Ulcerative colitis (UC) and Crohn s disease (CD) have

Transcription

1 GASTROENTEROLOGY 2006;130: Risk of Intestinal Cancer in Inflammatory Bowel Disease: A Population-Based Study From Olmsted County, Minnesota TINE JESS,* EDWARD V. LOFTUS JR, FERNANDO S. VELAYOS, W. SCOTT HARMSEN, ALAN R. ZINSMEISTER, THOMAS C. SMYRK, CATHY D. SCHLECK, WILLIAM J. TREMAINE, L. JOSEPH MELTON III, PIA MUNKHOLM,* and WILLIAM J. SANDBORN *Department of Medical Gastroenterology, Herlev University Hospital, Herlev, Denmark; Division of Gastroenterology and Hepatology, Mayo Clinic College of Medicine, Rochester, Minnesota USA, Division of Biostatistics, Mayo Clinic College of Medicine, Rochester, Minnesota USA, Division of Anatomic Pathology, Mayo Clinic College of Medicine, Rochester, Minnesota USA, and Division of Epidemiology, Mayo Clinic College of Medicine, Rochester, Minnesota USA See editorial on page Background & Aims: The risk for colorectal cancer in Crohn s disease and ulcerative colitis patients from the United States currently is unknown. We estimated the risk for small-bowel and colorectal cancer in a population-based cohort of 692 inflammatory bowel disease patients from Olmsted County, Minnesota, from 1940 to Methods: The Rochester Epidemiology Project was used to identify cohort patients with colorectal and small-bowel cancer. The cumulative probability of cancer and standardized incidence ratios (SIR) were estimated using expected rates from Surveillance, Epidemiology, and End Results, white patients from Iowa, from 1973 to 2000, and Olmsted County, from 1980 to Results: Colorectal cancer was observed in 6 ulcerative colitis patients vs 5.38 expected (SIR, 1.1; 95% confidence interval [CI], ), but 4 of these occurred among those with extensive colitis or pancolitis (SIR, 2.4; 95% CI, ). Six Crohn s disease patients (vs 3.2 expected) developed colorectal cancer (SIR, 1.9; 95% CI, ). Three Crohn s disease patients developed smallbowel cancer vs 0.07 expected (SIR, 40.6; 95% CI, ). Conclusions: The risk for colorectal cancer was not increased among ulcerative colitis patients overall, but appeared to be increased among those with extensive colitis. The colorectal cancer risk was increased slightly among Crohn s disease patients, who also had a 40-fold excess risk for small-bowel cancer. Ulcerative colitis (UC) and Crohn s disease (CD) have been recognized as disease entities for approximately a century, 1 and the first report of intestinal cancer occurrence in inflammatory bowel disease (IBD) was published 80 years ago. 2 Since then, numerous studies have addressed the issue, but the true risk for malignancy in IBD remains uncertain. The magnitude of risk observed in studies from referral centers 3 7 generally exceeds the risk reported in populationbased studies, 8 12 and in some population-based studies the overall risk for colorectal cancer (CRC) is even comparable with that of the background population A recent meta-analysis summarized the available data on CRC in UC patients and showed an increasing cumulative probability of CRC during the disease course. 18 However, the meta-analysis included a variety of studies with different designs, including referral center studies. It is essential to provide evidence from strictly population-based studies for prognostic information on unselected patients from the community and for reasonable life insurance estimates. 19 The risk for intestinal cancer among IBD patients in population-based cohorts from the United States was last assessed in the 1980s. 20,21 The aim of the present study was to describe the cumulative incidence and relative risk of CRC and smallbowel cancer in a population-based inception cohort of IBD patients from Olmsted County, Minnesota, and to evaluate the influence of age, sex, disease extent, and disease duration on the relative risk for cancer. Materials and Methods Patient Population A total of 692 Olmsted County, Minnesota, residents (378 UC, 314 CD) first were diagnosed with IBD in Olmsted County, Minnesota, during the period from 1940 to 2001, according to previously established well-defined criteria Olmsted County, situated in southeastern Minnesota, had a Abbreviations used in this paper: CI, confidence interval; CRC, colorectal cancer; IBD, inflammatory bowel disease; SEER, Surveillance, Epidemiology, and End Results; SIR, standardized incidence ratio; UC, ulcerative colitis; CD, Crohn s disease by the American Gastroenterological Association Institute /06/$32.00 doi: /j.gastro

2 1040 JESS ET AL GASTROENTEROLOGY Vol. 130, No. 4 population of 124,277 inhabitants in the 2000 US census. The majority of people reside in Rochester, which is the urban center of an otherwise rural county. Eighty-nine percent are non-hispanic white and a substantial portion of residents are of northern European heritage. Although 25% of county residents are employed in health care services (vs 8% nationwide), and the level of education is higher (30% of adults have completed college vs 21% nationwide), the residents of Olmsted County otherwise are similar socioeconomically to the US white population. 25 Medical care is practically self-contained within the region and is provided by the Mayo Clinic, Olmsted Medical Center, and their affiliated hospitals. Because of a unique medical record linkage system (the Rochester Epidemiology Project), 25 it is possible to conduct population-based research on IBD in this region, as previously described in detail. 22,23 Evaluation and Treatment Policy After diagnosis of IBD, patients were followed up by a combination of gastroenterologists, internists, pediatricians, and/or family practitioners at 1 of 2 medical centers within the county. Patients underwent periodic evaluations, usually once a year. Diagnostic evaluations in the earlier portion of the study period consisted of a combination of rigid proctoscopies, barium enemas, and small-bowel follow-throughs. In the 1970s, flexible fiberoptic endoscopy (both full colonoscopy and flexible sigmoidoscopy) began to replace rigid scopes and barium enemas as diagnostic tests. By the early 1990s, with the advent of video endoscopy, this transition largely was complete. Periodic surveillance colonoscopy with biopsy examination was performed in UC patients with long-standing and extensive disease beginning in the early 1980s. Although there is no formal institutional standard regarding the frequency of these surveillance examinations, an interval of 1 2 years between procedures has been recommended to patients by most gastroenterologists at the Mayo Clinic. Some, but not all, patients received maintenance therapy with sulfasalazine or an S-aminosalicylic acid derivative (olsalazine, mesalamine, or balsalazide). Azathioprine or 6-mercaptopurine was used as maintenance therapy for steroiddependent, steroid-refractory, or fistulizing patients in selected cases in the mid-1980s, but on a more widespread basis in the early 1990s. Short-term oral corticosteroid treatment was used for clinical exacerbations, usually at initial doses of mg of prednisone daily that was tapered and discontinued over 2 3 months. Methotrexate was used as a second-line immunosuppressive therapy beginning in the mid-1990s. Infliximab has been used for both induction and maintenance therapy in selected cases since Surgical resections were performed for emergent indications (eg, obstructive symptoms, hemorrhage), cancer prophylaxis (until the 1980s), and for failure to respond to medical therapy. Cancer Diagnosis The medical record linkage systems in Olmsted County contain information on CRC and small-bowel cancer diagnosed at out-patient clinic visits, hospitalizations, surgical interventions, and pathology examinations, including autopsy. Thus, all cases of cancer occurring in the cohort could be identified by cross-matching the registries with the medical record numbers of patients in the cohort. Medical records subsequently were reviewed to verify the diagnosis and obtain information on the cancer location, type, and stage. Statistical Analysis Person-years at risk were recorded from the date of IBD diagnosis until the date of intestinal cancer diagnosis, death, or end of the follow-up period (December 31, 2002) for each individual in the cohort. The risk for CRC and smallbowel cancer relative to the general population was estimated using standardized incidence ratios (SIRs, observed/expected numbers) with 95% confidence intervals (CIs) assuming a Poisson distribution for the observed number of cancers. Expected numbers were calculated using the observed age- and sex-specific person-years at risk in the cohort combined with age- and sex-specific intestinal cancer rates from the Surveillance, Epidemiology, and End Results (SEER) database (white patients from Iowa, ). In a secondary analysis, expected numbers also were derived from CRC rates in Olmsted County from 1980 to For purposes of estimating CRC risk, person-years of follow-up data were censored at total proctocolectomy. However, segmental bowel resection did not result in censoring of follow-up data. For estimation of smallbowel cancer risk, censoring at surgery did not occur. To determine the influence of disease extent or location on CRC risk, the SIR analysis was stratified by maximal disease extent into 3 categories for each condition (for UC, proctitis, leftsided colitis, and extensive colitis; and for CD, small bowel, ileocolonic, and colonic). To analyze the influence of disease duration on cancer risk, the SIR analysis also was stratified using a person-years approach into 4 categories of disease duration: 0 to less than 10 years, 10 to less than 20 years, 20 to less than 30 years, and 30 years or more. The cumulative incidence (1 survival free) of CRC and small-bowel cancer from the diagnosis of UC or CD was estimated using the Kaplan Meier product-limit method. These curves were compared with the expected cumulative incidence (based on SEER Iowa whites, ) by the 1-sample log-rank test. For the calculation of the cumulative incidence, cancers diagnosed within 30 days of the IBD diagnosis were excluded. The cumulative incidence of proctocolectomy from the diagnosis of IBD was estimated using similar techniques. Results The 692 IBD patients were followed up for a total of 10,470 person-years (median follow-up period, 14 years; range, 0 58 years). However, only 10,019 personyears occurred before total proctocolectomy and were available for analysis of CRC risk. A total of 118 patients (17%) died within the study period.

3 April 2006 INTESTINAL CANCER RISK IN IBD 1041 Figure 1. Cumulative incidence of proctocolectomy from time of IBD diagnosis: a population-based cohort study of patients with UC (n 378) and CD (n 314) from Olmsted County, Minnesota, Jagged line, Crohn s disease; solid line, ulcerative colitis. A total of 378 patients (44% women) with UC were followed up for a total of 5567 person-years (median follow-up period, 12.6 years; range, years) before proctocolectomy. The median age at diagnosis of UC was 33 years (range, 1 88 years). Altogether, 64 UC patients underwent total proctocolectomy, for a crude rate of 17%. The cumulative probability of proctocolectomy among UC patients was 9.0% (95% CI, ) after 5 years, 18.7% (95% CI, 14.1% 23.1%) after 15 years, and 20.8% (95% CI, ) after 25 years (Figure 1). The 314 patients (51% women) with CD were followed up for a total of 4908 person-years (median follow-up period, 10.8 y; range, y), of which 4456 person-years occurred before total proctocolectomy. The median age at diagnosis of CD was 29 years (range, 2 91 y). A total of 26 CD patients underwent total proctocolectomy, for a crude rate of 8%. The cumulative probability of proctocolectomy was 4.8% (95% CI, 2.3% 7.2%) after 5 years, 9.6% (95% CI, 5.8% 13.7%) after 15 years, and 10.4% (95% CI, 6.3% 15.7%) after 25 years (Figure 1). Colorectal Cancer in UC Among the 378 UC patients, a total of 6 CRCs were observed (Table 1), compared with 5.38 expected (SIR, 1.1; 95% CI, ) (Table 2). In a secondary analysis, expected numbers were derived from Olmsted County CRC incidence rates between 1980 and The expected numbers (4.08) and the SIR (1.5; 95% CI, ) did not change significantly from those obtained using SEER data. When the analysis was stratified by maximal extent of UC, the SIR increased from 0 (95% Table 1. Demographic and Clinical Data on 15 Patients With IBD and Intestinal Cancer: A Population-Based Cohort Study From Olmsted County, Minnesota, Patient Sex Year at IBD diagnosis Age at IBD diagnosis, y Maximal extent IBD Age at cancer diagnosis, y Cancer location Cancer type Dukes stage Age at death, y Cancer survival, y UC 1 F Pancolitis 37 Rectum ACA D F Left-sided 49 Sigmoid ACA A Alive 20 3 M Pancolitis 43 ACA M Left-sided 57 ACA M Pancolitis 66 Rectum ACA D M Pancolitis 52 Ascending ACA C 53 1 Small-bowel cancer 1 F Ileocolon 59 Ileum LMSA M Jejunum, 47 Jejunum, ileum LYMPH 47 0 ileocolon CD Small- and large-bowel cancer 1 F Ileocolon 34 Ileum, cecum ACA D 34 0 CRC 1 F Colon 42 Rectum ACA D F Ileum 76 Descending ACA B M Jejunum, 29 Rectum ACA C 30 1 ileocolon 4 M Ileum 22 Appendix CARC Alive 37 5 M Ileum 69 Cecum ACA C M Ileocolon 89 Cecum ACA A 93 4 Median NOTE. None of the cancer patients had a known history of hereditary CRC or primary sclerosing cholangitis. ACA, adenocarcinoma; LMSA, leiomyosarcoma; LYMPH, lymphoma; CARC, carcinoid.

4 1042 JESS ET AL GASTROENTEROLOGY Vol. 130, No. 4 Table 2. CRC Risk in 378 Patients With UC According to Age, Sex, and Calendar Year at Diagnosis: A Population-Based Cohort Study From Olmsted County, Minnesota, Patients, n Person-years at risk Observed cancers Expected number SIR 95% CI Total Extent a Proctitis Left-sided Extensive/pancolonic Duration, y b Sex Women Men Age at diagnosis, y Calendar year at IBD diagnosis a Does not include 1 person with unknown extent. b Does not include 2 cancers diagnosed within 30 days of IBD diagnosis. CI, ) among proctitis patients to 2.4 (95% CI, ) among those with extensive colitis or pancolitis (Table 2). By using a person-years approach, the analysis was stratified into 4 categories of UC duration but there was no clear relationship between disease duration and CRC risk (Table 2). All CRC patients had been diagnosed with UC before Although no significant calendar period effect was detected, the SIR for CRC among patients diagnosed before 1980 was 1.6 (95% CI, ) vs an SIR of 0 in those diagnosed after When stratifying patients by age at UC diagnosis and sex, no significantly increased risk was observed (Table 2). The median age at diagnosis of UC among cancer cases was 34 years (range, y), and the median age at CRC diagnosis was 51 years (range, years). None of the 6 CRC patients had primary sclerosing cholangitis, a family history of CRC, or had been diagnosed with dysplasia before cancer diagnosis. Only 1 patient was diagnosed with an early stage cancer (Dukes A), whereas 3 patients had either Dukes C or disseminated cancer at diagnosis. Two patients had unknown stages, but cancer survival of only 1 and 3 years. The overall median survival with CRC therefore was 1 year (range, 1 20 years) (Table 1). For the actuarial analysis of CRC incidence in UC, 2 cancers diagnosed within 30 days of IBD diagnosis were excluded. The cumulative incidence of CRC then was 0% at 5 years, 0.4% (95% CI, 0.0% 1.5%) at 15 years, and 2.0% (95% CI, 0.0% 4.9%) at 25 years after UC diagnosis (Figure 2). Similarly, the expected cumulative incidence of cancer based on Iowa SEER rates was 0.3% at 5 years, 1.1% at 15 years, and 2.3% at 25 years (P.55, log-rank test). Colorectal Cancer in CD Among the 314 CD patients, 6 cases of CRC were observed (Table 1), compared with 3.20 expected Figure 2. Cumulative incidence of CRC in patients with UC (n 378) compared with expected incidence of cancer (SEER data, Iowa whites, ) (P.55, log-rank). A population-based cohort study from Olmsted County, Minnesota, Jagged line, observed incidence; solid line, expected incidence.

5 April 2006 INTESTINAL CANCER RISK IN IBD 1043 Table 3. CRC and Small-Bowel Cancer Risk in 314 Patients With CD According to Age, Sex, and Calendar Year at Diagnosis: A Population-Based Cohort Study From Olmsted County, Minnesota, Patients, n Person-years at risk Observed cancers Expected cancers SIR 95% CI Small-bowel cancer a CRC Extent Small-bowel only Ileocolonic Colonic only Duration, y b Women, y Men, y * Calendar year at IBD diagnosis a 95% confidence intervals exclude 1.0 (P.05). b Does not include 2 cancers diagnosed within 30 days of IBD diagnosis. (SIR, 1.9; 95% CI, ) (Table 3). When Olmsted County CRC incidence rates ( ) were used, 26 the expected number of CRC (2.40) and the SIR (2.5; 95% CI, ) were not significantly different from those obtained using SEER data. The SIR analysis was stratified by maximal extent of CD. For CD patients with colonic involvement only, the SIR was 0.8 (95% CI, ), and this increased to 3.0 (95% CI, ) among patients with smallbowel involvement only. By using a person-years approach, the SIR analysis was stratified by duration of CD into 4 categories (Table 3). There was no clear relationship between cancer risk and duration of CD. Among men younger than 30 years at diagnosis of CD, a 22-fold increased risk for CRC was observed, yet the confidence interval was wide (95% CI, ). No increased risk for CRC was found when stratifying for calendar year at CD diagnosis (Table 3). Five cancers were adenocarcinomas and 1 was a carcinoid tumor (Table 1). None of the CRC patients had primary sclerosing cholangitis, a previous diagnosis of dysplasia, or a family history of CRC. The median age at CD diagnosis among the cancer cases was 49.5 years (range, years), and the median age at CRC was 55.5 years (range, years) (Table 1). Patients were diagnosed with CRC of Dukes stage A, B, or C, or with disseminated cancer, and the median survival time with CRC was 2.5 years (range, 1 37 years). For the actuarial analysis of CRC risk in CD, 2 cancers diagnosed within 30 days of the IBD diagnosis were excluded. The cumulative incidence of CRC then was 0.3% at 5 years (95% CI, 0.0% 1.1%), 1.6% at 15 years (95% CI, 0.0% 3.6%), and 2.4% at 25 years (95% CI, 0.0% 5.8%) after CD diagnosis (Figure 3). Likewise, the Figure 3. Cumulative incidence of CRC in patients with CD (n 314) compared with expected incidence of cancer (SEER data, Iowa whites, ) (P.66, log-rank). A population-based cohort study from Olmsted County, Minnesota, Jagged line, observed incidence; solid line, expected incidence.

6 1044 JESS ET AL GASTROENTEROLOGY Vol. 130, No. 4 expected incidence of cancer based on Iowa SEER rates was 0.2% at 5 years, 0.7% at 15 years, and 1.6% at 25 years (P.66, log-rank test). Small-Bowel Cancer Three small-bowel tumors were observed among the CD patients: an adenocarcinoma (which also involved the cecum), a leiomyosarcoma, and a local lymphoma (Table 1). Because the frequency of small-bowel cancer in the background population is very low, only 0.07 cancers were expected (SIR, 41.1; 95% CI, ) (Table 3). When the SIR analysis was stratified by duration of CD using a person-years approach, there was no clear relationship between small-bowel cancer risk and duration of CD (data not shown). The 3 cancers occurred in patients with a relatively long duration of CD (range, 9 27 years), were diagnosed at a young age (range, years), and were diagnosed before Patient survival after cancer diagnosis was 0 4 years (Table 1). The cumulative incidence of small-bowel cancer after CD diagnosis was 0% at 5 years, 0.5% at 15 years (95% CI, 0.0% 1.6%), and 1.7% at 25 years (95% CI, 0.0% 4.5%). The expected incidence based on Iowa white SEER rates was 0.005% at 5 years, 0.02% at 15 years, and 0.03% at 25 years (P.002, log-rank test). No small-bowel cancers were observed in the UC cohort. Discussion In this population-based study from North America, the CRC risk among UC patients overall was similar to that expected in the general population. Among patients with extensive colitis, the CRC risk was increased 2-fold, although this did not meet statistical significance. In CD, the risk for CRC was somewhat increased overall, but was not statistically significant. In a subset of male CD patients with young age at diagnosis, the CRC risk was increased. The risk for small-bowel cancer in the CD cohort was increased markedly. All patients derived from a well-defined region, Olmsted County, and had been included in the cohort according to strict international diagnostic criteria. 22,23 The regional medical record linkage system (the Rochester Epidemiology Project) provided access to inpatient and outpatient medical records in the community, and a high level of clinical detail therefore was available. However, the relatively small number of patients and the low frequency of events observed in the present study may have veiled differences in risk among subsets of patients. Moreover, the ethnic composition of the cohort, a relatively high percentage being white, may differ somewhat from the general US population. The finding of a CRC risk in UC patients similar to that of the background population is in accordance with a North American study from the 1980s 20 and with recent data from Denmark. 17 On the other hand, population-based studies from Canada, 12 Israel, 10 and Sweden 9,11 have shown increased relative risks of CRC in UC ranging from 1.4 to 6. The conflicting results may be owing to differences in local treatment policies. In Copenhagen County, Denmark, the SIR for CRC in UC was 1.1 (95% CI, ). As in Olmsted County, this risk was not increased significantly from expected. Maintenance treatment with S-aminosalicylates 27 has been advocated in both counties, and the cumulative proctocolectomy rate in Copenhagen was approximately 20% after 20 years among both UC and CD patients, 13 as compared with 19% among UC patients and 10% among CD patients from Olmsted County. However, surveillance colonoscopies were not performed as part of the standard follow-up regimen in Denmark. 13,17 Whether the low cancer risk observed in the 2 counties is owing to maintenance treatment with mesalamine, surgery, close follow-up evaluation of patients, or other factors remains to be investigated further. It is interesting to note that none of the UC patients in Olmsted County who were diagnosed with UC after 1980 has developed CRC, so perhaps the combination of more widespread maintenance therapy and surveillance colonoscopy has made a positive impact on CRC risk. When the CRC risk was stratified by extent of UC, we observed the lowest risk in proctitis patients (SIR, 0.0; 95% CI, ), the highest risk in patients with extensive UC or pancolitis (SIR, 2.4; 95% CI, ), and an intermediate risk in those with left-sided disease (SIR, 0.8; 95% CI, ). This risk gradient by extent is in keeping with the concept that extent of UC is an important risk factor for CRC. Although the 2-fold increase in relative risk among those with extensive UC was not significant, this may have been owing to the relatively small size of our cohort. However, the finding of disease extent as a risk factor for CRC has not been shown in all populationbased cohorts, 17 suggesting that factors as yet identified may influence cancer risk. When all of the personyears of observation in the cohort were stratified by duration from UC diagnosis, we could not show increasing relative risk for CRC over time. We observed a trend toward increased CRC risk in the CD cohort, and a significantly increased risk among men diagnosed with CD before the age of 30 years. Population-based studies from Denmark, 13 Israel, 14 and Sweden 15 have reported an overall normal

7 April 2006 INTESTINAL CANCER RISK IN IBD 1045 risk for CRC in CD, whereas studies from Canada 12 and Sweden 8 have reported an excess with relative risks of approximately 2.5. As in the present study, the Swedish study reported an increased risk for CRC in patients younger than 30 years at diagnosis of CD. In our analysis of CRC risk stratified by extent of CD, we found, to our surprise, that patients with smallbowel involvement alone had the highest SIR (3.0; 95% CI, ) and that those with colonic involvement only had the lowest SIR (0.8; 95% CI, ). We have no firm explanation for this unexpected result, except that the relatively small cohort size and number of cancers may have played a role. In contrast to the somewhat reassuring findings for CRC risk, especially in UC, the risk for small-bowel cancer in CD was increased 40-fold, although the CIs were wide. Nevertheless, the finding only confirms what already has been shown in larger population-based studies from Sweden 15 and Denmark, 13 where small-bowel cancer risk in CD was increased 16.6-fold and 66.7-fold, respectively. Still, one should remember that smallbowel cancer is a very rare condition in the background population, and for this reason the occurrence of only a few cancers (3 in the present study, 4 in Sweden, and 4 in Denmark) result in the worryingly high risk estimates observed. In the present investigation, the absolute cumulative incidence of small-bowel cancer after 25 years with CD was only 1.7%. In conclusion, this population-based study from North America showed that the risk for CRC was not increased among patients with UC overall, but appeared slightly increased among patients with extensive colitis. CD patients had a slightly increased risk for colorectal cancer (most pronounced among men aged 30 years at diagnosis) and a more than 40-fold increased risk for small-bowel cancer. References 1. Baron JH. Inflammatory bowel disease up to Mt Sinai J Med 2000;67: Crohn B, Rosenberg H. The sigmoidoscopic picture of chronic ulcerative colitis (non-specific). Am J Med Sci 1925;170: Weedon DD, Shorter RG, Ilstrup DM, Huizenga KA, Taylor WF. Crohn s disease and cancer. N Engl J Med 1973;289: Gyde SN, Prior P, Macartney JC, Thompson H, Waterhouse JA, Allan RN. Malignancy in Crohn s disease. Gut 1980;21: Gillen CD, Andrews HA, Prior P, Allan RN. Crohn s disease and colorectal cancer. Gut 1994;35: Greenstein AJ, Sachar DB, Smith H, Janowitz HD, Aufses AH Jr. A comparison of cancer risk in Crohn s disease and ulcerative colitis. Cancer 1981;48: Gillen CD, Walmsley RS, Prior P, Andrews HA, Allan RN. Ulcerative colitis and Crohn s disease: a comparison of the colorectal cancer risk in extensive colitis. Gut 1994;35: Ekbom A, Helmick C, Zack M, Adami HO. Increased risk of largebowel cancer in Crohn s disease with colonic involvement. Lancet 1990;336: Ekbom A, Helmick C, Zack M, Adami HO. Ulcerative colitis and colorectal cancer. A population-based study. N Engl J Med 1990; 323: Gilat T, Fireman Z, Grossman A, Hacohen D, Kadish U, Ron E, Rozen P, Lilos P. Colorectal cancer in patients with ulcerative colitis. A population study in central Israel. Gastroenterology 1988;94: Karlen P, Lofberg R, Brostrom O, Leijonmarck CE, Hellers G, Persson PG. Increased risk of cancer in ulcerative colitis: a population-based cohort study. Am J Gastroenterol 1999;94: Bernstein CN, Blanchard JF, Kliewer E, Wajda A. Cancer risk in patients with inflammatory bowel disease: a population-based study. Cancer 2001;91: Jess T, Winther KV, Munkholm P, Langholz E, Binder V. 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: Fireman Z, Grossman A, Lilos P, Hacohen D, Bar MS, Rozen P, Gilat T. Intestinal cancer in patients with Crohn s disease. A population study in central Israel. Scand J Gastroenterol 1989; 24: Persson PG, Karlen P, Bernell O, Leijonmarck CE, Brostrom O, Ahlbom A, Hellers G. Crohn s disease and cancer: a populationbased cohort study. Gastroenterology 1994;107: Brostrom O, Lofberg R, Nordenvall B, Ost A, Hellers G. The risk of colorectal cancer in ulcerative colitis. An epidemiologic study. Scand J Gastroenterol 1987;22: Winther KV, Jess T, Langholz E, Munkholm P, Binder V. Long-term risk of cancer in ulcerative colitis: a population-based cohort study from Copenhagen County. Clin Gastroenterol Hepatol 2004;2: Eaden JA, Abrams KR, Mayberry JF. The risk of colorectal cancer in ulcerative colitis: a meta-analysis. Gut 2001;48: Travis SP. Review article: insurance risks for patients with ulcerative colitis or Crohn s disease. Aliment Pharmacol Ther 1997; 11: Stonnington CM, Phillips SF, Zinsmeister AR, Melton LJ III. Prognosis of chronic ulcerative colitis in a community. Gut 1987;28: Gollop JH, Phillips SF, Melton LJ III, Zinsmeister AR. Epidemiologic aspects of Crohn s disease: a population based study in Olmsted County, Minnesota, Gut 1988;29: Loftus EV Jr, Silverstein MD, Sandborn WJ, Tremaine WJ, Harmsen WS, Zinsmeister AR. Ulcerative colitis in Olmsted County, Minnesota, : incidence, prevalence, and survival. Gut 2000;46: Loftus EV Jr, Silverstein MD, Sandborn WJ, Tremaine WJ, Harmsen WS, Zinsmeister AR. Crohn s disease in Olmsted County, Minnesota, : incidence, prevalence, and survival. Gastroenterology 1998;114: Loftus CG, Loftus EV Jr, Sandborn WJ, Tremaine WJ, Harmsen WS, Zinsmeister AR. Update on incidence and prevalence of Crohn s Disease (CD) and ulcerative colitis (UC) in Olmsted County, Minnesota (abstr). Gastroenterology 2003;124:A Melton LJ III. History of the Rochester Epidemiology Project. Mayo Clin Proc 1996;71: Gupta AK, Melton LJ III, Petersen GM, Timmons LJ, Vege SS, Harmsen WS, Diehl NN, Zinsmeister AR, Ahlquist DA. Changing trends in the incidence, stage, survival, and screen-detection of

8 1046 JESS ET AL GASTROENTEROLOGY Vol. 130, No. 4 colorectal cancer: a population-based study. Clin Gastroenterol Hepatol 2005;3: Eaden J, Abrams K, Ekbom A, Jackson E, Mayberry J. Colorectal cancer prevention in ulcerative colitis: a case-control study. Aliment Pharmacol Ther 2000;14: Received May 27, Accepted December 14, Address requests for reprints to: Edward V. Loftus Jr, MD, Division of Gastroenterology and Hepatology, Mayo Clinic, 200 First Street SW, Rochester, Minnesota loftus.edward@mayo.edu; fax: (507) Presented in part at the 106th Annual Meeting of the American Gastroenterological Association, May 14 19, 2005, Chicago, Illinois (Gastroenterology 2005;128(Suppl 2):A321). Supported by the Mayo Foundation for Medical Education and Research, AR30582 from the National Institutes of Health, the Astra Zeneca Travel Foundation (visit to Mayo Clinic by T.J.), Augustinus Foundation, Beckett Foundation, Danish Crohn Colitis Foundation, Jacob Madsen and Olga Madsen Foundation, and Sigrid R. Moran Foundation. Councilman of Councilman Bodies Copyright holder unknown. William Thomas Councilman ( ) was born on a farm near Baltimore where he spent an idyllic boyhood and was imbued by a lifelong passion for nature study. Seeking to emulate his father, a country doctor, he entered the medical school of the University of Maryland but completed only the first half of a 2-year course in order to accept a fellowship in the biology department of the newly founded Johns Hopkins University. This paved the way for further studies at the principal laboratories in Europe. In 1885, at the invitation of William Welch ( ) he joined the staff in pathology at the Johns Hopkins Hospital. Councilman and Welch formed an intimate friendship; together they were avid fans of the Baltimore Orioles baseball team. Councilman s speech was often affected by a stutter, described by his friend Harvey Cushing as his engaging hesitancy of speech. When times in the laboratory were trying and tiring Councilman might suggest, B-b-boys, let s go ch-ch-church ( church being the nearby Hanselmann s saloon). In 1891, Councilman defined a prevalent form of bloody flux and named it amebic dysentery. The following year he departed Baltimore to become the Shattuck Professor of Pathology at Harvard where he taught for 30 years. It was there he described eosinophilic aggregates in necrotic liver tissue that became known as Councilman bodies. For more on Councilman s career, see Cushing s charming account in Science 1933;77: Contributed by WILLIAM HAUBRICH The Scripps Clinic, La Jolla, California