Cholangiocarcinoma is relatively rare, but high incidence rates have been reported in Eastern Asia, especially in Thailand. The etiology of this cancer of the bile ducts appears to be mostly due to specific infectious agents. In 2009, infections with the liver flukes, Clonorchis sinensis or Opistorchis viverrini, were both classified as carcinogenic to humans by the International Agency for Research on Cancer for cholangiocarcinoma. In addition, a possible association between chronic infection with hepatitis B and C viruses and cholangiocarcinoma was also noted. The meta-analysis of published literature revealed the summary relative risks of infection with liver fluke (both Opistorchis viverrini and Clonorchis sinensis), hepatitis B virus, and hepatitis C virus to be 4.8 (95% confidence interval [95% CI]: 2.8–8.4), 2.6 (95% CI: 1.5–4.6), and 1.8 (95% CI: 1.4–2.4), respectively – liver fluke infection being the strongest risk factor for cholangiocarcinoma. Countries where human liver fluke infection is endemic include China, Korea, Vietnam, Laos, and Cambodia. The number of infected persons with Clonorchis sinensis in China has been estimated at 12.5 million with considerable variations among different regions. A significant regional variation in Opistorchis viverrini prevalence was also noted in Thailand (average 9.6% or 6 million people). The implementation of a more intensive preventive and therapeutic program for liver fluke infection may reduce incidence rates of cholangiocarcinoma in endemic areas. Recently, advances have been made in the diagnosis and management of cholangiocarcinoma. Although progress on cholangiocarcinoma prevention and treatment has been steady, more studies related to classification and risk factors will be helpful to develop an advanced strategy to cure and prevent cholangiocarcinoma.
Cholangiocarcinomas (CCAs) – primarily cancers of the epithelial cells (mostly adenocarcinoma) in the bile ducts arising anywhere along the intrahepatic or extrahepatic biliary tree(1,2)– are relatively rare, but high incidence rates have been reported in Eastern Asia, especially in Thailand.(3) CCAs are highly fatal tumours, as they are clinically silent in the majority of cases.(4–9) Fortunately, survival of this cancer is improving.(10)
CCA occurs with a highly varying frequency in different areas of the world. CCA is second-most common primary liver cancer and accounted for an estimated 15% of primary liver cancer worldwide;(3) however, it varies widely by region from 5% in Japan(11) and 20% in Pusan (Busan), Korea(12) to 90% in Khon Kaen in Thailand.(3)
Recently, a rising tendency of intrahepatic CCA incidence was reported in Western countries.(13–16) The reasons for this increase are not clear, but some of these increases were attributed to the switch between coding systems going from International Classification of Disease-Oncology-2 (ICD-O-2) to ICD-O-3.
The etiology of CCA in Asian countries appears to be mostly linked to infections, especially infections with the liver flukes Clonorchis sinensis (C. sinensis) and Opisthorchis viverrini (O. viverrini).(17–20) These liver flukes, two close members of the family Opistorchiidae,(21) are foodborne trematodes that chronically infect the bile ducts. Liver flukes induce chronic inflammation leading to oxidative DNA-damage of the infected biliary epithelium and malignant transformation.(21) Chronic infection with either of these two liver flukes is considered to be of major socioeconomic importance to humans and animals in Asian countries such as China, Korea, Vietnam, Laos, and Thailand.(21)
In 1994, infection with C. sinensis was classified by the International Agency for Research on Cancer (IARC) as ‘probably carcinogenic to humans’ (Group 2A) because there was limited evidence in humans, whereas infection with O. viverrini was already classified as ‘carcinogenic to humans’ (Group 1), based on its involvement in the etiology of CCA.(22) In February 2009, the IARC Monograph Working Group for Volume 100B reassessed the carcinogenicity of infection with liver flukes with a comprehensive review of all of the available published literature.(23) Since the 1994 Monograph, two papers showed that prevalence of infection with liver flukes (C. sinensis and O. viverrini) correlates with incidence of CCA,(19,20) and several case–control studies showed a high risk for this cancer.(24–27) Therefore, infections with C. sinensis or O. viverrini are now both classified in Group 1 by IARC, based on ‘sufficient evidence in humans’ for CCA. In addition, a possible association between chronic infection with hepatitis B and C viruses (HBV and HCV) – known to cause hepatocellular carcinoma – and cholangiocarcinoma was also reported by IARC as there is only limited human evidence.(23)
A few review papers on CCA in Asian countries have focused on the relationship of CCA with liver fluke infection,(28,29) and on a disclosed role of HCV infection.(30,31)
In the present paper, the updated CCA epidemiology includes (1) summary risk estimates of the risk factors of CCA (i.e. liver fluke, HBV, and HCV infection) in the form of a meta-analysis of published literature; and (2) epidemiology of liver fluke infection, focusing on East Asian countries.
Relative Risk Estimation of Risk Factors: Systematic Review and Meta-Analysis
Methods. Published literature showing association between CCA and infections with liver fluke, HBV, or HCV was retrieved and a full-text analysis was performed to estimate the summary odds ratio (OR) using a meta-analysis (Table 1). Pooled ORs with a 95% confidence interval (CI) on the basis of both fixed- and random-effect models were estimated. The pooled ORs were separately estimated for intrahepatic CCA or extrahepatic CCA and then the relative risks were compared separately for sensitivity analysis. The overall summary estimates were the combined effect for both intrahepatic CCA and extrahepatic CCA due to an overlap of confidence intervals of the pooled ORs of each CCA. Heterogeneity analysis was carried out to verify if there was a significant difference by studies. For statistical analysis, Stata version 10.0 (StataCorp, College Station, TX, USA) and Comprehensive Meta-Analysis version 2 (Biostat, Englewood, NJ, USA) were used.
Table 1. Summary of study characteristics according to liver fluke, HBV, and HCV infection included in meta-analysis
Matched case–control study. Three hospitals, Maharat Nakornratchasima Hospital (Korat) and Sappasithppasong Hospital, Ubonratchathani (Ubon) both in north-east Thailand, and the National Cancer Institute, Bangkok, Thailand, 1987–1988
Results. The overall summary relative risks of infection with liver fluke (both C. sinensis and O. viverrini), HBV, and HCV were 4.8 (95% CI: 2.8–8.4), 2.6 (1.5–4.6), and 1.8 (1.4–2.4), respectively (Figs 1–3). Heavy alcohol drinking had a two-fold increased risk for CCA and was not statistically significant (OR = 2.1, 95% CI: 0.4–10.2) (data not shown).
Issues for discussion. Many studies have investigated the mechanisms of carcinogenesis induced by O. viverrini infection,(29) but there is a paucity of data regarding mechanistic studies for clonorchiasis.
The relative risks of infection with HBV or HCV for the development of intrahepatic CCA and extrahepatic CCA vary widely according to the prevalence of each viral infection and are inconsistent in their significance through different studies. Even though the etiopathogenesis is still inconclusive, the overall risks of HBV- or HCV-associated CCA in meta-analysis are significant.
The overriding link between most known risk factors and CCA is chronic inflammation and chronic biliary irritation.(29,32,33) It is still unclear whether biliary neoplasms from different sites share common pathogenic features, even if they are linked anatomically and histopathologically;(34) only a few studies have been carried out to investigate mechanisms of carcinogenesis in CCAs according to anatomical location.(35,36) From the epidemiological point of view, the roles of risk factors of intrahepatic CCA and hilar CCA with Klatskin tumor coded as topographically intrahepatic have not yet been clearly elucidated. Further epidemiological research on risk factors according to the anatomical location (intrehepatic vs extrahepatic) and to the macroscopic appearance and/or new histological classification of CCA(37) is needed.
Liver Fluke Infection: Focusing on East Asian Countries
Prevalence, geographic distribution. Countries in which human liver fluke infection is endemic are China, Korea, Thailand, Vietnam, Laos, and Cambodia. Persons with C. sinensis infection were infrequently reported from Singapore and Malaysia, and many of them might have been infected in other countries when traveling or through eating imported fish. However, liver fluke infection occurs in all parts of the world where there are Asian immigrants from endemic areas.(38)
1. Clonorchis sinensis.
China. In 1994, some archaeologists found a large number of C. sinensis eggs in the content of the bowel from an ancient corpse buried at the middle stage of the Warring States Period (475–221 BC) in Hubei, China,(39) indicating that this disease has been present in this province for more than 2300 years. A comprehensive review paper of clonorchiasis in China shows a very wide range of variations of C. sinensis prevalence from 0.08% to 74.7% during 1960–2003.(21) In a survey over the period 2001–2004 in a Chinese endemic area for liver flukes, a total of 217 829 persons were investigated, and 5230 were found to be infected with C. sinensis, the infection rate being 2.4%. The number of infected persons was estimated to be 12.49 million with the highest prevalence in the provinces of Guangdong (16.4%, 2278/13 876) and Guangxi (9.8%, 1365/13 990) in southern China and Heilongjiang (4.7%, 636/13 458) in Korean (minority) communities in north-eastern China.(40) In 2001–2004, in a village of Zhaoyuan county, in Heilongjiang province, 87.8% of 1348 residents were found as being positive by stool examination.(41)
Since clonorchiasis patients were first diagnosed by Ohi in Taiwan, C. sinensis was one of the most important and most common food-borne parasites.(42) Infection with C. sinensis is not rare, but is ethnically and geographically associated.(43) The prevalence rates of infection with C. sinensis were 20–50% before 1980 in endemic areas of Miao-li in northern Taiwan, Sun-moon lake in the center, and Mei-nung in the south. In 2000, a prevalence survey in the above endemic areas showed that the C. sinensis infection rate of freshwater snails and fishes which are intermediate hosts was 0.9%, and high infection rate of Haplorchis infections was also observed.(44) From this result, it might be assumed that decrease of C. sinensis prevalence among residents in the endemic area is due to very low infection rate in intermediate hosts. But the reasons are not explained.
Korea.C. sinensis is currently the most prevalent human parasitic helminth detected by fecal examination in Korea. About 1.8 million people were estimated to be infected in 2004.(45) There has not been much decrease in the infection rate of C. sinensis for almost 30 years; the percentage of egg-positives was 4.6% in 1971, 2.6% (3.8% in males and 1.6% in females) in 1981, 1.4% (1.9% in males and 0.9% in females) in 1997, and 2.9% (3.2% in males and 1.6% in females) in 2004. Egg positivity in males was almost double than that of females.(45) Heavily endemic areas of clonorchiasis in Korea are scattered throughout the country, with the most extensive and intensive endemic regions found mainly along the Nakdong River and the lower reaches of the rivers.(46–48)
Vietnam. Due to a lack of available data from national surveys, there is no accurate estimate of the numbers of infected people in Vietnam, however, the World Health Organization estimated in 1995 that approximately 1 million people were infected with C. sinensis or O. viverrini liver flukes.(38)C. sinensis is prevalent in northern Vietnam, while O. viverrini is predominant in central and southern Vietnam.(49) A couple of publications have reported C. sinensis prevalence as low as 5% or less.(50,51) A study of 1155 villagers in two communes in northern Vietnam showed that the prevalence of C. sinensis infection was 26% (33.8% in males and 11.0% in females) in 1999–2000.(52) A survey on 771 stool samples showed that the positive rate of C. sinensis was 17.2% by the cellophane thick smear method.(53) There is a generally higher prevalence in males and this is often associated with male-oriented social gathering during which raw or pickled fish is consumed, and an increasing prevalence with increasing age was also reported.(51,54) The increased risks for human infections with fish-borne liver flukes (C. sinensis) are supposed to be due to the recent rapid development in aquaculture fish production.(54)
2. Opisthorchis viverrini.
Thailand. In 1980–1981 the prevalence of O. viverrini in the north, north-east, central, and south of Thailand was 5.6%, 34.6%, 6.3%, and 0.01% respectively, with an overall prevalence of 14% (or 7 million people infected).(55) Due to the intensive and continuous control activities, the prevalence of infection in north-east Thailand declined to 15.7% in 2001, and the rates in other areas were 19.3% (north), 3.8% (center), and 0% (south), with an average prevalence of 9.6% (or 6 million people infected).(56) The geographic pattern of O. viverrini infection within the endemic country is not uniform.(57)
Laos. It is estimated that over 2 million people are infected with O. viverrini in Laos,(38) mainly in the southern part of the country, in Saravane, Suvannakhet, and Khammuan along the Mekong River, stretching as far as in the lowlands of the country among people with close ethnic ties to the majority of the north-east Thai population. According to a national survey on primary school children in 17 provinces and the Ventiane Municipality, the rate for O. viverrini was 10.9% (29 846 participants). The regions along the Mekong River such as Khammuane, Saravane, or Savannakhet Province showed a higher prevalence of O. viverrini (32.2%, 21.5%, 25.9%, respectively).(58) More recently, a survey in Saravane district revealed a high prevalence of O. viverrini infection (58.5%) among 814 persons from 13 villages.(59)
Cambodia. In Cambodia, there are few official reports or published data on O. viverrini infection. In 2002, a small survey in primary school children from Kampongcham province showed a prevalence of Opisthorchis sp. of 4.0% from 251 fecal specimens.(60)
Vietnam.O. viverrini is reported to be endemic in the central and southern parts of the country.(49)
Risk factors for liver fluke infection. The main risk factors for liver fluke infection are: (1) consumption of raw or undercooked freshwater fish: many surveys show that people in Thailand,(28,29) Vietnam,(52) China,(21,40) Laos,(61) and Korea(20) have such habits. (2) Poor sanitation: it is common in some endemic regions in China, particularly in the provinces of Guangdong and Guangxi, that ‘lavatories’ are built adjacent to fish ponds, resulting in human excrement containing C. sinensis eggs ending up in the pond water.(21) It is also reported in Laos that 95.5% of houses in some rural villages in Bolikhamxay Province do not have a latrine. More than half of the people in the village use animal and/or human feces as fertilizer, and two-thirds of them eat raw fish.(61) (3) Fresh water aquaculture is developing rapidly, but food quality controls are not in place.(62) (4) Accidental ingestion of C. sinensis metacercariae via hands or utensils: contamination occurs as a consequence of not thoroughly washing after catching and handling freshwater fish in endemic areas.(21)
Transmission. The definitive host is infected by the liver fluke through ingestion of raw or undercooked (i.e. dried, pickled, or salted) infected fish which contain metacercariae – the infective stage of the liver fluke. In southern China and among the Cantonese populations in Hong Kong, raw fish is traditionally eaten after being dipped in rice porridge. Alternatively, large fish are sliced and eaten with ginger and garlic known as ‘yushen’, and with this route of transmission the infection rate tends to increase with age. Many children in the hilly areas of Guangdong and eastern China such as Jiangsu, Shandong, and Anhui provinces often catch fish during play-time and roast it incompletely before eating it. With this mode of transmission, the infection rate tends to decline with age.(40) Koreans eat raw fresh fish soaked in vinegar, red-pepper mash, or hot bean paste with rice wine at social gatherings.(46) The fact that men do so more frequently than women has been given as an explanation for the higher prevalence of the infection among men; however, in heavily endemic areas, no significant differences are observed between genders.(46) When fish is abundant, raw fish is eaten more frequently as opposed to being reserved for special occasions. Vietnamese people also eat raw fish.(52) In Thailand and lowland parts of Laos,(63,64) three types of uncooked fish preparations are noted: (1) koi pla, eaten soon after preparation; (2) moderately fermented pla som, stored from a few days to weeks; and (3) pla ra extensively fermented, highly salted fish, stored for at least 2–3 months.(57,65) At present, koi pla is probably the most infective dish, followed by fish preserved for <7 days, then pla ra and jaewbhong, in which viable metacercariae are rare.(57)
Other Risk Factors of Cholangiocarcinoma
Both intrahepatic CCA and extrahepatic CCA are well-known complications of primary scleosing cholangitis (PSC) in Western countries.(66) The other known risk factors for CCA include cirrhosis, chronic non-alcoholic liver disease, obesity, and hepatolithiasis. Other rarer conditions associated with the development of CCA are bile-duct adenoma, multiple biliary papillomatosis, choledochal cysts, congenital fibropolycystic liver, Caroli’s disease (cystic dilatation of intrahepatic bile ducts), and exposure to the radiopaque medium thorium dioxide (Thorotrast).(4–6,8,14) Approximately 90% of patients diagnosed with cholangiocarcinoma do not have a recognized risk factor in Western countries.(67)
Among these risk factors, hepatolithiasis is a very uncommon disease in the West; in contrast, intra- and extrahepatic bile duct stones are much more common in Eastern Asia.(30) Hepatolithiasis is strongly associated with cholangiocarcinoma with proportions varying from 10% in Japan to more than 60% in Taiwan.(68,69) Similarly, in the cholangiocarcinogenesis of other risk factors, bacterial infection and bile stasis, which are demonstrable in virtually all patients, underlie cholangiocarcinoma development.(30,70,71) This carcinogenic process is not limited to the intrahepatic lesions.
The etiology of CCA in Asian countries appears to be mostly linked to infections, especially infections with liver flukes C. sinensis and O. viverrini which are the strongest risk factor for CCA. In addition, infection with HBV and HCV are significantly associated with the development of CCA.
The mechanisms of adenocarcinoma carcinogenesis in the bile ducts caused by chronic infection with liver flukes include chronic inflammation and resulting oxidative stress. Established mechanistic events for hepatitis B or C viruses in the development of CCAs consist of inflammation, liver cirrhosis, chronic hepatitis, and liver fibrosis.(23)
In Asian countries, an improved recognition of etiological factors such as liver flukes, hepatitis B or C viruses, and environmental exposures has heightened awareness of CCA and improved surveillance in patients with these risk factors. Eating habits such as ingestion of raw or undercooked (dried, pickled, or salted) infected fish which contain metacercariae, however, are still prevalent in endemic areas. More attention should to be paid to primary prevention of liver fluke infection through health education and hygiene improvement. In high HBV endemic regions in Asia, universal vaccination in the 1990s showed long-lasting protection against becoming a carrier, and decreases of chronic liver disease and liver cancer are anticipated. Screening of HCV for blood transfusion has also prevented the transmission of HCV infection since the early 1990s. The burden of CCAs attributed with HBV and HCV are, however, not fully investigated. Recently, advances have been made in the diagnosis and management of CCA. Although progress on CCA prevention caused by liver fluke and treatment has been steady, more studies related to classification and risk factors will be helpful to develop an advanced strategy to control CCA.
The authors wish to thank Katiuska Veselinovic for her assistance in preparing this review and Laurent Galichet for manuscript editing.