The Risk of Liver and Bile Duct Cancer in Patients With Chronic Viral Hepatitis, Alcoholism, or Cirrhosis

Transcription

1 The Risk of Liver and Bile Duct Cancer in Patients With Chronic Viral Hepatitis, Alcoholism, or Cirrhosis HANNAH KUPER, 1 WEIMIN YE, 2 ULRIKA BROOMÉ, 3 ANDERS ROMELSJÖ, 4 LORELEI A. MUCCI, 5 ANDERS EKBOM, 2,5 HANS-OLOV ADAMI, 2,5 DIMITRIOS TRICHOPOULOS, 5 AND OLOF NYRÉN 2 No prospective study has analyzed simultaneously chronic viral hepatitis and alcoholism as risk factors for liver carcinogenesis, while taking into consideration the role of cirrhosis. Nor has the risk for hepatocellular among patients with chronic viral hepatitis been prospectively evaluated in a low-risk Western population. Last, the relationship between hepatocellular risk factors and bile duct cancer remains to be clarified. We analyzed prospectively the risk for primary liver and extrahepatic biliary tract cancer among 186,395 patients hospitalized with either chronic viral hepatitis, alcoholism, cirrhosis, or any combination of these conditions through linkages between national Swedish registers. Compared with the general population, the relative risk of hepatocellular was 34.4 for chronic viral hepatitis alone, 2.4 for alcoholism alone, and 40.7 for cirrhosis alone. Among patients with combinations of these risk conditions, the relative risk of hepatocellular was 27.3 for chronic viral hepatitis and alcoholism, for chronic viral hepatitis and cirrhosis, 22.4 for alcoholism and cirrhosis, and for all 3 conditions. We found limited evidence for an excess risk of intrahepatic, but not for extrahepatic, biliary duct cancer. Cirrhosis amplifies the risk of hepatocellular among patients with chronic viral hepatitis, but it is not a prerequisite for liver carcinogenesis. In contrast, cirrhosis may be a necessary intermediate for the development of hepatocellular among alcoholics. (HEPATOLOGY 2001;34: ) Abbreviations: ICD-7, seventh revision of the International Classification of Diseases; ICD-8, eighth revision of the International Classification of Diseases; ICD-9, ninth revision of the International Classification of Diseases; PAD, a histopathologic code according to ICD-7/WHO/HS/CANC/24.1 coding principles. From the 1 Department of Epidemiology and Public Health, University College London, London, UK; 2 Department of Medical Epidemiology, Karolinska Institutet, Stockholm, Sweden; 3 Department of Gastroenterology, Huddinge Hospital, Stockholm, Sweden; 4 Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden; and 5 Department of Epidemiology, Harvard School of Public Health, Boston, MA. Received June 21, 2001; accepted July 30, Address reprint requests to: Hannah Kuper, Sc.D., Department of Epidemiology and Public Health, University College London, 1-19 Torrington Place, London WC1E 6BT, England. hannahk@public-health.ucl.ac.uk; fax: (44) Copyright 2001 by the American Association for the Study of Liver Diseases /01/ $35.00/0 doi: /jhep Primary liver cancer is the fifth most common malignancy globally. 1 Interest in clarifying the pathogenesis of hepatocellular, by far the most common form of primary liver cancer, has grown since the recent report showing a significant increase in incidence in the United States. 2 Evidence has accumulated that chronic infection with hepatitis B and C viruses is causally related to hepatocellular, 3 and that excessive alcohol use elevates risk for this cancer. 4,5 However, the following questions about hepatocellular etiology have not been adequately addressed: Is cirrhosis a necessary intermediate in liver carcinogenesis related to chronic viral hepatitis or alcohol abuse? Is viral hepatocarcinogenesis enhanced in the presence of cirrhosis? Do chronic viral hepatitis and alcohol abuse act synergistically to increase hepatocellular risk over time? In addition, little is known about the role of chronic viral infection, alcoholism, and cirrhosis in the etiology of intrahepatic and various forms of extrahepatic bile duct cancer. We conducted a historical cohort study to analyze the risk for developing primary liver and extrahepatic biliary tract cancer among groups of patients with chronic viral hepatitis, alcoholism, cirrhosis, or any combination of these conditions. Complete long-term follow-up was achieved through linkage between nationwide Swedish databases. MATERIALS AND METHODS The Study Population. From 1965 onwards, the Swedish National Board of Health and Welfare began compiling data on individual hospital discharges from in-patient medical institutions as part of the In-Patient Register, as previously described. 6 Private in-patient treatment is rare in Sweden, and hospital-provided medical services are, in effect, population based. Each record in the Register, corresponding to one in-hospital episode, includes the national registration number (a unique identifier assigned to all Swedish residents), date of discharge, surgical procedures, and up to 8 medical conditions. Discharge diagnoses were coded according to the seventh revision of the International Classification of Diseases (ICD-7) until 1968, the eighth revision (ICD-8) until 1987, and the ninth revision (ICD-9) until the end of this study. The Cohorts. We considered all In-Patient Register records through 1994 with a hospital discharge diagnosis of chronic viral hepatitis, alcoholism, or cirrhosis using ICD codes specified in the footnotes of Table 1. Diagnosis of chronic viral hepatitis was defined as having at least 2 hospitalizations for viral hepatitis more than 6 months apart. After exclusion of records with erroneous national registration numbers, our search yielded complete records on 186,395 subjects who were diagnosed with at least one of the above conditions and did not have a prior history of cancer. Subjects were then classified into cohorts based on having 1 of 7 combinations of chronic viral hepatitis, alcoholism, or cirrhosis (Table 1). Entry into a cohort was defined as the date of discharge after the first hospitalization. By using record linkage to other national registries, we obtained additional information on study subjects, including dates of death from the Register of Causes of Death and dates of emigration from the Register of Population Migration. We used the Cancer Registry to identify all incident cancers diagnosed in the cohort during fol- 714

2 HEPATOLOGY Vol. 34, No. 4, 2001 KUPER ET AL. 715 TABLE 1. Descriptive Characteristics of the Study Cohorts Single Diagnoses Combination of Diagnoses Chronic Viral Hepatitis* Alcoholism Cirrhosis Hepatitis/ Alcoholism Chronic Viral Hepatitis/ Alcoholism/ Cirrhosis Cirrhosis All Three Conditions Total Number of patients 3, ,872 8,939 1, , ,395 Number of women 1,701 33,436 3, , ,162 Person-years at risk 32,775 1,767,669 61,605 20,934 1,245 75,582 2,415 1,962,225 Mean age at entry Mean duration of follow-up Mean calendar year at entry Cancer type or site (ICD-7 code; PAD code ) Primary liver and biliary duct cancer (ICD7:155) Primary liver cancer (ICD7:155.0) Hepatocellular (PAD 066) Intrahepatic bile duct (PAD 076) Mixed hepatocellular and bile duct (PAD 096) Gallbladder (ICD7:155.1) Extrahepatic bile duct (ICD7:155.2) Carcinoma of ampulla of Vater (ICD7:155.3) *Chronic hepatitis defined by ICD-7 code 092; ICD-8 code 070; ICD-9 code 070. Alcoholism defined by ICD-7 code 307, 322; ICD-8 code 291, 303; ICD-9 code 291, 303, 305A. Cirrhosis defined by ICD-7 code 58100, 58110, 58310; ICD-8 code 57100, 57101, 57190, 57198; ICD-9 code 571A, 571C, 571D, 571F, 571G. The observed number of cases by subclassification may not add up to total number of cases of primary liver and biliary duct cancers because of lack of subtype information on ICD or PAD codes. PAD is a histopathologic code according to ICD-7/WHO/HS/CANC/24.1 coding principles. low-up through Observation time was calculated from 1 year after date of entry into the cohort until the occurrence of a diagnosis of any first primary cancer, or censoring on account of emigration, death, or end of the observation period (December 31, 1995). In all analyses, we excluded all person-time that elapsed during the first year of follow-up, as well as all hepatobiliary cancers that occurred during that same period, to minimize selection bias. Such a bias could result in spurious relationships between nonmalignant and malignant conditions. 7 Statistical Analyses. The Swedish Cancer Registry has coded malignant diseases according to the ICD-7 classification scheme throughout the entire study period. Primary liver cancer, ICD-7: 155, is further classified anatomically, and then histologically based on a histopathologic diagnosis (PAD). By using these classifications, we distinguished the following primary liver and extrahepatic biliary duct cancer outcomes: primary hepatocellular, intrahepatic bile duct, mixed hepatocellular and bile duct, gallbladder, extrahepatic bile duct, and cancer of the ampulla of Vater (Table 1). The standardized incidence ratio, the ratio of observed number of cancers to those expected, was used as a measure of relative risk. To calculate the expected number of cancers, we multiplied the number of observed person-years in each of the 7 cohorts, (stratified by gender, 5-year age groups, and calendar year) by stratum-specific cancer incidence rates derived from the entire Swedish population. The 95% confidence interval for each ratio was calculated on the assumption that the observed cases follow a Poisson distribution. We excluded cancer cases that were detected incidentally at autopsy to avoid bias caused by differential autopsy rates of patients in these cohorts as compared with the general population. We further stratified the estimates by gender, calendar year of entry, duration of follow-up, and age at entry to assess whether these characteristics modified the association between the risk conditions on the one hand and liver and biliary duct cancer risk on the other. By using Kaplan-Meier methods, we calculated the cumulative probability of developing hepatocellular cancer over a 15-year period for subjects with chronic viral hepatitis or alcohol abuse, considering the conditions alone or with concomitant cirrhosis. In addition to Kaplan-Meier curves, we calculated point estimates and 95% confidence intervals for the 15-year cumulative incidence. RESULTS A total of 186,395 patients were followed for 1,962,225 person-years. The descriptive characteristics and distribution of cancer subtypes in the 7 subcohorts are given in Table 1. In the total cohort, 490 primary liver and biliary duct cancers were observed during follow-up an average of 10.5 years. Hepatocellular was by far the most common type of cancer, whereas few cases of cancer of other types were observed. The standardized incidence ratios showing the effect of the risk factors and their combinations on primary liver and biliary duct cancer are summarized in Table 2. Chronic viral hepatitis without evidence of cirrhosis was characterized by a 34.4 relative risk for hepatocellular. There was no indication that any type of extrahepatic or intrahepatic bile duct cancer was affected by chronic viral hepatitis because the mixed hepatocellular, for which there was a risk elevation, is likely to have originated from hepatic cells. Coexistence of chronic viral hepatitis and alcoholism without evidence of cirrhosis was not associated with additional risk for hepatocellular over and beyond that linked to chronic hepatitis alone, nor did the combination appear to affect risk for intrahepatic or extrahepatic bile duct cancer. In contrast, chronic viral hepatitis in the presence of cirrhosis interacted in a super-additive way, increasing the risk of HCC 119-fold and that of mixed hepatocellular and bile duct almost 115-fold. Single or combined exposures to these conditions did not appear to affect the risk for extrahepatic bile duct.

3 716 KUPER ET AL. HEPATOLOGY October 2001 TABLE 2. Standardized Incidence Ratios (SIR) for Primary Liver Cancer and Extrahepatic Biliary Tract Cancers Among Patients With Chronic Hepatitis, Alcoholism, or Cirrhosis, as Well as Patients With any Combination of These Conditions Single Diagnoses Combination of Diagnoses Hepatitis/Cirrhosis Alcoholism/Cirrhosis All Three Conditions Hepatitis/Alcoholism Hepatitis Alcoholism Cirrhosis SIR (95% CI) SIR (95% CI) SIR (95% CI) SIR (95% CI) SIR (95% CI) SIR (95% CI) SIR (95% CI) Cancer type or site Primary liver and biliary duct cancer 9.0 ( ) 1.6 ( ) 12.1 ( ) 13.0 ( ) 38.8 ( ) 7.9 ( ) 50.0 ( ) Primary liver cancer 23.3 ( ) 2.3 ( ) 31.2 ( ) 16.2 ( ) 96.5 ( ) 16.5 ( ) ( ) Hepatocellular 34.4 ( ) 2.4 ( ) 40.7 ( ) 27.3 ( ) ( ) 22.4 ( ) ( ) Intrahepatic bile duct 0 (0-37.5) 2.0 ( ) 8.9 ( ) 0 ( ) 0 ( ) 5.4 ( ) 0 ( ) 9.1 ( ) 2.5 ( ) 18.8 ( ) 0 ( ) ( ) 7.1 ( ) ) Mixed hepatocellular and bile duct Gallbladder 0 (0-6.2) 0.8 ( ) 0.7 ( ) 0 (0-68.0) 0 (0-78.7) 0 (0-1.3) 0 ( ) Extrahepatic bile duct 0 (0-19.0) 1.3 ( ) 0 (0-2.8) 0 ( ) 0 ( ) 3.4 ( ) 0 ( ) Carcinoma of ampulla of 0 (0-42.7) 1.4 ( ) 1.6 ( ) 0 ( ) 0 ( ) 1.8 (0-9.8) 0 ( ) Vater NOTE. SIR is stratified by age, gender, and calendar year and standardized to cancer rates in the general population. Abbreviation: 95% CI, 95% confidence interval. Alcoholism in the absence of chronic hepatitis and without evidence of cirrhosis was associated with excess risks of hepatocellular, intrahepatic bile duct, and mixed type liver cancer, but the increase was moderate and could be explained by the presence of preclinical cirrhosis. Alcohol-induced cirrhosis, on the other hand, was strongly associated with risk of hepatocellular as well as other types of liver cancer, and was even linked to a nonsignificant excess risk of extrahepatic bile duct. Cirrhosis in the absence of other conditions was linked with a more than 40-fold increased risk of hepatocellular, an almost 9-fold increased risk of intrahepatic bile duct cancer, as well as a 19-fold increased risk of mixed-type primary liver cancer. There was no indication that cirrhosis increased the risk of extrahepatic bile duct cancer. There were very few patients with concurrent chronic viral hepatitis, alcoholism, and cirrhosis, and among these patients the risk profile was similar to that of patients with chronic viral hepatitis combined with cirrhosis. In further stratification, we explored whether gender, calendar year of entry, duration of follow-up, or age at entry modified the effects of the 3 principal factors under consideration. No substantial effect of gender, calendar year of entry, and duration of follow-up was evident. The older the patient was at entry, however, the lower the excess risk for hepatocellular, particularly when considering chronic viral hepatitis and/or cirrhosis. Figure 1 shows Kaplan-Meier estimates of the cumulative risk of hepatocellular over a 15-year period among patients with chronic hepatitis alone, chronic hepatitis with cirrhosis, alcoholism alone, and alcoholism with cirrhosis. The probability of developing hepatocellular after 15 years of observation for subjects with chronic hepatitis alone was 0.5% (95% confidence interval, 0.2%-0.9%) and for subjects with alcoholism alone was 0.1% (95% confidence interval, 0.08%-0.13%). The cumulative risk for hepatocellular was substantially higher when cirrhosis complicated chronic hepatitis (6.2%; 95% confidence interval, 1%-12.5%) and moderately higher for alcohol-related cirrhosis (1.1%; 95% confidence interval, 0.8%-1.5%). Interpretation of these estimates of overall cumulative incidence should take into consideration the substantial age differences across the 4 subcohorts. In particular, the relative risk associated with chronic viral hepatitis alone (34.4) was higher than that associated with alcoholism and cirrhosis (22.4), yet the cumulative incidence for the former subgroup was only half of that for cirrhosis-related alcoholism. This is mostly because of the more than 14-year difference in mean age at entry into the respective subcohorts (mean age 34.3 and 48.7, respectively). DISCUSSION The present investigation has several advantages: it is large, of prospective nature, and essentially without losses to followup. Misclassification in the diagnosis of alcoholism, chronic viral hepatitis, and particularly cirrhosis is possible. Because such misclassification is essentially nondifferential, it would likely attenuate the true associations with the corresponding risk conditions. Moreover, the clinical value of these diagnoses as risk markers is unaffected because the same misclassification occurs in clinical practice. Use of nonroutine diagnostic methods is a prerequisite for ascertainment of cirrhosis, and thus it is conceivable that patients subject to such diagnostic efforts would remain under closer surveillance than

4 HEPATOLOGY Vol. 34, No. 4, 2001 KUPER ET AL. 717 FIG. 1. Kaplan-Meier estimates of the cumulative probability (per thousand) of developing hepatocellular among patients with chronic viral hepatitis, or alcoholism, with or without cirrhosis. patients with uncomplicated alcoholism or chronic viral hepatitis. This possible detection bias is, however, unlikely to explain excesses of risk of the magnitude we saw in our study. Moreover, we did not include any cancer cases found incidentally at autopsy, eliminating the possibility of this bias. In our study, the most important causes for hepatocellular in developed countries have been simultaneously investigated with adequate control of confounding. Assuming that the small excess risk of HCC among alcoholics without evident cirrhosis is explained mainly by residual confounding by unrecognized cirrhosis, our results indicate that cirrhosis might be a necessary intermediate in alcohol-dependent liver hepatocarcinogenesis. Viral hepatocarcinogenesis, on the other hand, can exist without overt cirrhosis, but subjects with concomitant cirrhosis may be at substantially higher risk. Alcoholism and cirrhosis, either separately or in combination, are not importantly associated with s of the gall bladder or the extrahepatic bile duct. Moreover, alcoholism appears unrelated to risk of of the ampulla of Vater. Despite the more than 50,000 person-years of observation, we did not detect a single case of nonhepatocellular in the chronic viral hepatitis cohort alone or in combination with the other risk conditions. Case-control and cohort studies are compatible with our findings that cirrhosis, alcoholism, and chronic viral hepatitis are strongly associated with hepatocellular. 3,4,8-11 Our results provide support to the hypothesis that cirrhosis is a necessary intermediate step in the pathway from alcoholism to hepatocellular. The increase in risk of hepatocellular among subjects with chronic viral hepatitis even in the absence of cirrhosis is reasonable, because hepatitis B, and conceivably hepatitis C, can cause hepatocellular through insertional mutagenesis or expression of oncogenic viral sequences Investigators have proposed the existence of effect modification between alcoholism and hepatitis viral infection in the causation of hepatocellular 14 possibly by increasing the progression to cirrhosis. 15 Our results do not support the existence of such an effect modification. Little is known about the cause of intrahepatic bile duct, outside of its established association with liver flukes, 16,17 and primary sclerosing cholangitis. 18,19 In the present investigation, we found a positive association with cirrhosis, an association that has also been considered by other investigators We have also found a weaker, albeit significant, association with alcoholism, even in the absence of cirrhosis. A relationship between intrahepatic bile duct and alcohol consumption has been found by some 19,20 but not all 18 investigators. It has been proposed that alcohol could increase the transition from primary sclerosing cholangitis to cancer. 19 There is, however, the possibility that the apparent effect in our study was caused by misclassification of hepatocellular to cholangio or mixed histologic type within the hepatic parenchyma, because hepatocellular is the dominating subtype. The etiology of extrahepatic biliary tract cancer is largely unknown, apart from the importance of gallstones. 23,24 Contradictory results have been reported concerning alcohol consumption in relation to cancers of the extrahepatic biliary tract, some studies suggesting a causal effect, 20,25 others indicating an inverse association, and still others pointing to a lack of association. 24,29 Our large study provides no evidence

5 718 KUPER ET AL. HEPATOLOGY October 2001 that alcoholism, either alone or in concert with cirrhosis, is an important cause of cancer of the extrahepatic bile duct. In conclusion, our study shows that while the cumulative risk of hepatocellular over a 15-year period is clinically insignificant in uncomplicated alcoholism, the presence of cirrhosis is associated with a substantial risk that requires clinical attention. Moreover, the risk among subjects with chronic viral hepatitis alone should be evaluated clinically in light of the young mean age at presentation and the strong potential for chronic viral hepatitis to develop cirrhosis with a further increase of risk for hepatocellular. Our findings illustrate that the link between alcoholism and hepatocellular is likely to be mediated through cirrhosis. Finally, our results indicate that alcoholism does not increase the relative risk linking chronic viral hepatitis to liver cancer unless and until alcoholism is complicated by cirrhosis. REFERENCES 1. Parkin DM, Pisani P, Ferlay J. Global cancer statistics. CA Cancer J Clin 1999;49: El-Serag HB, Mason AC. Rising incidence of hepatocellular in the United States. N Engl J Med 1999;340: IARC. Hepatitis viruses. IARC monographs on the evaluation of carcinogenic risks to humans. Volume 59. Lyon: IARC, IARC. Alcohol drinking. IARC monographs on the evaluation of carcinogenic risks to humans. Volume 44. Lyon: IARC, Adami HO, McLaughlin JK, Hsing AW, Wolk A, Ekbom A, Holmberg L, Persson I. Alcoholism and cancer risk: a population-based cohort study. Cancer Causes Control 1992;3: Mattsson B, Rutqvist LE, Wallgren A. Undernotification of diagnosed cancer cases to the Stockholm Cancer Registry. Int J Epidemiol 1985;14: Berkson J. Limitations of the application of fourfold table analysis to hospital data. Biomet Bull 1946;2: La Vecchia C, Negri E, Cavalieri d Oro L, Franceschi S. Liver cirrhosis and the risk of primary liver cancer. Eur J Cancer Prev 1998;7: Floreani A, Baragiotta A, Baldo V, Menegon T, Farinati F, Naccarato R. Hepatic and extrahepatic malignancies in primary biliary cirrhosis. HEPA- TOLOGY 1999;29: Donato F, Boffetta P, Puoti M. A meta-analysis of epidemiological studies on the combined effect of hepatitis B and C virus infections in causing hepatocellular. Int J Cancer 1998;75: Trichopoulos D, Day NE, Kaklamani E, Tzonou A, Muñoz N, Zavitsanos X, Koumantaki Y, et al. Hepatitis B vius, tobacco smoking and ethanol consumption in the etiology of hepatocellular. Int J Cancer 1987;39: Chung RT, Liang TJ. Hepatitis C virus and hepatocellular. In: Parsonnet J, ed. Microbes and Malignancy, Infections as a Cause of Human Cancers. New York: Oxford University Press, 1999; Robinson WS. Hepatitis B virus and hepatocellular. In: Parsonnet J, ed. Microbes and Malignancy, Infections as a Cause of Human Cancers. New York: Oxford University Press, 1999; Donato F, Tagger A, Chiesa R, Ribero ML, Tomasoni V, Fasola M, Gelatti U, et al. Hepatitis B and C virus infection, alcohol drinking, and hepatocellular : a case-control study in Italy. Brescia HCC Study. HEPATOLOGY 1997;26: Roudot-Thoraval F, Bastie A, Pawlotsky JM, Dhumeaux D. Epidemiological factors affecting the severity of hepatitis C virus-related liver disease: a French survey of 6,664 patients. The Study Group for the Prevalence and the Epidemiology of Hepatitis C Virus. HEPATOLOGY 1997;26: Kim YI, Yang DH, Chang KR. Relationship between Clonorchis sinensis infestation and cholangio of the liver in Korea epidemiological and pathologic reappraisal of 495 consecutive primary s of the liver in Seoul and Pusan areas. Seoul J Med 1974;15: Srivatanakul P, Parkin DM, Jiang YZ, Khlat M, Kao-Ian UT, Sontipong S, Wild C. The role of infection by Opisthorchis viverrini, hepatitis B virus, and aflatoxin exposure in the etiology of liver cancer in Thailand. A correlation study. Cancer 1991;68: Bergquist A, Glaumann H, Persson B, Broome U. Risk factors and clinical presentation of hepatobiliary in patients with primary sclerosing cholangitis: a case-control study. HEPATOLOGY 1998;27: Chalasani N, Baluyut A, Ismail A, Zaman A, Sood G, Ghalib R, McCashland TM, et al. Cholangio in patients with primary sclerosing cholangitis: a multicenter case-control study. HEPATOLOGY 2000;31: Altaee MY, Johnson PJ, Farrant JM, Williams R. Etiologic and clinical characteristics of peripheral and hilar cholangio. Cancer 1991; 68: Shin HR, Lee CU, Park HJ, Seol SY, Chung JM, Choi HC, Ahn YO, et al. Hepatitis B and C virus, Clonorchis sinensis for the risk of liver cancer: a case-control study in Pusan, Korea. Int J Epidemiol 1996;25: Okuda K, Kubo Y, Okazaki N, Arishima T, Hashimoto M. Clinical aspects of intrahepatic bile duct including hilar : a study of 57 autopsy-proven cases. Cancer 1977;39: Ekbom A, Hsieh CC, Yuen J, Trichopoulos D, McLaughlin JK, Lan SJ, Adami HO. Risk of extrahepatic bileduct cancer after cholecystectomy. Lancet 1993;342: Khan ZR, Neugut AI, Ahsan H, Chabot JA. Risk factors for biliary tract cancers. Am J Gastroenterol 1999;94: Kato I, Nomura AM, Stemmermann GN, Chyou PH. Prospective study of the association of alcohol with cancer of the upper aerodigestive tract and other sites. Cancer Causes Control 1992;3: Kraybill WG, Lee H, Picus J, Ramachandran G, Lopez MJ, Kucik N, Myerson RJ. Multidisciplinary treatment of biliary tract cancers. J Surg Oncol 1994;55: Chow WH, McLaughlin JK, Menck HR, Mack TM. Risk factors for extrahepatic bile duct cancers: Los Angeles County, California (USA). Cancer Causes Control 1994;5: Moerman CJ, Bueno de Mesquita HB, Runia S. Smoking, alcohol consumption and the risk of cancer of the biliary tract; a population-based case-control study in The Netherlands. Eur J Cancer Prev 1994;3: Zatonski WA, Lowenfels AB, Boyle P, Maisonneuve P, Bueno de Mesquita HB, Ghadirian P, Jain M, et al. Epidemiologic aspects of gallbladder cancer: a case-control study of the SEARCH Program of the International Agency for Research on Cancer. J Natl Cancer Inst 1997;89: