Risk of Liver and Other Types of Cancer in Patients With Cirrhosis: A Nationwide Cohort Study in Denmark
HEPATOLOGY, October 1998, p. 921-925, Vol. 28, No. 4
Henrik Toft Sørensen1, Søren Friis2, Jørgen H. Olsen2, Ane Marie Thulstrup3, Lene Mellemkjær2, Martha Linet4, Dimitrios Trichopoulos5, Hendrik Vilstrup1, and Jørn Olsen3
From the 1 Department of Medicine V, Aarhus University Hospital, Aarhus, Denmark; 2 The Danish Cancer Society, Division for Cancer Epidemiology, Copenhagen, Denmark; 3 The Danish Epidemiology Science Centre at the Department of Epidemiology and Social Medicine, Aarhus University, Aarhus, Denmark; 4 National Cancer Institute, Radiation Epidemiology Branch, Rockville, MD; and 5 Department of Epidemiology and the Harvard Center for Cancer Prevention, Harvard School of Public Health, Boston, MA
Cancer risk in patients with cirrhosis could be modified by factors such as changes in hormonal levels, impaired metabolism of carcinogens, or alteration of immunological status. We investigated the risk of liver and various forms of cancer in patients with cirrhosis in a follow-up study. We identified 11,605 1-year survivors of cirrhosis from the files of the Danish National Registry of Patients (NRP) from 1977 to 1989. Occurrence of cancer through 1993 was determined by linkage to the Danish Cancer Registry. For comparison, the expected number of cancer cases was estimated from national age-, sex-, and site-specific incidence rates. Overall, 1,447 cancers were diagnosed among the study subjects, as compared with 708.1 expected, to yield a standardized incidence ratio (SIR) of 2.0 (95% CI: 1.9 to 2.2). In all diagnostic subgroups of cirrhosis, the risk of primary liver cancer, mainly hepatocellular carcinoma, was markedly elevated, with 245 observed cases and an overall 36-fold elevated risk (59.9-fold elevated for hepatocellular carcinoma and 10-fold for cholangiocarcinoma). Substantial and persistent excesses during follow-up were seen for all types of cancer associated with tobacco and alcohol habits (cancer of the lung, larynx, buccal cavity, pharynx, pancreas, urinary bladder, and kidney), while moderate excesses were seen for cancers of the colon and breast. The latter, however, were not complemented by any decrease in the risk of prostate cancer (SIR: 1.0; 95% CI: 0.7 to 1.3). A slightly increased risk was seen for testis cancer, but disappeared after 10 years. We found evidence of an increased risk for liver and several extrahepatic cancers in patients with cirrhosis. Although part of this increase is likely attributable to alcohol and tobacco consumption, our study opens up the possibility that cirrhosis plays a role in the carcinogenesis of types of cancer other than liver cancer. (HEPATOLOGY 1998;28:921-925.)
The incidence of liver cancer varies from 100 in 100,000 per year in certain parts of Africa and Asia to less than 4 in 100,000 per year in most parts of Europe and North America.1 The incidence of various forms of cirrhosis in Denmark varies from 9 per million person-years for primary biliary cirrhosis to 137 per million person-years for alcoholic cirrhosis.2 In general, the incidences of liver cancer and cirrhosis follow similar patterns of variation, because these two diseases share etiologic factors. Important causes of liver cancer are alcohol and Hepatitis B and C viruses,3-11 which likely act indirectly through cirrhosis.3
Whether cirrhosis is a risk factor for cancer at sites other than the liver has not been well investigated, except for patients with primary biliary cirrhosis. 7,11-14 Changes in the metabolism of carcinogens or hormones could, however, affect cancer risk. It is known that cirrhosis induces changes in estrogen metabolism,15 and male patients with cirrhosis have hyperestrogenism, which may lead to gynecomastia and testis atrophy. Recent studies in women indicate that cirrhosis is associated with menstrual irregularity, increased frequency of spontaneous abortion, and early menopause.16 Moreover, the liver plays a central role in the metabolism of lipid and water-soluble drugs and other chemicals, and patients with cirrhosis have also been reported to experience alterations in immune functions and risk of infections.17
To examine the risk of liver cancer and extrahepatic cancers in patients with cirrhosis, we conducted a cohort study using nationwide registries.
PATIENTS AND METHODS
The Danish National Registry of Patients (NRP) contains information on virtually all hospital admissions in Denmark (population: 5.2 million) since 1977.18 Each admission record includes the personal identification number that is unique to every Danish resident, dates of admission, dates of discharge, and up to 20 discharge diagnoses. Diagnoses are classified according to the Danish version of the International Classification of Diseases, 8th edition (ICD-8).19 Persons were enrolled in the study if they had been discharged with alcoholic cirrhosis (ICD-8 = 571.09), primary biliary cirrhosis (571.90), nonspecified cirrhosis (571.92), chronic hepatitis (571.93), or “other types of cirrhosis, alcoholism not indicated” (571.99) at least once during the years 1977 to 1989. Cirrhosis was considered as a whole, but also as four separate types, largely following the ICD-8 codes given above, except that nonspecified cirrhosis and cirrhosis, alcoholism not indicated, were merged into one group termed “nonspecified cirrhosis” (571.92 and 571.99). If a patient had been discharged with more than one cirrhosis diagnosis, a prioritized order of diagnoses determined to which group the patient was assigned: 1) alcoholic, 2) primary biliary cirrhosis, 3) chronic hepatitis, and 4) nonspecified cirrhosis. Accordingly, those persons with alcoholic cirrhosis entered this category independently of whether they had any other cirrhosis diagnosis. Furthermore, cirrhosis patients with a diagnosis of alcoholism (ICD-8 = 303) at any time from 1977 to 1989 were always referred to as the alcoholic group, no matter which cirrhosis code was registered.
All members of the study cohort were linked through their personal identification number to the nationwide Danish Cancer Registry, which was initiated in 1943 and which has an almost complete coverage of incident cancers in the country.20 The Cancer Registry receives notifications of malignant and related diseases from hospital departments at the time of diagnosis, and if changes in the initial diagnosis occur. In addition, reports are received from pathology departments and departments of forensic medicine, giving the results of autopsies of cancer patients. Also, cases first diagnosed at autopsy, i.e., as an incidental finding, are included in the Registry’s material. This information is supplemented by a review of the files of the NRP and of all death certificates. Cancers were classified according to the modified Danish version of the International Classification of Diseases, 7th revision (ICD-7)20 and since January 1, 1978, also according to the International Classification of Diseases for Oncology (ICD-O).20 For liver cancers, the latter classification includes a subgrouping of cases according to tumor morphology, i.e., hepatocellular carcinomas, cholangiocarcinomas, mixed hepatocellular-cholangiocarcinomas and hemangiosarcomas. Liver cancers without specification for morphological type were only included in the site-specific cancer analysis if they were explicitly notified as “primary liver cancer.”
The follow-up period for cancer occurrence began at the date of discharge with cirrhosis from the hospital, irrespective of type, and ended at the date of death or December 31, 1993, whichever came first. Dates of death were obtained through linkage with the nationwide Cause of Death Registry.
Statistics. The number of cancer cases observed among study subjects was compared with the number of cases expected on the basis of rates from the Danish Cancer Registry, and the standardized incidence ratio (SIR) was calculated as the ratio of the observed to the expected number of cancer cases. The expected number of cancers was calculated by multiplying the number of person-years of study subjects by the site- and morphology-specific national incidence rates of each sex in 5-year age groups and calendar periods of observation. The statistical methods used assumed that the observed number of cases in any specific category followed a Poisson distribution. Tests of significance and confidence interval for the SIR were calculated based on Byar’s approximation; exact confidence limits were used if the observed number of cancers was less than 10.21
To avoid selection bias that might have been introduced by the entry of study subjects with cirrhosis and concurrent cancer, we computed risk estimates for cancer excluding observed and expected cancers during the first year of follow-up after discharge from the hospital for cirrhosis. Of the 18,358 patients with cirrhosis notified to the NRP, 6,753 (37%) died within the first year of follow-up, leaving 11,605 for cancer risk analysis.
table 1 shows characteristics of the 11,605 patients with cirrhosis included in the risk analysis. In both sexes combined, alcoholic cirrhosis was the most frequent specific subtype in Denmark (62%), followed by cirrhosis from chronic hepatitis (15%), and primary biliary cirrhosis (3%). In 21% of cases, the origin of cirrhosis remained nonspecified. There were more men than women among patients with alcoholic cirrhosis, whereas more women than men had primary biliary cirrhosis and cirrhosis from chronic hepatitis (table 1).
|table 1. Descriptive Characteristics of 11,605 1-Year Survivors With Cirrhosis in Denmark, 1977-1989|
Over the entire follow-up period, 1,447 cancers were diagnosed; 708.1 were expected in the combined cohort, which yields an SIR of 2.0 (95% CI: 1.9-2.2) (table 2). The risk for liver cancer was exceptionally high, with 245 observed cases against 6.8 expected (SIR: 36; 95% CI: 32-41). The excess risk was most pronounced for hepatocellular carcinoma. Significant excesses were seen for cancer related to tobacco, and alcohol habits and for cancer of the stomach (SIR: 1.9; n = 40) and colon (SIR: 1.5; n = 82); this pattern was evident in both sexes combined as well as separately (table 2). A slight increase in risk was observed for hormone-related cancers, in particular breast cancer in women and testis cancer in men. The number of observed prostate cancers was as expected. There were 20% more cases of nonmelanoma skin cancer than expected, which was exclusively a result of an excess among women, whereas there was a reduction in the number of malignant melanomas and of cancers of the brain/nervous system.
|table 2. Observed and Expected Numbers and SIR of Cancer in 11,605 Patients With Cirrhosis|
The risk for liver cancer was most pronounced during early follow-up (1-4 years; SIR: 42.8; 95% CI: 36.1-50) compared with late follow-up, though it remained high (10+ years; SIR: 28.8; 95% CI: 19.6-40.9) (data not shown). The excess risks for tobacco- and alcohol-related cancers, kidney, breast and colon cancer, also persisted during late follow-up, but excesses of testis and stomach cancer were not present after 10 years.
table 3 shows that the SIR for all cancers combined was increased in all types of cirrhosis, though only slightly in patients with primary biliary cirrhosis. The SIR of liver cancer was markedly increased in all subcohorts, ranging from 44.8 (alcoholic cirrhosis) to 18.5 (primary biliary cirrhosis). The cancer pattern was quite similar in alcoholic cirrhosis, chronic hepatitis, and nonspecified cirrhosis, with elevated risks of tobacco- and alcohol-related cancers, testicular cancer, stomach cancer, and colon cancer. An exception was the risk of breast cancer, which was increased only in alcoholic and nonspecified cirrhosis. Although the small size of the subcohort of primary biliary cirrhosis reduced the ability to detect risk deviations, there seemed to be a somewhat different cancer pattern among these patients; apart from liver cancer, only cancer of the colon was seen in an excess that reached a significant level (table 3).
|table 3. Observed and Expected Numbers and SIR of Selected Cancer Sites in Different Types of Cirrhosis|
The study confirmed that patients with cirrhosis have a considerably increased risk of primary liver cancer, mainly hepatocellular carcinoma, irrespective of the specific underlying cause of the cirrhosis. Our finding of a higher risk of liver cancer in alcoholic cirrhosis than in any other type of cirrhosis may suggest either a direct effect of alcohol on liver carcinogenesis or induction of a more severe type of cirrhosis carrying a higher risk of liver cancer, or both.
There is a strong association between alcohol intake, elevated liver enzymes, and smoking habits in the Danish population,22 and the markedly increased risks for tobacco-related cancer sites found in all cirrhosis patients except primary biliary cirrhosis were probably caused by excessive use of tobacco products. The increased risks of cancer of the stomach and large bowel may be more difficult to explain by confounding, although stomach cancer has recently been included in the group of tobacco-related cancers.23 Alcohol consumption has been associated with an increased risk of rectal cancer, 24,25 and Naveau et al.26 studied the effects of alcoholism and cirrhosis and found that alcoholics were three times more likely than nonalcoholics to have colon adenomatous polyps, controlling for cirrhosis. The same study showed that the prevalence of colon adenomatous polyps was over two times higher in patients with cirrhosis than in those without cirrhosis, after controlling for alcoholism. A meta-analysis based on 27 studies showed that for consumption of two alcoholic beverages daily, the average, relative risk of colorectal cancer was 1.1 (95% CI: 1.05-1.14).27 However, because the association was small, it was concluded that the findings regarding a causal role of alcohol were inconclusive.
The increased risk of breast cancer observed in patients with alcoholic and nonspecified cirrhosis, but not in chronic hepatitis, indicates that the excess is to some degree related to alcohol intake.28 Most patients with clinical cirrhosis have increased endogenous estrogen levels, 29,30 but the moderate size of the breast cancer finding, and the absence of an association with chronic hepatitis, suggests that these hormonal changes have very little effect on breast cancer risk. This is in line with the observation of no excess of prostate cancer. It has been hypothesized for this male type of cancer that cirrhosis patients have a reduced risk, again because of the disturbances in female hormones.31-33 However, the overall increased risk of testis cancers could be a result of the induced estrogen exposure.29
In contrast to other types of cirrhosis, less is known about the risk of hepatocellular carcinoma in primary biliary cirrhosis. The previous studies of patients with primary biliary cirrhosis have given conflicting cancer findings. 3,7,9,11-14,34,35 An increased risk of liver cancer has been reported in some studies, 9,11 but not in others. 3,12,35 We found a statistically significant increase in the risk of liver cancer, higher than in a recent large population-based study from Sweden that reported a nonsignificant relative risk of 2.9.7 However, our estimate relied on only four cases. We could not confirm the previously reported excess risk for the development of breast cancer in patients with primary biliary cirrhosis. 12,13,33
A limitation of our study is the lack of information on diagnostic criteria and potential confounding factors. A previous study indicated high data validity of the cirrhosis diagnosis in the NRP, but some misclassification between the different types of cirrhosis probably exists, particularly between alcoholic and nonspecified cirrhosis.36 Therefore, we also searched for diagnoses of alcoholism in the NRP for those not registered with alcoholic cirrhosis to reduce this misclassification. The subcohorts of chronic hepatitis and nonspecified cirrhosis are likely to include an overrepresentation of alcoholic patients relative to the general population as indicated by the increased risk for other alcohol-related cancers; however, these groups do not contain the same proportion of alcoholic patients as the subgroup of alcoholic cirrhosis. Misclassification of alcoholic cirrhosis may bias estimates across the different types of cirrhosis. The sex ratio and follow-up time in patients with chronic hepatitis and primary biliary cirrhosis indicate, however, that misclassification must be of minor importance in these groups. Denmark is a very-low-incidence area for infectious Hepatitis A, B, and C,37 and patients with autoimmune hepatitis account for a high proportion of the patients in the chronic hepatitis group. In the Danish population, approximately 100 cases of Hepatitis A and B are registered per year, and less than 80 are cases of Hepatitis C. The high frequency of Hepatitis B virus and Hepatitis C virus infections among individuals with alcoholic liver diseases and alcohol-associated hepatocellular carcinoma in some countries has led to the suggestion that these viruses may be implicated in the pathogenesis of hepatocellular carcinoma.38 However, these infections are rare in Denmark. A comprehensive screening of blood donors in Denmark has shown a prevalence of less than 0.04%,39 so it is not likely that hepatitis infections play an important role in the pathogenesis of hepatocellular carcinoma in the present study.
The clearest finding of the present study was the excessive risk of liver cancer found in all cirrhosis patients independent of the type of cirrhosis. Thus, cirrhosis seems to be an essential step in the induction of liver cancer. The hormonal and immune changes in cirrhosis did not affect the incidence of hormone cancers and cancers related to immunosuppression to any considerable degree. The elevated risk for breast cancer in alcoholic cirrhosis could be caused by a direct effect of alcohol rather than the elevated estrogen level. Cirrhosis patients experienced risk increase for many other cancers, but several are heavily influenced by alcohol and tobacco use.
Acknowledgement: The authors thank Anne Marie E. I. Larsen at the Division for Cancer Epidemiology for computer assistance.
Abbreviations:NRP, National Registry of Patients; ICD, International Classification of Diseases; SIR, standardized incidence ratio.
Supported by the Oncologic Research Unit at Aalborg Hospital, by the Ebba and Aksel Sjølin Foundation, by research grant MAO N01-CP-85639-04 from the National Cancer Institute, Bethesda, MD, and The Danish Medical Research Council (grant 9700766). The activities of the Danish Epidemiology Science Centre are financed by a grant from the Danish National Research Foundation
Received January 8, 1998; accepted May 4, 1998.
Address reprint requests to: Dr. Henrik Toft Sørensen, The Danish Epidemiology Science Centre at the Department of Epidemiology and Social Medicine, Aarhus University, DK-8000 Aarhus C, Denmark. Fax: 45-86-13-15-80.
1. Cook-Mozaffari P, Van Rensborg S. Cancer of the liver. Br Med Bull 1984; 40: 342-345 [Medline].
2. Almdal TP, Sørensen TIA. Incidence of parenchymal liver diseases in Denmark, 1981 to 1985: analysis of hospitalization registry. HEPATOLOGY 1991; 13: 650-655 [Medline].
3. Adami H-O, Hsing AW, McLaughlin JK, Trichopoulos D, Hacker D, Ekbom A, Persson I. Alcoholism and liver cirrhosis in the etiology of primary liver cancer. Int J Cancer 1992; 51: 898-902 [Medline].
4. Kobayaski K, Unoura M, Tanaka N, Hatton N. A comparison between hepatocellular carcinoma-developing and non-carcinoma-developing patients with cirrhosis over a long period. Hepatogastroenterology 1990; 37: 445-448 [Medline].
5. Hardell L, Bengtsson NO, Jonsson U, Eriksson S, Larsson LG. Aetiological aspects of primary liver cancer with special regard to alcohol organic solvents and acute intermittent porphyrinan epidemiological investigation. Br J Cancer 1984; 50: 389-397 [Medline].
6. La Vecchia C, Negri E, D’Avanzo B, Boyle P, Franceschi S. Medical history of primary liver cancer. Cancer Res 1990; 50: 6274-6277 [Medline].
7. Lööf L, Adami H-O, Sparén P, Eriksson LS, Hultcrantz R, Lindgren S, et al. Cancer risk in primary biliary cirrhosis: a population-based study from Sweden. HEPATOLOGY 1994; 20: 101-104 [Medline].
8. Ikeda K, Satoshi S, Koida I, Arase Y, Tsubota A, Chayama K, Kumada H, et al. A multivariate analysis of risk factors for hepatocellular carcinogenesis: a prospective observation of 795 patients with viral and alcoholic cirrhosis. HEPATOLOGY 1993; 18: 47-53 [Medline].
9. Johnson PJ, Krasner N, Portmann B, Eddleston ALWF, Williams R. Hepatocellular carcinoma in Great Britain: influence of age, sex, HBsAg status and aetiology of underlying cirrhosis. Gut 1978; 191: 1022-1026.
10. Davidson AR, Tomlinson S, Calne RY, Williams R. The variable course of primary hepatocellular carcinoma. Br J Surg 1974; 61: 349-352 [Medline].
11. Floreani A, Biagini MR, Chiaramonte M, Milani S, Surrenti C, Naccarato R. Incidence of hepatic and extra-hepatic malignancies in primary biliary cirrhosis (PBC). Ital J Gastroenterol 1993; 25: 473-476 [Medline].
12. Mills PR, Boyle P, Quiglely EMM, Birnie GG, Jarrett F, Watkinson G, MacSween RNM. Primary biliary cirrhosis. An increased incidence of extrahepatic malignancies. J Clin Pathol 1982; 35: 541-543[Medline].
13. Wolke AM, Schaffner F, Kapelman B, Sacks HS. Malignancy in primary biliary cirrhosis. High incidence of breast cancer in affected women. Am J Med 1984; 76: 1075-107 8[Medline].
14. Dam GM Van, Gips CH. Primary biliary cirrhosis in the Netherlands. Scand J Gastroenterol 1997; 32: 77-82 [Medline].
15. Becker U, Almdal T, Christensen E, Gluud C, Farholt S, Bennett P, Svenstrup B, et al. Hormones in postmenopausal women with primary biliary cirrhosis. HEPATOLOGY 1991; 13: 865-869 [Medline].
16. Becker U, Tønnesen H, Kaas-Claesson N, Gluud C. Menstrual disturbances and fertility in chronic alcoholic women. Drug Alcohol Depend 1989; 24: 75-78 [Medline].
17. Hara K, Kohno S, Koga H, Kaku M, Tomono K, Sakata S, Komatsu K, et al. Infections in patients with liver cirrhosis and hepatocellular carcinoma. Intern Med 1995; 34: 491-495 [Medline].
18. Danish National Board of Health. The activity in the hospital care system. Copenhagen: National Board of Health , 1981.
19. Danish National Board of Health. Classification of diseases. Copenhagen: National Board of Health , 1976.
20. Storm HH, Michelsen EV, Clemmensen IH, Pihl J. The Danish Cancer Registryhistory, content, quality and use. Dan Med Bull 1997; 44: 535-539 [Medline].
21. Rothman KJ, Boice JD. Epidemiologic analyses with a programmable calculator (DHHS Publication No [NIH] 79-1649). Washington DC: US Government Printing Office , 1979.
22. Steffensen FH, Sørensen HT, Brock A, Vilstrup H, Lauritzen T. Alcohol consumption and serum liver-derived enzymes in a Danish population aged 30-50 years. Int J Epidemiol 1997; 26: 92-99 [Medline].
23. Doll R. Cancers weakly related to smoking. Br Med Bull 1996; 52: 1-15.
24. Klatsky AL, Armstrong MA, Friedman GD, Hiatt RA. The relations of alcoholic beverage use to colon and rectal cancer. Am J Epidemiol 1988; 128: 1007-1015 [Medline].
25. Riboli E, Cornee J, Macguart-Moulin G, Kaaks R, Casagrande C, Guyader M. Cancer and polyps of the colorectum and lifetime consumption of beer and other alcoholic beverages. Am J Epidemiol 1991; 134: 157-166 [Medline].
26. Naveau S, Chaput JC, Bedossa P, Poynard T, Pauphilet C, Ink O, Houdayer C, et al. Cirrhosis as an independent risk factor for colonic adenomas. Gut 1992; 33: 535-540[Medline].
27. Longnecker MP, Orza MJ, Adams ME, Vioque J, Chalmers TC. A meta-analysis of alcoholic beverage consumption in relation to risk of colorectal cancer. Cancer Causes Control 1990; 1: 59-68 [Medline].
28. Smith-Warner SA, Spiegelman D, Yaun S-S, van den Brandt PA, Folsom AR, Goldbohm A, Graham S, et al. Alcohol and breast cancer in women. JAMA 1998; 279: 535-540 [Medline].
29. Gluud C. Testosterone and alcoholic cirrhosis. Epidemiologic, pathophysiologic and therapeutic studies in men. Dan Med Bull 1988; 35: 564-575 [Medline].
30. Becker U. The influence of ethanol and liver disease on sex hormones and hepatic oestrogen receptions in women. Dan Med Bull 1993; 40: 447-459 [Medline].
31. Glantz GM. Cirrhosis and carcinoma of the prostate gland. J Urol 1964; 91: 291-293.
32. Robson MC. Cirrhosis and prostatic neoplasms. Geriatrics 1966; 21: 150-154 [Medline].
33. Nomura AMY, Kolonel LN. Prostate cancer: a current perspective. Epidemiol Rev 1991; 13: 200-227 [Medline].
34. Goudie BM, Burt AD, Boyle P, MacFarlane G, Birnie GG, Mills PR, Gillis CR, et al. Breast cancer in women with primary biliary cirrhosis. BMJ 1985; 291: 1597-1598 [Medline].
35. Turissini SB, Kaplan MM. Hepatocellular carcinoma in primary biliary cirrhosis. Am J Gastroenterol 1997; 92: 676-678[Medline].
36. Vestberg K, Thulstrup AM, Sørensen HT, Ottosen P, Vilstrup H. Data quality of administratively collected hospital discharge data for liver cirrhosis epidemiology. J Med Syst 1997; 21: 11-20 [Medline].
37. The State Serum Institute. Hepatitis B 1995. Epinyt 1996; 49: 1-2.
38. Berchot C, Nalpas B, Feitelson MA. Interactions between alcohol and hepatitis viruses in the liver. Clin Lab Med 1996; 16: 273-287[Medline].
39. The State Serum Institute. Hepatitis C look-back. Epinyt 1996; 3: 1.
Copyright © 1998 by the American Association for the Study of Liver Diseases.