Canine biliary carcinoma: Epidemiological comparisons with man

J. COMP.

PATH.

1983.

VOL.

93.

99

CANINE EPIDEMIOLOGICAL

H.M.

HAYES,

Environmental

BILIARY CARCINOMA: COMPARISONS

JR, M.M.

Epidemiology

MORIN

Branch, .National Bethesda, Maryland

andD.A.

Cancer Institute, 20205, U.S.A.

WITH

MAN

RUBENSTEIN 3CO7 Landow

Building,

INTRODUCTION

Cancers of the biliary tract account for about 4200 human deaths annually in the United States; adenocarcinomas represent more than 95 per cent of these malignancies. Much is known about the descriptive epidemiology of these cancers in man, but the roles of a number of factors, in what appears to be a multifactorial aetiology, remain unclear (Fraumeni and Kantor, 1982). Carcinoma of the gallbladder, the most common cancer of this type (Young, Percy and Asire, 1981), has been associated with the presence of gallstones (Piehler and Crichlow, 1978). The cancer risk among women is about 2.5 times that of men (Young et al., 1981)) although there is no apparent sex difference when gallstones are absent. The malignancy occurs more frequently among south-western American Indians and Mexican Americans than among Anglo-Americans. Overall, a notable rise in incidence occurs in people over 60 years of age (Fraumeni and Kantor, 1982). Carcinomas of the extrahepatic ducts (i.e. hepatic, cystic, or common bile ducts) occur somewhat more frequently among men than women (Young et al., 1981) and are less often associated with gallstones or chronic gallbladder disease. Individuals with ulcerative colitis are at considerably greater risk (Ritchie, Allen, Macartney, Thompson, Hawley and Cooke, 1974). Congenital bile duct cysts have been associated with these tumours in a few cases (Todani, Tabuchi, Watanabe and Kobayashi, 1979). Other suggested factors affecting the incidence of carcinomas of the gallbladder and extrahepatic bile ducts include occupational exposures to certain chemicals, biliary stasis, and biliary infection or inflammation (Fraumeni and Kantor, 1982). Carcinoma of the intrahepatic bile ducts (cholangiocarcinoma) is relatively infected with rare in the U.S.A. (Young et al., 1981), except among immigrants the Chinese liver fluke, CZonorchis sine&s. Hou ( 1965)) in his extensive studies in Hong Kong, showed that C. sinensis produces virtually the same pathology in the intrahepatic ducts of man, dog and cat, which suggests that the latter 2 species might be appropriate models for testing suspected human biliary carcinogens. Histologically, biliary and gallbladder tumours in the dog are very similar to the corresponding cell types in man (Ponomarkov and Mackey, 1976). The biliary system of the dog is functionally similar to that of man, but anatomically there are some notable differences. Man has an ampulla (of Vater) 0 1983 Academic Press Inc. (London) Limited OOZl-9975/83/010099+09 $03.00/O

lop

H. M.

HAYES,

JR

et d.

at the distal ends of the common bile duct and pancreatic duct proximal to their opening into the duodenum (Warwick and Williams, 1973). The distal ends of the common bile duct and pancreatic duct are normally separated in the dog and each opens directly into the lumen of the duodenum (Evans and Christensen, 1979). In man, the hepatic bile ducts join the common bile duct proximal to the cystic duct, while the converse is usually true in the dog. Also, the dog has collateral channels which connect the intrahepatic bile ducts of adjacent lobes, thereby making possible continued drainage when primary passageways are obstructed. The human liver does not have a similar auxiliary bile network (Sleight and Thomford, 1970). Another major difference between the canine and human biliary systems is the composition of the bile. Normal canine bile contains a much higher proportion of cholesterol than human bile, so much so that human gallstones dissolved within a few weeks when placed in a dog’s gallbladder (Nakayama, 1969). Little is known about the epidemiology of spontaneous neoplasms of the canine biliary system. Rooney (1959), Rehmtulla (1974), Patnaik, Hurvitz, Lieberman and Johnson (198 1) and others have described about 130 dogs with biliary tumours. In none of these series were cases characterized in relation to the referent population, nor were associated conditions explored. Our study investigates the epidemiological features of biliary neoplasms in dogs seen between March 1964 and June 1978 at 15 North American veterinary medical teaching hospitals (VMTH) .

MATERIALS

AND

METHODS

We obtained computerized abstracts of medical records representing each clinical visit by patients to collaborating VMTH. By July Ist, 1978 abstracts on 1.8 million clinical visits were available, of which 1.1 million were dogs. Each abstract contains standard information (patient’s unique identity number, species, breed, age and sex) plus specific data on diagnoses, surgical procedures and diagnostic tests. A full description of the standardized abstraction procedure (Tjalma, Priester, Adelstein and Vander Zwaag, 1964) and the data collection programme (Hayes, Wilson and Moraff, 1979) have been given elsewhere. Dogs with microscopically confirmed primary tumours of the biliary system were identified. Cases were tabulated according to tumour site, cell type and malignancy behaviour. Case statistics by breed, age and sex were based on data collected at the time of first confirmed diagnosis. The abstracted medical history of each case was further reviewed to note the presence of biliary calculi, parasites and other neoplasms. For comparison, a hospital population at risk was obtained by counting each dog once for each year when it was seen for any reason (“patient years” at risk). If the animal appeared several times in the same year, one visit was randomly selected. No contribution was made to the patient years for those years when the animal did not visit the clinic. Once the diagnosis of biliary neoplasia had been confirmed, it was excluded from subsequent patient years tabulation. The population at risk was tabulated by age, breed and sex (based on gonadal presence or absence). The measure of association used to evaluate the effects of the independent variables of breed, sex and age was the maximum likelihood estimate of relative risk (RR), as approximated by the relative odds, a technique described by Gart (1970). RR is the ratio of the rate of disease among a test group (e.g. Labrador retrievers) to the rate in the control group (all dogs combined). Gart’s procedure provides adjustment for

BILIARY

101

CARCINOMA

confounding variables in assessing the summary RR of one factor across all strata (e.g. breed RR was calculated to allow control of variations in age and gonadal status among the test group and the control group). Ninety-five per cent confidence intervals (CI) were used to measure statistical significance. An RR of I.0 indicates that the test group has the same rate of disease as the comparison group; an RR less than or greater than 1.0 indicates a lower or higher risk, respectively. If the confidence interval for a calculated RR did not include 1.0 in its range, the RR was considered to differ significantly from the relative risk of disease in the comparison group at P-co-05 (two-tailed test). RESULTS

We identified 86 dogs with benign or malignant primary biliary tumours. Those originating in the intrahepatic biliary tree predominated (N=65), (M=5) (Table 1). followed by extrahepatic ducts (.YV=16) and gallbladder Of the 86 dogs, 77 had carcinomas originating from the intrahepatic (Jv=61) and extrahepatic (N=16) bile ducts; subsequent descriptive analyses focused on the combined cases of these malignant tumours. TABLE

1

DISTRIBUTION OF CANINE BILIARY SYSTEM NEOPLASMS DIAGNOSED AT 15 NORTH AMERICAN VETERINARY MEDICAL TEACHING HOSPITALS BETWEEN MARCH 1964 AND JUNE 1978

Site and tumour cell-bpe Intrahepatic bile ducts Adenocarcinoma Carcinoma, type not specified Squamous cell carcinoma Adenoma Leiomyoma Extrahepatic bile ducts Hepatic bile duct Adenocarcinoma Common bile duct Carcinoma, type not specified Adenocarcinoma Squamous cell carcinoma Gallbladder Adenocarcinoma Leiomyosarcoma Sarcoma, type not specified Adenoma Fibroma

Number

of cases 40 17 ii 1

Cases were distributed among 30 distinct breeds, with mongrels considered as a separate group in our statistical analyses. RR values were calculated for breeds represented by 5 or more cases (mongrels-18 cases; miniature poodle and Labrador retriever-5 cases each). None showed a relative risk that differed statistically from that in all dogs combined (RR=l) (Table 2). There were 37 entire male (non-castrated), 24 spayed (ovariectomized) female, and 16 intact female dogs in the series. Spayed female dogs, controlling for age, had 1.5 times the relative risk of intact female dogs and 1.1 times the risk of males (CI=O*7 to 2.9, O-6 to 1.9, respectively). RR values by age were calculated for each sex group with the ratio of cases to their hospital population at 4 to 6 years of age as a standardized reference point (RR=l). An increasing

102

H. M.

HAYES,

JR

TABLE ESTIMATED RELATIVE RISK BY 5 OR MORE CASES AMONG

(RR)

* OF CANINE BILE DUCT CARCINOMA PATIENTS7 SEEN AT 15 NORTH AMERICAN TEACHING HOSPITALS

Male Intact

retriever

All Miniature breeds

combined poodle

FOR BREEDS VETERINARY

REPRESENTED MEDICAL

Female

Breed Labrador Mongrel

et a/.

2

RR

Sgayed

95 per cent confidence interval

t

2”

A

2.4 1.1

O-85-6.13 0.61-l-82

372

162

241

i.9

0.33-2.38

* RR calculated with adjustment for age and sex, including gonadal t Thirty-one breeds were represented in the series of 77 dogs.

status.

hospital prevalence with age was evident for each sex group; the rate among spayed females was approximately twice that for ageing ( >lO years) males or intact females. These numbers were too small to make interval statistical comparisons. The medical history showed that 29 other primary neoplasms occurred in 21 carcinoma patients (M:F=ll: 10). F ive dogs had one or more testicular tumours; 3 others had mammary carcinoma (Table 3). Twenty-one of the 77 cases had metastases (M:F=lO: ll), with the lung the most common site (16 cases). Two cases had metastasized to the pancreas. TABLE NUMBER

Number 2 2 1 1

1

I

AND

TYPES

3

OF OTHER HISTOLOGICALLY CONFIRMED 77 DOGS WITH BILE DUCT CARCINOMA

NEOPLASMS

DIAGNOSED

IN

Neoplasm(s) Lipoma Mammary carcinoma Mammary carcinoma, mammary adenoma Mammary adenoma Sertoli cell tumour, adenocarcinoma of the third eyelid Sertoli cell tumour, mastocytoma Sertoli cell tumour. interstitial cell tumour. transitional cell carcinoma bladder ’ Seminoma, interstitial cell tumour, lipoma Oral sarcoma, generalized malignant melanoma, histiocytoma Chemodectoma Generalized sarcoma Interstitial cell tumour Leiomyoma of the stomach Leiomyosarcoma of the bladder Mastocytoma Meningioma Pancreatic adenoma Pituitary basophilic adenoma Snlenic carcinoma

of the

None of the 77 cancer cases had a previous history of biliary calculi, biliary stasis or known parasitic infections of the liver or biliary tract. Four cases with cholangiocarcinoma had a recent or concurrent infection with canine hookworms (Ancylostoma caninum) ; 3 of these dogs were > 10 years of age. Two of the 4 dogs with hookworm infection were also infected with canine whipworms ( Trichuris zd’is) ; another dog with cholangiocarcinoma (age 17 years) also had whipworm infection. To evaluate the relative frequency of these infestations,

BILIARY

CARCINOMA

103

we compared ratios of the number of parasitic infections per patient years at risk (adjusting for age) between cholangiocarcinoma cases and all dogs seen at VMTH during 1977, the last complete calendar year of available data. The observed frequency of hookworms among cholangiocarcinoma cases was while whipworm frequency was 3-O 2.2 times greater than expected (P=O*l), times greater than expected (P=O*O6). Infestation by either parasite was 2.5 times greater than expected (P=O.O2). We made further analyses to compare the prior and concomitant parasitic experience of cholangiocarcinoma cases with that of histologically confirmed stomach and small intestinal carcinoma cases (chosen as a control group because of the likelihood of examination of the intestinal tract diagnosed at the same VMTHs during the same study period (N=123). Cholangiocarcinoma cases experienced 3.3 times more infection from hookworms (P=O*2), 1-3 times more whipworm infection (P=O*5) and 2.0 times more infection with either parasite (P=O*2). DISCUSSION

The quality of hospital data for retrospective studies is always suspect. Medical records may be incomplete, misplaced, lost or inaccurately transcribed when abstracted. The coded abstracted data we analysed were edited for logical errors; we have no evidence that undetectable errors or omissions favour any particular age, sex or breed of animal. Veterinary hospital data may be further compromised because the owner determines if and when the pet is brought for medical attention. The effects of this uncontrollable influence should be minimal since our referent population was drawn from the same reporting facilities during the same period as the case series which was studied. Some tumours of the canine biliary system are undoubtedly missed since the clinical disease often lacks specific signs. Radiography is not particularly helpful in making the diagnosis (Weaver, 1976). Spontaneous neoplasms of the canine biliary system have been infrequently reported. Among the collaborating VMTH, carcinomas of the biliary system account for 0.9 per cent of the overall hospital prevalence of primary cancers, other than non-melanoma skin cancer, among dogs (Hayes, unpublished). Consistent with previous reports (Rooney, 1959; Rehmtulla, 1974; Patnaik et al., 1981), our survey indicates that among canine biliary cancers cholangiocarcinoma predominates (76 per cent), while cancers of the extrahepatic bile ducts and the gallbladder occur much less frequently (20 per cent and 4 per cent, respectively). In man, the converse applies, with cancer of the intrahepatic bile duct representing only 4 per cent of the incidence of biliary tract cancer and extrahepatic duct and gallbladder cancers representing 27 per cent and 52 per cent, respectively, in the U.S.A. Human biliary cancers account for about 0.8 per cent of the overall incidence of cancer (Young et al., 1981). Our survey failed to find any association between biliary calculi and canine bile duct carcinoma. Cholesterol-type gallstones predominate in human cholelithiasis, a disease which shares many of the same epidemiological characteristics as human gallbladder cancer (Diehl, 1980). Human gallstones form when

104

H. M.

HAYES,

JR

et

d.

the bile becomes supersaturated with cholesterol, due either to an increase in hepatic secretion of cholesterol or to a decrease in the bile acid pool which maintains its solubility (Bennion and Grundy, 1978). That the ratio of bile acids to cholesterol in canine bile is 6 times that of human bile (Nakayama, 1969) may explain the comparatively lower frequency of cholelithiasis among dogs (Schall, Chapman, Finco, Greiner, Mather, Rosin and Welser, 1973). This, in turn, is probably responsible for the lower frequency of gallbladder cancer. In man, associated gallstones are found in approximately 25 to 30 per cent of bile duct carcinoma cases (Shani, Hart and Modan, 1974; Thorbjamarson, 1975). This is a much weaker association than that found among gallbladder carcinoma cases, where 60 to 80 per cent have a history of cholelithiasis (Fraumeni and Kantor, 1982). We detected a suggestion of breed predisposition among Labrador retrievers for bile duct carcinoma (RR =2*4, P> O-05). However, mongrel dogs had about the same risk (RR= 1.1) as all dogs combined, which indicates other causal factors in addition to genetics. In man, familial cases of biliary neoplasms have been uncommon. A genetic component is suggested by the dichotomy of rates between American Indians (high), Mexican Americans (intermediate) and Anglo-Americans (low) in the south-west of the U.S.A., where admixture of the Indian and Mexican population has been common (Fraumeni and Kantor, 1982). There is a notable association, without obvious explanation, of biliary tract cancer among close relatives of patients with the immunodeficiency syndrome of ataxia-telangiectasia (Swift, Sholman, Perry and Chase, 1976). Environmental chemicals have also been implicated in human biliary carcinoma, particularly in workers in aircraft, chemical, rubber and woodfinishing industries. Some of these chemicals also have produced hepatobiliary tumours in laboratory animals (Fraumeni and Kantor, 1982). There was no difference in risk between male and sexually intact female dogs for bile duct carcinoma, but spayed females displayed 1.5 times the risk of intact females for this lesion (P >0*05), suggesting that ovarian hormones may, in some way, protect females of this species. We are not aware of data on the effect of oophorectomy on the incidence of bile duct cancer in women but the use of oral contraceptives and other exogenous oestrogens has been linked with an increased occurrence of gallstones which, in turn, is associated with an increased risk of gallbladder cancer (Fraumeni and Kantor, 1982). The positive association between intestinal blood-letting nematodes and cholangiocarcinoma in dogs offers a new clue to aetiology. Besides C. s&en.&, Opisthorchis felineus has been linked with canine biliary tract carcinoma ; similarly several parasites (C. sinensis, Opisthorchis spp., Giardia lamblia, and Ascaris lumbricoides) have also been suggested as associated with human cholangiocarcinoma (Schoenfield, 1977; Inouye and Whelan, 1978; Burton, 1982). The mechanisms of carcinogenesis are not fully understood. Experimental studies indicate that a prolonged, severe parasitic infection produces hyperplasia of the bile duct epithelium and adenomatous transformation (possibly by trauma or products from the parasite, or both), after which a carcinogen, such as a nitrosamine, may be necessary to promote carcinogenesis (Burton, 1982). Bile, or products of bacterial degredation, may also act as a promoter or co-

BILIARY

CARCINOMA

105

carcinogen. This hypothesis has been suggf%+d as a p&tial explanation for many cancers of the alimentary tract (Lowenfels, 1978). Ectopic migration into the biliary tract has not been reported for either hookworms or whipworms. Human infection with hookworms (Necator americanus) and whipworms (X trichiuru) is common, however, in the southern rural areas of the U.S.A. (e.g. Appalachia) (F ozard, 1978), where unexplained high rates for human biliary cancer have been reported (Fraumeni, 1975). The international pattern for these parasitic infections favours tropical and sub-tropical countries (Benenson, 1970), some of which have also reported very high rates of human biliary cancer (Doll, Muir and Waterhouse, 1970).

SUMMARY

A retrospective study of abstracted data representing 1.1 million hospital examinations of dogs at veterinary university teaching facilities identified 77 with bile duct carcinoma. Using the hospital population, based on patient years at risk, as the expected frequency of particular zoographic characteristics, there was a suggestion of a familial (breed) predisposition in Labrador retrievers and limited evidence of a sex differential. Because mongrel dogs displayed about the same frequency of bile duct cancer as all breeds combined, other causal factors, besides genetics, are implied. A possible association was detected between canine cholangiocarcinoma and host infection by canine hookworms and/or whipworms. Human infection from hookworms (Necator americanus) and whipworms (Trichuris trichiuru) is common in the southern rural areas of the U.S.A. (e.g. Appalachia), and tropical and sub-tropical countries of Africa, Central and South America, where there have been unexplained high rates reported for human biliary cancer. It is possible that the presence of these blood-letting nematodes, like a number of other parasites (ascarids, flukes and protozoa), may be related to the occurrence of human cholangiocarcinoma.

ACKNOWLEDGMENTS

The authors thank Drs A. Kantor, G. Burton and J. F. Fraumeni, comments and MS Stephanie Trimble for editorial assistance.

Jr for critical

REFERENCES

Benenson, A. S., Ed. (1970). Control of Communicable Disease in Man, 11th Edit., The American Public Health Association, New York. Bennion, L. J., and Grundy, S. M. (1978). Risk factors for the development of cholelithiasis in man. New England Journal of Medicine, 299, 1161-l 167. Burton, G. J. (1982). Parasites. In Cancer Epidemiology and Prevention, D. Schottenfeld, and J. F. Fraumeni, Jr, Eds, W. B. Saunders, Philadelphia, Pennsylvania. Diehl, A. K. (1980). Epidemiology of gallbladder cancer: a synthesisof recent data. Journal of the National Cancer Institute, 65, 1209-1214. Doll, R., Muir, C., and Waterhouse, J., Eds (1970). Cancer Incidence in Five Continents, VolumeII. International Union Against Cancer. Springer-Verlag, Berlin. Evdns, H. E., and Christensen, G. C., Eds (1979). Miller’s Anatomy of the Dog, 2nd Edit., W. B. Saunders, Philadelphia, Pennsylvania.

106

H. M.

HAYES,

JR

et d.

Fozard,

G. (1978). Major helminthic diseases of North America: a review. The of Family Practice, 6, 1195-1203. Fraumeni, J. F., Jr (1975). Cancers of the pancreas and biliary tract: epidemiologica considerations. Cancer Research, 35, 3437-3446. Fraumeni, J. F., Jr, and Kantor, A. F. (1982). Biliary Tract. In Cancer Ejidemiology and Prevention, D. Schottenfeld and J. F. Fraumeni, Jr, Eds, W. B. Saunders, Philadelphia, Pennsylvania. Gart, J. J. (1970). Point and interval estimation of the common odds ratio in the combination of 2 x 2 tables with fixed tables with fixed marginals. Biometrika, 57,471-475. Hayes, H. M., Jr, Wilson, G. P., and Moraff, H. (1979). The veterinary medical data program (VMDP): past, present, and future. In Proceedings of the International Symposium on Animal Health and Disease Data Banks, Technical Information Service, Science and Education Administration, U.S. Department ofAgriculture, Beltsville, Maryland. Hou, P. C. (1965). Hepatic clonorchiasis and carcinoma of the bile duct in a dog. Journal of Pathology and Bacteriology, 89, 365-368. Inouye, A. A., and Whelan, T. J. (1978). Carcinoma of the extrahepatic bile ducts. Journal

The American Journal

of Surgery, 136, 90-95.

Lowenfels, A. B. (1978), Does bile promote extra-colonic cancer ? Lancet, ii, 239-241. Nakayama, F. (1969). Composition of gallstone and bile: speciesdifference. Journal of Laboratory

and Clinical

Medicine,

73, 623-630.

Patnaik, A. K., Hurvitz, A. I., Lieberman, P. H., and Johnson, G. F. (1981). Canine bile duct carcinoma. Veterinary Pathology, 18, 439444. Piehler, J. M., and Crichlow, R. W. (1978). Primary carcinoma of the gallbladder. Surgery, Gynecology and Obstetrics, 147, 929-942.

Ponomarkov, V., and Mackey, L. J. (1976). Tumours of the liver and biliary system. Bulletin of the World Health Organization, 53, 187-194. Rehmtulla, A. J. (1974). Occurrence of carcinoma of the bile ducts: a brief review. Canadian Veterinary Journal, 15, 289-292. Ritchie, J. K., Allan, R. N., Macartney, J., Thompson, H., Hawley, P. R., and Cooke, W. T. (1974). Biliary tract carcinoma associatedwith ulcerative colitis. Quarterly Journal of Medicine, 43, 263-279. Rooney, J, R. (1959). Liver carcinoma in the dog. Acta Pathologica et Microbiologica Scandinavica. Section A. Pathology, 45, 321-330. Schall, W. D., Chapman, W. L., Jr, Finco, D. R., Greiner, T. P., Mather, G. W., Rosin, E., and Welser, J. R. (1973). Cholelithiasis in dogs. Journal of the American Veterinary Medical Association, 162, 469-472. Schoenfield, L. J. (1977). T umours of the biliary system. In Diseases of the Gallbladder and Biliary System, John Wiley and Sons, New York. Shani, M., Hart, J., and Modan, B. (1974). Cancer of the biliary system: a study of 445 cases.British Journal of Surgery, 61, 98-100. Sleight, D. R., and Thomford, N. R. (1970). Grossanatomy of the blood supply and biliary drainage of the canine liver. Anatomical Record, 166, 153-160. Swift, M., Sholman, L.., Perry, M., and Chase, C. (1976). Malignant neoplasmsin the families of patients with ataxia-telangiectasia. Cancer Research, 36, 209-2 15. Thorbjarnarson, B. (1975). Carcinoma of the biliary tree : carcinoma of bile ducts. New York State Journal of Medicine, 75, 705-707. Tjalma, R. A., Priester, W. A., Adelstein, E. H., and Vander Zwaag, R. (1964). Clinical records systemsand data retrieval function in veterinary medicine: a proposal for systematic data programming. Journal of the American Veterinary Medical Association, 145, 1189-l 197. Todani, T., Tabuchi, K., Watanabe, Y., and Kobayashi, T. (1979). Carcinoma arising in the wall of congenital bile duct cysts. Cancer, 44, 1134-l 141. Warwick, R., and Williams, P. L., Eds (1973). Gray’s Anatomy, 35th Edit., W. B. Saunders, Philadelphia, Pennsylvania.

BILIARY

CARCINOMA

107

Weaver, A. D. (1976). Radiological diagnosis and prognosis of primary abdominal neoplasia in the dog, Journal of Small Animal Practice, 17, 357-363. Young, J. L., Jr, Percy, C. L., and Asire, A. J., Eds (1981). Surveillance, Epidemiology, and End Results Incidence and Data, Mortality 1973-77 National Cancer Institute Monograph 57, U.S. Government Printing Office, Washington, D.C. [Received for publication,

November 18th, 1981]