Journal of the National Cancer Institute Advance Access originally published online on May 13, 2008
JNCI Journal of the National Cancer Institute 2008 100(10):745-750; doi:10.1093/jnci/djn102
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© The Author 2008. Published by Oxford University Press.
BRIEF COMMUNICATION |
Endoscopic Sphincterotomy and Long-Term Risk of Cholangiocarcinoma: A Population-Based Follow-up Study
Affiliations of authors: Departments of Surgical Gastroenterology L (FVM, PFJ, LSJ), Clinical Epidemiology (PJ, HTS), and Medicine V (Hepatology and Gastroenterology) (HTS), Aarhus University Hospital, Aarhus, Denmark; International Epidemiology Institute, Rockville, MD (RET)
Correspondence to: Frank Viborg Mortensen, DMSc, Department of Surgical Gastroenterology L, Aarhus University Hospital, Noerrebrogade 44, 8000 Aarhus C, Denmark (e-mail: fvmor{at}as.aaa.dk).
 |
ABSTRACT |
|---|
 Top Abstract Context and CaveatsFundingReferencesNotes |
|---|
Sphincterotomy of the ampulla of Vater—a common diagnostic and therapeutic procedure that is sometimes done during endoscopic retrograde cholangiography (ERC)—allows reflux of intestinal content into the biliary tree. The resulting inflammation may contribute to malignant transformation of the biliary epithelium and therefore increase the risk of cholangiocarcinoma. We used data from population-based Danish health-care registries to examine the incidence of cholangiocarcinoma after ERC for 10 690 ERC patients who underwent sphincterotomy between 1977 and 2003 and 10 690 ERC patients who did not undergo sphincterotomy. Patients with sphincterotomy were matched to patients without sphincterotomy by sex and age at, calendar year of, and indication for ERC. The cholangiocarcinoma incidence rate for sphincterotomy patients was 404 per 100 000 person-years during the first year after ERC and decreased progressively at later times after ERC (79, 42, and 27 per 100 000 person-years during years 2, 3–5, and >5, respectively). The corresponding rates for patients without sphincterotomy were 458, 12, 10, and 19 per 100 000 person-years, respectively. The gradual decrease in cholangiocarcinoma rate over time after ERC for sphincterotomy patients indicates that some of these patients had a cholangiocarcinoma that was present at the time of ERC but not diagnosed until 2–5 years later. The similar rates at the latest times after ERC suggest the lack of a causal association between sphincterotomy and cholangiocarcinoma.
Prior knowledge Sphincterotomy of the ampulla of Vater during endoscopic retrograde cholangiography (ERC) is a common diagnostic and therapeutic procedure that allows reflux of intestinal content into the biliary tree, which results in inflammation that may contribute to malignant transformation of the biliary epithelium and, therefore, increase the risk of cholangiocarcinoma. Study design A population-based cohort study using data from Danish health-care registries to examine the incidence of cholangiocarcinoma among ERC patients who did and did not undergo sphincterotomy. Contribution The cholangiocarcinoma incidence rates for sphincterotomy patients were 404 per 100 000 person-years during the first year after ERC and decreased progressively at later times after ERC (79, 42, and 27 per 100 000 person-years during years 2, 3–5, and >5, respectively). The corresponding rates for patients without sphincterotomy were 458, 12, 10, and 19 per 100 000 person-years, respectively. Implications Reflux after sphincterotomy does not appear to have a causal role in the development of cholangiocarcinoma. Limitations The number of cholangiocarcinomas was small. The association between sphincterotomy and risk of cholangiocarcinoma could be confounded by unknown or unmeasured risk factors for cholangiocarcinoma.
|
Endoscopic retrograde cholangiography (ERC) is a procedure used to diagnose disorders of the bile ducts, pancreas, and gallbladder. Upon diagnosis, treatment may involve sphincterotomy of the ampulla of Vater, which is done during the same procedure (1–3). However, sphincterotomy allows reflux of intestinal content into the biliary tree, which can lead to chronic inflammation and malignant transformation of the biliary epithelium (4–6). We compared cholangiocarcinoma incidence rates among ERC patients who had and had not undergone sphincterotomy.
We used the National Patient Registry to identify all Danish citizens who underwent an ERC in Denmark between January 1, 1977, and December 31, 2003 (see Appendix Table 1 for all procedure and diagnostic codes). This registry contains data from every hospitalization in Danish hospitals after 1977 and from outpatient visits after 1995. The data include the dates of hospital admission and discharge, surgical procedures, primary diagnosis, and secondary diagnoses. For patients who had undergone ERC multiple times, we considered only the earliest ERC. We excluded patients who had been diagnosed with cholangiocarcinoma before or at the time of ERC and patients who underwent ERC because cancer was suspected; all remaining patients were included in this study. Cholangiocarcinoma diagnoses recorded in the National Patient Registry were supplemented with those recorded in the Danish Cancer Registry. This registry, established in 1943, is notified of every cancer diagnosed in Denmark, including its diagnosis date, a diagnosis code, and topography and histology codes. Data from the two registries were linked through the Central Office of Civil Registration, which records the sex and dates of birth, death, and emigration of all Danish citizens.
We stratified ERC patients by sex, 5-year age interval at ERC, calendar year of ERC in 5- or 6-year intervals (1977–1982, 1983–1988, 1989–1993, 1994–1998, or 1999–2003), and indication for ERC (ie, first three characters of the primary diagnosis code given at the time of ERC). Within each stratum, each sphincterotomy patient was matched with a patient without sphincterotomy. All unmatched patients were excluded from the analyses, as were all patients in strata with fewer than 10 matched sphincterotomy patients.
We used the primary and secondary diagnoses recorded in the National Patient Registry to examine whether ERC patients had been diagnosed with primary sclerosing cholangitis, liver cirrhosis, chronic infection with hepatitis B or C virus, diabetes, or inflammatory bowel disease, all of which are risk factors for cholangiocarcinoma (7,8), before ERC.
For all ERC patients, follow-up began on the date of discharge after hospitalization for ERC and ended on the date of the first cholangiocarcinoma diagnosis, death from any cause, or emigration or on December 31, 2003, whichever occurred first. Patients who had ERC without sphincterotomy were censored at their first sphincterotomy during follow-up. Cholangiocarcinomas were categorized as histologically verified intrahepatic, histologically verified extrahepatic, or unspecified.
We computed the cumulative risks of cholangiocarcinoma and death from any cause for ERC patients and for 10 control subjects from the general population of Denmark per sphincterotomy patient (9); the control subjects were matched to the ERC patients by sex and brith year and were identified through the Central Office of Civil Registration. We used a proportional subdistribution hazards model (10) to estimate the cholangiocarcinoma incidence rate ratio for ERC patients with sphincterotomy compared with ERC patients without sphincterotomy. The model included sphincterotomy, age at ERC, sex, calendar year of ERC, indication for ERC, and whether ERC patients had primary sclerosing cholangitis, liver cirrhosis, chronic hepatitis B or C infection, diabetes, or inflammatory bowel disease. The assumption of proportional subdistribution hazards was examined with Schoenfeld residuals, as described by Fine and Gray (10), and tested by including the interaction between sphincterotomy and log follow-up time in the model. The examination with Schoenfeld residuals indicated that the hazards were not proportional over the entire follow-up period, as seen in the right-hand panel of Figure 1: the curve is not a horizontal line. Hence, we divided the follow-up period into four periods (year 1, year 2, years 3–5, and >5 years after ERC). The test based on the interaction term indicated that hazards were proportional within each of these four time periods. We examined whether any variable in the model had a statistically significant (P < .05) interaction with sphincterotomy. Finally, we repeated the matching of ERC patients with and without sphincterotomy and the subsequent analyses 1000 times and described the resulting distribution of cholangiocarcinoma incidence rate ratios.
|
A total of 10 690 ERC patients with sphincterotomy and 10 690 ERC patients without sphincterotomy were included in this analysis. Characteristics of the study population are given in Table 1. The median age of all patients at ERC was 67 years (range = 15–94 years); during follow-up, 7791 patients (36%) died and 111 patients (0.52%) were diagnosed with cholangiocarcinoma. Patients with and without sphincterotomy had similar all-cause mortality rates throughout follow-up (data not shown). During year 1 after ERC, the cholangiocarcinoma incidence rate was increased to a similar extent for all ERC patients compared with the rate for the matched population control subjects (>400 vs 12 per 100 000 person-years) (Table 2). During year 2 and years 3–5 after ERC, the cholangiocarcinoma incidence rate for patients without sphincterotomy was similar to the general population rate (12 and 10, respectively, vs 12 per 100 000 person-years), whereas the rate for sphincterotomy patients remained elevated although it decreased from 79 per 100 000 person-years during year 2 after ERC to 42 per 100 000 person-years during years 3–5 after ERC (Table 2). More than 5 years after ERC, the cholangiocarcinoma incidence rate for sphincterotomy patients was similar to that for ERC patients without sphincterotomy (27 vs 19 per 100 000 person-years) and slightly higher than the rate for the matched population control subjects (Table 2 and Figure 1). Such similar incidence rates for ERC patients with and without sphincterotomy were seen for all three categories of cholangiocarcinoma at more than 5 years after ERC. Confounding was minimal on account of the matching of ERC patients and the small number of cholangiocarcinomas among patients with risk factors (Table 1), and no interactions were statistically significant. Consequently, the adjusted incidence rate ratios for sphincterotomy patients vs ERC patients without sphincterotomy were similar to the ratio of their crude incidence rates, and they were therefore dependent on the time since ERC, as indicated by the Schoenfeld residuals (Table 2 and Figure 1). Subdistribution hazards were proportional within all follow-up periods. In the 1000 samplings, the median (5th percentile, 95th percentile) adjusted incidence rate ratios of cholangiocarcinoma during years 1, 2, 3–5, and more than 5 years after ERC were 0.88 (0.76, 1.01), 6.14 (5.14, 7.87), 3.99 (3.06, 5.26), and 1.67 (1.19, 2.23), respectively.
|
|
Our finding of an elevated cholangiocarcinoma incidence rate during years 2–5 after sphincterotomy differs from results of previous smaller studies [reviewed in (11)]. The largest study to date (12) followed 878 Swedish patients starting at 1 year after sphincterotomy; none of the patients developed cholangiocarcinoma during a median follow-up time of 6 years. Given our incidence estimates for sphincterotomy patients, we would have expected two or three cholangiocarcinomas to have been diagnosed in that cohort.
Our findings depend on the validity of registry data. Virtually all surgical procedures and cancer diagnoses are registered correctly, and the small proportion of registered procedures and cancer diagnoses that are incorrect is considered to be acceptable (13–15). Erroneous sphincterotomy or cholangiocarcinoma registration would bias incidence rate ratios toward 1 (16) and thus cannot explain the elevated cholangiocarcinoma incidence rates we observed for the sphincterotomy patients (vs ERC patients without sphincterotomy) in this study.
The very high cholangiocarcinoma incidence rate during year 1 after ERC indicates that many ERC patients had a preexisting but undiagnosed cholangiocarcinoma at the time of their ERC. The elevated but decreasing rate for sphincterotomy patients during years 2–5 after ERC indicates that a sizable proportion of these existing undiagnosed cholangiocarcinomas was not diagnosed until year 2 after ERC and that a smaller proportion was not diagnosed until years 3–5 after ERC. It is unclear why undiagnosed cholangiocarcinomas might be more frequent among sphincterotomy patients, but they could not have been caused by reflux after sphincterotomy. The fact that the cholangiocarcinoma incidence rate for ERC patients without sphincterotomy declined during years 2–5 after ERC and then increased at later times indicates that diagnostic procedures that were performed during year 1 after ERC led to an earlier diagnosis of cholangiocarcinomas that otherwise would have been diagnosed during years 2–5 after ERC. The fact that cholangiocarcinoma incidence rates were similar for ERC patients with and without sphincterotomy from 5 years after ERC indicates that reflux after sphincterotomy does not have a causal role in the development of these cancers.
The strengths of our study were the large number of patients, the complete long-term follow-up, and the population-based, matched study design. The major limitation was the small number of cholangiocarcinomas. In addition, the association between sphincterotomy and risk of cholangiocarcinoma could be confounded by unknown or unmeasured risk factors for cholangiocarcinoma, but the lack of an association between sphincterotomy and mortality rules out any risk factor that is also associated with mortality. Therefore, smoking and obesity, which may increase the risk of cholangiocarcinoma (7,8), cannot explain our findings.
In conclusion, sphincterotomy did not appear to be causally related to cholangiocarcinoma in this study.
|
Funding |
|---|
|
TopAbstractContext and CaveatsFundingReferencesNotes |
|---|
Karen Elise Jensens Fond; Western Danish Research Forum for Health Sciences (Vestdansk Forskningsforum).
Appendix Table 1. Definitions of procedures and diseases as coded in registry data*
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Procedure codes in the National Patient Registry are coded according to a Danish classification of surgical procedures through 1995 (17) and according to the Danish version of the Nordic Classification of Surgical Procedures thereafter (18). Hospital diagnosis codes in the National Patient Registry are coded according to the International Classification of Diseases (ICD), 8th Edition, through 1994 (19), and according to the 10th edition thereafter (20). In the Cancer Registry, cancer topography and histology codes are coded according to the 1st edition of the International Classification of Diseases for Oncology (ICD-O-1) (21), and cancer diagnosis codes are coded according to the 7th edition of the ICD (22). For example, we defined histologically verified intrahepatic cholangiocarcinoma by topography code 155.0 in combination with histology code 8160.3 or 8161.3 or topography code 155.1 in combination with histology code 8140.3, 8160.3, 8161.3, 8020.3, or 8010.3. ERC = endoscopic retrograde cholangiography.
|
NOTES |
|---|
|
TopAbstractContext and CaveatsFundingReferencesNotes |
|---|
The sponsors had no role in the design of the study; the collection, analysis, or interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication.
|
REFERENCES |
|---|
|
TopAbstractContext and CaveatsFundingReferencesNotes |
|---|
1. McCune WS, Shorb PE, Moscovitz H. Endoscopic cannulation of the ampulla of Vater: a preliminary report. Ann Surg (1968) 167(5):752–756.[Web of Science][Medline]
2. Kawai K, Akasaka Y, Murakami K, Tada M, Koli Y. Endoscopic sphincterotomy of the ampulla of Vater. Gastrointest Endosc (1974) 20(4):148–151.[Web of Science][Medline]
3. Classen M, Demling L. Endoskopische Sphincterotomie der Papilla Vateri und Stein Extraktion aus dem Ductus Choledochus [in German]. Dtsch Med Wochenschr (1974) 99(11):496–497.[Medline]
4. Bergman JJ, van Berkel AM, Groen AK, et al. Biliary manometry, bacterial characteristics, bile composition, and histologic changes fifteen to seventeen years after endoscopic sphincterotomy. Gastrointest Endosc (1997) 45(5):400–405.[CrossRef][Web of Science][Medline]
5. Kurumado K, Nagai T, Kondo Y, Abe H. Long-term observations on morphological changes of choledochal epithelium after choledochoenterostomy in rats. Dig Dis Sci. (1994) 39(4):809–820.[CrossRef][Web of Science][Medline]
6. Berthiaume EP, Wands J. The molecular pathogenesis of cholangiocarcinoma. Semin Liver Dis (2004) 24(2):127–137.[CrossRef][Web of Science][Medline]
7. Welzel TM, Graubard BI, El-Serag HB, et al. Risk factors for intrahepatic and extrahepatic cholangiocarcinoma in the United States: a population-based case-control study. Clin Gastroenterol Hepatol (2007) 5(10):1221–1228.[CrossRef][Web of Science][Medline]
8. Welzel TM, Mellemkjær L, Gloria G, et al. Risk factors for intrahepatic cholangiocarcinoma in a low-risk population: a nationwide case-control study. Int J Cancer (2007) 120(3):638–641.[CrossRef][Web of Science][Medline]
9. Marubini E, Valsecchi MG. Competing risks. In: Analysing Survival Data From Clinical Trials and Observational Studies (1995) 1st ed. Chichester, UK: John Wiley & Sons. 331–363.
10. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. (1999) 94(446):496–509.[CrossRef][Web of Science]
11. Frimberger E. Long-term sequelae of endoscopic papillotomy. Endoscopy (1998) 30(9):A221–A227.[Medline]
12. Karlson BM, Ekbom A, Arvidsson D, Yuen J, Krusemo UB. Population-based study of cancer risk and relative survival following sphincterotomy for stones in the common bile duct. Br J Surg (1997) 84(9):1235–1238.[CrossRef][Web of Science][Medline]
13. Nickelsen TN. Data validity and coverage in the Danish National Health Registry. A literature review [in Danish]. Ugeskr Laeger (2001) 164(1):33–37.[Medline]
14. Wille-Jørgensen PA, Meisner S. The validity of data in registration of operations. A quality analysis [in Danish]. Ugeskr Laeger (1997) 159(49):7328–7330.[Medline]
15. Storm HH, Michelsen EV, Clemmensen IH, Pihl J. The Danish Cancer Registry—history, content, quality and use. Dan Med Bull. (1997) 44(5):535–539.[Web of Science][Medline]
16. Rothman KJ, Greenland S. Precision and validity in epidemiologic studies. In: Modern Epidemiology—Rothman KJ, Greenland S, eds. (1998) Philadelphia, PA: Lippincott Williams & Wilkins. 115–134.
17. Danish National Board of Health. Operations—og behandlingsklassifikation (1988) 3rd ed. Copenhagen, Denmark: Danish National Board of Health.
18. Danish National Board of Health. Klassifikation af operationer (1995) 1st ed. Copenhagen, Denmark: Danish National Board of Health.
19. Danish National Board of Health. Klassifikation af sygdomme—dansk-latinsk udgave af verdenssundhedsorganisationens internationale klassifikation af sygdomme, skader og dødsårsager (8. Revision 1965) (1971) Copenhagen, Denmark: Danish National Board of Health.
20. Danish National Board of Health. Klassifikation af sygdomme (10. Revision) (1992) Copenhagen, Denmark: Danish National Board of Health.
21. World Health Organization. Manual of the International Classification of Diseases for Oncology (1976) 1st ed. Geneva, Switzerland: World Health Organization.
22. World Health Organization. Manual of the International Statistical Classification of Diseases, Injuries and Causes of Death. Based on the Recommendations of the Seventh Revision Conference, 1955, and Adopted by the Ninth World Health Assembly Under the WHO Nomenclature Regulations (1957) Geneva, Switzerland: World Health Organization.
Manuscript received October 18, 2007; revised February 22, 2008; accepted March 11, 2008.
Related Article in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2008 100: 687.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||