, et al. Infliximab in inflammatory bowel 524

7 Inflamm Bowel Dis Volume 14, Number 4, April 2008 Survival in Danish Patients disease: clinical outcome in a population based cohort from Stockholm County. Gut. 2004;53: Chon BH, Loeffler JS. The effect of nonmalignant systemic disease on tolerance to radiation therapy. Oncologist. 2002;7: Goessling W, Mayer RJ. Systemic treatment of patients who have colorectal cancer and inflammatory bowel disease. Gastroenterol Clin North Am. 2006;35: Palli D, Trallori G, Saieva C, et al. General and cancer specific mortality of a population based cohort of patients with inflammatory bowel disease: the Florence Study. Gut. 1998;42: Frank L. Epidemiology. When an entire country is a cohort. Science. 2000;287: Storm HH, Michelsen EV, Clemmensen IH, et al. The Danish Cancer Registry history, content, quality and use. Dan Med Bull. 1997;44: Danish National Board of Health. Cancer incidence in Denmark Health statistics 1999, No. 3. Copenhagen: Danish National Board of Health; Andersen TF, Madsen M, Jorgensen J, et al. The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull. 1999;46: Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40: Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45: Thomsen RW, Hundborg HH, Lervang HH, et al. Diabetes and outcome of community-acquired pneumococcal bacteremia: a 10-year population-based cohort study. Diabetes Care. 2004;27: Young JL Jr, Roffers SD, Ries LAG. SEER Summary Staging Manual 2000: Codes and Coding Instructions, National Cancer Institute, NIH Pub. No Bethesda, MD; Harlan LC, Clegg LX, Trimble EL. Trends in surgery and chemotherapy for women diagnosed with ovarian cancer in the United States. J Clin Oncol. 2003;21: Potosky AL, Harlan LC, Kaplan RS, et al. Age, sex, and racial differences in the use of standard adjuvant therapy for colorectal cancer. J Clin Oncol. 2002;20: Jess T, Riis L, Vind I, et al. Changes in clinical characteristics, course, and prognosis of inflammatory bowel disease during the last 5 decades: a population-based study from Copenhagen, Denmark. Inflamm Bowel Dis. 2007;13: Wolters FL, Russel MG, Sijbrandij J, et al. Crohn s disease: increased mortality 10 years after diagnosis in a Europe-wide population based cohort. Gut. 2006;55: Hoie O, Schouten LJ, Wolters FL, et al. Ulcerative colitis: no rise in mortality in a European-wide population based cohort 10 years after diagnosis. Gut. 2007;56: Jess T, Loftus EV Jr, Harmsen WS, et al. Survival and cause specific mortality in patients with inflammatory bowel disease: a long term outcome study in Olmsted County, Minnesota, Gut. 2006; 55: Masunaga Y, Ohno K, Ogawa R, et al. Meta-analysis of risk of malignancy with immunosuppressive drugs in inflammatory bowel disease. Ann Pharmacother. 2007;41: Copeland KT, Checkoway H, McMichael AJ, et al. Bias due to misclassification in the estimation of relative risk. Am J Epidemiol. 1977;105: Jensen AR, Overgaard J, Storm HH. Validity of breast cancer in the Danish Cancer Registry. A study based on clinical records from one county in Denmark. Eur J Cancer Prev. 2002;11: Fonager K, Sorensen HT, Rasmussen SN, et al. Assessment of the diagnoses of Crohn s disease and ulcerative colitis in a Danish hospital information system. Scand J Gastroenterol. 1996;31: Wilchesky M, Tamblyn RM, Huang A. Validation of diagnostic codes within medical services claims. J Clin Epidemiol. 2004;57: Lash TL, Fox MP, Thwin SS, et al. Using probabilistic corrections to account for abstractor agreement in medical record reviews. Am J Epidemiol. 2007;165: Greenland S, Finkle WD. A critical look at methods for handling missing covariates in epidemiologic regression analyses. Am J Epidemiol. 1995;142: