GASTROENTEROLOGY
The Risk of Upper Gastrointestinal Familial Adenomatous Polyposis G. JOHAN A. OFFERHAUS, SUSAN V. BOOKER, ANNE and
STANLEY
1992;102:1960-1962
Cancer in
FRANCIS M. GIARDIELLO, ANNE J. KRUSH, C. TERSMETTE, N. CHRISTOPHER KELLEY,
R. HAMILTON
Departments of Medicine and Pathology and Oncology Center, The Johns Hopkins University School of Medicine and Hospital, Baltimore, Maryland
Adenomas with potential for malignancy occur frequently in the upper gastrointestinal tract of patients with familial adenomatous polyposis. However, an assessment of relative risk of upper gastrointestinal cancer in patients with adenomatous polyposis has never been performed. Therefore, the incidence rate of upper gastrointestinal cancer in patients with familial adenomatous polyposis in The Johns Hopkins Registry was compared with the rate of the general population through person-year analysis with adjustment for demographics. There was an increased relative risk of duodenal adenocarcinoma (relative risk, 330.82; 95% confidence limits, 132.86 and 681.49; P < 0.001) and ampullary adenocarcinoma (relative risk, 123.72; 95% confidence limits, 33.65 and 316.72; P < 0.001). No significant increased risk was found for gastric or nonduodenal small intestinal cancer. These results indicate that periodic surveillance of the upper gastrointestinal tract for duodenal and periampullary cancer is needed in patients with familial adenomatous polyposis. Prophylactic duodenectomy is a consideration when large adenoma with high-grade dysplasia are identified but awaits risk benefit analysis. amilial adenomatous polyposis (FAP) is an autosomal-dominant disease characterized by the development of hundreds of colorectal adenomas in young adults.’ If prophylactic colectomy is not performed, colorectal cancer develops in virtually all affected individuals by the 5th decade of life. Patients with FAP frequently have upper gastrointestinal tract adenomas, particularly in the stomach and duodenum.2*3 The duodenum and periampullary region are second only to the colorectum as a site of malignancy in patients with FAPs4z5 A high percentage of gastric cancer has also been reported by some polyposis registries throughout the world.4 Concern over upper gastrointestinal tract cancer has prompted recommendations for upper digestive tract
F
surveillance by endoscopy,204 but a formal risk assessment has never been performed. Therefore, to understand the magnitude of risk and the appropriateness of surveillance, the risk of invasive upper gastrointestinal cancer in individuals at risk for and affected with familial adenomatous polyposis in The Johns Hopkins Polyposis Registry was compared with the risk of the general population of the United States through person-year analysis. Patients and Methods Data were collected from The Johns Hopkins Polyposis Registry, which contains 157 pedigrees with familial adenomatous polyposis. Patient information was obtained on all registry subjects through chart review and was subsequently computerized using a dBase III plus software program (Ashton-Tate, Torrance, CA). For risk assessment, patients with FAP were considered at risk for upper gastrointestinal cancer from birth until death, and first-degree relatives of patients with FAP were considered at risk from birth through age 55; both groups contributed person-years accordingly. Observation time was until date of last contact, death, date of diagnosis of upper gastrointestinal tract cancer, or closing date of the study. The analysis started on January 1, 1969, and ended December 31,1987. These dates were chosen because population rates for comparison were available only for this period. Person-years at risk were calculated according to sex-, race-, and age-specific categories through 89 years of age during subsequent 5-year calendar time periods of observation using a Fortran computer program for cohort analysis.6 The ratio of observed carcinomas over the expected number was computed with a test of significance and 95% confidence limits assuming a Poisson distribution.‘*’ This ratio forms the relative risk and compares the cancer risk in the study population with that in the general population. The expected numbers of upper gastrointestinal tract cancers were calculated by multiplying the person-years with corresponding age-, race-, sex-, and calendar-time specific incidence rates obtained from Surveillance, Epide0 1992 by the American
Gastroenterological 0016-5065/92/$3.00
Association
UPPER GASTROINTESTINAL
June 1992
miology, and End Results (SEER) data.* Thus, the relative risk is adjusted for race, age, sex, and also calendar time (to account for possible secular trends). The absolute risk was defined as the number of observed carcinomas per total person-years. A risk assessment was performed for gastric adenocarcinoma [International Classification of Diseases (ICD) 151.0adenocarcinoma (ICD 152.0, 0.9, 9th revision], duodenal 9th revision], ampullary adenocarcinoma (ICD 156.2, 9th revision), and nonduodenal small intestinal cancer (ICD 152.1-0.9, 9th revision).
Results One thousand subjects
participated
three hundred and ninety-one in the person-year analysis;
they contributed 18,679 person-years of follow-up. These subjects included 604 white men, 646 white women, 76 black men, and 65 black women. Table 1 lists the patients with carcinomas of the upper gastrointestinal tract. There were two gastric adenocarcinomas, seven duodenal adenocarcinomas, four ampullary adenocarcinomas, and one jejunal adenocarcinoma. All of these patients were affected with adenomatous polyposis of the large bowel. The results of the person-year analysis are shown in Table 2. A high statistically significant relative risk of duodenal and ampullary adenocarcinoma was observed. No statistically significant increased relative risk for gastric adenocarcinoma and nonduodenal small bowel cancer was found. The absolute risk of duodenal and ampullary cancer combined was 11 cases per 18,679 personyears (l/1698 person-years).
Table 1. Patients With Familial Adenomatous Polyposis in the Johns Hopkins Polyposis Registry Who Developed Upper Gastrointestinal Adenocarcinoma Age at diagnosis of upper GI cancer (Yd 64 34 50 44 45 43 66 58 63 58 42 43 51 64
Race
Sex
Location of upper GI cancer
B w W W W B W W W W W W W w
M F M M M F M F M M M F F F
Stomach Stomach Duodenum Duodenum Duodenum Duodenum Duodenum Duodenum Duodenum Ampulla Ampulla Ampulla Ampulla Jejunum
CANCER IN POLYPOSIS
1981
Discussion The present study compares the incidence of upper gastrointestinal cancer in patients with familial adenomatous polyposis from a large populationbased registry with the incidence of the general population of the United States. A striking, multifold, and high statistically significant increased relative risk of duodenal and ampullary adenocarcinoma was found. This estimate of relative risk is conservative, because the analysis included unaffected first-degree relatives of FAP patients from birth to withdrawal at the relatively late age of 55 years. To the best of our knowledge, this is the first investigation that calculates the relative risk for upper gastrointestinal tract cancer in FAP, although increased risk has been appreciated clinically for several years.2-5 In our analysis, SEER data were used for the reference population and compared with the data from The Johns Hopkins Polyposis Registry. There may be differences in factors that pertain to precision in diagnosis between the two groups, raising the question of ascertainment bias. In older patients, the precise primary site of ampullary and periampullary cancer may be less accurate and relatively underscored in SEER data compared with the Hopkins Registry data. However, in the relatively young patients relevant to our analysis, aggressive diagnostic work-up with precision in the localization of tumor would be the standard of care in institutions from which SEER data is accumulated. Moreover, the magnitude of the increased relative risk in our study makes it unlikely that these results are attributable to merely differences in the precision of diagnosis. Finally, an increased relative risk of duodenal and periampullary adenocarcinoma corresponds to the most common location of upper gastrointestinal tract adenomas in patients with FAP.‘s3 This correlation also strongly supports our findings and points to the adenoma as the precursor lesion. Thus, these adenomas appear to be appropriate targets for surveillance. Based on the striking magnitude of relative risk and the level of absolute risk of upper gastrointestinal tract cancers occurring in patients with FAP at relatively young age, close surveillance with periodic upper endoscopy with biopsies and cytological brushing is warranted.2 Prophylactic duodenectomy (Whipple procedure) for patients who have adenoma with high-grade dysplasia may be a consideration. However, this decision involves a difficult balance between the benefits and the relatively high morbidity and mortality of this surgical procedure. Further recommendations on this issue await risk benefit analysis. No increased risk of gastric adenocarcinoma was found in our study. This agrees with the observation
1982
OFFERHAUS
ET AL.
Table
2. Relative Risk of Stomach, Duodenal, and Ampullary Carcinoma Polyposis From the Johns Hopkins Polyposis Registry Compared
GASTROENTEROLOGY
in Patients
With Familial
With the Risk of General
Vol. 102, No. 6
Adenomatous
Population
of the
United States Site, ICD 9th revision
Patient no. (patient-years)
Observed no. of carcinomas
Relative risk (observed/expected)
95% confidence limits
P
Duodenal 152.0
1391 (18,679)
7
330.82
132.66-681.49
< 0.001
1391 (18,679)
4
123.72
33.65-316.72
< 0.001
1391 (18,679)
2
2.44
1391 (18,679)
1
12.74
Ampullary 156.2
Gastric 151.0-0.9
0.29-6.81
NS
0.32-70.96
NS
Nonduodenal Small bowel 152.1-0.9
of others that gastric polyps in patients with FAP are most often nonneoplastic fundic gland polyp~.‘*~ Some investigators have postulated that a genetically determined mucosal growth abnormality interacts with environmental factors.‘*’ Thus, geographic variation in the incidence and location of upper digestive tract cancers may existm3 Interestingly, in Japan gastric cancer is more common than duodenal cancer in patients with FAP.” Also, in a longitudinal evaluation of upper gastrointestinal polyps, 50% of Japanese patients with FAP had gastric adenomas, an observation which further supports this notion.” This may be related to the high overall incidence of gastric cancer in Japan. In contrast, the Leeds Castle Polyposis Group (10 registries from Western countries) reported that duodenal cancer was much more common than stomach cancer.4*‘o In this regard, comparison of the relative risk of upper gastrointestinal tract cancers in Japanese patients with FAP with patients in Western countries would be of interest. The risk of gastric cancer for Japanese FAP migrants to Western countries would also be intriguing because in “conventional” gastric cancer, Japanese migrants adopt the risk of their new environment.” References Bussey HJR. Familial polyposis coli. Family studies, histopathology, differential diagnosis, and results of treatment. Baltimore, MD: Johns Hopkins University Press, 1975. Spigelman AD, Williams CB, Talbot IC, Domizio P, Phillips RKS. Upper gastrointestinal cancer in patients with familial adenomatous polyposis. Lancet 1989;2:783-785. Domizio P, Talbot IC, Spigelman AD, Williams CB, Phillips RKS. Upper gastrointestinal pathology in familial adenomatous polyposis: results from a nonprospective study of 102 patients. J Clin Path01 1990;43:738-743. Jagelman DG, DeCosse JJ, Bussey HJR. Upper gastrointestinal
cancer
in familial
adenomatous
polyposis.
Lancet
1988;
1:1149-1151. 5. Arvanitis
ML, Jagelman DG, Fazio VW, Lavery IC, McGannon E. Mortality in patients with familial adenomatous polyposis. Dis Colon Rectum 1990;33:639-642. 6. Coleman M, Doublas A, Hermon C, Peto L. Cohort study analysis with a Fortran computer program. Int J Epidemiol 1986;15:698-703. 7. Breslow NE, Day NE. Statistical
8.
9.
10.
11.
12.
methods in cancer research. Volume II. The design and analysis of cohort studies. IARC Scientific Publications, No. 82. Lyon, France: International Agency for Research on Cancer, 1987. Surveillance, Epidemiology, and End Results (SEER] Program public use tape. National Cancer Institute, DCPC, Surveillance Program. Cancer Statistics Branch, 1990. Spigelman AD, States DK, Venitt S, Phillips RKS. Excess of DNA adducts in the foregut of patients with familial adenomatous polyposis (abstr). Gut 1990;31:A1200. Utsunomiya J. The concept of hereditary colorectal cancer and the implications of its study. In: Utsunomiya J, Lynch HT, eds. Hereditary colorectal cancer. Tokyo: Springer-Verlag, 1990:3-16. Iida M, Yao T, Itoh H, Watanabe H, Matsui T, Iwashita A, Fujishima M. Natural history of gastric adenomas in patients with familial adenomatous coli/Gardner’s syndrome. Cancer 1988;61:605-611. Howson CP, Hyama T, Wynder EL. The decline in gastric cancer: epidemiology of an unplanned triumph. Epidemiol Rev 1986;8:1-27.
Received August 16,1991. Accepted November 19,1991. Address requests for reprints to: Francis M. Giardiello, M.D., Gastroenterology Division, Blalock 935, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, Maryland 21205. Supported in part by The Clayton Fund, The McAshan Fund, and grant CA44555 from the National Institutes of Health, Department of Health and Human Services. Dr. Offerhaus is a clinical fellow in gastrointestinal pathology supported by The Netherlands Organization for Scientific Research (NWO) and by The Netherlands Digestive Disease Foundation (NLD). Dr. Hamilton is a recipient of a young faculty award from E. I. Du Pont de Nemours & Co. The authors thank Linda Welch for secretarial support.
The Risk of Upper Gastrointestinal Familial Adenomatous Polyposis G. JOHAN A. OFFERHAUS, SUSAN V. BOOKER, ANNE and
STANLEY
1992;102:1960-1962
Cancer in
FRANCIS M. GIARDIELLO, ANNE J. KRUSH, C. TERSMETTE, N. CHRISTOPHER KELLEY,
R. HAMILTON
Departments of Medicine and Pathology and Oncology Center, The Johns Hopkins University School of Medicine and Hospital, Baltimore, Maryland
Adenomas with potential for malignancy occur frequently in the upper gastrointestinal tract of patients with familial adenomatous polyposis. However, an assessment of relative risk of upper gastrointestinal cancer in patients with adenomatous polyposis has never been performed. Therefore, the incidence rate of upper gastrointestinal cancer in patients with familial adenomatous polyposis in The Johns Hopkins Registry was compared with the rate of the general population through person-year analysis with adjustment for demographics. There was an increased relative risk of duodenal adenocarcinoma (relative risk, 330.82; 95% confidence limits, 132.86 and 681.49; P < 0.001) and ampullary adenocarcinoma (relative risk, 123.72; 95% confidence limits, 33.65 and 316.72; P < 0.001). No significant increased risk was found for gastric or nonduodenal small intestinal cancer. These results indicate that periodic surveillance of the upper gastrointestinal tract for duodenal and periampullary cancer is needed in patients with familial adenomatous polyposis. Prophylactic duodenectomy is a consideration when large adenoma with high-grade dysplasia are identified but awaits risk benefit analysis. amilial adenomatous polyposis (FAP) is an autosomal-dominant disease characterized by the development of hundreds of colorectal adenomas in young adults.’ If prophylactic colectomy is not performed, colorectal cancer develops in virtually all affected individuals by the 5th decade of life. Patients with FAP frequently have upper gastrointestinal tract adenomas, particularly in the stomach and duodenum.2*3 The duodenum and periampullary region are second only to the colorectum as a site of malignancy in patients with FAPs4z5 A high percentage of gastric cancer has also been reported by some polyposis registries throughout the world.4 Concern over upper gastrointestinal tract cancer has prompted recommendations for upper digestive tract
F
surveillance by endoscopy,204 but a formal risk assessment has never been performed. Therefore, to understand the magnitude of risk and the appropriateness of surveillance, the risk of invasive upper gastrointestinal cancer in individuals at risk for and affected with familial adenomatous polyposis in The Johns Hopkins Polyposis Registry was compared with the risk of the general population of the United States through person-year analysis. Patients and Methods Data were collected from The Johns Hopkins Polyposis Registry, which contains 157 pedigrees with familial adenomatous polyposis. Patient information was obtained on all registry subjects through chart review and was subsequently computerized using a dBase III plus software program (Ashton-Tate, Torrance, CA). For risk assessment, patients with FAP were considered at risk for upper gastrointestinal cancer from birth until death, and first-degree relatives of patients with FAP were considered at risk from birth through age 55; both groups contributed person-years accordingly. Observation time was until date of last contact, death, date of diagnosis of upper gastrointestinal tract cancer, or closing date of the study. The analysis started on January 1, 1969, and ended December 31,1987. These dates were chosen because population rates for comparison were available only for this period. Person-years at risk were calculated according to sex-, race-, and age-specific categories through 89 years of age during subsequent 5-year calendar time periods of observation using a Fortran computer program for cohort analysis.6 The ratio of observed carcinomas over the expected number was computed with a test of significance and 95% confidence limits assuming a Poisson distribution.‘*’ This ratio forms the relative risk and compares the cancer risk in the study population with that in the general population. The expected numbers of upper gastrointestinal tract cancers were calculated by multiplying the person-years with corresponding age-, race-, sex-, and calendar-time specific incidence rates obtained from Surveillance, Epide0 1992 by the American
Gastroenterological 0016-5065/92/$3.00
Association
UPPER GASTROINTESTINAL
June 1992
miology, and End Results (SEER) data.* Thus, the relative risk is adjusted for race, age, sex, and also calendar time (to account for possible secular trends). The absolute risk was defined as the number of observed carcinomas per total person-years. A risk assessment was performed for gastric adenocarcinoma [International Classification of Diseases (ICD) 151.0adenocarcinoma (ICD 152.0, 0.9, 9th revision], duodenal 9th revision], ampullary adenocarcinoma (ICD 156.2, 9th revision), and nonduodenal small intestinal cancer (ICD 152.1-0.9, 9th revision).
Results One thousand subjects
participated
three hundred and ninety-one in the person-year analysis;
they contributed 18,679 person-years of follow-up. These subjects included 604 white men, 646 white women, 76 black men, and 65 black women. Table 1 lists the patients with carcinomas of the upper gastrointestinal tract. There were two gastric adenocarcinomas, seven duodenal adenocarcinomas, four ampullary adenocarcinomas, and one jejunal adenocarcinoma. All of these patients were affected with adenomatous polyposis of the large bowel. The results of the person-year analysis are shown in Table 2. A high statistically significant relative risk of duodenal and ampullary adenocarcinoma was observed. No statistically significant increased relative risk for gastric adenocarcinoma and nonduodenal small bowel cancer was found. The absolute risk of duodenal and ampullary cancer combined was 11 cases per 18,679 personyears (l/1698 person-years).
Table 1. Patients With Familial Adenomatous Polyposis in the Johns Hopkins Polyposis Registry Who Developed Upper Gastrointestinal Adenocarcinoma Age at diagnosis of upper GI cancer (Yd 64 34 50 44 45 43 66 58 63 58 42 43 51 64
Race
Sex
Location of upper GI cancer
B w W W W B W W W W W W W w
M F M M M F M F M M M F F F
Stomach Stomach Duodenum Duodenum Duodenum Duodenum Duodenum Duodenum Duodenum Ampulla Ampulla Ampulla Ampulla Jejunum
CANCER IN POLYPOSIS
1981
Discussion The present study compares the incidence of upper gastrointestinal cancer in patients with familial adenomatous polyposis from a large populationbased registry with the incidence of the general population of the United States. A striking, multifold, and high statistically significant increased relative risk of duodenal and ampullary adenocarcinoma was found. This estimate of relative risk is conservative, because the analysis included unaffected first-degree relatives of FAP patients from birth to withdrawal at the relatively late age of 55 years. To the best of our knowledge, this is the first investigation that calculates the relative risk for upper gastrointestinal tract cancer in FAP, although increased risk has been appreciated clinically for several years.2-5 In our analysis, SEER data were used for the reference population and compared with the data from The Johns Hopkins Polyposis Registry. There may be differences in factors that pertain to precision in diagnosis between the two groups, raising the question of ascertainment bias. In older patients, the precise primary site of ampullary and periampullary cancer may be less accurate and relatively underscored in SEER data compared with the Hopkins Registry data. However, in the relatively young patients relevant to our analysis, aggressive diagnostic work-up with precision in the localization of tumor would be the standard of care in institutions from which SEER data is accumulated. Moreover, the magnitude of the increased relative risk in our study makes it unlikely that these results are attributable to merely differences in the precision of diagnosis. Finally, an increased relative risk of duodenal and periampullary adenocarcinoma corresponds to the most common location of upper gastrointestinal tract adenomas in patients with FAP.‘s3 This correlation also strongly supports our findings and points to the adenoma as the precursor lesion. Thus, these adenomas appear to be appropriate targets for surveillance. Based on the striking magnitude of relative risk and the level of absolute risk of upper gastrointestinal tract cancers occurring in patients with FAP at relatively young age, close surveillance with periodic upper endoscopy with biopsies and cytological brushing is warranted.2 Prophylactic duodenectomy (Whipple procedure) for patients who have adenoma with high-grade dysplasia may be a consideration. However, this decision involves a difficult balance between the benefits and the relatively high morbidity and mortality of this surgical procedure. Further recommendations on this issue await risk benefit analysis. No increased risk of gastric adenocarcinoma was found in our study. This agrees with the observation
1982
OFFERHAUS
ET AL.
Table
2. Relative Risk of Stomach, Duodenal, and Ampullary Carcinoma Polyposis From the Johns Hopkins Polyposis Registry Compared
GASTROENTEROLOGY
in Patients
With Familial
With the Risk of General
Vol. 102, No. 6
Adenomatous
Population
of the
United States Site, ICD 9th revision
Patient no. (patient-years)
Observed no. of carcinomas
Relative risk (observed/expected)
95% confidence limits
P
Duodenal 152.0
1391 (18,679)
7
330.82
132.66-681.49
< 0.001
1391 (18,679)
4
123.72
33.65-316.72
< 0.001
1391 (18,679)
2
2.44
1391 (18,679)
1
12.74
Ampullary 156.2
Gastric 151.0-0.9
0.29-6.81
NS
0.32-70.96
NS
Nonduodenal Small bowel 152.1-0.9
of others that gastric polyps in patients with FAP are most often nonneoplastic fundic gland polyp~.‘*~ Some investigators have postulated that a genetically determined mucosal growth abnormality interacts with environmental factors.‘*’ Thus, geographic variation in the incidence and location of upper digestive tract cancers may existm3 Interestingly, in Japan gastric cancer is more common than duodenal cancer in patients with FAP.” Also, in a longitudinal evaluation of upper gastrointestinal polyps, 50% of Japanese patients with FAP had gastric adenomas, an observation which further supports this notion.” This may be related to the high overall incidence of gastric cancer in Japan. In contrast, the Leeds Castle Polyposis Group (10 registries from Western countries) reported that duodenal cancer was much more common than stomach cancer.4*‘o In this regard, comparison of the relative risk of upper gastrointestinal tract cancers in Japanese patients with FAP with patients in Western countries would be of interest. The risk of gastric cancer for Japanese FAP migrants to Western countries would also be intriguing because in “conventional” gastric cancer, Japanese migrants adopt the risk of their new environment.” References Bussey HJR. Familial polyposis coli. Family studies, histopathology, differential diagnosis, and results of treatment. Baltimore, MD: Johns Hopkins University Press, 1975. Spigelman AD, Williams CB, Talbot IC, Domizio P, Phillips RKS. Upper gastrointestinal cancer in patients with familial adenomatous polyposis. Lancet 1989;2:783-785. Domizio P, Talbot IC, Spigelman AD, Williams CB, Phillips RKS. Upper gastrointestinal pathology in familial adenomatous polyposis: results from a nonprospective study of 102 patients. J Clin Path01 1990;43:738-743. Jagelman DG, DeCosse JJ, Bussey HJR. Upper gastrointestinal
cancer
in familial
adenomatous
polyposis.
Lancet
1988;
1:1149-1151. 5. Arvanitis
ML, Jagelman DG, Fazio VW, Lavery IC, McGannon E. Mortality in patients with familial adenomatous polyposis. Dis Colon Rectum 1990;33:639-642. 6. Coleman M, Doublas A, Hermon C, Peto L. Cohort study analysis with a Fortran computer program. Int J Epidemiol 1986;15:698-703. 7. Breslow NE, Day NE. Statistical
8.
9.
10.
11.
12.
methods in cancer research. Volume II. The design and analysis of cohort studies. IARC Scientific Publications, No. 82. Lyon, France: International Agency for Research on Cancer, 1987. Surveillance, Epidemiology, and End Results (SEER] Program public use tape. National Cancer Institute, DCPC, Surveillance Program. Cancer Statistics Branch, 1990. Spigelman AD, States DK, Venitt S, Phillips RKS. Excess of DNA adducts in the foregut of patients with familial adenomatous polyposis (abstr). Gut 1990;31:A1200. Utsunomiya J. The concept of hereditary colorectal cancer and the implications of its study. In: Utsunomiya J, Lynch HT, eds. Hereditary colorectal cancer. Tokyo: Springer-Verlag, 1990:3-16. Iida M, Yao T, Itoh H, Watanabe H, Matsui T, Iwashita A, Fujishima M. Natural history of gastric adenomas in patients with familial adenomatous coli/Gardner’s syndrome. Cancer 1988;61:605-611. Howson CP, Hyama T, Wynder EL. The decline in gastric cancer: epidemiology of an unplanned triumph. Epidemiol Rev 1986;8:1-27.
Received August 16,1991. Accepted November 19,1991. Address requests for reprints to: Francis M. Giardiello, M.D., Gastroenterology Division, Blalock 935, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, Maryland 21205. Supported in part by The Clayton Fund, The McAshan Fund, and grant CA44555 from the National Institutes of Health, Department of Health and Human Services. Dr. Offerhaus is a clinical fellow in gastrointestinal pathology supported by The Netherlands Organization for Scientific Research (NWO) and by The Netherlands Digestive Disease Foundation (NLD). Dr. Hamilton is a recipient of a young faculty award from E. I. Du Pont de Nemours & Co. The authors thank Linda Welch for secretarial support.
