1.1.1. Taxonomy

Three of the human liver flukes, Opisthorchis viverrini, O. felineus and Clonorchis sinensis, are Pathologically important food-borne members of the class Trematoda (Beaver et al., 1984). These flukes establish a chronic infection within the smaller intrahepatic bile ducts and occasionally in the pancreas and gall-bladder of humans and other fish-eating mammals.

The life cycle of food-borne trematodes is complex, involving two more intermediate hosts (the first always a snail) and several morphological stages. The consumption of raw or incompletely cooked foods which contain the infective stages is the major risk factor for these infections. As a result, people who enjoy a variety of raw foods often harbour several trematodes (liver and intestinal flukes). Similarities in egg morphology and cross-reactive antigens complicate both parasitological and immunological diagnosis and may confound clinical and epidemiological research on the liver flukes.

Fish-eating mammals, for example dogs, cats, pigs, mink, weasels and civets, may become infected with human liver flukes, and some may act as reservoir hosts (Beaver et al., 1984).

1.1.2. Structure

In humans, Clonorchis measures 8–15 mm long and 1.5–5 mm wide, while the two Opisthorchis species are somewhat smaller—3–12 mm by 1–3 mm. They are covered by a tegument and have an oral sucker at the anterior end and a ventral sucker or acetabulum located posterior at about one-third to one-fifth of the body length. Adult worms are hermaphroditic, with reproductive organs occupying much of the body (Sadun 1955; Komiya, 1966; Beaver et al., 1984; Rim, 1986).

The morphology of the adult worms of O. viverrini, O. felineus and C. sinensis is similar but can be distinguished at the cercarial stage by the bilateral pattern of excretory flame cells (Vajrasthira et al., 1961; Wykoff et al., 1965). The adults differ in the shape of testicular lobes, their location relative to the ovary and the appearance of vitelline glands (Sadun 1955; Wykoff et al., 1965). The metacercariae and juvenile worms bear spines.

The yellowish-brown eggs have a distinct operculum, which opens to release the miracidia when the eggs are ingested by an appropriate species of snail. The posterior end of the egg has a small protuberance, or knob. Eggs average 29 μm long by 15 μm wide for C. sinensis (Ditrich et al., 1992), 27 by 15 μm for O. viverrini (Sadun, 1955; Kaewkes et al., 1991) and 30 by 11 μm for O. felineus (Ditrich et al., 1990), with differences in shape between species.

1.1.3. Life cycle and biology of the adult worm

The life cycle of liver flukes is illustrated in Figure 1.

Figure 1.

Life cycle of liver flukes

Infection with Opisthorchis and Clonorchis is acquired through the consumption of raw or undercooked fish containing the infective stage, called metacercaria. The metacercariae leave the cyst in the duodenum and migrate through the ampulla of Vater via the common and extrahepatic bile ducts to the smaller, proximal bile ducts under the surface of the liver, where they mature. Although most adult worms are found in these ducts, in heavy infections they can be found in the extrahepatic bile ducts, pancreatic ducts and, rarely, the gall-bladder (Hou, 1955; Sithithaworn et al., 1991a). Infection is confined to the lumen of the hepatobiliary tract; there is no phase of tissue migration, even when the common bile duct is severed (Sun et al., 1968).

Clonorchis moves up the biliary tract by attaching and detaching its two suckers and extending and contracting its body. Its attachment to the wall of the bile duct may be secured by adherence of the ventral sucker to the biliary epithelium, leaving the oral sucker free for feeding (Hou, 1955).

About one month after the metacercariae have been ingested, adult worms begin producing eggs, which pass down the bile duct and are excreted in the faeces. Eggs can also be found in gall-bladder bile. The average egg output per gram of faeces per adult Opisthorchis worm ranges from 15 to 180 (Elkins et al., 1991; Sithithaworn et al., 1991b). Density-dependent decreases in fecundity have been documented, which partially explain the wide variation in estimates between these studies (Ramsay et al., 1989). Estimated fecundity in infected people and animals is generally in the range of 1000–5000 eggs per day (Komiya, 1966; Wykoff & Ariyaprakai, 1966; Rim, 1986). The average lifespan of the worms, inferred from epidemiological data, is about 10 years, while the maximal lifespan in the absence of reinfection may exceed 25 years (Attwood & Chou, 1978; Zhu, 1984).

If the eggs reach a body of freshwater (small ponds, streams and rivers, flooded rice fields and large reservoirs), they are ingested by snails. Asexual reproduction in the snail results in the release daily of thousands of cercariae one to two months after infection of the snail. The free-swimming cercariae penetrate the tissue of fish and encyst, becoming fully infective metacercariae after 21 days. Over 80 species of the Cyprinidae family and some 13 species of other families, and possibly freshwater prawns, can serve as the second intermediate host (Komiya, 1966; Vichasri et al., 1982; Rim, 1986; Joo, 1988).

In any defined freshwater body, only 1–3% of snails maybe infected, while up to 100% of fish may contain metacercariae (Vichasri et al., 1982; Rim, 1986; Brockelman et al., 1986). The distribution patterns of metacercariae in fish determine patterns of human exposure in endemic areas. The intensity of liver fluke infection in fish varies from one to hundreds, depending on season, type of water body, species and individual. Transmission is seasonal, as a result of patterns of human faecal contamination, water temperature and snail or fish density.

The prevalence of infection in reservoir hosts varies by area and is not closely associated with human infection patterns. In endemic areas, transmission to snails is due mainly to human eating patterns, poor sanitation and high egg excretion (Sadun, 1955; Rim, 1986); the importance of reservoir hosts is limited.

1.1.4. Immune response to infection

The existence of protective immune responses to liver fluke infections remains unclear (Sirisinha, 1984). Several experimental studies have suggested that small decreases in the establishment or fecundity of worms can be observed in animals that receive repeated infections, spleen cells or serum from infected donors and immunization with parasite antigens (Flavell et al., 1980; Flavell, 1982; Sirisinha et al., 1983; Sirisinha, 1984; Kwon et al., 1987). Flavell and Flavell (1986) reported that animals deprived of T cells had similar worm burden and egg output to intact animals after an equivalent challenge. Wongratanacheewin et al. (1987) reported that O. viverrini infection was associated with reduced immuno-responsiveness to red blood cells and mitogens, an effect that was reversed by praziquantel treatment.

In humans, the parasites clearly survive high levels of parasite-specific immunoglobulins G, A and E in both serum and bile (Wongratanacheewin et al., 1988). While experimental studies suggest that parasites may induce immunosuppression, no evidence of suppressed skin test reactivity or reduced responsiveness during infection has been observed in humans (Wongratanacheewin et al., 1988; Haswell-Elkins et al., 1991a). Epidemiological patterns reveal little evidence of, but do not rule out, protective immunity. There appears to be no decline in prevalence of infection among individuals exposed to infection for decades, and rapid rates of reinfection have been reported after treatment in areas of heavy infection (Upatham et al., 1988).

High levels of parasite-specific antibodies have been reported in people with severe hepatobiliary disease and cholangiocarcinoma (Srivatanakul et al., 1990; Haswell-Elkins et al., 1991a). Antibody levels are correlated more closely with clinical indicators of infection, such as gall-bladder size and function, wall abnormalities and ultrasound echogenicity of the portal triad, than is egg count (Haswell-Elkins et al., 1991a; Mairiang et al., 1992).

1.2.1. Qualitative faecal examination for eggs

Detection of liver fluke infection is most often based on the observation of eggs in faeces. The techniques that have been used, in increasing order of sensitivity, are: direct smear, sedimentation, Kato technique, Stoll's technique and formol-ethyl acetate/ether concentration (FECT) (Viyanant et al., 1983; Feng & Chen, 1985; Zavoikin et al., 1985, 1986; Sithithaworn et al., 1991b; Chen et al., 1994). FECT has been used for quantitative measurements, and the Kato technique in large-scale surveys.

Qualitative diagnosis based on a single reading with Stoll's technique and FECT is highly sensitive (nearly 100%) in people with 20 worms or more, but the sensitivity of a single reading drops to as low as 20% in people with fewer than five worms (Sithithaworn et al., 1991b). Multiple reading of the same sample increases sensitivity up to 20% (Haswell-Elkins et al., 1994a). The sensitivity of the diagnostic techniques used in epidemiological studies determines accurate assessment of prevalence and the effects of intervention.

In patients whose bile ducts are completely obstructed, eggs occur in the gall-bladder bile, and serological methods may be used to determine infection (Kurathong et al., 1985; Pungpak et al., 1985).

Eggs of minute intestinal flukes, e.g. species of Heterophyes, Phaneropsolus and Haplorchis, can be confused with those of Opisthorchis and Clonorchis (Ditrich et al., 1990; Giboda et al., 1991a; Kaewkes et al., 1991; Ditrich et al., 1992).

1.2.2. Quantitative faecal assessment of intensity of infection

The combination of egg counts with worm recovery after treatment is the optimal procedure for assessing intensity of infection (Haswell-Elkins et al., 1994a), as there is a close relationship (Radomyos et al., 1984; Sithithaworn et al., 1991b). Studies of autopsy specimens show, however, that people with high egg counts sometimes do not expel eggs (Ramsay et al., 1989; Elkins et al., 1991).

Daily variation in faecal egg output appears to be minimal and does not greatly affect the accuracy of different techniques (Viyanant et al., 1983; Kurathong et al., 1984; Pungpak et al., 1990).

1.2.3. Serological tests for helminth-specific antibody and antigen

Immunodiagnostic tests for liver fluke infections are considered to be supplementary tools rather than definitive diagnostic assays (Rim, 1986; Sirisinha, 1986). Their use in epidemiological surveys is limited, owing to lack of specificity, lack of differentiation of past and present infections and limited sensitivity (Viyanant et al., 1985; Chen et al., 1987; Hong, 1988; Thammapalerd et al., 1988; Wongratanacheewin et al., 1988).

Serological methods may be preferable in some circumstances, as they indicate exposure that occurred before antihelminthic treatment.

Comparisons between the enzyme-linked immunosorbent assay (ELISA), immuno-fluorescence, complement fixation and indirect haemagglutination for the detection of antibodies against Opisthorchis and Clonorchis show that ELISA is usually the most sensitive and specific. A sensitivity of 90.2% and a specificity of 84% was seen for Clonorchis in a comparison of infected people with people outside an endemic area (Chen et al., 1988). Cross-reactions in crude ELISAs have been reported in sera from patients with intestinal nematodes, schistosomiasis, angiostrongyloidiasis, paragonimiasis and minute intestinal fluke infection (Chen et al., 1988; Poopyruchpong et al., 1990; Ditrich et al., 1991).

Comparisons of antibody responses to crude somatic extracts among infected and uninfected individuals within endemic communities show significant associations between infection and the frequency and level of helminth-specific antibody (Janechaiwat et al., 1980; Srivatanakul et al., 1985; Poopyruchpong et al., 1990; Haswell-Elkins et al., 1991a). The sensitivity of antibody tests in light infections is limited (Haswell-Elkins et al., 1991a), and intensity of infection cannot be inferred from antibody level (Rim, 1986; Haswell-Elkins et al., 1991a).

Chen et al. (1987) described a sandwich ELISA for detecting circulating antigen in sera of people infected with C. sinensis. ELISAs that include a mixture of helminth-specific monoclonal antibodies can be used to detect O. viverrini antigens in faeces, while a 340-base-pair DNA probe can be used to detect helminthic DNA in faeces (Sirisinha et al., 1991, 1992).

1.2.4. Intradermal tests

Skin testing with a diluted extract of adult C. sinensis antigens has been used widely in China and the Republic of Korea in epidemiological and surveillance studies, but this method is less commonly used today. The estimated specificity and sensitivity of the reaction was reported to be 98% by Komiya (1966), but lower values (83 and 78%) were reported subsequently (Rim, 1986).

1.3.1. Geographical distribution

The worldwide distribution of O. viverrini, O. felineus and C. sinensis is shown in Figures 2 and 3.

Figure 2.

Worldwide distribution of Opisthorchis viverrini and O. felineus

Figure 3.

Worldwide distribution of Clonorchis sinensis

Countries in which human liver fluke infection is endemic are China, Japan, the Republic of Korea, Laos, Thailand, Viet Nam, the Russian Federation, the Ukraine and parts of eastern Europe. A very crude estimate of the global number of infections is of the order of 17 million, comprising seven million with Clonorchis, nine million with O. viverrini and 1.5 million with O. felineus (WHO, 1994). Regional and global migration of peoples has broadened the distribution of the helminths. Since their life cycles usually do not become established, this widened distribution has limited epidemiological relevance, but, given the potential severity of disease, it is of clinical importance (Chan & Lam, 1987). As details of the sampling methods and examination techniques used are sometimes omitted from survey reports, the sensitivity and representativeness of the measurements cannot be evaluated.

1.3.2. Risk factors for infection

1.3.3. Age- and sex-related patterns of infection

While the levels of O. viverrini infection vary considerably between villages in Thailand, the patterns of infection are fairly similar. In general, the youngest age groups (often 0–4 years) show low prevalence and intensity, while these increase in the pre- and early teens and often reach a plateau in late teenage groups (i.e. 15–19). In some areas, the intensity of egg excretion continues to increase with age (Upatham et al., 1982), while the worm burden may decline (Haswell-Elkins et al., 1991b; Sithithaworn et al., 1991a).

Anecdotal descriptions have been reported of mothers in the Republic of Korea and Thailand feeding raw fish to their infants (Choi, 1984), and infections have been observed in young infants (Sadun, 1955; Harinasuta & Vajrasthira, 1960; Upatham et al., 1982, 1984). The reported intensities of infection in children under the age of four are, however, invariably very low, and there is little evidence that young children have ever had frequent, intense exposure to infection.

The prevalence and average intensity of O. viverrini infection do not usually differ, or are slightly higher, among males than females (Wykoff et al., 1966; Upatham et al., 1982, 1984; Haswell-Elkins et al., 1991b; Elkins et al., 1994). Even in areas where these measurements do not differ significantly with sex, higher frequencies of heavy infection may be observed among males (Haswell-Elkins et al., 1991b; Elkins et al., 1994).

In general, the prevalence of Clonorchis infection appears to rise at later ages, and differences in prevalence and intensity between the sexes are more pronounced than those of Opisthorchis (Figure 4). For example, in several river basins in the Republic of Korea, large increases in prevalence are observed between the ages of 10–19 and 20–29. Sometimes this is apparent only in males, and females maintain relatively low prevalences throughout life, while in other areas the two sexes have virtually identical age-related patterns of infection. Most studies in Japan show maximal prevalences at 30–50 years of age (e.g. 39–67% and 8.8–46%) (Rim, 1986). This finding appears to be generally true in China, except in areas where children become infected by catching and eating undercooked fish during play (Chen et al., 1994).

Figure 4.

Prevalence (a) and intensity (b) of infection with Clonorchis sinensis in an area of low intensity in the Republic of Korea; prevalence (c) and intensity (d) of infection with C. sinensis in an area of high intensity in the Republic of Korea; prevalence (more...)

1.3.4. Aggregation of infection

The population of O. viverrini, and probably all three liver flukes, is highly aggregated within a small minority of people who are heavily infected. For example, Haswell-Elkins et al. (1991b) observed that 81% of 11 027 worms recovered after treatment of 246 village residents were expelled by just 27 individuals with burdens of over 100 worn Similarly, Sithithaworn et al. (1991a) reported that 30 of 181 cadavers examined contained 66% of all worms recovered at autopsy.

The levels of infection vary considerably between communities within the same province and district, for unknown reasons. Tesana et al. (1991) found higher prevalences of infection in six villages located far from a river than in villages situated along river banks. This observation is in contrast to the patterns usually reported for Clonorchis infection and may reflect variation in the habitats of infected fish.

2.1.1. Opisthorchis viverrini

All of the available studies are from Thailand, where there is substantial geographical variation in the prevalence of infection, increasing from the south to the north, the highest rates being observed in Khon Kaen Province in North-east Thailand (see section 1.3.1a). In incidence data from the national cancer registry, the highest frequency was observed in North-east Thailand in 1980–82 (Srivatanakul et al., 1988) and again, especially in Khon Kaen Province, in 1988–91 (Vatanasapt et al., 1993). In the earlier period, the proportionate incidence ratio was 3.1 (95% confidence interval [CI], 2.8–3.5) for cholangiocarcinoma and was 1.2 (95% CI, 1.1–1.4) for hepatocellular carcinoma (Srivatanakul et al., 1988). In Khon Kaen Province around 1985, the age-standardized incidence rate of cholangiocarcinoma was 84.6 per 100 000 per year in men and 36.8 per 100 000 per year in women. Outside of Thailand, the incidence of cholangiocarcinoma shows little variation (range, 0.2–2.8 per 100 000 per year in men, and 0.1–4.8 per 100 000 per year in women) (Parkin et al., 1993). Thus, the incidence in the area of highest incidence in Thailand is at least 40 times as high as that in the area of highest incidence elsewhere.

Within Khon Kaen Province, during the period 1985–88, Vatanasapt et al. (1990) observed the highest incidence and mortality rates of liver cancer in three adjacent districts; studies in two of the districts had shown high prevalences of infection and heavy infection (Upatham et al., 1984). Subsequently, Sriamporn et al. (1993) showed that there was no difference in the overall prevalence of infection between the districts of highest and lowest incidence of liver cancer within the Province during the period 1988–90; however, 9% of 331 subjects from randomly selected villages in the district of highest incidence had > 10 000 fluke eggs/g of stool, while only 3.7% of 296 subjects in villages in the district of low incidence had the same level of infection.

Srivatanakul et al. (1991a) carried out a correlation analysis of liver cancer incidence, titre of antobodies to O. viverrini and faecal egg count (determined in healthy volunteers who had been born and resided in the area) in five regions with different frequencies of the two main histological types of liver cancer: Chiang Mai in the north, Nakon Ratchasima and Ubon Ratchathani in the north-east (but not in Khon Kaen Province), Bangkok in the centre and Songkhla in the south. The correlation between the incidence of cholangiocarcinoma and the proportion of subjects with an antibody titre ≥ 1:40 was 0.98 (p = 0.004), and that with faecal egg count was 0.53 (p = 0.35). For hepatocellular carcinoma, which showed little geographical variation in incidence, the correlations were −0.37 (p = 0.54) and 0.02 (p = 0.96), respectively. [The weaker association between cholangiocarcinoma and faecal egg count may reflect the introduction of effective therapy; antibody titre is thought to provide a more valid indicator of past infection, but cross-reactivity with other parasites common in the region may have been involved.]

These studies are summarized in Table 1.

Table 1.

Descriptive studies of Opisthorchis viverrini and liver cancer in Thailand.

2.1.2. Opisthorchis felineus

In the T'umen' region in western Siberia (an area of O. felineus endemicity), Shain (1971) related the prevalence of infected people in four subregions as reported by local health centres with the incidence of liver cancer observed in the same period, 1960–69. The correlation computed by the Working Group from the tabulated data was [0.98; p < 0.05]. A similar analysis in seven cities within one of the regions confirmed this correlation [0.77]. No information was given on the relative frequency of histological types.

2.2.1. Opisthorchis viverrini

All of the available reports are from Thailand. The earliest case reports are of a papillary adenocarcinoma of the liver and an adenocarcinoma of the bile duct (Viranuvatti & Mettiyawongse, 1953) and a retention cyst of the liver caused by opisthorchiasis associated with carcinoma of the liver (Viranuvatti et al., 1955); O. viverrini infection was detected at autopsy in each case. Subsequent case series are summarized in Table 2. Among patients from the area in which O. viverrini is endemic, cases of cholangiocarcinoma outnumber cases of hepatocellular carcinoma, in contrast to other series.

Table 2.

Case series of patients with liver cancer associated with Opisthorchis viverrini infection in Thailand.

Cancers other than of the liver have been reported in association with this infection, but no particular type has predominated (Koompirochana et al., 1978; Pungpak et al., 1985).

2.2.2. Opisthorchis felineus

Three studies on the presence of O. felineus infection in liver cancer cases were conducted in western Siberia (Table 3). One of the regions, T'umen', is reported to be an area of high endemicity. The prevalence of infection in 250 histologically verified cases of liver cancer was 52% in the study of Shain et al. (1971). The prevalence of infection in 44 cases of liver cancer detected in 657 autopsies performed in the same region was 95% (Glumov et al., 1974). The first study also reported a higher frequency of cholangiocarcinoma among infected liver cancer cases and a difference in the sex ratios between the two main histological types [no information was provided about the sex ratio of infection].

Table 3.

Prevalence of Opisthorchis felineus in case series of liver cancer in western Siberia in the Russian Federation.

In a similar study conducted in a region of intermediate endemicity, 7 liver cancers out of 103 detected at autopsy were infected with O. felineus (Iablokov et al., 1980). Similar proportions of cases of cholangiocarcinoma (4/47) and hepatocellular carcinoma (3/56) were infected.

2.2.3. Clonorchis sinensis

The earliest case reports of primary liver cancer concerned Chinese subjects (Watson-Wemyss, 1919; Bentham, 1920; Nauck & Liang, 1928; Ch'in et al., 1955). Subsequent case series, from Hong Kong and the Republic of Korea and among Asian subjects in the USA, are summarized in Table 4. Cases have also been described in immigrants to North America from China (Schwartz, 1986; Colquhoun & Visvanathan, 1987) and Laos (Drinka & Sheehy, 1985; Sher et al., 1989; Ona & Dytoc, 1991). The only other population in which cases have been reported is that of Japan (Nakashima et al., 1977).

Table 4.

Case series of patients with cancer of the liver associated with Chlonorchis sinensis infection.

2.3.1. Opisthorchis viverrini

Kurathong et al. (1985) assessed the prevalence of cholangiocarcinoma and hepatocellular carcinoma during 1981–83 in 551 (47%) patients from the north-east (49.8% of those attending a hospital in Bangkok) who agreed to provide stool specimens, on the basis of which they were characterized for the presence of O. viverrini eggs. All 551 were screened for hepatobiliary tract diseases. Nineteen of 25 cases of cholangiocarcinoma and 9 of 12 of hepatocellular carcinoma had ova in the stools. The cases were diagnosed by a variety of methods, including ultrasound biopsy and hepatic angiography. The crude prevalence odds ratios were [1.3 (0.5–3.6)] for cholangiocarcinoma and [1.3 (0.3–4.7)] for hepatocellular carcinoma. [Use of controls with other hepatobiliary disease may have biased the results.]

A hospital-based case–control study of cholangiocarcinoma (Parkin et al., 1991) and hepatocellular carcinoma (Srivatanakul et al., 1991b) was carried out in Thailand, in which 103 cholangiocarcinoma patients and 65 hepatocellular carcinoma patients living in and originating from North-east Thailand were recruited in 1987–88 from among patients whose disease was diagnosed sequentially in three hospitals. One control was matched to each case for sex, age (within five years), residence and hospital of recruitment. Controls were selected from among patients affected by a variety of non-malignant diseases, considered not to be related to the consumption of alcohol or tobacco. Infection with O. viverrini was assessed in terms of an increase in titre of antibodies to O. viverrini in serum as observed by ELISA (Srivatanakul et al., 1985). For cholangiocarcinoma, the matched estimate of the odds ratio obtained from the final multivariate model, including adjustment for consumption of ‘sticky’ rice and betel-nut chewing, was 5.0 (95% CI, 2.3–11.0). No association was seen with chronic carriage of hepatitis B virus nor with recent aflatoxin intake (Parkin et al., 1991). O. viverrini infection was not significantly associated with the risk of developing a hepatocellular carcinoma. The observed odds ratio was 1.7 (0.8–3.7). In a multivariate analysis, there was a strong association with chronic carriage of hepatitis B virus (Srivatanakul et al., 1991b).

Haswell-Elkins et al. (1994a) conducted a cross-sectional population-based survey in 1990–91 of subjects aged 25 or more from 46 villages in two districts of Khon Kaen Province and 39 villages in Maha Sarakham Province, within the endemic area of O. viverrini infection in North-east Thailand. Stool specimens were obtained from 7727 subjects (participation rate, 72%) in Khon Kaen Province and 4585 subjects (participation rate, 79%) in Maha Sarakham Province after a health education programme about liver fluke infection. A 15% random sample of 1807 uninfected and lightly infected (< 3000 fluke eggs/g) subjects and all subjects with higher intensities of infection were invited to undergo an ultrasound examination. Among the 78% of subjects who complied, 44 had evidence of cholangiocarcinoma without overt symptoms. In nine of these, the diagnosis was corroborated by endoscopic retrograde cholangiopancreatography; a further six who died before they could undergo the procedure or who declined it were strongly suspected to have cholangiocarcinoma. Thus, there was a total of 15 cases, seven in patients who died with jaundice and hepatomegaly in 1991–92. Among 410 uninfected subjects, one case occurred. The multivariate prevalence odds ratios, accounting for age, sex and district of residence, were 1.7 (95% CI, 0.2–16.3) for subjects with up to 1500 fluke eggs/g, 3.2 (0.4–30) for subjects with 1501–6000 eggs/g and 14 (1.7–119) for more heavily infected subjects.

2.3.2. Clonorchis sinensis

In a consecutive series of 1484 autopsies in a single hospital in Hong Kong during the period 1964–66, clonorchiasis was found on gross examination in 11 of 17 (65%) cases of cholangiocarcinoma and in 24 of 83 (29%) cases of hepatocellular carcinoma. The expected proportions infected, on the basis of the whole series and adjusted for age and sex, were 38 and 35%, respectively. [The odds ratios, adjusted for age and sex, calculated by the Working Group, were 3.1 (95% CI, 1.1–8.4) for cholangiocarcinoma and 0.73 (0.45–1.2) for hepatocellular carcinoma] (Gibson, 1971).

Kim et al. (1974) studied records of autopsy and surgical specimens from one hospital in an area of low prevalence of C. sinensis (Seoul) and one hospital in an area of high prevalence (Pusan) in the Republic of Korea during the period 1961–72. In the area of low prevalence, a total of 386 histologically proven cases of primary liver cancer were identified among 1447 subjects with liver disease, and in the area of high prevalence, there were 109 cases of primary liver cancer among 396 subjects with liver disease. C. sinensis infection was determined by examination of liver tissue or stool samples. Comparison of cases of liver cancer with subjects with liver disease in whom cancer was not found showed a weak positive association between the cancer and C. sinensis infection [odds ratio, 1.7; 95% CI, 1.2–2.3]. The corresponding odds ratio for cholangiocarcinoma, based on 54 cases, was [6.5 (95% CI, 3.7–12)] and that for hepatocellular cancer, based on 423 cases, was [1.2, 0.80–1.7].

In Pusan, Republic of Korea, one of the areas of highest prevalence of C. sinensis infection, the occurrence of clonorchiasis was determined in stool specimens from 206 of a consecutive series of 368 cases of primary liver carcinoma diagnosed mainly in two hospitals during the period 1963–74 (Chung & Lee, 1976). [The Working Group noted that as one of these hospitals had been included in the study of Kim et al. (1974), there is some overlap with that study.] The control series comprised 559 subjects admitted to these hospitals with diseases other than of the liver; again, the presence of clonorchiasis was determined from stool specimens [no further details]. The crude odds ratio for cholangiocarcinoma, based on 36 cases, was [6.0 (95% CI, 2.8–13)]; the odds ratio was unchanged after adjustment for age and sex. The crude odds ratio for hepatocellular carcinoma, based on 170 cases, was 1.1 (95% CI, 0.65–1.7).

These studies are summarized in Table 5.

Table 5.

case–control studies of the association between Chlonorchis sinensis infection and cholangiocarcinoma and hepatocellular carcinoma.

3.2.1. N-Nitrosodimethylamine

Hamster: Male Syrian golden hamsters, aged three to four weeks, were divided into four groups: 18 animals served as untreated controls; 21 animals received 0.0025% [25 mg/L] N-nitrosodimethylamine (NDMA) in the drinking-water starting from seven to eight weeks of age; 18 animals were infected with 100 O. viverrini metacercariae by intragastric intubation; and 21 animals were infected with O. viverrini and, four weeks later (as soon as the parasitic eggs were detected in faeces), received NDMA in the drinking-water. NDMA treatment was discontinued after 10 weeks, and animals were killed eight weeks thereafter (at 23 weeks). All of the animals that received NDMA and were infected developed cholangiocarcinoma and cholangiofibrosis. No such tumour was observed in the group that received either NDMA or parasite alone [p < 0.001; Fisher exact test], although cholangiofibrosis was found in some NDMA-treated animals (Thamavit et al., 1978).

A total of 130 male Syrian golden hamsters, six to eight weeks of age, were divided into three groups: 50 animals were infected with 50 O. viverrini metacercariae by intragastric intubation, followed 41 days later by a single oral dose of 1.6 mg NDMA; 30 animals received a single oral dose of 1.6 mg NDMA on day 41; and 50 animals were infected with 50 O. viverrini metacercariae. Animals were maintained for 70 weeks or were killed when moribund. Cholangiocarcinomas developed in 5/50 infected animals given NDMA at latent periods of 18, 21, 29 (two animals) and 42 weeks after NDMA treatment. No malignant bile-duct tumour was found in any of the hamsters given either NDMA or metacercariae alone, but benign cystic cholangiomas [numbers not specified] were found commonly in these animals (Flavell & Lucas, 1982). [The Working Group noted that the authors did not report cholangiofibrosis in any of the groups. They also noted the single treatment and small dose of the carcinogen.]

A total of 176 male Syrian golden hamsters, six to eight weeks of age, were divided into four groups: 50 animals were infected with 50 O. viverrini metacercariae by intragastric intubation, followed 41 days later by a single oral dose of 1.6 mg NDMA; 46 animals received a single oral dose of 1.6 mg NDMA, followed 96 h later by infection with 50 O. viverrini metacercariae; 30 animals received a single oral dose of 1.6 mg NDMA; and 50 animals were infected with 50 O. viverrini metacercariae. Animals were killed when in poor condition or at the end of the 490-day experimental period. Mortality was highest in infected animals that received NDMA. Cholangiocarcinomas were observed in 5/50 animals (10%) that were first infected and then received NDMA and in 9/46 animals (20%) that received NDMA and were then infected. The difference between these two groups was not significant [Fisher's exact test]. None of the animals given NDMA alone or only infected with parasites developed malignant bile-duct tumours. The mean tumour latency was 249 days (range, 124–346 days) for the group that was first infected and then received NDMA, and that for the group that first received NDMA and were then infected was 308 days (range, 184–393 days); the difference was not significant. Tumours were most frequently found in the right liver lobe, the lobe in the hamster that also contains the largest proportion of O. viverrini worms (Flavell & Lucas 1983).

A total of 280 male Syrian golden hamsters, three to four weeks of age, were divided into four main groups: one remained untreated; others were infected with 12, 25, 50 or 100 O. viverrini metacercariae by intragastric intubation; further groups were administered NDMA at 3, 6 or 12 mg/L in the drinking-water at four to five weeks of age for 10 weeks; and others were infected with 12, 25, 50 or 100 metacercariae two weeks before administration of NDMA at 3, 6 or 12 mg/L in the drinking-water for 10 weeks. All animals were then maintained on basal diet until the end of the experiment at week 40, at which time they were killed. Only 2/17 animals (12%) in the group that received NDMA at 12 mg/L had detectable cholangiocarcinomas. No neoplastic lesion was seen in those that received NDMAat 6 mg/L or 3 mg/L, in those only infected or in untreated controls. In contrast, significant increases in the incidence of cholangiocarcinomas were seen in animals given both NDMA and metacercariae: 14/15, 10/17, 13/19, 7/10 [p < 0.01; Fisher's exact test]; and cholangiofibrotic lesions were observed (Thamavit et al., 1987a).

Nitrite and aminopyrine can form NDMA in the stomach under certain conditions. A total of 150 male Syrian hamsters, three to four weeks of age, were divided into eight groups: one group was untreated; a second received 0.1% sodium nitrite in the drinking-water; one received 0.1% aminopyrine in the drinking-water; one received sodium nitrite and aminopyrine in the drinking-water; one was infected with 100 O. viverrini metacercariae by a single intragastric intubation; one was similarly infected and four weeks later received sodium nitrite in the drinking-water for 8 or 10 weeks; one was infected and four weeks later received aminopyrine in the.drinking-water for 8 or 10 weeks; and the last was infected and four weeks later received sodium nitrite and aminopyrine in the drinking-water for 8 or 10 weeks. Hamsters that received the eight-week drinking-water treatment were killed 12 weeks later, and animals that received the treatment for 10 weeks were killed 20 weeks later. Combined administration of nitrite and aminopyrine for 8–10 weeks resulted in development of two hepatocellular nodules, seven cholangiofibrotic lesions and three cholangiocellular carcinomas. Prior infection with O. viverrini metacercariae induced inflammatory and proliferative changes in the livers of infected hamsters and was associated with a significant increase in the incidences of hepatocellular nodules (8; p < 0.05), cholangiofibrosis (18; p < 0.05) and cholangiocarcinomas (14; p < 0.01) (Thamavit et al., 1988a).

A total of 105 male Syrian hamsters, six to eight weeks of age, were divided into four groups: 50 animals received a single intraperitoneal injection of 20 mg/kg bw NDMA, followed 19 days later by infection with 80 O. viverrini metacercariae by single intragastric intubation; 25 animals received the intraperitoneal dose of NDMA only; 15 animals were infected with O. viverrini only; and 15 animals served as untreated controls. Hamsters were killed when they became moribund or at the end of the experiment at 45 weeks. Among the 43 animals treated with both NDMA and O. viverrini, 19 developed cholangiocarcinomas, 40 developed cholangiofibrosis, 15 developed mucinous cystadenomas, 2 developed hepatocellular carcinomas and 42 developed hepatocellular nodules. Although 17/20 (85%) of the hamsters treated with NDMA alone developed hepatocellular nodules, with an average of 3.0 nodules per animal, there was an average of 9.5 nodules per animal in the combined treatment group. No lesion was observed in untreated controls, and 2/15 animals only infected with the parasite developed cholangiofibrosis. The difference in incidence of cholangiocarcinomas between the combined group (19/45) and the group only infected with O. viverrini (0/20) was significant (p < 0.001; Fisher's exact test) (Thamavit et al., 1994).

3.2.2. N-Nitrosodiethylamine

Hamster: A total of 180 female Syrian hamsters, three to four weeks of age, were divided into eight groups: 20 animals served as untreated controls; 20 animals were infected by gastric intubation with 60 O. viverrini metacercariae only; groups of 20–30 animals were infected with 60 O. viverrini metacercariae, followed four weeks later by administration of 10, 20 or 40 mg/L N-nitrosodiethylamine (NDEA) in the drinking-water for 12 weeks; and groups of 20–25 animals were administered only 10, 20 or 40 mg/L NDEA in the drinking-water for 12 weeks. The animals were killed at week 32. Infection with 60 metacercariae four weeks before administration of 20 or 40 mg/L NDEA resulted in significantly (p < 0.01) increased incidences of hepatocellular nodules in the groups also receiving NDEA (12/19 and 23/25, with 2.5 and 7.1 nodules/animal) when compared with the groups that received NDEA alone (3/19 and 9/21 with 0.2 and 0.9 nodules/animal). A high incidence of cholangiofibrosis was seen in animals receiving the combined treatment (Thamavit et al., 1987b).

In a further study, 95 female Syrian golden hamsters, six to eight weeks of age, were divided into five groups: a group of 20 animals received a single intraperitoneal injection of 150 mg/kg bw NDEA dissolved in saline, and two groups of 20 animals each received NDEA followed 18 days later by infection with 50 or 100 O. viverrini metacercariae by intragastric intubation; 20 animals received 100 metacercariae without prior treatment with NDEA, and 15 animals were untreated. The animals were killed at the end of week 41. Infection with either 50 or 100 metacercariae of O. viverrini after NDEA injection resulted in significantly (p < 0.01) enhanced incidences of hepatocellular nodules/animal: 4.3 and 6.8 versus 1.4 in animals treated with NDEA alone (Thamavit et al., 1992a).

3.2.3. N-Nitrosodihydroxydi-n-propylamine

Hamster: A total of 75 male Syrian golden hamsters, three to four weeks of age, were divided into four groups: 25 animals were infected with 100 metacercariae of O. viverrini per animal by gastric intubation and two and four weeks later received intraperitoneal injections of 1000 mg/kg bw N-nitrosodihydroxydi-n-propylamine (NDHDPA); 20 animals were treated with NDHDPA alone; 15 animals were infected with O. viverrini alone; and 15 animals served as untreated controls. Animals were killed at week 22. In the group treated only with NDHDPA, 2/20 animals had basophilic hepatocellular foci. Among 19 animals receiving combined treatment with NDHDPA and O. viverrini, six developed cholangiocarcinomas [p = 0.02], 18 developed cholangiofibrosis [p = 0.001] and nine developed hepatocellular nodules [p = 0.002] [all Fisher's exact test]; all 19 had hepatocellular basophilic foci, and eight had atypical proliferation of the pancreatic duct. Two of 20 animals given NDHDPA alone had hepatocellular basophilic foci (Thamavit et al., 1988b).

A total of 100 male Syrian hamsters, three to four weeks of age, were divided into four groups: 10 animals served as untreated controls; 20 animals were infected with 80 O. viverrini metacercariae by intragastric intubation; 30 animals received three intraperitoneal injections of 500 mg/kg bw NDHDPA at weeks 16, 17 and 18; and 40 animals were infected with 80 O. viverrini and received similar NDHDPA treatment. Animals were maintained on basal diet until they were killed, at week 52, when they were examined histologically. Cholangiocarcinomas occurred in 8/16 animals in the combined treatment group and 0/16 in that receiving NDHDPA alone [p = 0.001; Fisher's exact test]. Liver foci were seen in 16/16 hamsters in the combined treatment group and in 14/16 of those given NDHDPA, but the group receiving the combined treatment had a significantly increased number of foci per cm² (23.4 ± 7.5 versus 3.5 ± 2.6; p < 0.001) (Moore et al., 1991).

3.5.1. Rat

As part of a combination study (see section 3.6.1), a control group of 25 male Wistar albino rats, 8–10 weeks of age, was administered 50 Clonorchis sinensis metacercariae by intragastric intubation. A few hepatic necrotic foci and mild inflammatory cell changes were seen in animals from each group killed at 4, 8, 12, 16, 20, 24 and 28 weeks after infection. Neither bile-duct lesions nor liver tumours were observed (Park, 1989). [The Working Group noted the short duration of the study and the inadequate reporting.]

In a further combination study, a control group of 10 male Fischer 344 rats, six weeks old, were each infected with 60 C. sinensis metacercariae by intragastric intubation and killed after 40 weeks. The infected animals developed cholangiocellular lesions, including bile-duct proliferation, periductal inflammation, fibrosis with occasional mucinous metaplasia, particularly at the main duct, and extensive areas of ductular proliferation. No tumour was observed (Jang et al., 1990). [The Working Group noted the short duration of the study.]

3.5.2. Cat

Three cases of cholangiocarcinoma associated with C. sinensis infection were reported in cats (Felis catus) (Hou, 1964). Two of the cases were found at necropsy in two approximately four-year-old, well-developed, well-nourished cats out of a total of 215 obtained at random. The two cats harboured 150 and 200 adult C. sinensis in the liver. The third case was also in a four-year-old cat, which was one of 26 infected experimentally by feeding a diet of fish (Ctenopharyngodon idellus, Hypophthalmichthys nobilis and Mylopharyngodon aethiops) flesh containing metacercarial cysts of C. sinensis for 28 feedings. The animal died of bronchopneumonia; 105 C. sinensis were recovered from the bile ducts. The authors reported that the histopathological features of cholangiocarcinoma in the three cats were similar to those of many forms of bile-duct cancer found in humans infected with C. sinensis (Hou, 1956).

3.5.3. Dog

Cholangiocarcinoma associated with C. sinensis infection was also reported in one well-developed, well-nourished eight-year-old female chow dog, which had suffered from abdominal enlargement for an unknown period before death (Hou, 1965a). The histopathological features of the cholangiocarcinoma were reported to be similar to those of a form of bile-duct cancer found in humans infected with C. sinensis (Hou, 1956).

3.6.1. Aflatoxin B₁

A total of 75 male Wistar albino rats, 8–10 weeks old were divided into three groups: 25 rats were fed aflatoxin B₁ at 1 mg/kg diet for 12 weeks; 25 rats were infected by administration of 50 C. sinensis metacercariae by intragastric intubation; and 25 animals were infected with C. sinensis and fed aflatoxin B₁ in the diet concomitantly. Three rats from each group were killed at four-week intervals up to 28 weeks after the beginning of treatment. Well-differentiated hepatocellular carcinomas were detected in two of three rats given the combined treatment and alive at 28 weeks; such tumours were not seen in rats treated with aflatoxin B₁ alone and killed at the same intervals (Park, 1989). [The Working Group noted the inadequate reporting of the study and the small comparison groups in the serial killings.]

3.6.2. N-Nitrosodimethylamine

Rat: A total of 101 male Fischer 344 rats, six weeks of age, were divided into six groups: 20 animals were each infected with 60 C. sinensis metacercariae by single intragastric intubation four weeks before receiving NDMA at 25 mg/L in the drinking-water for eight weeks; 20 animals were infected with C. sinensis while receiving NDMA at 25 mg/L in the drinking-water for eight weeks; 20 animals were infected with C. sinensis one week after NDMA treatment; 19 animals received NDMA in the drinking-water alone for eight weeks; 10 animals were infected with C. sinensis alone; and 12 animals served as untreated controls. The animals were killed at week 40, and all were found to have heavy helminthic loads. Livers were examined immunohistochemically for foci of the placental form of glutathione S-transferase. Animals infected before NDMA administration had significantly (p < 0.05) increased numbers of foci. No such effect was seen when animals were infected with C. sinensis during or after exposure to NDMA (Jang et al., 1990).

Hamster: A total of 48 Syrian golden hamsters [sex unspecified], three to four weeks old, were divided into four groups: 12 animals received NDMA at 15 mg/L in the drinking-water for eight weeks and were given 10 metacercariae of C. sinensis suspended in saline by intragastric intubation seven days after the beginning of NDMA administration; 12 animals received the NDMA treatment alone; 12 received the helminthic treatment alone; and 12 animals served as untreated controls. After 11 weeks, 6/8 (75%) infected animals given NDMA developed cholangiocarcinomas, 8/8 developed cholangiofibrosis and 8/8 developed cholangiofibroma. Of the 12 animals given NDMA alone, two developed cholangiofibrosis and cholangiofibroma; of those given the helminth alone, 5/12 developed cholangiofibrosis. No lesions were observed in the 12 untreated controls (Lee et al., 1993).

A total of 90 Syrian golden hamsters [sex unspecified], weighing 50–60 g, were divided into six groups of 15 animals each: one group received NDMA at 15 mg/L in the drinking-water for four weeks, followed one week later by administration of 15 metacercariae of C. sinensis suspended in saline by intragastric intubation; five weeks later the animals received oral administrations of 200 mg/kg bw praziquantel daily for three days. Another group was similarly infected with C. sinensis metacercariae but was treated with praziquantel for three days before treatment with NDMA. A further group received concomitant administration of NDMA and infection with C. sinensis. One control group received NDMA and another was infected with the helminth only. A final group served as untreated controls. At the end of 13 weeks, the group that had received concomitant treatment with NDMA and C. sinensis had 11/15 cholangiocarcinomas, 3/15 cholangiofibromas and 1/15 cholangiofibroses. In the group infected one week after NDMA treatment and given praziquantel, 3/15 had cholangiocarcinomas, 3/15 had cholangiofibromas and 6/15 had cholangiofibroses. In the group with combined treatment but given praziquantel three days before NDMA, 11/15 animals developed cholangiofibroses. In the group given NDMA alone, 4/15 animals had cholangiofibroma and 5/15 had cholangiofibroses; and in the group receiving only helminthic infection, 12/15 animals developed cholangiofibroses. No cancerous or precancerous lesion of the bile duct was found in the untreated control group (Lee et al., 1994).

3.6.3. 2-Acetylaminofluorene

Hamster: Groups of 50 and 60 female Syrian golden hamsters, 8–10 weeks old, received 0 or 40 C. sinensis metacercariae per animal orally and were fed diets containing 0.03% 2-acetylaminofluorene for 40 weeks. After this time, all surviving animals were fed normal diets without carcinogen. Small numbers of animals from both groups were killed every three to four weeks from 0 up to 54 weeks, at which time the experiment was terminated. In animals that lived beyond 25 weeks, the incidence of cholangiocarcinomas was significantly (p < 0.05) higher in the infected group (11/14 animals) than in the uninfected group (6/17 animals). Metastases to other organs were observed only in infected animals with cholangiocarcinomas. The first bile-duct tumours were noted at 25 weeks in the infected group and at 35 weeks in the uninfected group (Iida, 1985).

4.1.1. Humans

4.1.2. Experimental systems

4.1.3. Comparison of humans and experimental animals

The general pathological features of clonorchiasis and opisthorchiasis are similar in human cases and experimental animals, including the Syrian hamster, which is the most commonly used species in carcinogenicity studies. The changes involve predominantly the intrahepatic bile ducts and pancreatic ducts. The initial changes, such as desquamation of the bile-duct lining cells, are followed by hyperplasia of the cells lining the intrahepatic bile ducts and are identical in humans and in the acute stage of experimental infections. In chronic infections in humans and experimental animals, adenomatous hyperplasia of the bile ducts, heavy eosinophilic infiltration and periductal fibrosis occur. Secondary bacterial infections, especially ascending infection with E. coli, result in multiple hepatic abscesses and cholangiohepatitis in livers infected by both Opisthorchis and Clonorchis.

4.2.1. Humans

Inflammatory responses in host tissues challenged by infections or inflammatory agents have been postulated to play a role in the development of cancers which arise in infected organisms (for reviews, see Gentile & Gentile, 1994; Ohshima & Bartsch, 1994). Reactive oxygen species and nitrates, nitrites and various nitrosating agents are produced to kill invading microorganisms and helminths. Polymorphonuclear leukocytes play a prominent role in the production of these host defence agents (for reviews, see Preussmann & Eisenbrand, 1984; Shepard et al., 1987). The radicals have been shown to induce genetic damage in normal host tissues adjacent to the site of inflammation, producing DNA strand breaks, mutations and chromosomal aberrations (Weitzman & Stossel, 1981; Birnboim, 1982). While no data on humans are available to verify these observations, increased levels of urinary nitrates and salivary nitrites are found in O. viverrini-infected individuals.

Srianujata et al. (1984) reported significantly higher concentrations of nitrate (2–2.8 times) and nitrite (2–5.6 times) in the saliva of inhabitants in a high-risk area for cholangiocarcinoma in North-east Thailand than in subjects in Bangkok (low-risk area). Nitrate levels in urine were also significantly higher (1.5–3 times) in the subjects from the high-risk areas. Srianujata et al. (1987) also reported higher levels of nitrite (1.8 times) and N-nitrosoproline (2.6 times) in the urine of subjects infected with O. viverrini than in uninfected subjects from the same area of North-east Thailand. Haswell-Elkins et al. (1994b) confirmed these observations in a study in North-east Thailand in which diet and smoking were controlled for; they also demonstrated decreased concentrations of nitrates and nitrites in these subjects after treatment with praziquantel. Srivatanakul et al. (1991c), in a study in which diet and smoking were not controlled for, reported that subjects living in high-risk areas for fluke infection who had antibodies to O. viverrini had a 10-fold greater potential for endogenous nitrosation, measured on the basis of urinary levels of N-nitrosoproline after proline ingestion, than individuals who did not have antibodies.

Cholangiocarcinomas from O. viverrini-infected patients differed from those in uninfected patients with respect to point mutations in the c-Ki-ras proto-oncogene: mutations were found at codon 12 of this gene in five of nine individuals in Japan who had cholangiocarcinoma but no concomitant fluke infection, but not in six patients from Thailand who harboured both cholangiocarcinoma and fluke infection (Tsuda et al., 1992). Similar results were reported by Kiba et al. (1993), who found, however, that a mutation at the p53 tumour suppressor gene was similar in the two sets of cholangiocarcinoma patients, all but one being GC→AT transitions in a highly conserved GpG site.

4.2.2. Experimental systems

In male Syrian hamsters and jirds (Meriones unguiculatus), 220 days after experimental infection with O. viverrini, marked proliferation of smooth endoplasmic reticulum was observed in hepatocytes, and lobed and enlarged nuclei and mitochondria were seen which showed significant pathological degeneration, up to lysis. There was also accumulation of intermediate filaments in adjacent bile-duct epithelia and in the epithelium of the gall-bladder (Adam et al., 1993). Depressed lymphoproliferative response to phytohaemag-glutinin stimulation has also been described in Syrian hamsters infected with O. viverrini, suggesting an immunodepressive effect (Wongratanacheewin et al., 1987).

The role of O. viverrini in enhancing host response to chemical carcinogens (particularly nitrosamines) has been well documented in Syrian hamster models (see section 3.2). A significant increase in the proportion of water-soluble aflatoxin B₁ metabolites was found in hamsters infected with liver flukes over that measured in uninfected animals (Makarananda et al., 1991), suggesting increased expression of enzymes that metabolize aflatoxin B₁. A cytochrome P450 isozyme(s) (CYP2A) has been identified in the livers of hamsters infected with O. viverrini, the activity of which increased nonuniformly in male but not female animals, the highest levels of activity occurring in hepatocytes immediately adjacent to areas of inflammation. This increase occurred in spite of a decrease in the total hepatic P450 content. The enzyme was shown to contribute up to 50–60% of the metabolism of hepatic aflatoxin B₁ and N-nitrosodiethylamine in infected males and 20–40% in infected females (Kirby et al., 1994).

Immunohistochemical analysis of aflatoxin B₁-DNA adducts in parasitized animals indicated that the greatest numbers of adducts occurred in the regions of highest CYP2A activity. Studies with a related liver fluke, Fasciola hepatica, also showed enhanced cytochrome P450-related activation of aflatoxin B₁ into a mutagen by liver extracts from fluke-infected mice over that with extracts prepared from livers of uninfected animals (Gentile & DeRuiter, 1981).

Nitrosamine and nitrate biosynthesis mediated by nitric oxide synthase was found to be increased in O. viverrini-infected Syrian hamsters, and nitric oxide synthase activity in liver cytosol was twice as high in infected as in untreated hamsters The enzyme was located in macrophages and eosinophils which accumulated at the site of the infection (Ohshima et al., 1994).

Opisthorchis viverrini

Within Thailand, the highest proportional incidence rate of cholangiocarcinoma is observed in the north-east region of the country where the prevalence of infection with O. viverrini is also highest. In this region, the incidence of cholangiocarcinoma is about 40 times the highest incidence outside Thailand. A formal analysis across five regions of the country showed a strong correlation between proportional incidence of cholangiocarcinoma and estimated average titres of antibodies to O. viverrini and, to a lesser degree, faecal egg count. Correlations with proportional incidence rates of hepatocellular carcinoma were much weaker.

Many cases of liver cancer arising in patients with O. viverrini infection have been reported from Thailand. In most regions of the world, cholangiocarcinoma is a very rare tumour. In areas where O. viverrini is endemic, however, the numbers of cases of cholangiocarcinoma generally outnumber those of hepatocellular carcinoma.

Three cross-sectional or case–control studies of the association between infection with O. viverrini and cancer of the liver have been reported from Thailand. In the earliest and smallest of these studies, the estimated relative risks for cholangiocarcinoma and hepatocellular carcinoma in association with the presence of O. viverrini eggs in faeces were each 1.3. In the second study, the estimated relative risk for the association between cholangiocarcinoma and the presence of O. viverrini antibodies in serum was 5.0, which was significant. The association was not explained by possible confounding with hepatitis B virus infection or estimated recent intake of aflatoxins. The estimated relative risk for the association with hepatocellular carcinoma was 1.7 (not significant). In the third study, based on 15 cases of cholangiocarcinoma, estimated relative risks of 1.7, 3.2 and 14.1 were calculated for categories of faecal excretion of increasing numbers of O. viverrini eggs. This trend was highly significant.

Opisthorchis felineus

The incidence of liver cancer was observed to be correlated with the prevalence of infection with O. felineus across four areas in the T'umen' region of north-west Siberia. Cases of both cholangiocarcinoma and hepatocellular carcinoma have been reported in people infected with O. felineus.

Clonorchis sinensis

Cases of cancer of the liver in association with infection with C. sinensis have been reported from China, Hong Kong, the Republic of Korea and Japan and in immigrants to North America from China and Laos.

Two case–control studies of the relationship between C. sinensis infection and liver cancer, with partially overlapping case series, have been carried out in the Republic of Korea. Significantly increased estimated relative risks of 6.5 and 6.0 were seen for an association with cholangiocarcinoma, but no significant association was seen with the occurrence of hepatocellular carcinoma. In a third case–control study, in Hong Kong, the estimated relative risk for cholangiocarcinoma, after adjustment for age and sex, was 3.1, while that for hepatocellular carcinoma was 0.7.

Overall evaluations

Infection with Opisthorchis viverrini is carcinogenic to humans (Group 1).

Infection with Opisthorchis felineus is not classifiable as to its carcinogenicity to humans (Group 3).

Infection with Clonorchis sinensis is probably carcinogenic to humans (Group 2A).