The removal of the gallbladder, cholecystectomy, will. Intestinal Cancer After Cholecystectomy: Is Bile Involved in Carcinogenesis?

Transcription

1 GASTROENTEROLOGY 2001;121: Intestinal Cancer After Cholecystectomy: Is Bile Involved in Carcinogenesis? JESPER LAGERGREN,*, WEIMIN YE, and ANDERS EKBOM *Department of Surgery, Karolinska Hospital, Stockholm; Department of Medical Epidemiology, Karolinska Institutet, Stockholm, Sweden Background & Aims: Results concerning an association between cholecystectomy and right-sided colon cancer are inconsistent. Little is known about the relation between cholecystectomy and small bowel cancer. Therefore, we evaluated cholecystectomy and risk of bowel cancer. Methods: Cholecystectomized patients, identified through the Swedish Inpatient Register, from 1965 through 1997, were followed up for subsequent cancer. The standardized incidence ratio (SIR) estimated relative risk. Results: In total, 278,460 cholecystectomized patients, contributing 3,519,682 person-years, were followed up for a maximum of 33 years after surgery. Cholecystectomized patients had an increased risk of proximal intestinal adenocarcinoma, which gradually declined with increasing distance from the common bile duct. The risk was significantly increased for adenocarcinoma (SIR, 1.77; 95% confidence interval [CI], ) and carcinoids of the small bowel (SIR, 1.71; 95% CI, ), and right-sided colon cancer (SIR, 1.16; 95% CI, ). No association was found with more distal bowel cancer. The gradient was further pronounced when surgery of the common bile duct was included. The associations remained increased up to 33 years after cholecystectomy. No differences between sexes were found. Conclusions: Cholecystectomy increases the risk of intestinal cancer, a risk that declines with increasing distance from the common bile duct. Changes in the intestinal exposure to bile might be the underlying biological mechanism. The removal of the gallbladder, cholecystectomy, will result in a more continuous flow of bile to the duodenum with a less effective physiological periodic release at mealtimes. Hence, the dilution of bile with foods and gastric juices only occurs during a limited time of the total intestinal exposure to bile. Such changes in the intestinal exposure to bile leads to an increased bacterial degradation of bile acids to form secondary bile acids, 1,2 which in turn has been proposed to increase the risk of colorectal cancer. 3,4 Although numerous epidemiological investigations and 2 meta-analyses have addressed the relation between cholecystectomy and colorectal cancer, this possible association has still not been firmly established or refuted. 5,6 However, results from meta-analyses and large cohort studies indicate that cholecystectomy might increase the risk of colon cancer moderately, particularly for right-sided tumors 5,6 among women, 7,8 and after a latency period of 15 years or more after surgery. 5,8 The distribution of small bowel adenocarcinoma, with a clear predominance in the proximal part, also indicates that bile might be of importance in the carcinogenesis. 9,10 The rarity of small bowel cancer has been a major obstacle for epidemiological studies, but the increasing incidence of these tumors during recent decades underlines the need for etiological research. Although 2 previous studies have reported a positive association between cholecystectomy and small bowel cancer, 15,16 a causal association still remains to be established. Our aim was to determine whether cholecystectomy is linked to an increased risk of intestinal adenocarcinoma of any site. We used Swedish Registry data, which enabled us to gain sufficient statistical power, even for rare outcomes such as cancer of the small bowel, and to evaluate duration of follow-up of more than 3 decades after cholecystectomy. Methods Design We conducted a nationwide, retrospective populationbased cohort study in Sweden. The cohort consisted of patients who had undergone cholecystectomy between the 33-yearperiod 1965 through Cholecystectomized patients were identified through the nationwide Swedish Inpatient Register. The Swedish Inpatient Register From , the National Board of Health and Welfare established the Swedish Inpatient Register. Information about the National Registration Number, a unique 10- Abbreviations used in this paper: CI, confidence interval; ICD, International Classification of Diseases; SIR, standardized incidence ratio by the American Gastroenterological Association /01/$35.00 doi: /gast

2 September 2001 CHOLECYSTECTOMY AND INTESTINAL CANCER 543 digit identification number assigned to all Swedish residents, main and co-discharge diagnoses and surgical procedures were available. The 7th revision of the International Classification of Diseases (ICD-7) was used for coding diagnoses from 1964 to 1968, the 8th revision (ICD-8) for 1969 to 1986, and the 9th revision (ICD-9) thereafter. The surgical procedures were coded according to the Swedish Classification of Operations and Major Procedures. The percentage of all Swedish hospitals included in the Inpatient Register was 60% in 1969, 75% in 1978, 85% in 1983, and 100% from 1987 and thereafter. Because there is virtually no private medical care in Sweden, a study based on the Swedish Inpatient Register is, therefore, considered to be population-based. Follow-up The National Registration Number was used for record linkage to the nationwide Registry of Causes of Death and the Migration Register for information on the date of death or emigration among those who died or emigrated during the study period. All incident cancers were ascertained through linkage to the National Swedish Cancer Register, founded in 1958 and considered to be 98% complete. 17 Exclusions We identified a total of 304,815 unique and correct National Registration Numbers with a recorded cholecystectomy, among which 6,073 records with inconsistencies or invalid dates uncovered during the record linkage were deleted from analysis. Furthermore, 20,282 records were excluded because of prevalent cancers. The final cohort remaining for analysis consisted of 278,460 cholecystectomized patients. Anatomic Subsite of Cancer The anatomic subsite of about 50% of small bowel cancer was not specified in the Cancer Register. However, the percentage of small bowel cancer that was histologically verified was close to 100% We used histologic type of small bowel cancer (adenocarcinoma and carcinoids) as a proxy for anatomic subsite. Adenocarcinomas of the small bowel are in 75% 80% located in the duodenum or proximal jejunum, whereas 90% of small bowel carcinoids are located in the ileum. 21 Cancer of the colon was subdivided by subsite into cancer of the cecum or ascending colon (right-sided colon cancer) (ICD-7 code 153.0), transverse colon (ICD-7 code 153.1), and descending and sigmoid colon (left-sided colon cancer) (ICD-7 code and 153.3). Statistical Analyses Person-years were calculated from the first discharge with a cholecystectomy until the occurrence of any first cancer, death, emigration, or end of observation (December 31, 1997), whichever occurred first. Primary cancers that occurred after a first cancer or that were found first at autopsy were excluded. The standardized incidence ratio (SIR), estimated as the ratio of the observed to the expected number of incident cancers, was used to estimate relative risk. The expected number of cancers was calculated by multiplying the observed personyears by age (in 5-year groups), sex, and calendar year-specific cancer incidence rates. The expected rates were derived from the entire Swedish population who were without a reported cancer and aggregated by 5 calendar years to avoid instability across calendar year. CIs of SIRs were calculated assuming that the observed number of events followed a Poisson distribution 22 with the mean given by the expected number of events. Analyses were also stratified by latency interval after cholecystectomy, surgery on the common bile duct, age at cholecystectomy, sex, and laparoscopic or open surgery. We excluded cancers and person-years accrued at the first year of observation from all cohort members in the analyses to avoid the possible influence of reversed causality or selection bias. Such bias could arise if subjects with a subclinical cancer were more likely to be hospitalized for cholecystectomy than subjects without a subclinical cancer. 23 Results The Cohort Some characteristics of the cohort members are presented in Table 1. In total, 278,460 cholecystectomized patients were followed-up for an average of 12.1 years. The cohort members contributed 3,519,682 person-years of follow-up. The average age at entry into the cohort, which was equal to the discharge after cholecystectomy, was 52 years. Cholecystectomy was more common among women than among men and the women were, on average, younger than men at the time of surgery. Risk of Intestinal Cancer After Cholecystectomy by Distance From the Common Bile Duct There was an association between cholecystectomy and the risk of cancer of the bowel, which decreased with increased distance from the source of intestinal bile, i.e., the common bile duct (Table 2). Among patients Table 1. Description of the Cohort of Cholecystectomized Persons in Sweden During Characteristics Men Women All Number of cholecystectomized patients 90, , ,460 Mean age (and range) at cholecystectomy (yr) 56.8 (0 99) 49.9 (0 100) 52.2 (0 100) Mean follow-up duration after cholecystectomy (yr) Range of follow-up after cholecystectomy (yr) Number of person-years of follow-up 1,034,651 2,485,030 3,519,682

3 544 LAGERGREN ET AL. GASTROENTEROLOGY Vol. 121, No. 3 Table 2. Risk of Cancer of the Small and Large Intestine After Cholecystectomy With Increasing Distance From the Common Bile Duct Location of the intestinal tumor Cholecystectomy only Cholecystectomy and surgery of the common bile duct SIR (95% CI) Proximal small bowel (adenocarcinomas) ( ) ( ) Distal small bowel (carcinoids) ( ) ( ) Cecum and ascending colon ( ) ( ) Transverse colon ( ) ( ) Descending colon ( ) ( ) Sigmoid colon ( ) ( ) Rectum ( ) ( ) NOTE. First year after cholecystectomy excluded. Included were 3,519,682 person-years of follow-up. with a cholecystectomy only, the risk of proximal small bowel cancer (adenocarcinoma) was increased nearly 2-fold, whereas cholecystectomy was significantly but somewhat less strongly associated with the risk of more distal location of the small bowel cancer (carcinoids). A 16% significantly increased risk was also found in the proximal part of the colon (cecum and ascending colon combined). The gradient was even more pronounced among patients who underwent surgery on the common bile duct as part of the cholecystectomy. In this group, the risk of small bowel adenocarcinoma, with a predominantly proximal location (duodenum or jejunum), was 3.14 (95% CI, ). The risk was then gradually reduced with increasing distance from the common bile duct (Table 2). We did not have site-specific data of small bowel adenocarcinomas of more than 50% of the cases, but among cholecystectomized patients, SIR was 1.80 (95% CI, ) for adenocarcinomas with a defined location in the duodenum, and the corresponding SIR among patients in whom surgery of the common bile duct was performed, the SIR for duodenal adenocarcinoma was 4.49 ( ). Risk of Small Bowel Adenocarcinoma After Cholecystectomy The risk of small bowel adenocarcinoma was nonsignificantly higher among men than women, but was significantly increased in both sexes (Table 3). Patients older than 70 years at cholecystectomy had a higher risk as compared with younger age groups. The risk remained increased during the entire follow-up period, with an SIR of 1.63 (95% CI, ) 15 years or more after surgery. However, the risk did not increase with increasing latency time after surgery. Table 3. SIR and 95% CI for Adenocarcinoma and Carcinoids of the Small Bowel Among Cholecystectomized Persons Adenocarcinoma of the small bowel Carcinoids of the small bowel Variable No. person-years Latency interval after cholecystectomy (yr) ,519, ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( ) Common bile duct surgery No 3,014, ( ) ( ) Yes 505, ( ) ( ) Age at cholecystectomy 50 years 1,811, ( ) ( ) years 757, ( ) ( ) years 617, ( ) ( ) 70 years 333, ( ) ( ) Sex Men 1,034, ( ) ( ) Women 2,485, ( ) ( ) NOTE. First year after cholecystectomy excluded.

4 September 2001 CHOLECYSTECTOMY AND INTESTINAL CANCER 545 Table 4. SIR and 95% CI for Colorectal Cancer Among Cholecystectomized Persons Cecum and ascending colon cancer Colorectal cancer distal to the ascending colon Variable No. personyears Latency interval after cholecystectomy (yr) ,519, ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( ) , ( ) ( ) 25 96, ( ) ( ) Common bile duct surgery No 3,014, ( ) ( ) Yes 505, ( ) ( ) Age at cholecystectomy 50 years 1,811, ( ) ( ) years 757, ( ) ( ) years 617, ( ) ( ) 70 years 333, ( ) ( ) Sex Men 1,034, ( ) ( ) Women 2,485, ( ) ( ) NOTE. First year after cholecystectomy excluded. Risk of Small Bowel Carcinoids After Cholecystectomy In the group of carcinoids of the small bowel, with a 90% predominance for location within ileum, the association with cholecystectomy was significantly increased (SIR, 1.71; 95% CI, ) (Table 3). Sex and age at cholecystectomy and bile duct surgery did not affect the risk estimates. The risk remained increased after 19 years of follow-up, but not thereafter. The SIR was 1.55 (95% CI, ) 10 years or more after surgery. Proximal Colon Cancer The increased risk of cancer of the cecum and ascending colon remained virtually at the same level with increasing latency time after surgery (Table 4). The association was of similar strength and statistically significant in both sexes, and there were no important differences between age groups at cholecystectomy. Surgery of the common bile duct as part of the cholecystectomy did not affect the results. Colorectal Cancer Distal to the Ascending Colon There was no association with any site of the colon or rectum distal to the ascending colon. This lack of association was independent of sex, age at cholecystectomy, latency period after surgery, and the occurrence of bile duct surgery (Table 4). Laparoscopic Cholecystectomy Among patients who had been cholecystectomized with a laparoscopic technique, we did not find any differences in the associations with bowel cancer at any site compared with open cholecystecomy (data not shown). Gallstone Cohort To evaluate whether the cholecystectomy, and not the gallstone disease per se, was the factor associated with adenocarcinoma of the bowel, we identified a second cohort consisting of 167,646 patients in the Swedish Inpatient Register who all had a recorded gallstone. Some of the patients did not undergo cholecystectomy, whereas the remaining part of the cohort was censored at cholecystectomy. In this comparison cohort, we found a tendency of an increased risk of small bowel adenocarcinoma that did not reach statistical significance (SIR, 1.64; 95% CI, ). There was no association with cancer of the cecum or ascending colon (SIR, 1.00; 95% CI, ) or the remaining parts of the colon or rectum (SIR, 1.00; 95% CI, ). Discussion In the present study, we could demonstrate an association between cholecystectomy and cancer of the small bowel and proximal colon, with a gradient and a decreasing risk with increasing distance from the com-

5 546 LAGERGREN ET AL. GASTROENTEROLOGY Vol. 121, No. 3 mon bile duct. A history of surgery on the common bile duct as part of the cholecystectomy enhanced the association between cholecystectomy and proximal small bowel cancer. More than 60 epidemiologic studies have analyzed the association between cholecystectomy and the risk of colorectal cancer with conflicting results. However, the results from 2 meta-analyses indicated an 11% 34% increased risk of colorectal cancer, most pronounced for proximal colon cancer with an 86% 88% increase in risk. 5,6 The majority of previous studies were case-control studies, which are in general more susceptible to bias, in this case particularly selection bias, than cohort studies. The association between cholecystectomy and colorectal cancer was less pronounced in population-based casecontrol studies compared with hospital-based studies, 5,6 which might be explained by methodological limitations inherent for hospital-based studies. Among the previous cohort studies, 7,8,16,24 28 most did not find any association, possibly caused by insufficient duration of followup. However, the largest previous cohort study, which was based on a cohort that was partly included also in our study, found a 54% increased risk of proximal colon cancer among women 15 years or more after surgery, but no other associations were detected. 8 Results from casecontrol studies also indicate that the increase in risk might not occur until years after cholecystectomy. 5 Our results of a weak association between cholecystectomy and proximal colon cancer, and a lack of association with more distally located cancers, are therefore in agreement with the conjunction of previous cohort studies and some case-control studies. Small bowel cancer is a rare tumor, which is probably the underlying reason for the scarcity of previous studies of the risk of small bowel cancer after cholecystectomy. In one hospital-based case-control study in the United States, 3 of 19 patients with small bowel adenocarcinoma (P 0.004) and 2 of 17 patients with carcinoids of the small bowel (P 0.02) had a history of cholecystectomy, compared with none of the 52 control subjects. 15 However, confounding by indication might explain the positive results of this study. In our study, we found a highly increased risk of bowel cancer during the first year after the cholecystectomy (data not shown), which seems to point to the need to evaluate confounding by indication. In addition, a Danish cohort study reported a significantly increased risk factor of 2.6 for small bowel malignancies in cholecystectomized patients. 16 Although neither of these 2 studies allowed detailed analyses, they support our findings of a positive association between cholecystectomy and small bowel cancer. Because of its size, population-based design, and long and virtually complete follow-up, the present study has some advantages compared with previous studies. However, there is a risk of selection bias in our study because of a previously unrecognized bowel cancer as a result of the gallbladder surgery, or a risk that the symptoms of the tumor were confused with gallstone disease. To avoid such bias, we excluded all person-years of follow-up during the first year after surgery. Tobacco smoking may act as a confounder. Therefore, we analyzed the association between cholecystectomy and lung cancer. We found no association (SIR, 1.02; 95% CI, ), which indicates that smokers are not more likely to undergo cholecystectomy. We also attempted to evaluate other potential confounding variables, such as obesity, Crohn s disease, peptic ulcer disease, and diabetes. However, in the case of small bowel cancer, the number of patients with any of these conditions was too small to allow meaningful analyses and none of the 4 listed conditions influenced the association between cholecystectomy and colorectal cancer. Because of a decreased statistical power in the stratification into smaller subsets of data, the CIs included 1.0 in some of the smaller strata. This was particularly evident for small bowel cancer, because of the rarity of this cancer. The weak association between gallstone disease and adenocarcinoma of the small bowel in our comparison cohort might indicate that any changes of the bile associated with gallstone disease, and not cholecystectomy, could have an influence on the risk of cancer. However, we did not find any association with proximal colon cancer. Furthermore, members of this cohort are likely to suffer from more coexisting diseases than the population at large, or members of the cholecystectomy cohort, because patients in the gallstone cohort may not have been suited for surgery. Coexisting diseases might also lead to an increased risk of gastrointestinal cancer. Therefore, we find it probable that the increased risk of adenocarcinoma of the bowel is caused by the cholecystectomy per se and not any changes of the bile associated with gallstone disease. The biological mechanism by which cholecystectomy increases the risk of intestinal cancer might be the elevated intestinal bile concentration and the following increased metabolism of bile. After the removal of the gallbladder, the bile entering the duodenum directly from the liver is less dependent on the intake of food, and gastric juice and food only occasionally dilute the bile. Hence, the concentration of bile is increased locally in the small bowel, particularly in the proximal part. Bacterial degradation of bile salts to secondary bile acids might be pathogenic to the intestinal

6 September 2001 CHOLECYSTECTOMY AND INTESTINAL CANCER 547 mucosa. 9 This change may increase the occurrence of mucosal damage leading to cellular proliferation and eventually an invasive cancer. The increased risk of duodenal adenocarcinoma, and the gradual decrease in risk with increasing distance from the source of bile indicate that bile concentrations might be the major reason for the development of adenocarcinoma of the bowel after cholecystectomy. The more pronounced risk of small bowel adenocarcinoma after surgery of the common bile duct gives further support to this hypothesis. The flow of bile is more unregulated after cholecystectomy following the sphincter damage of the common bile duct (sphincter of Oddi) during the surgery. In conclusion, our study showed an increased risk of cancer of the small bowel and proximal colon after cholecystectomy. No association was found with cancer of the distal colon or rectum. The increasing risk of intestinal cancer with decreasing distance from the common bile duct indicates that high local bile concentrations might be carcinogenic to the intestinal mucosa. From an individual perspective, the rarity of cancer of the small intestine and the weak association with proximal colon cancer makes the absolute risk of developing intestinal cancer after cholecystectomy low. References 1. Almond HR, Vlahcevic ZR, Bell CC, Gregory DH, Swell L. Bile acid pools, kinetics and biliary lipid composition before and after cholecystectomy. N Engl J Med 1973;289: Hepner GW, Hofmann AF, Malagelada JR, Szczepanik PA, Klein PD. Increased bacterial degradation of bile acids in cholecystectomized patients. Gastroenterology 1974;66: Narisawa T, Magadia NE, Weisburger JH, Wynder EL. Promoting effect of bile acids on colon carcinogenesis after intrarectal instillation of N-methyl-N -nitro-n-nitrosguanidine in rats. J Natl Cancer Inst 1974;53: Werner B, deheer K, Mitschke H. Cholecystectomy and carcinoma of the colon: an experimental study. Z Krebsforsch 1977; 88: Giovannucci E, Colditz GA, Stampfer MJ. A meta-analysis of cholecystectomy and risk of colorectal cancer. Gastroenterology 1993;105: Reid FD, Mercer PM, Harrison M, Bates T. Cholecystectomy as a risk factor for colorectal cancer: a meta-analysis. Scand J Gastroenterol 1996;31: Linos DA, Beard CM, O Fallon WM, Dockerty MB, Beart RW Jr, Kurland LT. Cholecystectomy and carcinoma of the colon. Lancet 1981;2: Ekbom A, Yuen J, Adami HO, McLaughlin JK, Chow WH, Persson I, Fraumeni JF Jr. Cholecystectomy and colorectal cancer. Gastroenterology 1993;105: Lowenfels AB. Does bile promote extra-colonic cancer? Lancet 1978;2: Ross RK, Hartnett NM, Bernstein L, Henderson BE. Epidemiology of adenocarcinomas of the small intestine: is bile a small bowel carcinogen? Br J Cancer 1991;63: Weiss NS, Yang CP. Incidence of histologic types of cancer of the small intestine. J Natl Cancer Inst 1987;78: Chow JS, Chen CC, Ahsan H, Neugut AI. A population-based study of the incidence of malignant small bowel tumours: SEER, Int J Epidemiol 1996;25: Severson RK, Schenk M, Gurney JG, Weiss LK, Demers RY. Increasing incidence of adenocarcinomas and carcinoid tumors of the small intestine in adults. Int J Epidemiol 1996;25: DiSario JA, Burt RW, Vargas H, McWhorter WP. Small bowel cancer: epidemiological and clinical characteristics from a population-based registry. Cancer Epidemiol Biomarkers Prev 1996; 5: Chen CC, Neugut AI, Rotterdam H. Risk factors for adenocarcinomas and malignant carcinoids of the small intestine: preliminary findings. Cancer Epidemiol Biomarkers Prev 1994;3: Johansen C, Chow WH, Jorgensen T, Mellemkjaer L, Engholm G, Olsen JH. Risk of colorectal cancer and other cancers in the patients with gallstones. Gut 1996;39: Mattsson B, Rutqvist LE, Wallgren A. Undernotification of diagnosed cancer cases to the Stockholm Cancer Registry. Int J Epidemiol 1985;14: Cancer Registry. Cancer incidence in Sweden National Board of Health and Welfare, Stockholm Cancer Registry. Cancer incidence in Sweden National Board of Health and Welfare, Stockholm Cancer Registry. Cancer incidence in Sweden National Board of Health and Welfare, Stockholm Schottenfeld D, Islam SS. Cancers of the small intestine. In: Schottenfeld D, Fraumeni JF Jr, eds. Cancer epidemiology and prevention. New York: Oxford University, 1996: Bailar JC III, Ederer F. Significance factors for the ratio of a Poisson variable to its expectation. Biometrics 1964;20, Berkson J. Limitations of the application of fourfold table analysis to hospital data. Biomet Bull 1946;2: Adami HO, Meirik O, Gustavsson S, Nyren O, Krusemo UB. Colorectal cancer after cholecystectomy: absence of risk increase within years. Gastroenterology 1983;85: Maringhini A, Moreau JA, Melton LJ, Hench VS, Zinsmeister AR, DiMagno EP. Gallstones, gallbladder cancer, and other gastrointestinal malignancies. Ann Intern Med 1987;107: Adami HO, Krusemo UB, Meirik O. Unaltered risk of colorectal cancer within years of cholecystectomy: updating of a population-based cohort study. Br J Surg 1987;74: Gudmundsson S, Moller TR, Olsson H. Cancer incidence after cholecystectomy a cohort study with 30 years of follow-up. Eur J Surg Oncol 1989;15: Nielsen GP, Theodors A, Tulinius H, Sigvaldason H. Cholecystectomy and colorectal carcinoma: a total-population historical prospective study. Am J Gastroenterol 1991;86: Received December 14, Accepted May 2, Address requests for reprints to: Jesper Lagergren, M.D., Ph.D, Department of Surgery, Karolinska Hospital, SE Stockholm, Sweden. Jesper.Lagergren@ks.se; fax: (46) The authors thank Li Yin for skillful help with programming.